Advances in Experimental Medicine and Biology 1087

Junjie Xiao Editor

Circular RNAs
Biogenesis and Functions
Advances in Experimental Medicine
and Biology

Volume 1087

Editorial Board:

IRUN R. COHEN, The Weizmann Institute of Science, Rehovot, Israel

ABEL LAJTHA, N.S. Kline Institute for Psychiatric Research,
Orangeburg, NY, USA
JOHN D. LAMBRIS, University of Pennsylvania, Philadelphia, PA, USA
RODOLFO PAOLETTI, University of Milan, Milan, Italy
NIMA REZAEI, Tehran University of Medical Sciences, Children’s Medical
Center Hospital, Tehran, Iran
More information about this series at http://www.springer.com/series/5584
Junjie Xiao
Editor

Circular RNAs
Biogenesis and Functions
Editor
Junjie Xiao
School of Life Science,
Institute of Cardiovascular Sciences
Shanghai University
Shanghai, China

ISSN 0065-2598     ISSN 2214-8019 (electronic)

Advances in Experimental Medicine and Biology
ISBN 978-981-13-1425-4    ISBN 978-981-13-1426-1 (eBook)
https://doi.org/10.1007/978-981-13-1426-1

Library of Congress Control Number: 2018956322

© Springer Nature Singapore Pte Ltd. 2018

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or
part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way,
and transmission or information storage and retrieval, electronic adaptation, computer software,
or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are
exempt from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in
this book are believed to be true and accurate at the date of publication. Neither the publisher nor
the authors or the editors give a warranty, express or implied, with respect to the material
contained herein or for any errors or omissions that may have been made. The publisher remains
neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore
189721, Singapore
Contents

Part I Overview
1 An Overview of Circular RNAs����������������������������������������������������    3
Rajendra Awasthi, Anurag Kumar Singh, Gaurav Mishra,
Anand Maurya, Dinesh Kumar Chellappan, Gaurav Gupta,
Philip Michael Hansbro, and Kamal Dua

Part II Bioinformatics for Circular RNAs

2 RNA sequencing and Prediction Tools for Circular
RNAs Analysis��������������������������������������������������������������������������������   17
Elena López-Jiménez, Ana M. Rojas,
and Eduardo Andrés-León
3 Online Databases and Circular RNAs ����������������������������������������   35
Seyed Hamid Aghaee-Bakhtiari

Part III Biogenesis of Circular RNAs

4 Circular RNA Splicing������������������������������������������������������������������   41
Nicole Eger, Laura Schoppe, Susanne Schuster,
Ulrich Laufs, and Jes-Niels Boeckel
5 Circular RNAs Biogenesis in Eukaryotes Through
Self-­Cleaving Hammerhead Ribozymes��������������������������������������   53
Marcos de la Peña

Part IV Molecular Mechanisms and Gene Regulation

of Circular RNAs
6 Circular RNAs Act as miRNA Sponges ��������������������������������������   67
Amaresh Chandra Panda
7 Regulation of Transcription by Circular RNAs��������������������������   81
Rumela Bose and Rupasri Ain

v
vi Contents

8 Functional Analysis of Circular RNAs����������������������������������������   95

Shanmugapriya, Hisham Alkatib Huda,
Soundararajan Vijayarathna, Chern Ein Oon, Yeng Chen,
Jagat R. Kanwar, Mei Li Ng, and Sreenivasan Sasidharan

Part V Circular RNAs as Potential Disease Biomarkers

9 Circular RNA in Exosomes ���������������������������������������������������������� 109
Daniele Fanale, Simona Taverna, Antonio Russo,
and Viviana Bazan
10 Circular RNAs in Blood���������������������������������������������������������������� 119
Angela Vea, Vicenta Llorente-Cortes,
and David de Gonzalo-Calvo
11 Circular RNA in Saliva������������������������������������������������������������������ 131
Farinaz Jafari Ghods
12 Emerging Role of Circular RNAs as Potential
Biomarkers for the Diagnosis of Human Diseases���������������������� 141
Rupal Ojha, Raj Nandani, Nina Chatterjee,
and Vijay Kumar Prajapati
13 Circular RNAs as Novel Biomarkers
for Cardiovascular Diseases���������������������������������������������������������� 159
Qiulian Zhou, Zhongrong Zhang, Yihua Bei, Guoping Li,
and Tianhui Wang
14 Circular RNAs as Biomarkers for Cancer���������������������������������� 171
Lu Xia, Meiyi Song, Mengxue Sun, Fei Wang,
and Changqing Yang

Part VI Circular RNAs and Human Diseases

15 Circular RNAs in Cardiovascular Diseases�������������������������������� 191
Lijun Wang, Xiangmin Meng, Guoping Li, Qiulian Zhou,
and Junjie Xiao
16 Circular RNAs and Neuronal Development�������������������������������� 205
Lena Constantin
17 Circular RNAs in Cancer�������������������������������������������������������������� 215
Susanne Lux and Lars Bullinger
18 Circular RNAs in Brain Physiology and Disease������������������������ 231
S. Gokul and G. K. Rajanikant
19 Circular RNA and Alzheimer’s Disease�������������������������������������� 239
Rumana Akhter
20 Circular RNA in Liver: Health and Diseases������������������������������ 245
Meiyi Song, Lu Xia, Mengxue Sun, Changqing Yang,
and Fei Wang
Contents vii

21 Circular RNAs in Organ Fibrosis������������������������������������������������ 259

Jianhua Yao, Qiying Dai, Zhuyuan Liu, Lei Zhou,
and Jiahong Xu
22 Circular RNAs in Metabolic Diseases������������������������������������������ 275
Tianhui Wang, Wen Pan, Jun Hu, Zhongrong Zhang,
Guoping Li, and Yajun Liang
23 Circular RNAs in Vascular Functions and Diseases ������������������ 287
Shengguang Ding, Yujiao Zhu, Yajun Liang, Haitao Huang,
Yiming Xu, and Chongjun Zhong
24 Functional Role of Circular RNA
in Regenerative Medicine�������������������������������������������������������������� 299
Richard Y. Cao, Qiying Dai, Qing Li, and Jian Yang
25 The Role of Circular RNAs in Cerebral Ischemic
Diseases: Ischemic Stroke and Cerebral
Ischemia/Reperfusion Injury�������������������������������������������������������� 309
Jian Yang, Mengli Chen, Richard Y. Cao, Qing Li,
and Fu Zhu

Part VII Circular RNAs in Plants and in Archaea

26 CircRNAs in Plants������������������������������������������������������������������������ 329
Xuelei Lai, Jérémie Bazin, Stuart Webb, Martin Crespi,
Chloe Zubieta, and Simon J. Conn
27 Circular RNAs and Plant Stress Responses�������������������������������� 345
Celso Gaspar Litholdo Jr.
and Guilherme Cordenonsi da Fonseca

Part VIII Future Prospects

28 Prospective Advances in Circular RNA Investigation���������������� 357
Siti Aishah Sulaiman, Nor Azian Abdul Murad,
Ezanee Azlina Mohamad Hanif, Nadiah Abu,
and Rahman Jamal
 Part I
Overview
 An Overview of Circular RNAs
1
Rajendra Awasthi, Anurag Kumar Singh,
Gaurav Mishra, Anand Maurya,
Dinesh Kumar Chellappan, Gaurav Gupta,
Philip Michael Hansbro, and Kamal Dua

Abstract Keywords
Circular RNAs (cirRNAs) are long, noncoding cirRNA · Circular RNAs · Gene expression ·
endogenous RNA molecules and covalently Translation
closed continuous loop without 5′–3′ polarity
and polyadenylated tail which are largely con-
centrated in the nucleus. CirRNA regulates gene
expression by modulating microRNAs and 1 Introduction
functions as potential biomarker. CirRNAs can
translate in vivo to link between their expression In 1976 Sanger and coworkers proposed that the
and disease. They are resistant to RNA exonu- viroids are single-stranded structures covalently
clease and can convert to the linear RNA by bound to circular RNAs (cirRNAs). These are
microRNA which can then act as competitor to pathogenic to certain plants of higher class. It
endogenous RNA. This chapter summarizes the was primitively reported as a viroid, consisting of
evolutionary conservation and expression of cir- a covalently closed cirRNA molecule, and patho-
RNAs, their identification, highlighting various genic to particular higher plants [1].
computational approaches on cirRNA, and CirRNAs, a class of noncoding endogenous
translation with a focus on the breakthroughs RNA, regulate gene expression in mammals at
and the challenges in this new field. the transcriptional or posttranscriptional level by

R. Awasthi (*)
Amity Institute of Pharmacy, Amity University,
Noida, Uttar Pradesh, India
A. K. Singh
Centre of Experimental Medicine & Surgery, Institute G. Gupta
of Medical Sciences, Banaras Hindu University, School of Pharmaceutical Sciences, Jaipur National
Varanasi, Uttar Pradesh, India University, Jaipur, India
G. Mishra · A. Maurya P. M. Hansbro · K. Dua (*)
NKBR College of Pharmacy and Research Centre, School of Biomedical Sciences and Pharmacy,
Meerut, Uttar Pradesh, India University of Newcastle, Hunter Medical Research
Institute, Newcastle, Australia
D. K. Chellappan
Department of Life Sciences, School of Pharmacy, Discipline of Pharmacy, Graduate School of Health,
International Medical University, University of Technology Sydney,
Kuala Lumpur, Malaysia Sydney, NSW, Australia

© Springer Nature Singapore Pte Ltd. 2018 3

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_1
4 R. Awasthi et al.

interacting with microRNAs [2–6]. For many RNA detection [12]. CirRNAs have great poten-
years, cirRNAs were overlooked as rare isoforms tial as clinical diagnostic markers and new
that result from splicing artifacts or gene rear- therapeutic molecules for the disease therapy. Till
rangements [7]. These rediscovered RNA today few reports have been published on cir-
­molecules mainly arise from exon circularization RNAs due to low expression level. Originally
or intron circularization and covalently joined 3′ these molecules were considered as by-products
and 5′ ends of a single-stranded RNA molecule of alternative splicing and were named as a
by backsplice events (an upstream splice accep- genetic accident or experimental errors [1].
tor is joined to a downstream splice donor), thus
presenting as covalently closed continuous loops
[1, 7–9]. 2 Identification
CirRNAs are misinterpreted as splicing errors. and Appropriate Validation
Recently, cirRNAs have shown to be widespread of cirRNA
and diverse in eukaryotic cells [10]. CirRNAs are
relatively stable in the cytoplasm [5]. These are The cirRNA was firstly recognized in the early
produced by a backsplicing process, wherein 1990s. The recognition on a large scale was not
downstream exons are spliced to upstream exons focused in the early stages because of its tedious
in reverse order [2]. CirRNAs are more stable traditional method of study and due to the lack of
than linear RNA isoforms due to the lack of useful information. Therefore, the developed
accessible ends, which resist exonucleases. recent method of study and identification brings a
However, the mechanism of cirRNA formation very precise way to explore cirRNA. Moreover, a
and their cellular function are still unclear. key element of cirRNA is out of-order arrange-
Relating to the function of cirRNA, it is hypoth- ment of exons kenned as a backsplice (described
esized that these molecules are epigenetic beneath) is not one of a kind to cirRNAs. An
microRNA sponges [7]. Human CDR1as/ciRS-7 early RNA-seq mapping algorithm filtered out
are examples of functional exonic cirRNAs such sequences. These issues have been tended to
which have been experimentally validated to through the improvement of exonuclease-based
function as miRNA sponges and involve in gene enhancement approaches. Novel bioinformatic
expression regulation [10]. However, it is not devices such as sequencing with longer reads and
clear whether all cirRNA molecules work as higher throughput and sequencing of ribosomal
miRNA sponges or not [7]. The stable nature of RNA (rRNA)-depleted RNA libraries (as opposed
cirRNAs makes these moieties intriguing candi- to poly(A)-advanced libraries) make easy to sep-
dates as functional molecules in circulating body arate cirRNAs from other RNAs and also main-
fluid [11]. However, currently, there is no system- tain its circularity [13].
atic approach available for identifying exonic cir- The identification of cirRNAs is exceptionally
RNAs in the human transcriptome [10]. valuable for understanding the regulatory mecha-
Various challenges associated with the detec- nisms and for potential ramifications for remedial
tion of cirRNA include exclusion of sequencing applications, for instance, working as miRNA
errors, unfair treatment between exonic cirRNAs, sponges for oncogenic miRNAs. lncRNA is
and other types of RNAs (e.g., trans-spliced effectively recognized from other little ncRNA,
RNAs and genetic rearrangements) on the basis such as miRNA, siRNA, and snoRNA, by utiliz-
of prejudice, adjustment of errors, in vitro arti- ing straightforward property transcript size.
facts, and the reconciliation of heterogeneous However, for cirRNA identification from differ-
results [10]. ent lncRNAs, it has been nearly unrealistic to dis-
CirRNAs are specific to certain diseases such tinguish them just on simple features. cirRNA
as neuronal disorders and atherosclerosis [4, 5]. has shown some extraordinary succession attri-
Our insensitivity about cirRNAs is due to an butes from different lncRNAs, for example,
insufficiency of available sequencing data for cir- GT-AG match of sanctioned graft locales, com-
1 An Overview of Circular RNAs 5

bined Alu rehash, and backsplice [14]. Sequence backsplicing utilizing the genome browser. To
features cumulating with machine learning are get a general summary of the required exons
accounted to be puissant to prognosticate gene in cirRNA, an entire backsplice sequence is
regulation, splicing sites, and chromatin 18. They inserted in the BLAT implement by using
promote sequence-based strategy possibly used https://genome.ucsc.edu/cgi-bin/hgBlat. The
to recognize cirRNA from different lncRNAs corresponding exon sequences are fetched to
efficaciously [15]. the respective gene and species using www.
Discovery of cirRNA articulation can be ensembl.org. The exon order is reversed,
accomplished utilizing various techniques such keeping 5′→3′ orientation of both exons. The
as polymerase chain reaction (PCR) of the sequence has to be pasted into the correspond-
Northern blot, two-dimensional gel electrophore- ing box using http://primer3.ut.ee/. The box is
sis, gel trap electrophoresis, in situ hybridization, changed from a product size in the range of
and RNase degradation assay [16]. 70–150 bp and cull pick primers. Cull primer
pair ascertains amplified region covering the
backsplice site and controls the presaged
2.1 Identification and Validation primer tm to 60 °C. It is suggested that the
of cirRNA by PCR primers should not overlap the backsplice site.
The primer sequence is examined by UCSC in
PCR is the speediest and most effortless tech- silico PCR implement to check the amplifica-
nique to distinguish the expression of cirRNAs. tion in genome assembly and the UCSC-­
Primers are utilized as a part of PCR for recogni- annotated genes; no presaged amplification is
tion of protein-coding or noncoding RNAs. These expected.
are essentially planned and focused to permit b. Semiquantitative PCR: The accompanying
enhancement of the primer-flanked nucleic acid convention is depicted for the enhancement of
region. The utilization of different oriented cirRNAs. For reference, articulation of the
primer sets is fundamental for the recognition of straight RNA of the quality of intrigue ought
cirRNA articulation utilizing PCR [17]. Sanger to be evaluated. Briefly, the buffer concentrate
sequencing is the fundamental technique to iden- is defrosted, dNTP is mixed, and random
tify various circular transcripts by semiquantita- hexamer-­primed cDNA is kept on the ice. The
tive or quantitative PCR. Sanger sequencing is quantity of responses is computed, and 1–2
also useful for further refinement of the PCR extra responses are incorporated to make up
product to validate backsplice site. Backsplice for inevitable misfortune by pipetting out.
sequence information can be generated from PCR Master is mixed maintaining Taq
RNA sequencing data or publicly accessible sets Reaction Buffer (10×) 2.5 μL, 1 μL of Taq
of non-poly(A)-culled RNA sequencing data polymerase (1 U/μL), 0.5 μL of 10 mM
from the National Center for Biotechnology dNTPs, forward and reverse primers (1 μL
Information – Gene Expression Omnibus each), and 14 μL of RNase/DNase-free water.
(NCBI – GEO) database. Primers to categorically PCR Master Mix (20 μL) is distributed for
detect cirRNAs by PCR should be planned diver- each reaction, and 5 μL of random hexamer-
gently which can be straightforwardly achieved primed cDNA (an RNA/cDNA equivalent of
utilizing free online implements such as Primer3. >10 ng per reaction is recommended) is added.
Identification of circular RNA by PCR is done Control PCR Master (15 μL) containing H2O
in the following steps: and 5 μL of RNase-/DNase-free H2O is added
and mixed. PCR is carried out at 95 °C for
a. Primer design: For the determination of chro- 2 min, 95 °C for 10 s, 60 °C for 20 s at 30–35
mosome position of the terminuses presaged PCR cycles, and 72 °C for 15 s. The time and
to pair for backsplicing, we need to determine temperature are subject to the individual poly-
which exons/introns are to be included in the merase and the item estimate. PCR items are
6 R. Awasthi et al.

broken down by gel electrophoresis utilizing of the PCR. An equivalent volume of phenol/
2% agarose gels. chloroform/isoamyl liquor (25:24:1 (v/v/v)) is
c. The accompanying convention is portrayed added to the PCR item in a 1.5 mL response
utilizing the SYBR Green Master Mix for a tube mixed and centrifuged for 5 min at
standard 96-well qPCR: Heumuller and 12,000 × g at room temperature (25 °C). The
Boeckel described specificity of the PCR tube is handled deliberately and abstained
examine for cirRNA discovery utilizing semi- from irritating stage partition-exchange, the
quantitative PCR and gel electrophoresis pre- upper (watery) stage to another 1.5 mL
ceding qPCR. Besides, the qPCR item ought response tube. It is recommended not to
to dependably be prepared by liquefying bend aggravate the lower (natural) stage. Tainting
investigation and in any event once by conse- with the lower stage can bring about dimin-
quent gel electrophoresis. Dissolve bend ished extraction productivity. The tube con-
investigation is not vital when utilizing hydro- taining the lower stage (phenol squander) is
lysis test-based qPCR. In this situation, a disposed of. The product was blended with
hydrolysis test ace blend is utilized rather than 2.5 mL of ice-cold ethanol and mixed for
the SYBR Green Master Mix (SYBR-GMM) 15 min at 4 °C to hasten the DNA. The super-
in the accompanying convention. SYBR-­ natant is evacuated using a pipette. The pellet
GMM is defrosted, and irregular hexamer is dried on a warm obstruct with open cover at
cDNA is prepared on ice. The quantity of 37 °C for 0.5–2 min. The pellets are resus-
responses is figured out, and 1–2 extra pended in 10 μL TE cushion. DNA sum ought
responses are incorporated to make up for to be resolved, and the test is sent to PCR
inevitable misfortune by pipetting. This is fol- sequencing utilizing the disparate forward and
lowed by planning qPCR Master Mix. The invert preliminary.
qPCR Master Mix contains 10 μL of SYBR-­
GMM, 3 μL of water, and 1 μL of the forward
and switch 10 μM groundwork stock. To per- 2.2 I dentifying cirRNAs by RNA
form hydrolysis test-based qPCR, it is recom- Fluorescence In Situ
mended to utilize 10 μL of hydrolysis test ace Hybridization (FISH)
blend, 2 μL of water, 1 μL of the forward and
turnaround 10 μM preliminary stock each, and FISH permits the representation of various RNA
1 μL of the hydrolysis test. PCR ace blend species inside the cell. This section contains an
(15 μL) is circulated for every response, and 5 all-around appropriate technique to recognize
μL irregular hexamer-prepared cDNA is cirRNA through a junction specific test.
incorporated. This is followed by the incorpo- CirRNAs are set apart by making a beeline for a
ration of H2O control comprising of 15 μL of tail-ligated intersection that is not found in some
the PCR ace blend and includes 5 μL H2O. It other RNAs. Till today, this convention is very
is ensured to liquefy bend for every ground- strong and delicate. Numerous tests marked by
work. The information is broken down utiliz- an alternate fluor can be taken into consider-
ing the 2-CT strategy or the 2-∆CT technique ation to achieve synchronous identification of
when a housekeeping quality (e.g., the mRNA different targets [18]. RNA FISH depends on
of RPLP0) has been estimated. the straightforward idea of uncovering settled
d. PCR items increasing the back-grafted area cells or tissues to short DNA oligonucleotides in
ought to be filtered utilizing phenol/chloro- adequately high fixations to enable blending
form/isoamyl liquor precipitation: In this way, with corresponding RNA molecules to frame
Sanger sequencing (PCR sequencing) is uti- stable DNA-RNA half-­breeds. The tests com-
lized to approve the presence of the back-join prise of a pooled set of ~32–48 DNA oligos of
site and to control the specificity of the differ- various arrangements, every 20 nucleotides in
ently oriented preliminaries utilized as a part length and named with a solitary fluorophore at
1 An Overview of Circular RNAs 7

its 3′ end. The convention can recognize single tests give an extremely profitable and exceed-
RNA molecule with high specificity (a couple of ingly useful approach.
false positives) and high affectability (a couple
of false negatives) and does not require flag
intensification steps, which tend to render sin- 2.4 Portrayal of cirRNA
gle-atom identification approaches less quanti- Concatemers
tatively [19].
The model on cirRNA biogenesis suggests that
the rearranged rehashes take part in base blend-
2.3  orthern Blot Analysis
N ing, accordingly situating the two splice sites in
of cirRNAs nearness. Wang and his colleagues outlined the
embodiment of exon 2 from beta-globin (HBB)
Northern smear hybridization makes the strat- in the middle of modified components. As far as
egy for the decision to convincingly show round anyone is concerned, this exon is not creating cir-
setup of putative cirRNAs. CirRNA identifica- RNA in its normal setting. However, when
tion can be proficient by short tests spreading flanked with transformed rehashes, the exon pro-
over the round graft intersection or by longer duces one particular cirRNA, as well as a step of
tests covering as much as a whole circularized cirRNA-like items (cirRNA concatemers) [21].
exon. This alternative winds up significantly if The identification and profiling of cirRNA are
the specificity for roundabout isoform is not normally done by cutting-edge sequencing
fundamental (for instance, if the straight struc- (NGS) or by qRT-PCR [22]. The cirRNA in the
tures do not enter the gel, if both direct and first place contained exon rehashes or whether
roundabout isoforms ought to be identified in the monotony was presented by the RT chemical.
parallel, or if there should be an occurrence of Barrett et al. presented a blueprint of basic bio-
solely roundabout RNAs). Northern blots are chemical tests projected by northern smearing to
thus basic part of any cirRNA portrayal, because ponder the idea of these cirRNAs and demon-
of their incredible flexibility. To start with the strated that they are made out of exon rehashes
decision of test districts (round or straight joint (cirRNA concatemers). To recognize concate-
intersection or exonic areas) and identification mers and interwoven cirRNAs (topologically
standard (digoxigenin or 32P-named tests) bolted single exon cirRNAs), three particular
decides the specificity for roundabout versus examinations such as (1) RNase R absorption to
direct isoforms. The decision of gel network approve the roundabout structure of the cirRNA
includes greater adaptability in northern smudge species, (2) RNase H absorption to decide the
examine. Agarose gels are reasonable for cir- structure of exons by crumbling the cirRNAs into
RNAs from 0.2 kb up to a few kb. In agarose their exon units, and (3) a soluble treatment to
gels, round and straight RNAs of a similar size tenderly scratch the cirRNA into a relating direct
cannot be recognized by their running conduct. RNA have been suggested [23].
Actually, in denaturing polyacrylamide gels,
direct RNA keeps running at the normal size,
while cirRNAs have a lower evident versatility 3 Computational Approaches
with respect to straight markers; this hindrance on cirRNA
impact is upgraded by expanding acrylamide
fixations [20]. Due to this impediment, cirRNAs The hereditary data streams of life, in which
up to 1 kb can be examined by polyacrylamide DNA and protein are considered as primary on-­
gel electrophoresis. Thus, at any rate for a far- screen characters of cell life, retain RNA as basic
reaching investigation of one or a couple of part of protein synthesis. However, this perspec-
putative cirRNAs, not for a medium- to high- tive of the organic part of RNA experienced vari-
throughput screening endeavors, Northern blot ous challenges [24]. Computational approaches
8 R. Awasthi et al.

to deal with RNA tertiary structure expectation identification calculations, promising computa-
are based on the examination of RNA tertiary tional techniques have been developed for the
themes, and diagram hypothesis for RNA and downstream investigations of cirRNAs. However,
RNA endeavors plan to enhance the in vitro test new computational techniques to remake full
choice for aptamer outline. Thus, the examina- length of cirRNAs and measure their demeanor
tion of RNA basic correlation is important are critically required. Late examinations have
thought of root-mean-square deviation (RMSD) shown that cirRNAs are universal and have dif-
since the forecasts are not exact for RNA during ferent capacities and systems of biogenesis. In
this phase [25]. There are numerous different such investigations, computational profiling of
zones of advancement in RNA bioinformatics, cirRNAs has been pervasively utilized as an irre-
for instance, auxiliary structure forecasts [26]. placeable strategy to give high-throughput ways
Current discoveries in the field of noncoding to deal with identifying and breaking down of
RNA are focused on cirRNA [27] which are pro- cirRNAs. In any case, without a general compre-
duced by nonlinear backsplicing linked to a hension of the basic methodologies, these com-
downstream splice donor and upstream splice putational techniques may not be exactly chosen
acceptor. CirRNA is present in all eukaryotic or utilized for a particular research reason, and a
clades, including insects, fungi, and plants, and it few misguided judgments may bring about pre-
also exists in humans to establish several thou- dispositions in the examinations. Gao and Zhao
sand different cirRNAs. In the immense majority, reviewed the key advances and abridged trade-off
the function of cirRNA is not clear. A small sub- of various systems, covering every single promi-
set of cirRNAs has been reported to act as steer- nent calculation for cirRNA discovery and differ-
ers for miRNAs [28, 29] or to bind and regulate ent downstream investigations [39].
protein function [30, 31]. The diversity of cir- The computational approach plays an impor-
RNAs can be explicated based on the gene frac- tant role in high-throughput RNA-seq data exam-
tion and antisense strand of some genes and from ination and in expression of cirRNA profiling.
intergenic regions [32–34]. The length of cirRNA Till today about 11 computational approaches for
ranges from 100 bp to 4 kb [35]. cirRNAs may cirRNA detection have been reported. CIRI,
hold multiple or single exon [36] and are present CIRCexplorer [40], and KNIFE [41] are more
in different cell lines, tissues, and extracellular functional than other approaches. All the reported
exosomes. Biogenesis of cirRNA is based on computational approaches have their own advan-
lariat-driven, intron-pairing-driven, and RNA-­ tages and essential point sensitivity, precision,
binding protein-driven circularization mecha- and computational cost (Tables 1.1 and 1.2).
nisms [37, 38]. It has been proposed that cirRNA Downstream computational approaches are sig-
acts as microRNA sponges and regulates multi- nificant due to their primary detection results.
ple gene expressions. The source quality, mode These approaches are linked to the quantification
of exon creation, biogenesis, and capacity make and differential expression analysis (Table 1.3)
them different than other RNAs. Comment-free [42, 43].
recognition calculations can be utilized as a part
of an extensive variety of living beings. However,
it requires more careful systems to guarantee 3.1  etection of cirRNA Using
D
unwavering quality. Annotation, Genome
The majority of the location techniques are Reference, and GT-AG Splicing
upgraded for their assigned aligners, and these Signals
can be additionally partitioned into joint mindful
aligners and adaptable read mappers. Paired-end Genomes are essential for algorithm sensing and
sequencing gives more data to diminish false can be used in the detection workflows. It mainly
positives for discovery techniques that receive works for the direct alignment of sequencing reads
sifting in light of paired-end mapping. In view of against the standard genome. UROBORUS [44],
1 An Overview of Circular RNAs 9

Table 1.1 Compilation of 11 cirRNA detection methods

Mapper
Category Method Mapper(s) Characteristics type
Split-alignment-­ CIRI BWA-MEM Filtering stringent PEM Versatile
based Restore of unbalanced BSJ read multiple
seed matching
Split-alignment-­ CIRCexplorer TopHat/STAR Noncollinearity detection Splice-­
based aware
Split-alignment-­ DCC STAR GT-AG splice sites Splice-­
based aware
Split-alignment-­ cirRNA_ STAR GT-AG splice sites Splice-­
based finder aware
Split-alignment-­ MapSplice Bowtie Embedded in algorithm to detect cirRNA Versatile
based
Pseudoreference-­ KNIFE Bowtie, Bowtie 2 De novo detection as remedy Versatile
based
Split-alignment-­ Find circ Bowtie 2 No PEM filtering Versatile
based Two 20bp anchors for noncollinearity
detection
Split-alignment-­ segemehl Per se Few of the filters adopted Versatile
based
Pseudoreference-­ NCLscan BWA, BLAT, Trans-spliced transcript detection in Mixed
based Novoalign addition to cirRNA detection
Split-alignment-­ UROBORUS TopHat Storage of unbalance BSJ reads Mixed
based
Pseudoreference-­ PTESFinder Bowtie, Bowtie 2 No PEM filtering Versatile
based

Table 1.2 Performance comparison among 11 cirRNA detection methods by third-party evaluation
Method Hs68 true positive Hs68 precision (%) HeLa true positive HeLa precision (%)
KNIFE 2359 66.53 2055 44.26
MapSplice 1854 76.33 1766 54.11
DCC 2107 63.08 1760 45.22
UROBORUS 279 19.73 761 31.00
CIRI 3400 69.49 3210 54.20
PTESFinder 2474 63.29 2054 35.65
Segemehl 3094 8.74 2506 14.32
NCLscan 892 64.73 954 45.06
Find_circ 2377 59.75 2092 39.99

Table 1.3 Summary of computational methods for downstream analysis of cirRNAs

Method Language Input requirement Function
FUCHS Python BAM/SAM formatted alignment miRNA seed analysis
CIRI-AS Perl cirRNA, references of SAM formatted Detection of internal structure
CircView Java CirRNA list Visualization
Sailfish-cir Python GTF format annotation Very close quantification
CirPro Perl FASTQ-formatted sequencing reads Protein-coding potential estimation
CircTest R Parental gene with read-count Differential expression test
10 R. Awasthi et al.

Fig. 1.1 Pseudoreference-based approach for the cirRNA detection

The reference genome is combined with the corresponding genome annotation to build pseudo-sequence

CIRCexplorer [45], find circ [46], and CIRI [47] read mappers which are generally applicable in
are the examples of detection algorithms. The cir- reference-based RNA/DNA sequence studies.
cularity pathway of cirRNA is different from other The simplest way is splice-aware aligners which
categories of RNAs, and thus an ­evident feature were developed for RNA-seq reads across intron-­
can be captured from the circle junction alignment sized gaps on genome references such as
which is used as a backsplice junction (BSJ). By Novoalign, STAR [51], and TopHat [52].
contrast, forward-spliced junction (FSJ) in mRNA, Detection algorithms, such as CIRCexplorer and
the developed sequencing reads that are collin- DCC [53], depend on this type of aligner. An evi-
early aligned on the genome, interprets spanning dent advantage is that the aligners are optimized
BSJs are divided into segment and are aligned to according to eukaryotic transcription which is
the address/reference sequence in reverse order. easier than various read mappers. The flexibility
Hence, detection algorithms in this category can of splice-aware aligners is less than the versatile
be termed as split-­alignment-­based approaches. mappers, which have developed both end-to-end
For different algorithms, such as NCLscan and local alignment. In other condition, nearly all
[48] and KNIFE [49], the reference genome is splice-aware aligners are based on versatile read
combined with the corresponding genome anno- mappers.
tation to build pseudo-sequence around putative
BSJs in the first few steps (Fig. 1.1). Consequent
steps are centered on the complete alignment of 4 Translation of cirRNA
sequencing reads against such pseudo-sequences
to identify BSJ reads. In addition to a BSJ The introduction of nearly all cirRNAs from
pseudo-­ sequence database, KNIFE also con- exons and their confinement in cytoplasm increase
structed a FSJ sequence information according the probability of cirRNA translation [54].
to the annotation to remove candidate reads with Translation of several viral proteins in many
high-score alignment in both databases. In this organisms, including humans, depends on inter-
category, detection algorithms may be termed nal ribosome entry site (IRES) and cellular IRESs.
pseudoreference-­based approaches. However, their mechanism of action remains con-
The application of annotation is much useful. As troversial [55]. Based on this theory, artificial cir-
an example, a comprehensive evaluation of RNA- RNAs with IRES have been translated [56, 57]. It
seq aligners which actively addressed that annota- has also been noticed that principal cirRNAs can
tion can help to increase the sensitivity for junction be translated in vitro and in vivo [58]. Translation
reorganization versus de novo detection [50]. of cirRNAs in living human cells is based on roll-
ing circle amplification mechanism. The elements
including IRES are not required for the translation
3.2  arious Read Mappers or
V of cirRNA in eukaryotic translation system [59].
Specified Splice-Aware Aligner The cirRNAs have been endogenously translated
and indirectly tested [59].
BSJ reads such as split-alignment and pseudoref- Binding of open pre-initiation complex con-
erence approaches are identified with the help of taining small ribosome is the beginning step for
alignment of transcriptomic read. Many algo- canonical translation process in eukaryotes [59].
rithms, for example, KNIFE and CIRI, prefer The interaction between the cap-binding protein
1 An Overview of Circular RNAs 11

(CBP) poly(A) regions leads to the ­circularization ZNF609 was indicated to sediment with heavy
of mRNA competent for translation [60]. Small polysome. Treatment with puromycin disrupted
ribosomal subunits and mRNAs are further active translation of ribosomes which shifted
scanned by small ribosomal subunits for the start circ-ZNF609 to lighter polysomes. p-circ3XF
codon. After that the 60S ribosomal subunit is containing 3XFLAG-coding sequence of stop
required. Also, the internal start codons can recir-codon has been tested to express circ-ZNF609
culate the ribosomes internally by an IRES-­ protein-coding ability. The study resulted in two
dependent mechanism. flagged isoforms. On the other side, p-lin3XF
containing circ-ZNF609 ORF was also produced
and expressed same proteins more efficiently.
4.1 Translation of cirMbl The RNA amount was normalized in both cir-
c3XF and p-lin3XF. The results suggested that
Till date, the investigations are limited to the the translation efficiency of p-circ3XF was lower
in vivo translation of endogenous cirRNAs. Ribo-­ than that of p-lin3XF. CRISPR/Cas9 has been
cirRNAs have specific ribosome profiling and are utilized to introduce 3XFLAG-code in endoge-
denoted as cirRNAs. Ribosome profiling datasets nous ZNF609 gene which translated circZNF609
of Drosophila have shown presence of cirRNA-­ from the chromosomal gene. Positive clone with
specific junctions [61]. Expression of cirRNA in alleles (1) holding expected flag and (2) holding
Drosophila S2 cells depends on the presence of a deletion that prevents circ-ZNF609 production
intron-exon-intron minigenes. The minigenes have has been reported. However, only clone carrying
been reported to express V5-tagged proteins. flagged allele can develop circ-ZNF609.
V5-tagged proteins have been reported from the In an investigation the positive clone cell
cells transfected with circCdiV5, circPde8V5, or lysates were immunoprecipitated with anti-­
circMblV5. This was not observed with the cells FLAG antibody and subject to mass spectrome-
which are not transfected with minigenes of cir- try. One peptide mapping to the cir-ORF was
cHaspinV5 or circCamKIV5. The protein of observed in the positive clone cell, whereas sev-
desired size has been observed using antiMBL eral peptides were present in the protein from
antibody. For cirMbl, cirRNA has been established cells overexpressing p-circ3XF and
as the main source of detected protein. Transgenic p-lin3XF. This resulted in the formation of lower
flies containing MBL-­immunoreactive bands have cir-RNAs from the chromosomal gene and lower
been originated to express in vivo cirMbl minigene translational capacity. The heat shock resulted to
[62]. In ribosome footprinting (RFP) reads of fly increased translation of circZNF609. In this pro-
heads about 122 ribo-cirRNAs have been identi- cess no cap structure was present in cirRNA. The
fied. Protein domains have been identified in many study proposed the possibility of translation
ribo-­cirRNA-­encoded proteins. Protein expression through sequences with internal ribosome entry
of V5-tagged cirRNA minigenes was not affected site (IRES) activity [62].
by co-expression of 4E-BP that inhibits cap-­
dependent translation. circMbls are translated in a
cap-independent manner, and the untranslated 5 m6 A Modification in cirRNA
region (UTR) sequence of circMbl is capable of Translation
facilitating cirRNA translation [61].
Yang and coworkers discovered sequences to
induce translation of cirRNA [56]. RRACH frag-
4.2 Translation of circ-ZNF609 ment involves in the methylation of N6 position
of adenosine (m6 A). m6 Higher peak density of
Circ-ZNF609 codon starts with a linear transcript m6 A has been observed in cirRNA when com-
and terminates at a stop codon. It contains 753-nt pared to mRNAs [63, 64] suggesting its role in
ORF [62]. A substantial percentage of circ-­ mRNA translation [65, 66]. These findings
12 R. Awasthi et al.

s­ upport the hypothesis that m6 A plays an impor- 2. Li J, Yang J, Zhou P, et al (2015a) Circular RNAs in
cancer: novel insights into origins, properties, func-
tant role in cirRNA translation. To validate this tions and implications. Am J Cancer Res 5(2):472
hypothesis, a short fragment containing m6 A 3. Li P, Chen S, Chen H et al (2015b) Using circular
motif was introduced before the start codon in RNA as a novel type of biomarker in the screening of
cirRNA reporter. The level of GFP protein output gastric cancer. Clin Chim Acta 444:132–136
4. Greene J, Baird AM, Brady L et al (2017) Circular
was assessed in all the transfected cells [57], and RNAs: biogenesis, function and role in human dis-
it was observed that cirRNA containing m6 A eases. Front Mol Biosci 4:38
motif was translated expeditiously. The methyla- 5. Hansen TB, Kjems J, Damgaard CK (2013)
tion of cirRNAs with m6 A motif is associated Circular RNA and miR-7 in cancer. Cancer Res
73(18):5609–5612
with RNA-immunoprecipitation (RNA-IP). 6. Bachmayr-Heyda A, Reiner AT, Auer K et al (2015)
Negative effect of m6 ademethylase FTO co-­ Correlation of circular RNA abundance with prolifer-
expression on the amount of immunoprecipitated ation–exemplified with colorectal and ovarian cancer,
RSV-containing cirRNA and GFP translation idiopathic lung fibrosis, and normal human tissues.
Sci Rep 5:8057
from cirRNA has been observed. eIF4G2, a non- 7. Barrett SP, Salzman J (2016) Circular RNAs: analy-
canonical protein which recognizes IRES, also sis, expression and potential functions. Development
plays an important role in cirRNA translation and 143(11):1838–1847
partakes in cap-independent translation and m6 A 8. Lasda E, Parker R (2014) Circular RNAs: diversity of
form and function. RNA 20(12):1829–1842
reader protein YTHDF3 [56, 67]. N6-methylation 9. Chen I, Chen CY, Chuang TJ (2015) Biogenesis,
requires the translation of mRNA and leads to identification, and function of exonic circular RNAs.
GFP protein translation from m6 A-containing Wiley Interdiscip Rev RNA 6(5):563–579
cirRNA by heat shock stress [56, 65, 67]. 10. Bahn JH, Zhang Q, Li F et al (2015) The landscape of
microRNA, Piwi-interacting RNA, and circular RNA
in human saliva. Clin Chem 61(1):221–230
11. Gao Y, Wang J, Zhao F (2015) CIRI: an efficient and
6 Conclusions unbiased algorithm for de novo circular RNA identifi-
cation. Genome Biol 16(1):4
12. Jeck WR, Sharpless NE (2014) Detecting and charac-
It is evident that the number of cirRNAs with terizing circular RNAs. Nat Biotechnol 32(5):453
known functions is expanding during the last few 13. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular
years. However, the function of various cirRNAs RNAs are abundant, conserved, and associated with
remains unknown, and very little information is ALU repeats. RNA 19(2):141–157
14. Xiong HY, Alipanahi B, Lee LJ et al (2015) The human
available about the control of backsplicing pro- splicing code reveals new insights into the genetic
cess of cirRNA generation. This could be due to determinants of disease. Science 347(6218):1254806
the availability of limited and challengeable meth- 15. Hansen TB (2018) Characterization of circular RNA
ods to detect and characterize cirRNAs. A detailed concatemers. In: Circular RNAs. Humana Press,
New York, pp 143–157
study of cirRNA biogenesis and in vivo research 16. Hansen TB, Wiklund ED, Bramsen JB (2011) miRNA-­
may allow testing for functional consequences of dependent gene silencing involving Ago2-mediated
cirRNA expression and will provide novel insights cleavage of a circular antisense RNA. EMBO
into cellular development human disease. J 30(21):4414–4422
17. Gingeras TR (2009) Implications of chimaeric non-­
co-­linear transcripts. Nature 461(7261):206
Competing Financial Interests The authors declare no
18. Itzkovitz S, Van Oudenaarden A (2011) Validating
competing financial interests.
transcripts with probes and imaging technology.
Nature Methods 8(4s):S12
19. Günzl A, Palfi Z, Bindereif A (2002) Analysis of
RNA–protein complexes by oligonucleotide-targeted
RNase H digestion. Methods 26(2):162–169
References 20. Tabak HF, Van der Horst G, Smit J (1988)
Discrimination between RNA circles, interlocked
1. Liu L, Wang J, Khanabdali R et al (2017) Circular RNA circles and lariats using two-dimensional poly-
RNAs: isolation, characterization and their potential acrylamide gel electrophoresis. Nucleic Acids Res
role in diseases. RNA Biol 14(12):1715–1721 16(14):6597–6605
1 An Overview of Circular RNAs 13

21. Wang PL, Bao Y, Yee MC et al (2014) Circular RNA 40. Gao Y, Zhao F (2018) Computational strate-
is expressed across the eukaryotic tree of life. PloS gies for exploring circular RNAs. Trends Genet
one 9(3):e90859 34(5):389–400
22. You X, Vlatkovic I, Babic A (2015) Neural circular 41. Zhang XO, Wang HB, Zhang Y et al (2014)
RNAs are derived from synaptic genes and regulated Complementary sequence-mediated exon circulariza-
by development and plasticity. Nat Neurosci 18(4):603 tion. Cell 159(1):134–147
23. Barrett SP, Wang PL, Salzman J (2015) Circular RNA 42. Szabo L, Morey R, Palpant NJ et al (2015)
biogenesis can proceed through an exon-containing Statistically based splicing detection reveals neural
lariat precursor. Elife 4 enrichment and tissue-specific induction of circu-
24. Jakobi T, Dieterich C (2018) Deep computational lar RNA during human fetal development. Genome
circular RNA analytics from RNA-seq data. Methods Biol 16(1):126
Mol Biol 1724:9–25 43. Glažar P, Papavasileiou P, Rajewsky N (2014)
25. Reznichenko A (2012) Translational renal genetics. circBase: a database for circular RNAs. RNA
University Library Groningen Host. Humana Press, 20(11):1666–1670
New York 44. Meng X, Chen Q, Zhang P et al (2017) CircPro:
26. Shapiro BA, Yingling YG, Kasprzak W et al (2007) anintegrated tool for the identification of circRNAs
Bridging the gap in RNA structure prediction. Curr with protein-coding potential. Bioinformatics
Opin Struct Biol 17(2):157–165 33:3314–3316
27. Ebbesen KK, Kjems J, Hansen TB (2016) Circular 45. Memczak S, Jens M, Elefsinioti A (2013) Circular
RNAs: identification, biogenesis and function. RNAs are a large class of animal RNAs with regula-
Biochim Biophys Acta 1859:163–168 tory potency. Nature 495(7441):333
28. Hansen TB, Jensen TI, Clausen BH et al (2013) 46. Zhang XO, Wang HB, Zhang Y et al (2014)
Natural RNA circles function as efficient microRNA Complementary sequence-mediated exon circulariza-
sponges. Nature 495:384–388 tion. Cell 159(1):134–147
29. Memczak S, Jens M, Elefsinioti A et al (2013) 47. Gao Y, Zhang J, Zhao F (2017) Circular RNA iden-
Circular RNAs are a large class of animal RNAs with tification based on multiple seed matching. Brief
regulatory potency. Nature 495:333–338 Bioinform. https://doi.org/10.1093/bib/bbx014
30. Zheng Q, Bao C, Guo W et al (2016) Circular RNA 48. Song X, Zhang N, Han P et al (2016) Circular RNA
profiling reveals an abundant circHIPK3 that regu- profile in gliomas revealed by identification tool
lates cell growth by sponging multiple miRNAs. Nat UROBORUS. Nucleic Acids Res 44(9):e87–e87
Commun 7:11215 49. Chuang TJ, Wu CS, Chen CY et al (2015) NCLscan:
31. Ashwal-Fluss R, Meyer M, Pamudurti NR et al (2014) accurate identification of non-co-linear transcripts
circRNA biogenesis competes with pre-mRNA splic- (fusion, trans-splicing and circular RNA) with a good
ing. Mol Cell 56:55–66 balance between sensitivity and precision. Nucleic
32. Du WW, Yang W, Liu E et al (2016) Foxo3 circular Acids Res 44(3):e29–e29
RNA retards cell cycle progression via forming ter- 50. Szabo L, Morey R, Palpant NJ et al (2015)
nary complexes with p21 and CDK2. Nucleic Acids Statistically based splicing detection reveals neural
Res 44:2846–2858 enrichment and tissue-specific induction of circu-
33. Gao Y, Wang J, Zhao F (2015) CIRI: an efficient and lar RNA during human fetal development. Genome
unbiased algorithm for de novo circular RNA identifi- Biol 16(1):126
cation. Genome Biol 16(1):4 51. Baruzzo G, Hayer KE, Kim EJ et al (2017) Simulation-­
34. Memczak S, Jens M, Elefsinioti A et al (2013) based comprehensive benchmarking of RNA-seq
Circular RNAs are a large class of animal RNAs with aligners. Nat Methods 14(2):135
regulatory potency. Nature 495(7441):333 52. Dobin A, Davis CA, Schlesinger F et al (2013) STAR:
35. Salzman J, Chen RE, Olsen MN et al (2013) Cell-type ultrafast universal RNA-seq aligner. Bioinformatics
specific features of circular RNA expression. PLoS 29(1):15–21
Genet 9(9):e1003777 53. Trapnell C, Pachter L, Salzberg SL (2009) TopHat:
36. Lasda E, Parker R (2014) CircularRNAs: diversity of discovering splice junctions with RNA-Seq.
form and function. RNA 20:1829–1842 Bioinformatics 25(9):1105–1111
37. Li Z, Huang C, Bao C et al (2015) Exon-intron cir- 54. Cheng J, Metge F, Dieterich C (2015) Specific iden-
cular RNAs regulate transcription in the nucleus. Nat tification and quantification of circular RNAs from
Struct Mol Biol 22(3):256 sequencing data. Bioinformatics 32(7):1094–1096
38. Conn SJ, Pillman KA, Toubia J (2015) The RNA 55. Jackson RJ (2013) The current status of vertebrate
binding protein quaking regulates formation of cir- cellular mRNA IRESs. Cold Spring Harb Perspect
cRNAs. Cell 1(6):1125–1134 Biol 5:a011569
39. Hansen TB, Jensen TI, Clausen BH et al (2013) 56. Yang Y, Fan X, Mao M (2017) Extensive translation
Natural RNA circles function as efficient microRNA of circular RNAs driven by N-6-methyladenosine.
sponges. Nature 495(7441):384 Cell Res 27:626–641
14 R. Awasthi et al.

57. Wang Y, Wang Z (2015) Efficient backsplicing pro- 63. Legnini I, Di Timoteo G, Rossi F et al (2017)
duces translatable circular mRNAs, RNA-Publ. RNA CircZNF609 is a circular RNA that can be translated
Soc 21:172–179 and functions in myogenesis. Mol Cell 66:22–37
58. Chen CY, Sarnow P (1995) Initiation of protein syn- 64. Dominissini D, Moshitch-Moshkovitz S, Schwartz
thesis by the eukaryotic translational apparatus on cir- S (2012) Topology of the human and mouse m(6)
cular RNAs. Science 268:415–417 A RNA methylomes revealed by m(6)A-seq. Nature
59. Abe N, Matsumoto K, Nishihara M (2015) Rolling 485:201–284
circle translation of circular RNA in living human 65. Meyer KD, SaletoreY ZP et al (2012)
cells. Sci Rep 5:16435 Comprehensive analysis of mRNA methyla-
60. Guo JU, Agarwal V, Guo H et al (2014) Expanded tion reveals enrichment in 3' UTRs and near stop
identification and characterization of mammalian cir- codons. Cell 149:1635–1646
cular RNAs. Genome Biol 15:409 66. Meyer KD, PatilDP ZJ et al (2015) 5′ UTR m(6)
61. Aitken CE, Lorsch JR (2012) A mechanistic overview A promotes cap-independent translation. Cell
of translation initiation in eukaryotes. Nat Struct Mol 163:999–1010
Biol 19:568–576 67. Fu Y, Dominissini D, Rechavi G (2014) Gene expres-
62. Pamudurti NR, Bartok O, Jens M et al (2017) sion regulation mediated through reversible m(6)A
Translation of CircRNAs. Mol Cell 66:9–21 RNA methylation. Nat Rev Genet 15:293–306
 Part II
Bioinformatics for Circular RNAs
 RNA sequencing and Prediction
Tools for Circular RNAs Analysis 2
Elena López-Jiménez, Ana M. Rojas,
and Eduardo Andrés-León

Abstract approaches (RNase R) have taken even further

Circular RNAs (circRNAs) are noncoding and the possibility of detecting circRNAs.
single-stranded RNA transcripts able to form Besides, the identification of circRNAs
covalently circular-closed structures. They are presence requires the development of specific
generated through alternative splicing events bioinformatics tools to detect junction-­
and widely expressed from human to viruses. spanning sequences from transcriptome
CircRNAs have been appointed as potential deep-­sequencing samples. Thus, recently
regulators of microRNAs (miRNAs), RNA-­ established bioinformatics’ approaches have
binding proteins (RPBs), and lineal protein-­ permitted the discovery of an elevated number
coding transcripts. Although their mechanism of different circRNAs in diverse organisms. In
of action remains unclear, the deregulation of that sense, recent studies have compared dif-
circular RNAs has been confirmed in different ferent methods and advocate the simultaneous
diseases such as Alzheimer or cancer. use of more than one prediction tool. For that
The introduction of high-throughput next-­ reason, we want to highlight pipelines such as
generation sequencing (NGS) technology pro- miARma-Seq that is able to execute various
vides millions of short RNA sequences at circular RNA identification algorithms in an
single-nucleotide level, allowing an accurate easy way, without the tedious installation of
and proficient method to measure circular third-party prerequisites.
RNAs. Novel protocols based on non-­
polyadenylated RNAs, rRNA-depleted, and
RNA exonuclease-based enrichment

Keywords
E. López-Jiménez CircRNAs · CircRNA RNA-seq · CircRNA
Imperial College London, London, UK
prediction tools
A. M. Rojas
Computational Biology and Bioinformatics Group,
Institute of Biomedicine of Seville, Seville, Spain
E. Andrés-León (*)
Bioinformatics Unit, Instituto de Parasitología y
Biomedicina “López-Neyra”, Consejo Superior de
Investigaciones Científicas (IPBLN-CSIC),
Granada, Spain
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 17

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_2
18 E. López-Jiménez et al.

1 Introduction Due to their circular shape, circRNAs are mol-

ecules missing the 3′ polyadenylated tail and
Circular RNAs (circRNAs) are a class of single-­ hence, resistant to RNA-degrading enzymes,
stranded RNA transcripts able to form covalently which increases their cellular half-life to approxi-
circular-closed structures. This type of RNA mol- mately 48 h, while linear RNAs have a roughly
ecule has been considered as noncoding RNA half-life of 10 h [7]. However, circRNAs are not
because no protein product is expressed, although stable in circulating serum exhibiting a short
recent studies point out in an opposite direction. half-life lower than 15 s, apparently due to RNA
The existence of circular transcripts was asserted endonucleases [11]. The amount of circular RNA
almost three decades ago [1], but they were con- molecules is frequently small, constituting
sidered as RNA splicing artifacts [2]. between 5% and 10% of their linear
Nevertheless, current next-generation sequenc- counterparts.
ing (NGS) techniques of non-polyadenylated
RNAs have revealed large numbers of wide-
spread circRNAs highly and stably expressed in 1.2 Biogenesis
cells and tissues [3]. Successive studies had
revealed that the expression level of circRNAs is CircRNAs, like the majority of RNAs, are tran-
rigorously controlled and specific among the dif- scribed by the RNA polymerase II (Pol II)
ferent tissues and even between the different cell enzyme [12], as a pre-messenger RNA (pre-­
types. For instance, in humans they are remark- mRNA). These pre-mRNAs are the principal
ably expressed in the brain, exosomes, and product of transcription, which frequently under-
peripheral blood [4]. Moreover, circRNAs are take a splicing process to harvest linear mRNAs.
present in numerous organisms throughout evo- In the case of circRNAs, they suffer from a back-
lution [5] such as viruses, bacteria, and plants [6]. splice event, promoting the circularization pro-
Functional studies unveil that circular RNAs cess. Consequently, there is a frequently reduced
can control the translation of lineal RNA protein expression of linear mRNAs when they are circu-
transcripts and microRNAs (miRNAs) and con- larized, which produces an inverse correlation
sequently encompass a key role in gene expres- between the number of regular and circular RNA
sion regulation [7]. During the last decade, molecules [12]. Two different processes have
several studies related with this class of noncod- been suggested for mammalian exonic circRNA
ing RNAs have emerged, pointing out possible circularization by the spliceosome machinery
relationships with relevant diseases such as can- [13]. The first proposed mechanism ensues if a
cer, neurodegenerative diseases, and cardiovas- downstream splice donor pair having a non-­
cular disorders [7–9]. spliced upstream splice acceptor and the contrib-
uting RNA are covalently closed. The second
one, named “exon skipping” mechanism, involves
1.1 Basic Characteristics a splicing event within loop structures (lariats)
of Circular RNAs formed from the process that consists of avoiding
exons [14].
CircRNAs are comprised mainly by sequences Introns close to backsplice sites tend to be
coming exonic or intronic regions having 5′ and larger than regular introns; nevertheless, contigu-
3′ ends covalently closed as a result of a backs- ous introns can be lesser than average [15].
plicing event [10]. This event occurs between a Furthermore, the size of an exon appears to be
splice donor site followed by a preceding splice related also with the circularization process as it
acceptor site, whereas in a conventional linear was previously described that the average size of
splicing, this happens preferably to a rearward exons which constitute a circRNA for their own
acceptor. has a mean size three times longer in comparison
with all expressed exons [3]. It has been also
2 RNA sequencing and Prediction Tools for Circular RNAs Analysis 19

described the existence of paired Alu repeats linked with INK4/ARF expression levels. This
(repetitive sequences typical in the human DNA fact has been described as a risk factor for athero-
and specific for primates) localized close to back- sclerosis disease [19]. Finally, the circular RNA
splice regions, where there is an elevated occur- ciRS-7 which functions as a sponge, is expressed
rence of human exonic circRNA generation [16]. in neuronal tissues and disposes more than 70
In such a way, longer exons than average, sur- binding sites for miR-7 [10]. This microRNA is
rounded by small introns having reversed Alu drastically repressed in patients exhibiting spo-
repeats, appear to be main features present in the radic Alzheimer disease [9].
RNA circularization mechanism.
1.3.2 Gene Expression Regulation
Nowadays, most of the identified circRNAs is
1.3 Putative Roles of CircRNAs demonstrated to be derived from exons, even
though there are some intron-containing circular
The function of circular RNAs still remains RNAs (named ciRNAs). They are characterized
unclear. Various analyses have identified numer- by their constrained expression in the nucleus
ous exonic circRNAs having conserved circular- [20], as most linear RNAs having retained introns
ization sites in orthologous exons [15, 16]. This are usually confined in the nucleus of the cells
evolutionary conservation suggests the execution [21]. Several evidences suggest that these ciR-
of important roles in an organism; hence, numer- NAs allow the transcription regulation of genes
ous possible functions have been suggested. in cis. Particularly, they can promote the RNA
Among these plausible functions, we can high- polymerase II (Pol II) transcription activity of
light ongoing functional studies indicating that their host genes. Nevertheless, the causal mecha-
circular RNAs are able to regulate the expression nism remains unclear [15, 22].
of linear mRNA transcripts. Moreover, the generation of circRNAs via cir-
cularization of exons has been suggested to be a
1.3.1 MiRNA Sponges process that competes with the splicing machin-
The vast majority of known circRNAs ery as they perform their function on the same
are enclosed in the cellular cytoplasm [3, 16] as splice-sequence sites. This fact was observed in
they are transported outside the nucleus. It has neural tissue, in which an inverse expression
been demonstrated that circular RNAs contain level was described, being the circRNAs more
abundant miRNA binding sites to interact with. abundantly expressed than their linear counter-
This fact has allowed them to be called “miRNAs parts [22]. Similarly, in brain tissue during the
sponges” as they bind miRNAs, preventing them aging process, there is a raised expression of cir-
from executing their regulatory roles [17]. A cRNAs opposite to the low levels of linear RNAs
well-studied example comes from the Sry gene, [8]. This high level of circRNAs in certain tissues
discovered in 1993, and belonging to the sex-­ sustains the idea that RNA circularization can
responsible region Y. In specific conditions when control gene expression by displacing the canoni-
miR-138 is overexpressed, it coprecipitates with cal splicing of linear RNAs [23].
argonaute 2 (AGO2) and with the Sry circular
transcript due to the presence of 16 binding sites 1.3.3 Interaction with RNA-Binding
for miR-138 within the circRNA sequence. Proteins
Besides, in mouse cells, the expression level of In a similar manner to some other no protein-­
miR-138 is negatively correlated with Sry; there- coding linear RNA transcripts, circRNAs are able
fore, miR-138 expression is reduced while Sry to interact with RNA-binding proteins, for
expression is increased [17]. ANRIL, an antisense instance, AGO [10]. It has been also suggested
RNA from the tumor suppressor INK4 locus, that they could serve as “scaffolding” for RNA-­
contributes in transcription inhibition [18]. The binding proteins interacting with numerous pro-
expression of circular ANRIL RNA is directly teins that increase the stability of the circRNA
20 E. López-Jiménez et al.

transcript [13]. Another example of interaction and with the additional experimental validation
has been shown in Foxo3, a tumor suppressor that they perform, they were able to provide a
gene [24]. Circ-Foxo3 has been implicated in cell strong evidence of the presence of translation
cycle due to its interactions with some proteins associated with circular RNAs. Their conclusions
involved on that pathway, regulating and prevent- pointed out to the presence of a particular
ing an abnormal proliferation. In detail, the cell sequence to allow the translation process in a
division protein kinase 2 (CDK2) and cyclin-­ regular endogenous framework. Even more, they
dependent kinase inhibitor 1 (p21) interact with reported strong evidences of the fact that the
circ-Foxo3 to establish a RNA-protein complex translation of this subset of circRNAs is not by
that reduces cell cycle progression. Consequently, chance and presented these results suggesting a
it produces a cell cycle arrest as a consequence of specific and regulated effect. Simultaneously,
CDK2 and p21 proteins depletion, and the cell is Legnini et al. were able to describe an example of
retained in the G1/S phase. a eukaryotic protein-coding circRNA called circ-­
ZNF609. In this work, they also conclude that the
translation of this RNA is splicing-related and 5′
1.4 Losing the Identity: Could cap-independent. After that, Yang et al. finally
CircRNAs be still considered as revealed that a single m6A motif is sufficient to
Noncoding RNAs? lead translation initiation in human cells.
Furthermore, he also discovered that circRNAs
The possibility of the translation of the circRNAs have an elevated number of m6A sites [26, 28].
emerged more than two decades ago as a conse- Besides, these studies reported strong evidences
quence of the existence of an internal ribosome showing that the translation of these circRNAs is
entry site (IRES) that could allow it [25]. not a random effect and presented their results
Theoretically, if a circRNA owns an IRES and an suggesting that it is a specific fact.
ATG sequence, it would be able to be translated. This discovery could allow us to have a deeper
Chen et al. corroborated that idea using the hepa- knowledge of the possible regulatory functions
titis δ agent by a noncanonical mechanism, and it over gene expression levels that circRNAs could
was thought to have probably been specific for perform. However, more information is needed
some viral agents. After that, Jerk et al. and other about the mechanisms of translation of circRNAs
authors took into consideration the protein-­ to establish this fact as other layer of control for
coding features of numerous ATG-containing genomic regulation. Therefore, with the rapid
exonic circRNAs, but they could not identify any advance of current molecular and sequencing
naturally protein produced from a circRNA [15, techniques and the development of new bioinfor-
16]. matic methodologies, novel functions will be dis-
More recent studies experimentally demon- covered, and some of the actual unresolved
strated the translation into protein of some cir- questions will be determined in the near future.
cRNAs [26, 27]. Pamudurti et al. described a
group of circRNAs that were translated in vivo in
Drosophila melanogaster. They showed that 2 Experimental Methodologies
these circRNAs commonly presented the main for CircRNA Discovery
features related with the translation process, for and Characterization
instance, they encode proteins having specific
domains although are translated in a non-5′ cap 2.1 Sample Treatment
mode. Besides, they shared the start codon with
the hosting RNA. In this work, none of the pro- The presence of circular RNAs is not easy to
cessed sequences or the results that they obtained detect and distinguish from other small RNAs
could separately support the existence of cir- and miRNAs due to their size or mobility proper-
cRNA translation. But in a combinatorial way ties. Nowadays, the most frequently used
2 RNA sequencing and Prediction Tools for Circular RNAs Analysis 21

­ ethodology requires destroying the circularity

m 2.2 Microarrays from CircRNA
of these RNA species that could allow to their Identification
identification, because of the amplification and/
or fragmentation steps performed. The only commercially available circRNA micro-
Some techniques, such as “rapid amplification array for human has been developed by Arraystar
of cDNA ends” (RACE) or poly(A) enrichment company to facilitate the analysis of circular
of the samples for NGS transcriptome studies, RNA data. It is also available for mouse and rat.
cannot be effective in this case due to circRNAs This platform contains a total of 13,617 different
not having neither a defined end nor a free 3′ or human probes, matching the circRNA-specific
5′ that could be modified for allowing to the junctions, and distributed in an 8*15K format
detection. Furthermore, one of the main features platform. These probes were selected from a total
of circRNAs, their generation by a “backsplice” of six different recent studies describing the data-
process, is not exclusive of these species of small sets [10, 15, 16, 29–31].
RNAs, and initial RNA-seq aligners tools elimi- They offer a highly sensitive and specific plat-
nated those sequences. form for circRNA discovery, providing a service
Recently, with the development of new meth- with circular junction sequences, linear RNA
odology that improves the selection of circRNAs digestion by RNase R enzyme (pre-treatment of
during the processing of the samples like the RNA samples), and an efficient circRNA
exonuclease-­enrichment approaches, as well as labeling system.
sequencing of ribosomal RNA (rRNA)-depleted In comparison with the RNA sequencing for
libraries instead of poly(A)-enriched libraries detecting circRNAs, they point out some of the
with longer reads and higher coverage and the specific features of the circular RNA, arguing
generation of novel bioinformatic tools, this that (1) junction-spanning sequences are only a
problem has been sorted out. small portion of the circular RNA compared with
In the study described by Jeck et al. [13], a linear RNA at the similar expression level, so it
new biochemical protocol called Circle-Seq was could be not detected by the conventional RNA-­
introduced. This methodology involves the treat- seq methods [31]; (2) in order to perform a dif-
ment of RNA samples with an exonuclease ferential expression analysis in this type of data,
enzyme (RNAse R). In this way, linear RNAs are not enough numbers of sequences from circular
processed leaving the circRNAs intact. RNA are achieved; and (3) for the detection of
Nevertheless, it has recently been claimed that the presence of circRNAs, only few reads are
resistance to this enzyme alone is not sufficient to needed, whereas for a reliable quantification, a
determine the circularity of a RNA transcript, due greater number of read counts are required.
to the fact that some circRNAs were sensitive to Moreover, the protocol for preparing the RNA
this enzyme. This strategy then could interfere in samples adds a group of exogenous RNA con-
the global selection of the circRNAs within a trols developed by the External RNA Controls
sample, generating a bias, as well as not be able Consortium (ERCC) as spike-in controls.
to completely eliminate other RNA species still Including this in the protocol, RNA amplifica-
resistant to this process. Other studies suggest the tion, labeling, or hybridization procedural effects
employment of additional biochemical proce- can be corrected for obtaining an accurate and
dures for the isolation of circRNAs, like the use reliable result across the samples.
of a 2D (two-dimensional) denaturing polyacryl- One of the disadvantages of using this plat-
amide gel electrophoresis or ribosomal RNA form is the high input of total RNA needed for
(rRNA) depletion and poly(A)-depletion for preprocessing the samples. Depending on the
increasing the amount of circRNAs in sequenc- field of the study, it could be a limiting condition
ing samples [15]. due to the extremely low amount of material
available for each sample (e.g., human patient
22 E. López-Jiménez et al.

samples, early embryonic stages, etc.). However, provoke. These library preparation strategies
Arraystar also offers the use of an amplification have been combined with two main different
step in the preparation of samples with a low-­ approaches to identify the precise candidate junc-
input material, which unfortunately could tion for every circRNA: (1) a large candidate
increase the noise on the results and the cost of junction-based approach, based on existing tran-
the process. script models, and (2) the identification of these
In summary, the use of this platform is highly junctions searching for those sequences able to
recommended in those cases in which a candidate-­ target directly to the genome (as it was previously
based approach could be applied (accepting the applied for spliced alignment algorithms). These
bias that this kind of platform could introduce in methods have been able to identify circRNAs that
the data) for a faster and reliable acquisition of were experimentally confirmed by sequencing,
the data. The use of the RNA sequencing is rec- RNase R protocol, and other different
ommended for the novel discovery of circular techniques.
molecules since only a small number of read Analyzing in depth the combination of these
counts are required, but it is inappropriate for an methodologies, we can observe some different
accurate analysis of differential expression or aspects: the first option, using a large candidate
specific quantification of circRNAs, taking into junction-based approach has obtained reliable
account the current methodological procedures. results and is the fastest one for applying in ribo-
somal RNA-depleted libraries. This selection
generates a bias in the results that is adding the
2.3 RNA Sequencing (RNA-Seq) inconvenient of being unable to detect novel cir-
cRNAs present in the samples. Also, this strategy
2.3.1 Genomic Detection doesn’t provide evidences of circularity of the
and Isolation Methods species detected. Within this option, there are two
In the last two decades, important changes have different ways to perform this approach: (a)
occurred in the scenario of the circular RNA applying an RNase R enrichment that eliminates
genome-wide studies. The methodology for dis- the linear RNA species and (b) without RNase R
criminating the different RNA species has been pre-treatment, in which a 75 nt paired-end
improved since the discovery of the intrinsic sequencing has to be performed and the pairs of
characteristics of circRNAs (circularity, absent of reads, containing one of the read the splice junc-
3′ or 5′ ends, non-polyadenylated 3′, cytoplasmic tion, will be divided in a group in which the
location, etc.) in several previous studies [3, 13, paired read without the subsequent splice was
16, 32, 33]. aligned to a coding exonic region amid back-
Based on these features, recent high-­ spliced exons (which can arise and be explained
throughput studies have been performed using a by circRNAs) and a group in which the pair mate
deeper sequencing with longer reads strategy that is located in an exon, separated of the backspliced
allows the detection of circular RNAs. Moreover, exons (which are considered as artifacts of
there was an improvement in the algorithms used sequencing). After the sequencing, statistics has
for mapping these reads appropriately, and ribo- to be performed with these two groups of reads in
somal RNA depletion was used in order to order to calculate an accurate score for every sin-
enhance the sequencing of non-polyadenylated gle junction detected. It has the advantage of pro-
RNA species. viding a false discovery rate (FDR) cutoff instead
Focusing in the library preparation for of a random read depth-based threshold.
sequencing, researchers have developed and A combined approach employing a qPCR
compared different strategies for enriching the assay and the addition of RNase exonuclease was
libraries in circRNAs, trying to avoid the noise of used to validate those newly discovered cir-
the presence of other similar RNA species or the cRNAs. This first methodology showed the resis-
bias that the elimination of some of them could tance to RNAse R enzyme of the transcripts, and
2 RNA sequencing and Prediction Tools for Circular RNAs Analysis 23

other properties of backsplice-containing linear these lariat RNAs are easily distinguishable from
RNA were missed. circRNAs, their branch region sequence resem-
The second option consists in using rRNA-­ bles backsplice read in these parts of the sequence,
depleted and RNase R-treated libraries for high-­ and besides they are also disordered in compari-
throughput sequencing, and after that, mapping son with their genomic annotation.
reads directly to de novo genomic positions and Although the Circle-Seq protocol could gen-
discovers backspliced reads in specific sequences. erate a high depth of circular and lariat products,
This method avoids the bias of a candidate-based it has some limitations, for instance, as we com-
approach, allowing the identification of novel cir- mented before regarding the microarrays plat-
cRNAs. This method consists selecting the reads form, for performing Circle-Seq, a higher amount
that could not be directly aligned to the genome of total RNA than in a regular sequencing proto-
and take the two terminations of a single read and col is needed, without any enrichment, which is
map them separately, based on the backsplice more prone to suffer from endonuclease contami-
properties of the sequences (has to be flanked by nation. One important point to keep in mind is
GT/AG splice site in the genome context). This that this process can generate a bias on the results
method is less accurate than a candidate-based due to the possible elimination of longer cir-
approach but allows the detection of unannotated cRNA products, as a single nicking event would
splice junctions. confer exonuclease sensitivity.
This methodological system, which combines However, for detecting backsplicing alterna-
the biochemical properties of the circRNAs for tive events in circRNAs, non-poly(A) (without
improving the library preparation with the enrich- RNase R) or rRNA-depleted (without RNase R)
ment on circular RNA species, and posterior samples are recommended. Both of these meth-
sequencing, was named as Circle-Seq for Jeck ods have discovered different circRNAs that were
et al. [13], and it was described in archaea studies later validated by other alternative detection
[33] and in mammals [16]. This library prepara- methods, as an example sequencing or RNAse R
tion is followed by the application of a mapping testing. This is a key step for the detection of
algorithm called MapSplice [34] that is able to putative circRNAs by bioinformatic algorithms.
identify apparent backsplice sequences. Depleting highly expressed RNA species with
In this strategy, they performed an RNase R different splicing patterns such as ribosomal
treatment of the samples before the library prepa- RNAs or linear mRNAs favors the identification
ration. For mammalian samples, the rRNA deple- and quantification of the expression of circular
tion step is required, but not in archaea. This RNAs (Fig. 2.1).
technique is based on the use of two exonic cir-
cRNA features for the identification: (1) the
inclusion of the backsplice junction reads, by 2.4 CircRNA Validation
means of a segmented mapping approach; and (2)
the samples have been pre-treated with a step of The identification and validation of circRNAs are
RNase R, eliminating the linear RNA species and required from several specific methods based on
enriching in circular ones (in comparison with biochemical approaches. One of the most basic
the mock-treated control). An example of the tools that can be used for validating them is the
data obtained using this methodology was exhib- reverse transcription PCR (RT-qPCR). Following
ited in the study that described cANRIL [16], the these assays, as the cDNA is going to arise from
circular RNA from the ANRIL gene. the circRNA, the sequence should comprehend
The use of the RNase R treatment doesn’t the “exon junctional” region which is not present
avoid the presence of lariat RNAs (circRNAs in the canonical spliced mRNA. To achieve this
mostly intronic that were formed during the goal, primers have to be designed to detect and
canonical RNA splicing and possess a 2′–5′ car- amplify this indicative region. The specific design
bon linkage at the splicing fork site). Although of these primers called inverse or outward-facing
24 E. López-Jiménez et al.

Fig. 2.1 Schematic workflow for sample preparation. (a) ration and enrichment of circular RNAs (with or without
Schematic workflow of the process for a microarray circu- RNAse R treatment) for sequencing processing and sub-
lar RNA samples processing and data acquisition. (b) sequently data analysis
Schedule of different options for total RNA sample prepa-

primers prevents the alignment and amplification ity of the different species) to confirm the results
of other RNA (such as mRNA) species or DNA in a qualitative manner.
containing the diagnostic sequence. This is a In situ hybridization (ISH) techniques allow to
quantitative approach that can be used in order to confirm the presence of circRNAs in addition to
obtain the relative abundance of circRNAs in a identify the specific expression patterns of both
biological context. RT-qPCR could also have cells and explicit tissues. To do this, specific
artifacts and biases [35], so the result should be probes are included, which are able to bind back-
confirmed by means of sequencing the PCR spliced junction sites [38]. In contrast, specific
products to check the presence of the junctional exonic sequences only existent in mRNAs are
sequence [36]. Even though, it is possible to used as a control of the presence of the counter-
detect other species in the sequencing that make partying linear transcripts expression, belonging
us aware of the level of noise of the experimental to the same locus as circRNA.
procedure for such specific condition. Less frequent system to check the circularity
Other important molecular technique for the and validate circRNA presence involves the use
validation is Northern blot [37]. The probes have of RNase H (an endoribonuclease protein able to
to be designed for targeting the circular sequence cleave RNA and RNA-DNA double strands)
or the specific junction sequence, and additional [37]. This method is based on the different bind-
probes for the same circular RNA can be used in ing patterns that two short DNA probes generate
individual blottings for ensuring the presence of a when they bind with the RNA of interest in the
specific circRNA. This is a very simple and spe- presence of each probe separately. In addition, a
cific procedure (because it is based on the mobil- different migratory pattern can be observed
2 RNA sequencing and Prediction Tools for Circular RNAs Analysis 25

depending if the species are circular or linear, Therefore, several algorithms have been
because they exhibit a different behavior in a already designed; thus, a vast amount of cir-
polyacrylamide gel. cRNAs resulting from exonic, intergenic,
Two-dimensional denaturing polyacrylamide intronic, and UTR has been identified [3, 10, 41]
gel electrophoresis (2D gel) can be also used to and deposited in specialized databases such as
discriminate among linear and circular RNAs due circBase [42] or CIRCpedia [43].
to the different migratory patterns of both types Currently, there are several algorithms able to
of molecules. The specific pattern in 2D gel of process RNA-seq samples in order to identify cir-
the linear RNA is along the diagonal trajectory in cular RNAs. Most of them have been bench-
the gel (and depending on the size), and circular marked recently [44, 45], and the results are quite
RNA exhibits an accurate pattern. The gel trap comparable. Besides, they conclude that although
method is other possibility for being used in particular methods perform better than others, the
RNAse R-treated samples [39]. Gel-trapping highest percentage of true positives is obtained
technique holds the pool enriched for circRNAs when results are combined from various methods
in the well of an electrophoresis gel, and at the and removing those circRNAs that do not appear
same time, linear RNAs migrate away. The result in at least two different methods [44]. Because of
could be directly extracted and sequenced using this relevant conclusion and the rapid develop-
NSG technology [40]. ment of new prediction tools, we will present
The emerging results related to the circular diverse softwares and discuss in detail their
transcriptome due to the revolution of the advantages and disadvantages (Table 2.1).
sequencing techniques in the last years come to The majority of algorithms responsible for the
light the need of an effective method for validat- identification of circular RNAs are divided in two
ing in silico results and predictions. None of the different types with regard to the implemented
strategies explained before alone have the com- methodology to discover circRNAs. The first
plete accuracy for ensuring the validation of the group of programs is based on a “pseudo-­
results. In that sense, a combinatorial validation reference” approach; briefly, they build a putative
strategy should be taken in consideration. circRNA sequence reference from a gene annota-
tion repository, to subsequently identify junction-­
spanning reads. The other strategy is called
3 Computational Predictions “segmented-based” and relies on the identifica-
of CircRNAs tion of backsplicing junctions from the mapping
information provided by aligning reads to the ref-
As it was mentioned, circRNAs are distinguished erence genome or transcriptome (Fig. 2.2).
by a “backspliced” process that occurs between a Among the pseudo-reference algorithms, we
splice donor site and an upstream splice acceptor highlight KNIFE [46] and PTESFinder [47].
site. Therefore, the identification of circRNA
presence requires the development of specific
bioinformatic tools to detect junction-spanning 3.1 KNIFE
sequences that reveal this backspliced from tran-
scriptome deep-sequencing samples [10, 16]. KNIFE [46] starts by mapping independently
The introduction of this high-throughput next-­ each paired-end read to the genome, ribosomal
generation sequencing technology and an accu- RNA sequences, lineal or scrambled exon-exon
rate protocol to reduce lineal mRNA (RNase R or junction indexes, using Bowtie2 [48]. It rejects
non-polyadenylated procedures) has improved potential backspliced junction sequences if they
the description of numerous circular RNAs in dif- also align with abnormal scores to lineal and
ferent organisms. scramble junction sequences. Those reads are
26 E. López-Jiménez et al.

Table 2.1 CircRNA prediction tools

Tool name Category Aligners References
KNIFE Pseudo-­reference Bowtie 1 and Bowtie 2 [46]
PTESFinder Pseudo-­reference Bowtie 1 and Bowtie 2 [47]
MapSplice Segmented-­based Bowtie 1 [34]
CIRCexplorer Segmented-­based TopHat and TopHat-­fusion (Bowtie 1 and Bowtie 2), STAR [30, 43]
CIRI Segmented-­based BWA-­MEM [50]
Acfs Pseudo and BWA-­MEM [51]
segmented-­based
List of well-recognized circRNA prediction tools. They are organized according to a category (pseudo-reference, junc-
tion-spanning reads from potential circular RNAs are used to build a putative circRNA sequence which will be used as
a reference; segmented-base, reads are aligned against a reference genome/transcriptome, and short segments from
those reads are inspected to find backsplicing junctions and appropriate mapping signals supporting circRNA struc-
tures). The table also shows the mapper utility needed for each prediction tool and the research article that includes
further information.

Fig. 2.2 Classification of circular RNA prediction tools gene annotation information. Reads are studied under this
according to the strategy. RNA samples of interest are reference to identify junction-spanning reads. (b) The
sequenced. Those RNA-seq results can be processed other strategy is called “segmented-based” and relies on
for different methods; these can be classified into two the identification of backsplicing junctions from the map-
groups: (a) the first group is composed of methods that ping information provided by aligning reads to the refer-
rely on a “pseudo-reference” approach based on the gen- ence genome or transcriptome
eration of a putative circRNA sequence reference using
2 RNA sequencing and Prediction Tools for Circular RNAs Analysis 27

considered false positives and used to model all An important improvement included in this
false positives according to two classes: real method is that it allows a “guided” evaluation by
alignments (mapped to lineal mRNAs) or arti- providing previously discovered PTES struc-
facts, when the paired-end reads alignment orien- tures, hence avoiding in this way to perform the
tations are not coherent with neither a linear nor finding phase again. However, one of the weak-
a circular RNA (named “decoy” alignments). The nesses is that it does not use the information
statistic model is based in a generalized model obtained from paired-end mapping (PE) since it
(GLM) focused on alignments scores, mapping interferes with the filtering phase affecting the
quality, and offset position (category 1 or 2, specificity.
“decoy” alignment). Among the “segmented-based” strategies, we
One of the advantages of this algorithm is that will emphasize MapSplice [34], CIRCexplorer
they compute a posterior probability for each [30], CIRI [50], and Acfs [51] (which also uses
junction consistent or not, with decoy reads. This a “pseudo-reference” approach as a filtering
approach highlights those reads belonging to a step).
circular RNAs.
Finally, unmapped reads are incorporated in a
de novo algorithm aiming to discover unanno- 3.3 MapSplice
tated splice sites responsible of RNA
circularization. MapSplice [34] is an algorithm for the detection
of backsplice junction sequences which is inde-
pendent of the splice-site information. This
3.2 PTESFinder approach allows to discover novel splicing events
along with noncanonical junctions in any tran-
PTESFinder is a software that identifies putative scriptome sample. Even more, it also identifies
posttranscriptional exon shuffling (PTES) struc- canonical junctions. MapSplice is splitted into
tures from RNA-seq reads [47]. This tool is based two phases, the “tag alignment” step, where
in the assumption that circRNAs are transcripts mRNA tags are mapped to a reference genome.
characterized by the existence of exonic junction The identification of candidate tag alignments is
reads having an incongruous orientation accord- performed, in turn, in three parts: first, tags are
ing to their location in the genome. This program subdivided into successive shorter pieces, which
is divided into three consecutive phases: a dis- are aligned to the reference. In the next stage,
covery phase, an evaluation phase, and a filtering sections lacking from an exonic alignment are
phase. In the first step, short sequences (20 base considered for a spliced alignment method using
pairs by default) of each read end is aligned a splice junction exploration procedure which
against the reference transcriptome using Bowtie includes adjacent pieces previously aligned.
[49]. A pair of sequences from the same read that Finally, in the final step, tag alignments are com-
map to the same gene but in reverse positions bined to trace global candidate alignments for
from their original order in the read sequence is every single tag. Although, tags including splice
potentially recognized as PTES. In the following junctions should involve a gapped alignment that
phase, all initial reads are realigned to newly ought to correspond to a removed intron by the
classified PTES structures using Bowtie2 [48]. splicing machinery in the transcription step. The
Accordingly, it permits to obtain mapping scores second step, called the “splice inference phase,”
from PTES in order to compare with those scores examines splice junctions that appear in the
coming from lineal transcriptomic alignments. alignments of each tag to infer a splice signifi-
This information is used in the filtering stage to cance value based on the quality and variety of
remove presumably false positives having higher the alignments. The goal of this step is to help
scores using genomic or transcriptomic align- with the selection process of the most reliable
ments rather than PTES mapping. alignments for each tag, based on a mixture of
28 E. López-Jiménez et al.

quality alignment values and implication of the 3.5 CIRI

splicing event and based on that criteria, and dis-
card spurious sequences. CIRI [50] uses the underlying strategy employed
by BWA-MEM [55] which incorporates a local
alignment with an affine-gap algorithm by
3.4 CIRCexplorer ­seeding with a maximal exact match, typical of
spanning junction reads in circular RNAs. This
CIRCexplorer [30, 43] is a tool capable of identi- tool uses a method which relies on paired chiastic
fying alternative backsplice and canonical splice clipping (PCC) signal recognition in BWA
junctions from single and paired-end reads. This aligned files, combined with various filtering
software uses TopHat [52] coupled with TopHat-­ phases aimed to eliminate false positives.
Fusion [53], although it optionally supports mul- Therefore, it collects PCC signals from aligned
tiple circular RNA aligners such as STAR [54], files which support the backsplicing junctions
BWA, [55] or segemehl [56]. Their approach is a and the proper paired-end mapping (PEM) pro-
two-step mapping strategy, where reads are first files coherent with circRNA structures. CIRI
mapped against the reference sequence genome employs PEM information if existing, for an ini-
and later, nonaligned sequences are remapped tial filter of spurious PCC signals. According to
using the TopHat-Fusion utility. New reads the authors, as two segments of a reliable junc-
extracted from the fusion alignment in a non-­ tion sequence theoretically point out to the ends
collinear order coming from the same chromo- where all circRNA reads aligned, a probable
some are remapped to a combination of (novel or junction indicates a circRNA if its paired mate
known) gene annotation to conclude the exact read is mapped within the putative circular RNA
location of backsplice sites from reliable cir- region when supplied by the fragments of the
cRNA structures. In the upgraded CIRCexplorer junction site. Subsequently, it searches for those
version [43], sequences aligned to the reference junction sites with known GT/AG splice signals.
genome and collinear exon junction reads are Finally, it clusters the unbalanced junction reads
now studied further in a de novo assembly, which by employing a dynamic alignment methodology
permits to track down novel exons and therefore to remove putative false-positive junctions result-
new splicing processes. Interestingly, in accor- ing from repetitive or homologous regions.
dance with the authors and in comparison with Furthermore, if an accurate gene annotation is
their linear equivalent RNAs, circRNAs seem to available for the organism of interest, it could be
present different alternative splicing and backs- of interest to expand the search to other possible
plicing patterns. contiguous exon boundaries splice-site signals.
This program has many advantages, for
instance, it is capable of working with single and
paired-end reads. Besides, it also permits the 3.6 ACFS
usage of different kinds of aligners which allows
a very detailed study of circular RNAs based on Acfs [51], similar to CIRI, employs the BWA-­
different mappers. In addition, it is able to predict MEM mapper to identify and quantify the abun-
circRNAs with high reliability while allowing the dance of circRNAs from single or paired-ended
study of splicing patterns to identify new tran- reads, although this tool is mainly designed to
scripts and exons. Among the disadvantages we determine backsplice junctions from single-end
can indicate that since each aligner requires dif- transcriptome data. Acfs methodology is divided
ferent indexes, specific parameters and inputs in three different steps: preprocessing phase,
files, it is a package indicated for researchers identification phase, and quantification phase.
with high knowledge in the field. In the first part, reads are mapped to the genome.
2 RNA sequencing and Prediction Tools for Circular RNAs Analysis 29

If paired-end reads are included, they are pro- 3.7 MiARma-Seq

cessed and treated as single-end sequences. In the
second step, potentially originated reads from the MiARma-Seq [58] (miRNA-Seq And RNA-Seq
backsplice junctions are scrutinized by selecting Multiprocess Analysis) is a tool designed to find
those that align in a genomic position on the mRNAs, miRNAs, and circRNAs, as well as for
same chromosome and strand. Subsequently, differential expression, target prediction (using
Acfs inspects the strength of each backsplice the miRGate database [59] and its application
junction alignment using a maximum entropy programming interface [60]), and functional
model [57] in order to identify the exact genomic analysis in transcriptome samples. This software
position. In the latest phase, Acfs implements a intends to reduce some of the principal difficul-
supplementary alignment approach to precisely ties that researchers may face when analyzing
measure the abundance of the inferred circular next-generation sequencing data such as (1) easy
RNAs. So, for each potential circRNA, a pseudo- installation, removing third-party requisites that
circular reference is generated; then, the tool make the installation and configuration hard for
aligns each sequence to the genome, and before researchers with little experience in the field; (2)
reporting the circRNA expression level, it speed of execution, allowing analysis in a stan-
examines alignments spanning the backsplice
­ dard computer or in a high-performance comput-
junctions. ing infrastructure (taking advantage of its
This software is also capable of identifying parallelization); and (3) consistency, as the pipe-
circRNAs originating from noncanonical gene line includes most of the standard tools available,
structure such as fusion genes, besides from the in order to perform all calculations.
detection of circRNAs derived from regular MiARma asks for a configuration file with
genes. In that sense, as the splicing machinery is general information about the experiment, and
involved in the generation of circular RNAs, it upon request, it can perform a quality step
would be probable that fusion gene loci could using FastQC [61] and a trimming phase using
also produce these circular molecules. cutadapt [62] or kraken [63]. Subsequently, the
Most of the programs discussed here, with the user can perform an identification study of cir-
exception of Acfs, have recently been bench- cRNAs based on different methods that miARma-­
marked using RNase R and poly(A)-depleted seq includes, such as CIRI version 1.x and 2.x
samples to measure the level of false positives [50], KNIFE [46], CIRCexplorer [30, 43], and
[44, 45]. However, it has been claimed that resis- PTESFinder [47]. The possibility of using up to
tance to the RNase R enzyme as a unique factor four different circRNA prediction methods in a
cannot be employed to conclude whether an single package without any prerequisite installa-
mRNA is circular or not, since it was observed tion makes miARma one of the easiest tools in
that some circRNAs were susceptible to exonu- the identification of circular RNAs field.
clease degradation [16, 46]. Nevertheless, Interestingly, this pipeline offers the possibility
KNIFE, CIRI, and CIRCexplorer achieved higher of carrying out a differential expression analysis
values of precision and sensitivity compared to (using edgeR [64] or NOISeq [65]) among two
other tools. As these studies highlight, given that conditions. In that sense, it could be possible, for
the highest percentage of true positives is instance, to compare transcriptome data from
achieved from the intersection of predictions healthy controls and patient samples and corre-
resulting from two different methods, here we late the presence/absence of circular RNAs with
present the new version of the miARma-seq pipe- the onset or the prognosis of the disease. But
line, which is, as far as we know, the only one also, it could be useful to compare RNase
framework that includes several circRNA predic- R-treated samples against untreated samples,
tion tools in a single bundle. given that this approach would allow to assess the
30 E. López-Jiménez et al.

Fig. 2.3 Overview of the integrated tools available in available modules. Hence, users can perform a complete
miARma for circRNA identification and quantification. analysis, from raw reads to circRNA identification, quan-
This workflow illustrates all software included in tification, and differential expression
miARma-seq (only related to circRNA study) and all

capacity of the method to accurately identify cir- databases such as CIRCpedia [43] or circBase
cRNAs. An overview of the integrated tools and [42]. The existence of this circRNA information,
the all-possible workflows available in miARma many of them experimentally validated, has
for circRNA identification is shown in detail in allowed the collection of a bona fide dataset to
Fig. 2.3. facilitate the development of new stringent and
miARma-Seq is freely available at http:// reliable bioinformatics tools.
miarmaseq.com along with a complete documen- Currently, CIRCpedia contains circRNA
tation and diverse examples of usage. backsplicing and alternative splicing data from
13 human cell lines, tissue, and species samples.
This database provides query support by gene
3.8  ircRNA Databases (CircBase
C names and includes a helpful table with genomic
and CIRCpedia) coordinates, circRNA accession names, host
gene names, relative expression values, and alter-
To date, numerous researchers have reported an native (back)-splicing sequences from circRNAs
elevated number of circular RNAs (circRNAs) along with exon identity. Links are also offered to
expressed in diverse organisms and under differ- download all the information for additional
ent conditions. All information coming from studies.
these RNAs along with the identified alternative On the contrary, circBase [42] contains identi-
splicing or backsplicing events, and newly dis- fied circRNAs in six organisms: human, mouse,
covered exons, are available in the specialized C. elegans, D. melanogaster, L. chalumnae, and
2 RNA sequencing and Prediction Tools for Circular RNAs Analysis 31

L. menadoensis. For each of these species, the As a summary, the identification and charac-
repository stores accurate material from different terization of new circRNAs, together with the
tissues, organs, and cell lines and allows explor- modification of protocols, such as Circle-Seq, to
ing public circRNA datasets and downloading increase the presence of these circular molecules
the scripts needed to discover and annotate your versus their linear counterparts, can help the
own circRNAs. development of more reliable tools. Without any
doubt, this will respond unanswered questions to
allow the use of the circRNAs as disease
4 Future Perspectives biomarker.

The massive generation of high-throughput data Acknowledgments We wish to thank the No Surrender
by means of the new sequencing technologies Cancer Trust for supporting the position and projects of
ELJ at Imperial College London.
unveils the lack of an efficient and accurate
methodology for the isolation and validation of Competing Financial Interests The authors declare no
circRNAs. Here we highlight that any of the competing financial interests.
existent methodologies is enough to assort the
presence of this RNA species; they should be
used in a combinatorial way to analyze them. It
could be essential to standardize the procedure References
for detecting the circRNA using only one method
that allows us to isolate and enrich only the cir- 1. Nigro JM, Cho KR, Fearon ER et al (1991) Scrambled
cular RNA fraction within the RNA, in a simple exons. Cell 64(3):607–613
2. Cocquerelle C, Mascrez B, Hetuin D et al (1993)
and accurate way, avoiding the bias that other Mis-splicing yields circular RNA molecules. FASEB
techniques could generate in different databases J 7(1):155–160
in order to be able to compare them. Thus, novel 3. Salzman J, Gawad C, Wang PL et al (2012) Circular
RNA-seq protocols and newly developed bioin- RNAs are the predominant transcript isoform from
hundreds of human genes in diverse cell types. PLoS
formatics methodologies have emerged as a One 7(2):e30733
source of discovery for thousands of circRNAs 4. Abu N, Jamal R (2016) Circular RNAs as promising
in diverse organisms. Although dozen of projects biomarkers: a mini-review. Front Physiol 7:355
based on circRNAs are published daily, several 5. Morris KV, Mattick JS (2014) The rise of regulatory
RNA. Nat Rev Genet 15(6):423–437
questions related to biogenesis and function of 6. Lasda E, Parker R (2014) Circular RNAs: diversity of
circRNA remain uncovered. Some research in form and function. RNA 20(12):1829–1842
cancer has revealed how the identification of 7. Chen LL (2016) The biogenesis and emerging roles of
upregulated and downregulated circRNA levels circular RNAs. Nat Rev Mol Cell Biol 17(4):205–211
8. Gruner H, Cortes-Lopez M, Cooper DA et al (2016)
can be employed as a biomarker for specific CircRNA accumulation in the aging mouse brain. Sci
types of cancer like in the case of laryngeal can- Rep 6:38907
cer [66]. Furthermore, the content of exosomes, 9. Lukiw WJ (2013) Circular RNA (circRNA) in
which are highly enriched in circular RNAs, is Alzheimer’s disease (AD). Front Genet 4:307
10. Memczak S, Jens M, Elefsinioti A et al (2013)
altered by the generation and progression of can- Circular RNAs are a large class of animal RNAs with
cer, as well as neurodegenerative and infectious regulatory potency. Nature 495(7441):333–338
diseases [67]. The recent discovery of the trans- 11. Umekage S, Uehara T, Yoshinobu F et al (2012) In
lation of a subset of circRNAs in eukaryotes vivo circular RNA expression by the permuted intron-­
exon method. Innov Biotechnol 76:1
opens the door to a more direct and specific 12. Ashwal-Fluss R, Meyer M, Pamudurti NR et al (2014)
functionality of the circular RNAs mediated by circRNA biogenesis competes with pre-mRNA splic-
protein production [26–28], and their possible ing. Mol Cell 56(1):55–66
implication in human diseases has been pointed 13. Jeck WR, Sharpless NE (2014) Detecting and charac-
terizing circular RNAs. Nat Biotechnol 32(5):453–461
out in several studies [68].
32 E. López-Jiménez et al.

14. Zaphiropoulos PG (1996) Circular RNAs from tran- cision in quantitative transcript expression profiling.
scripts of the rat cytochrome P450 2C24 gene: cor- Bioinformatics 27(13):i383–i391
relation with exon skipping. Proc Natl Acad Sci U S 32. Suzuki H, Zuo Y, Wang J et al (2006) Characterization
A 93(13):6536–6541 of RNase R-digested cellular RNA source that con-
15. Salzman J, Chen RE, Olsen MN et al (2013) Cell-type sists of lariat and circular RNAs from pre-mRNA
specific features of circular RNA expression. PLoS splicing. Nucleic Acids Res 34(8):e63
Genet 9(9):e1003777 33. Danan M, Schwartz S, Edelheit S et al (2012)
16. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular Transcriptome-wide discovery of circular RNAs in
RNAs are abundant, conserved, and associated with Archaea. Nucleic Acids Res 40(7):3131–3142
ALU repeats. RNA 19(2):141–157 34. Wang K, Singh D, Zeng Z et al (2010) MapSplice:
17. Hansen TB, Jensen TI, Clausen BH et al (2013) accurate mapping of RNA-seq reads for splice junc-
Natural RNA circles function as efficient microRNA tion discovery. Nucleic Acids Res 38(18):e178
sponges. Nature 495(7441):384–388 35. Roy CK, Olson S, Graveley BR et al (2015) Assessing
18. Yap KL, Li S, Munoz-Cabello AM et al (2010) long-distance RNA sequence connectivity via RNA-­
Molecular interplay of the noncoding RNA ANRIL templated DNA-DNA ligation. elife 4:e03700
and methylated histone H3 lysine 27 by polycomb 36. Cocquet J, Chong A, Zhang G et al (2006) Reverse
CBX7 in transcriptional silencing of INK4a. Mol Cell transcriptase template switching and false alternative
38(5):662–674 transcripts. Genomics 88(1):127–131
19. Burd CE, Jeck WR, Liu Y et al (2010) Expression 37. Capel B, Swain A, Nicolis S et al (1993) Circular
of linear and novel circular forms of an INK4/ARF-­ transcripts of the testis-determining gene Sry in adult
associated non-coding RNA correlates with athero- mouse testis. Cell 73(5):1019–1030
sclerosis risk. PLoS Genet 6(12):e1001233 38. Braissant O, Wahli W (1998) A simplified in situ
20. Li Z, Huang C, Bao C et al (2015) Exon-intron cir- hybridization protocol using non-radioactively
cular RNAs regulate transcription in the nucleus. Nat labeled probes to detect abundant and rare mRNAs on
Struct Mol Biol 22(3):256–264 tissue sections. Biochemica 1:6
21. Braunschweig U, Barbosa-Morais NL, Pan Q et al 39. Schindler CW, Krolewski JJ, Rush MG (1982)
(2014) Widespread intron retention in mammals Selective trapping of circular double-stranded
functionally tunes transcriptomes. Genome Res DNA molecules in solidifying agarose. Plasmid
24(11):1774–1786 7(3):263–270
22. Rybak-Wolf A, Stottmeister C, Glazar P et al (2015) 40. Awan AR, Manfredo A, Pleiss JA (2013) Lariat
Circular RNAs in the mammalian brain are highly sequencing in a unicellular yeast identifies regulated
abundant, conserved, and dynamically expressed. alternative splicing of exons that are evolutionarily
Mol Cell 58(5):870–885 conserved with humans. Proc Natl Acad Sci U S A
23. Huang S, Yang B, Chen BJ et al (2017) The emerg- 110(31):12762–12767
ing role of circular RNAs in transcriptome regulation. 41. Chen BJ, Mills JD, Takenaka K et al (2016)
Genomics 109(5–6):401–407 Characterization of circular RNAs landscape
24. Du WW, Yang W, Liu E et al (2016) Foxo3 circular in multiple system atrophy brain. J Neurochem
RNA retards cell cycle progression via forming ter- 139(3):485–496
nary complexes with p21 and CDK2. Nucleic Acids 42. Glazar P, Papavasileiou P, Rajewsky N (2014)
Res 44(6):2846–2858 circBase: a database for circular RNAs. RNA
25. Chen CY, Sarnow P (1995) Initiation of protein syn- 20(11):1666–1670
thesis by the eukaryotic translational apparatus on cir- 43. Zhang XO, Dong R, Zhang Y et al (2016) Diverse
cular RNAs. Science 268(5209):415–417 alternative back-splicing and alternative splic-
26. Legnini I, Di Timoteo G, Rossi F et al (2017) Circ-­ ing landscape of circular RNAs. Genome Res
ZNF609 is a circular RNA that can be translated and 26(9):1277–1287
functions in myogenesis. Mol Cell 66(1):22–37 e29 44. Hansen TB, Veno MT, Damgaard CK et al (2016)
27. Pamudurti NR, Bartok O, Jens M et al (2017) Comparison of circular RNA prediction tools. Nucleic
Translation of CircRNAs. Mol Cell 66(1):9–21 e27 Acids Res 44(6):e58
28. Yang Y, Fan X, Mao M et al (2017) Extensive 45. Zeng X, Lin W, Guo M et al (2017) A comprehensive
translation of circular RNAs driven by N(6)- overview and evaluation of circular RNA detection
methyladenosine. Cell Res 27(5):626–641 tools. PLoS Comput Biol 13(6):e1005420
29. Guo JU, Agarwal V, Guo H et al (2014) Expanded 46. Szabo L, Morey R, Palpant NJ et al (2015) Statistically
identification and characterization of mammalian cir- based splicing detection reveals neural enrichment
cular RNAs. Genome Biol 15(7):409 and tissue-specific induction of circular RNA during
30. Zhang XO, Wang HB, Zhang Y et al (2014) human fetal development. Genome Biol 16:126
Complementary sequence-mediated exon circulariza- 47. Izuogu OG, Alhasan AA, Alafghani HM et al (2016)
tion. Cell 159(1):134–147 PTESFinder: a computational method to identify
31. Labaj PP, Leparc GG, Linggi BE et al (2011) post-transcriptional exon shuffling (PTES) events.
Characterization and improvement of RNA-Seq pre- BMC Bioinform 17:31
2 RNA sequencing and Prediction Tools for Circular RNAs Analysis 33

48. Langmead B, Salzberg SL (2012) Fast gapped-read 59. Andres-Leon E, Gonzalez Pena D, Gomez-Lopez G
alignment with bowtie 2. Nat Methods 9(4):357–359 et al (2015) miRGate: a curated database of human,
49. Langmead B, Trapnell C, Pop M et al (2009) mouse and rat miRNA-mRNA targets. Database
Ultrafast and memory-efficient alignment of short (Oxford) 2015:bav035
DNA sequences to the human genome. Genome Biol 60. Andres-Leon E, Gomez-Lopez G, Pisano DG (2017)
10(3):R25 Prediction of miRNA-mRNA interactions using miR-
50. Gao Y, Wang J, Zhao F (2015) CIRI: an efficient and Gate. Methods Mol Biol 1580:225–237
unbiased algorithm for de novo circular RNA identifi- 61. Anders S (2010) FastQC: a quality control tool for
cation. Genome Biol 16:4 high throughput sequence data. Available online at:
51. You X, Conrad TO (2016) Acfs: accurate circRNA http://www.bioinformatics.babraham.ac.uk/projects/
identification and quantification from RNA-Seq data. fastqc
Sci Rep 6:38820 62. Martin M (2011) Cutadapt removes adapter sequences
52. Trapnell C, Pachter L, Salzberg SL (2009) TopHat: from high-throughput sequencing reads. EMBnet
discovering splice junctions with RNA-Seq. J 17:1
Bioinformatics 25(9):1105–1111 63. Davis MP, van Dongen S, Abreu-Goodger C et al
53. Kim D, Salzberg SL (2011) TopHat-fusion: an (2013) Kraken: a set of tools for quality control and
algorithm for discovery of novel fusion transcripts. analysis of high-throughput sequence data. Methods
Genome Biol 12(8):R72 63(1):41–49
54. Dobin A, Davis CA, Schlesinger F et al (2013) STAR: 64. Robinson MD, McCarthy DJ, Smyth GK (2010)
ultrafast universal RNA-seq aligner. Bioinformatics edgeR: a Bioconductor package for differential
29(1):15–21 expression analysis of digital gene expression data.
55. Li H, Durbin R (2009) Fast and accurate short Bioinformatics 26(1):139–140
read alignment with Burrows-Wheeler transform. 65. Tarazona S, Garcia-Alcalde F, Dopazo J et al (2011)
Bioinformatics 25(14):1754–1760 Differential expression in RNA-seq: a matter of depth.
56. Otto C, Stadler PF, Hoffmann S (2014) Lacking align- Genome Res 21(12):2213–2223
ments? The next-generation sequencing mapper sege- 66. Xuan L, Qu L, Zhou H et al (2016) Circular RNA: a
mehl revisited. Bioinformatics 30(13):1837–1843 novel biomarker for progressive laryngeal cancer. Am
57. Yeo G, Burge CB (2004) Maximum entropy J Transl Res 8(2):932–939
modeling of short sequence motifs with appli-
­ 67. Lin J, Li J, Huang B et al (2015) Exosomes: novel bio-
cations to RNA splicing signals. J Comput Biol markers for clinical diagnosis. ScientificWorldJournal
11(2–3):377–394 2015:657086
58. Andres-Leon E, Nunez-Torres R, Rojas AM (2016) 68. Lyu D, Huang S (2017) The emerging role and clini-
miARma-Seq: a comprehensive tool for miRNA, cal implication of human exonic circular RNA. RNA
mRNA and circRNA analysis. Sci Rep 6:25749 Biol 14(8):1000–1006
 Online Databases and Circular
RNAs 3
Seyed Hamid Aghaee-Bakhtiari

Abstract CircInteractome, CircNet, Circ2Traits,

circRNAs are a novel class of ncRNAs that CircR2Disease, TCSD, and CSCD. In this
unlike other ncRNAs are not linear and have a chapter, we have an overview on these main
circular structure. These valuable ncRNAs circRNA databases and introduce key features
have been detected in a wide range of organ- of each database.
isms from plants to animals and in all cell
lines and tissues. Commonly, circRNAs have Keywords
several functions as gene expression regula- circRNA · Online databases · Web-accessible
tion at transcriptional or posttranscriptional databases · Online resources
level, miRNA partnership, and splicing inter-
cede. Currently, circRNAs are roughly
remarked in a widespread collection of dis-
eases, and circRNAs simply can be recog- 1 Introduction
nized in liquid samples for disease detection
and progression assessment. Considering Noncoding transcriptome comprises a collection
these features of circRNAs, these molecules of RNAs that have many controlling and funda-
are evolving the impeccable collection of mental functions [1]. MicroRNAs (miRNAs) and
original biomarkers for disease therapy and circRNAs are two classes of noncoding RNAs
diagnosis. As the critical role of these mole- that their significance is completely demonstrated
cules in different aspects medicine and biol- in numerous biological processes and several dis-
ogy, circRNAs are considered as key and eases [2, 3]. miRNAs are small endogenous
critical class of ncRNA in the current ncRNA RNAs that regulate gene expression, and numer-
search field. To simplify the assessment of ous studies have established the important role of
diverse features of circRNAs, several data- miRNAs in usual actions, such as cell propaga-
bases have been established such as circBase, tion [4], cell differentiation [5], and cell cycle [6].
Furthermore, miRNAs have essential roles in
human diseases and have a countless significance
S. H. Aghaee-Bakhtiari (*) as biomarkers [7], therapeutic agents [8], and dis-
Biotechnology Research Center, Mashhad University ease advancement assessment [9]. circRNAs are
of Medical Sciences, Mashhad, Iran
a group of newly discovered endogenous ncRNAs
Department of Medical Biotechnology, Faculty of [10] and have been determined in many tissues
Medicine, Mashhad University of Medical Sciences,
Mashhad, Iran and cell lines across most creatures [11].
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 35

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_3
36 S. H. Aghaee-Bakhtiari

Generally, circRNAs control gene expression CircInteractome is established by Dudekulay

transcriptionally or posttranscriptionally by tran- et al. and is openly available at http://
scription regulation, intervening with splicing, circinteractome.nia.nih.gov. This database
and collaboration with miRNAs [12]. Today, cir- simplifies the investigation of circRNAs and their
cRNAs are broadly noticed in an extensive range relations with more binding factors, principally
of diseases [13]. Furthermore, circRNAs have miRNAs and RBPs [18]. CircInteractome offers
the cell or tissue specificity and stability and users treasured features about circRNAs and their
also easily can be detected in saliva [14] or potential character in isolating miRNAs or RBPs
blood samples [15]. Consequently, circRNAs and thus lessens their accessibility for mRNAs.
are becoming the perfect group of novel bio- CircInteractome similarly simplifies the primer
markers for disease diagnosis and therapy. design to assess circRNA by RT-Qpcr
ncRNA online resources are essential tools for investigation. Furthermore, CircInteractome can
investigators to obtain data, and the number of be utilized to calculate RBP binding to sequences
them has been rapidly growing [16]. Considering of the transcript, consequently possibly clarify
the critical role of these molecules in molecular the production of circRNAs. In addition,
biology, circRNA has been converted to the hub CircInteractome combines numerous features
in the current ncRNA exploration field. To from other websites, such as StareBase 2.0,
facilitate the study of the different aspects of circBase, Primer3, and TargetScan 7.0. By
circRNAs, numerous databases have been combining these databases, CircInteractome
developed such as circBase, CircInteractome, allows researchers to realize the circRNA
CircNet, Circ2Traits, CircR2Disease, TCSD, and sequence and genomic site, circRNA-binding
CSCD. associates, primers, and siRNAs to analyze
circRNA ranks, activity, and localization.
Because all the information provided in
2 circRNA Databases CircInteractome are anticipated on the basis of
sequence similarity and existence of circRNA
circBase database covers circRNA data available structures, experimental confirmation is
up to 2013 and commonly becomes up to date necessary to validate functional positions [19].
with new published records [17]. circBase is CircNet database is assembled by transcriptome
established by Glažar P et al. and is accessible by sequencing datasets and is compiled by Liu et al.,
the web server at http://www.circbase.org/. Now and the website is available at http://circnet.mbc.
circBase presents data from human, mouse, C. nctu.edu.tw/ [20]. In CircNet, human circRNAs
elegans, and Latimeria organisms. The sequence are arranged with circRNA expression profiles
and the supportive evidence of their expression of through 464 human transcriptome samples. It
circRNAs can be retrieved, downloaded, and provides circRNA-­miRNA gene controlling net-
searched within the genomic context. Simple works and tissue-­ specific circRNA expression
search, list search, and table browser are three dif- profiles. Moreover, it presents a combined con-
ferent methods to search circBase. Simple search trolling system that shows the arrangement
is used for search by sequence, gene explanation, among circRNAs, miRNAs, and genes. Generally,
or genomic location. List search helps users to CircNet offers collective tools for researchers to
find joint of a big number of query terms with cir- simply access widespread data about genome loca-
cBase contents. The last method of search is table tion, expression analysis in different situations, and
browser which can be used for conditional data controlling complexes. In CircNet, researchers can
recovery. After choosing the desired animal and select a desired miRNA or gene, and presented data
experiment, users can additionally improve the containing the expression analysis, genomic loca-
results by some possibilities, like existence in a tion, and cohesive mRNA-miRNA-­circRNA con-
specific sample and quantity of supporting reads trolling system were gathered. In addition, CircNet
of the head-to-­tail splice intersection [17]. would be an advantageous tool to investigate cir-
3 Online Databases and Circular RNAs 37

cRNA relationship to disease and also tissue spe- platform to browse, explore, and transfer in
cialized action [20]. addition to submit new disease-related circRNAs.
Circ2Traits, a database of circRNAs hypothet- CircR2Disease could be very helpful for users
ically related to diseases in human, is developed who study the process of disease-related
by Ghosal et al. and is freely reachable at http:// circRNAs and discover the proper procedures for
gyanxet-beta.com/circdb/ [21]. The current ver- foreseeing novel relations [22].
sion of this database has classified 1951 human TSCD (tissue-specific circRNA database) is
circRNAs possibly connected to 105 diverse dis- compiled by Xia et al., and this resource is freely
eases. circRNAs and their stored data in accessible at http://gb.whu.edu.cn/TSCD [23].
Circ2Traits are classified conferring to their pos- This database is the first overall observation of
sible relationship with diseases which was tissue specificity for circRNAs. At this point,
detected from the GWAS-associated SNPs. In TSCD accomplished the full investigation to
addition, Circ2Traits stocks the whole putative distinguish the characteristic of human and
miRNA-circRNA-mRNA-lncRNA interaction mouse tissue-specific circRNAs. This database
network for any disease. Users have several recognized altogether 302 853 tissue-specific
search options in this database. At first, the user circRNAs in the human and mouse genome and
can select a disease and observe a list of circRNAs exhibited that the brain has the uppermost
related to the disease and moreover present the plethora of tissue-specific circRNAs. This
interaction table and the interaction network for resource additional established the presence of
each disease. The other search options are circRNAs by RT-PCR. TSCD similarly
keyword search for circRNAs, miRNAs, categorized the genomic position and preservation
lncRNAs, and protein-coding genes and search of these tissue-specific circRNAs and revealed
for GWAS traits connected to circRNAs [21]. that the mainstream of tissue-specific circRNAs
CircR2Disease is a manually analyzed is produced from exon areas. RNA binding
resource which is developed by Fan et al. and is protein and microRNAs which might bind to
openly reachable at http://bioinfo.snnu.edu.cn/ tissue-specific circRNA were recognized to more
CircR2Disease/ [22]. This tool delivers a broad comprehend the possible activity of tissue-­
database for circRNA dysregulation in different specific circRNAs. This procedure recommended
diseases, and collective indications have revealed their participation in progress and organ
that circRNAs have a key character in different development [23].
levels of gene regulation including transcription, CSCD (cancer-specific circRNA database) is
posttranscription, and translation levels. Based developed by Xia et al. and is openly reachable at
on previous studies, the irregular expression of http://gb.whu.edu.cn/CSCD. CSCD is the first
circRNAs has been related with a set of diseases. resource which fully explore the cancer-specific
Considering the enormous number of deregulated circRNAs [24]. This database recognized 272 152
circRNAs in various diseases, it is necessary to cancer-specific circRNAs from 228 total RNA
make a superior resource to gather the circRNAs samples from cancer together normal cell lines.
in diseases. The present version of CircR2Disease 170 909 circRNAs were recognized in normal and
covers 725 relations among 100 diseases and 661 tumor samples which could be utilized as non-
circRNAs by studying current articles. Every tumor background, and 950 962 circRNAs were
item in the CircR2Disease includes exhaustive known in normal samples only. CSCD foresees
data for the circRNA-disease association, the RNA binding protein sites and microRNA
comprising name of circRNA, name of disease, response element sites for every circRNA to com-
expression levels of circRNA, investigational prehend the practical properties of circRNAs. In
methods, a concise explanation of the circRNA-­ addition, this database predicted possible ORFs
disease connection, publication year, and the (open reading frames) to recognize translatable
PubMed ID. CircR2Disease offers an easy-to-use circRNAs. Considering the properties of CSCD
38 S. H. Aghaee-Bakhtiari

database could meaningfully provide to the inves- 10. Chen L, Huang C, Wang X et al (2015) Circular RNAs
in eukaryotic cells. Curr Genomics 16(5):312–318
tigation for the activity and control of cancer- 11. Holdt LM, Kohlmaier A, Teupser D (2018) Molecular
related circRNAs [24]. roles and function of circular RNAs in eukaryotic
cells. Cell Mol Life Sci 75(6):1071–1098
Acknowledgments This research is supported by 12. Hansen TB, Jensen TI, Clausen BH et al (2013)
Mashhad University of Medical Sciences (No. 941245, Natural RNA circles function as efficient microRNA
950909). sponges. Nature 495(7441):384–388
13. Haque S, Harries LW (2017) Circular RNAs (cir-
cRNAs) in health and disease. Genes (Basel)
Competing Financial Interests The authors declare no
8(12):353
competing financial interests.
14. Bahn JH, Zhang Q, Li F et al (2015) The landscape of
microRNA, Piwi-interacting RNA, and circular RNA
in human saliva. Clin Chem 61(1):221–230
15. Memczak S, Papavasileiou P, Peters O et al (2015)
Identification and characterization of circular RNAs
References as a new class of putative biomarkers in human blood.
PLoS One 10(10):e0141214
1. Santosh B, Varshney A, Yadava PK (2015) Non-­ 16. Aghaee-Bakhtiari SH, Arefian E, Lau P (2017)
coding RNAs: biological functions and applications. miRandb: a resource of online services for miRNA
Cell Biochem Funct 33(1):14–22 research. Brief Bioinform. https://doi.org/10.1093/
2. Han B, Chao J, Yao H (2018) Circular RNA and bib/bbw109
its mechanisms in disease: from the bench to the 17. Glazar P, Papavasileiou P, Rajewsky N (2014)
clinic. Pharmacol Ther. https://doi.org/10.1016/j. circBase: a database for circular RNAs. RNA
pharmthera.2018.01.010 20(11):1666–1670
3. Esteller M (2011) Non-coding RNAs in human dis- 18. Dudekula DB, Panda AC, Grammatikakis I et al
ease. Nat Rev Genet 12(12):861 (2016) CircInteractome: a web tool for exploring
4. Rahimian A, Soleimani M, Kaviani S et al (2011) circular RNAs and their interacting proteins and
Bypassing the maturation arrest in myeloid cell microRNAs. RNA Biol 13(1):34–42
line U937 by over-expression of microRNA-424. 19. Panda AC, Dudekula DB, Abdelmohsen K et al
Hematology 16(5):298–302 (2018) Analysis of circular RNAs using the web tool
5. Fallah P, Arefian E, Naderi M et al (2013) miR-­ circInteractome. Methods Mol Biol 1724:43–56
146a and miR-150 promote the differentiation of 20. Liu YC, Li JR, Sun CH et al (2016) CircNet: a
CD133+ cells into T-lymphoid lineage. Mol Biol Rep database of circular RNAs derived from tran-
40(8):4713–4719 scriptome sequencing data. Nucleic Acids Res
6. Attar M, Arefian E, Nabiuni M et al (2012) MicroRNA 44(D1):D209–D215
17-92 expressed by a transposone-based vector 21. Ghosal S, Das S, Sen R et al (2013) Circ2Traits:
changes expression level of cell-cycle-related genes. a comprehensive database for circular RNA poten-
Cell Biol Int 36(11):1005–1012 tially associated with disease and traits. Front Genet
7. Aghaee-Bakhtiari SH, Arefian E, Soleimani M et al 4:283
(2016) Reproducible and reliable real-time PCR 22. Fan C, Lei X, Fang Z, et al (2018) CircR2Disease:
assay to measure mature form of miR-141. Appl a manually curated database for experimentally sup-
Immunohistochem Mol Morphol 24(2):138–143 ported circular RNAs associated with various dis-
8. Aghaee-Bakhtiari SH, Arefian E, Naderi M et al eases. Database (Oxford) 2018.
(2015) MAPK and JAK/STAT pathways targeted 23. Xia S, Feng J, Lei L et al (2017) Comprehensive
by miR-23a and miR-23b in prostate cancer: com- characterization of tissue-specific circular RNAs
putational and in vitro approaches. Tumour Biol in the human and mouse genomes. Brief Bioinform
36(6):4203–4212 18(6):984–992
9. Arefian E, Kiani J, Soleimani M et al (2011) Analysis 24. Xia S, Feng J, Chen K et al (2018) CSCD: a database
of microRNA signatures using size-coded ligation-­ for cancer-specific circular RNAs. Nucleic Acids Res
mediated PCR. Nucleic Acids Res 39(12):e80 46(D1):D925–D929
 Part III
Biogenesis of Circular RNAs
 Circular RNA Splicing
4
Nicole Eger, Laura Schoppe, Susanne Schuster,
Ulrich Laufs, and Jes-Niels Boeckel

Abstract lar RNAs and highlight the derivation of dif-

Circular RNAs (circRNAs) are covalently ferent types of circular RNAs.
closed single-stranded RNA molecules
derived from exons by alternative mRNA Keywords
splicing. Circularization of single-stranded circRNA · RNA splicing · Circular RNA ·
RNA molecules was already described in Backsplicing
1976 for viroids in plants. Since then several
additional types of circular RNAs in many
species have been described such as the circu-
lar single-stranded RNA genome of the hepa- 1 Introduction
titis delta virus (HDV) or circular RNAs as
products or intermediates of tRNA and rRNA Circular RNAs (circRNAs) are circular single-­
maturation in archaea. CircRNAs are gener- stranded RNA (ssRNA) molecules formed from
ally formed by covalent binding of the 5′ site exons of genes by alternative mRNA splicing
of an upstream exon with the 3′ of the same or (Figs. 4.1 and 4.3). Circular RNAs in eukaryotes
a downstream exon. Meanwhile, two different were first detected by electron microscopy in
models of circRNA biogenesis have been human HeLa cells in 1979 [1]. However, at that
described, the lariat or exon skipping model time the origin of these circular RNA molecules
and the direct backsplicing model. In the lariat was not known. The discovery of inverted orien-
model, canonical splicing occurs before back- tated exons in ssRNA, referred to as “scrambled
splicing, whereas in the direct backsplicing exons” in humans [2–4], gave first evidence that
model, the circRNA is generated first. In this these molecules might be originated from
chapter, we will review the formation of circu- protein-­coding open reading frames [2, 5].
CircRNAs are generally formed by covalent link-
age of the 5′ splice site of an upstream exon with
Author contributed equally with all other contribu- the 3′ site of the same or a downstream exon [6].
tors. Nicole Eger and Laura Schoppe However, also introns located in the open reading
N. Eger · L. Schoppe frame (ORF) of a protein-coding gene can circu-
University of Heidelberg, Heidelberg, Germany larize [7]. Spliced introns forming circular RNAs
S. Schuster · U. Laufs · J.-N. Boeckel (*) are referred to as circular intronic RNAs (ciR-
Clinic and Polyclinic for Cardiology, University NAs) and are distinguished from those formed by
Hospital Leipzig, Leipzig, Germany exons from the open reading frame of a
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 41

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_4
42 N. Eger et al.

Fig. 4.1 Circular single-stranded RNA variants and their lengths between 21 and 105 nt. The tRNA-splicing endo-
derivation nuclease (TSEN) complex in archaea, consisting of a spe-
Various circular RNAs (here illustrated in red) have been cific endonuclease and ligase, has been shown to produce
reported. The single-­stranded RNA circles are covalently circularized excised tRNA introns. Circular rRNA (cir-
closed, while sequence-dependent double-­stranded base crRNA) precursors were reported in archaea containing
pairing can appear at multiple regions of the RNA circles. the premature rRNA sequence, which is later spliced out
The genome of viroids, which does not code for proteins during the process of rRNA maturation. This splicing is
but has catalytic ribozyme properties, was shown to be autocatalytic and accompanied by a guanosine addition to
formed of single-stranded RNAs with sizes of ~220–400 the 5′ end. Mechanistically, rRNA splicing occurs in a
nucleotides length. Viroids replicate utilizing the host similar fashion as splicing of circular tRNA introns.
cell’s machinery via a rolling-circle mechanism resulting Circular intronic RNAs (ciRNA) can appear within
in a concatemeric linear sequence which is then cleaved eukaryotes as either products of spliceosomal splicing or
by endonuceases into multiple linear replicates and finally intermediates of self-splicing introns. In canonic splicing
circularized by end-to-end ligation. The Hepatitis delta intron sequences are removed by the spliceosome in lariat
(δ) virus (HDV) was discovered as the first animal virus form, containing an internal 2′ to 5′ phosphodiester bond.
having a single-stranded circular RNA molecule as carrier Introns of group I and II are mobile genetic elements that
of its genetic information with a length of 1750 nt. The catalyze their own splicing from DNA via their ribozymal
host cell machinery synthesizes the circular viral RNA properties. Circular exonic RNAs (circRNA) are pro-
genome via a rolling-circle mechanism, while finally the duced via pre-mRNA processing in eukaryotic cell nuclei
host cell’s ligases form 3′ to 5′ or 2′ to 5′ phosphodiester by an alternative splicing process referred to as backsplic-
bonds between the respective ends of the HDV genome. ing. A downstream exon’s 3′ splice site performs a nucleo-
tRNA introns occur within eukaryotes and prokaryotes philic attack on its own or a different upstream exon’s 5′
alike, while tRNA intronic circular (tric)RNAs can splice acceptor site, resulting in a 3′ to 5′ phosphodiester
appear in archaea as a product of tRNA splicing with linked circular RNA molecule.

p­rotein-­coding gene, which are called circular and single-stranded DNA [13, 14]. The hepatitis
RNAs (circRNAs) [8, 9] (Figs. 4.1 and 4.2b). delta virus (HDV) was discovered in 1986 as the
Several circular RNA species have been reported first animal virus having a circular single-­
in eukaryotes, prokaryotes, viruses, and subviral stranded RNA genome [15] (Fig. 4.1). Studies in
agents and were found to be synthesized by vari- archaea revealed the existence of circular RNAs
ous ligation reactions. Circularization of as by-product and end product of tRNA matura-
­single-­stranded RNA (ssRNA) molecules in gen- tion called tricRNAs [16–19] (Fig. 4.1). Archaea
eral was already described in 1976 for the genome also give rise to circular single-stranded pre-­
of some subviral agents named “viroids” [10–12] rRNA intermediates of the 16S and 23S RNA
(Fig. 4.1). In contrast, the genomes of viruses can [19, 20]. The predominant mechanism of circular
consist of several different nucleotide structures, RNA formation is 3′ to 5′ end ligation found in
such as linear single-stranded RNA (ssRNA), lin- viroids, hepatitis delta virus, and the products or
ear or circular double-stranded DNA (dsDNA), intermediates of tRNA and rRNA maturation in
4 Circular RNA Splicing 43

Pre-mRNA 5’ 3’

A B C

A A A

CiRNA

CircRNA

Linear mRNA Linear mRNA Linear mRNA CircRNAs

Fig. 4.2 Derivation of circular RNA molecules by alter- from the mRNA. (b) Alternative splicing of pre-mRNA
native pre-mRNA splicing may also result in circular RNAs. Base pairing between
Circular RNAs in eukaryotes can be derived from pri- intronic inverted repeat regions thereby connecting the
mary RNA transcripts via alternative splicing. (a) During two regions promotes circular RNA derivation.
canonical splicing by the spliceosome, intronic regions This internal secondary stem-loop structures within the
are removed in a two-­step transesterification reaction pre-mRNA sterically facilitates backsplicing. Besides
with a lariat intermediate. The branch point adenosine’s linear mRNA, circular intronic (ciRNA) and circular
2′ hydroxyl group performs a nucleophilic attack on an exonic RNA (circRNA) can be spliced from the same
upstream exon’s 3′ splice site within the same intron, primary transcript. (c) CircRNAs can also be spliced in
forming an intronic 2′ to 5′ phosphodiester. In a second multiple different alternative variants. Up to six exons
step, the upstream exon’s 3′ hydroxyl group binds after were reported to be spliced in inverted order within one
deprotonation to the downstream exon’s 5′ phos- circRNA. Interceding intronic regions may also be
phate end, thereby displacing the intron in its lariat form included in the final splice product.

archaea. However, also 2′ to 5′ end ligations known since the early twentieth century [11]. The
involving a nucleophilic attack of the branch first covalently closed RNA circles were discov-
point were reported for ciRNAs in eukaryotes ered in the potato spindle tuber viroid in 1967
[8]. Two models were proposed for the biogene- [23]. The genome of this viroid was first pro-
sis of circRNAs as a product of alternative exon posed to be a double-stranded RNA molecule but
splicing: the lariat model and the backsplicing later on shown to be single-stranded [12]. These
model [21, 22] (Fig. 4.3). The essential differ- circular RNA containing viroids are in general
ence between these two reactions is the order in small 220–400 nucleotides RNA-only plant
which splicing events do occur. pathogens which replicate autonomously within
In the following chapter, we will provide an the nucleus or chloroplasts of their host’s cells
overview about the different circular RNA classes [10, 24, 25]. Their genome exist in circular con-
arising from different species and their biogene- formation and further form self-complementary
sis by splicing and ligation. base-paired rod-like structures [10] (Fig. 4.1).
Viroid RNA does not code for proteins; how-
ever their circular single-stranded genomic RNA
2  he Origin and Splicing
T has catalytic ribozyme properties [12, 26]. The
of Different Circular RNA ribozyme is responsible for the infection of and
Species proliferation within host cells [24]. After infec-
tion of the plant hosts, viroids also utilize the
2.1 Viroid Circular RNA plant’s enzymatic machinery in order to ensure
their own functionality and the replication of
The existence of viroids as proto-organisms con- their circular ssRNA genome [27]. Viroid repli-
taining circular RNAs as genomes has been cation ensues via a rolling-circle mechanism
44 N. Eger et al.

Pre-mRNA 5’ 3’

A Lariat splicing model B Direct backsplicing model

A A

O
H
OH

A A
A
OH

HO
A

A
Linear mRNA CircRNa Double lariat Linear mRNA CircRNA

Fig. 4.3 Different backsplicing models nosine’s 2′ hydroxyl group nucleophilically attacks the 3′
Two different mechanisms for derivation of circRNA have splice site of the downstream exon, forming a double lar-
been described so far. In both models internal base pairing iat structure. The downstream exon’s 3′ hydroxyl group is
between inverted repeat regions facilitates the alternative now free to attack the upstream 5′ splice acceptor of the
splicing. The main difference is whether circularization of exon, thereby circularizing it. (b) Direct Backsplicing
the circRNA is after canonical splicing or before. (a) Model In the direct backsplicing model, an upstream
Lariat Splicing Model According to the lariat splicing branch point adenosine 2′ hydroxyl group initiates the
model, backsplicing occurs after canonical splicing. process by attacking a downstream exon’s 3′ splice site,
During canonical splicing alternative exons may be resulting in a Y-shaped intermediate. The free 3′ hydroxyl
spliced out of the final mRNA product and end up con- then attacks an upstream 5′ splice site resulting in
tained within the excised lariat. The lariat then undergoes circularization
internal backsplicing. First an upstream branch point ade-

enabled by their circular genomic structure [28, 2.2 Viral Circular RNA
29]. The circular genome is transcribed times in
an iterative fashion forming one long, linear tran- Viruses are small nucleic acid-based intracellular
script of concatemers, which may later on be pathogens. Outside a host they are covered by a
cleaved by endonucleases into multiple linear protein coat encoded within their own genome
RNA replicates which are in turn circularized by which they shed during host infection. Viral
end-to-end ligation [30, 31]. The common genomes come in many different forms and con-
hypothesis explaining the genomic structure of formations of both DNA and RNA molecules,
viroids is that the circularity of their RNA allows while circular single-stranded RNA is one of
for faster replication, since the rate-limiting ini- them.
tiation step must only be performed once, but cre- In 1986 the hepatitis delta virus (HDV) was
ates multitudes of offspring. discovered as the first animal virus having a cir-
4 Circular RNA Splicing 45

cular RNA molecule as carrier of its genetic cularity may protect from heat denaturation of
information [15] (Fig. 4.1). Comparable to the molecules since it does not offer a free ter-
viroids, HDV uses the host cell machinery for minal helix with steric flexibility at either its 3′
production of circular RNA molecules via the or 5′ end [42].
rolling-circle mechanism [30, 32]. Viruses uti-
lize their host cell’s ligases to form 3′ to 5′ or 2′
to 5′ phosphodiester bonds between the respec- 2.4 Circular rRNA (CircrRNA)
tive ends of their genomes [33]. This circular-
ization process in turn protects the viral genome In 1981 Grabowski and colleagues discovered
from digestion by intracellular exonucleases the existence of a circular RNA molecule in
[34]. It further protects the viral RNA from Tetrahymena thermophila as an intermediate of
other host immunity mechanisms, such as the rRNA maturation [43]. Here the premature
MDA-5 and RIG-I receptors, which bind free rRNA is part of a larger circular single-stranded
nonhost cytoplasmic ssRNA in higher verte- ribosomal RNA precursor and later spliced out
brates [35, 36]. during proceeding rRNA maturation (Fig. 4.1).
There have been several reports about ribosomal
RNA maturation in archaea involving a proce-
2.3 t RNA Intronic Circular (tric) dure wherein rRNA precursors are first cleaved
RNAs to derive linear pre-16S and pre-23S rRNAs [18,
20, 44, 45] (Fig. 4.1). The ends of these RNAs
Transfer RNA (tRNA) introns occur within are then ligated, forming circular single-
eukaryotes and prokaryotes alike. Archaeal stranded RNA molecules, which are processed
tRNA primary transcripts can contain intronic further until becoming mature, functional
regions that require splicing before acquiring rRNAs [18]. This underlying splicing reaction is
functionality [37]. Also in yeast tRNA splic- autocatalytic and accompanied by a guanosine
ing occures but generates linear RNA intermedi- addition to the 5′ end [46]. Mechanistically, this
ates [38]. Interestingly, tRNA intronic circular process is comparable to the previously
(tric)RNAs can appear in archaea as a product described processing of the circular tRNA
of tRNA splicing [39] (Fig. 4.1). The tRNA- introns, where recognition occurs via a bulge-
splicing endonuclease (TSEN) complex in helix-bulge motif [20]. In accordance, the TSEN
archaea [37], consisting of a specific endonucle- complex is also responsible for the final ligation
ase and ligase, has been shown to produce circu- of the circular molecules [37].
larized excised tRNA introns [18, 19]. This
complex recognizes the cleavage sites by a
bulge-helix-bulge motif at exon-intron junctions 2.5  ircular Intronic RNAs
C
[17, 18, 37]. One may hypothesize that circular- (CiRNAs)
ization of functional noncoding RNAs in gen-
eral may increase their stability and thereby It was initially believed that spliced out introns
prolong their regulatory effects within the cells. were the main source of circular RNAs in higher
Like other circular RNAs, they are resistant to eukaryotes [47–49]. Different forms and origin
digestion by cytoplasmic exonucleases like of intronic RNAs have been described so far.
RNAse R and therefore less likely to be degraded Circular intronic RNAs (ciRNAs) are defined as
[40]. This effect has been used to the advantage intronic sequences forming RNA circles [8],
of sequencing circular RNAs as well [21, 41]. unlike circRNAs which can contain both exonic
Circular noncoding RNAs have been most and intronic sequences. A regulatory function of
prominently described in thermophilic archaea ciRNA in relation to polymerase II transcription
like Haloferax volcanii [18, 19]. Here, their cir- is suggested since knockdown of ciRNA showed
46 N. Eger et al.

altered expression at the corresponding gene 2.6 Circular RNA Spliced Exons
locus [8].
Circular exonic RNA molecules (circRNA) were
2.5.1  roup I and Group II Self-­
G first detected in 1991 in human cells and initially
Splicing-­Derived Circular believed to be linear “scrambled exons” [2]
Intronic RNAs (CiRNAs) (Fig. 4.1). Shortly afterward circRNA molecules
The self-splicing group I introns are mainly were described as RNA products derived from the
located within genomic ribosomal RNA regions testis-determining gene Sry located to the cyto-
of eukaryotic microorganisms [50]. Unlike in plasm of murine cells [55], as well as from the Ets-
spliceosomal or group II intron splicing, where 1 gene in human cells [4, 5]. For many years they
the RNA hydroxyl groups act as internal nucleo- were described as rare events of transcription,
philes, group I introns recruit an external guano- being merely by-products of linear mRNA splic-
sine as a nucleophile to initiate splicing [34]. ing [2, 4, 5, 55–58]. But in the last decade, evi-
During the process first a linear excised RNA is dences increase that circRNAs are derived from
separated, which can then undergo 3′ to 5′ circu- thousands of expressed protein-­coding genes and
larization [43] (Fig. 4.1). moreover have functional roles in eukaryotic cells.
Ribosomal introns of group II are mobile Tissue- and cell-specific alternative mRNA splic-
genetic elements that autocatalytically splice ing is a hallmark of cell specialization found in
themselves from precursor RNAs by using a many eukaryotic organs and highly specialized
transesterification reaction similar to the classical cell types such as cardiomyocytes and neurons
spliceosomal splicing reaction [42, 51, 52]. Their [59]. Alternative splicing was already reported in
structure is made up of catalytically active RNAs 1992 by the group of Bernard Bailleul for two cir-
and an intron-encoded reverse transcriptase [51]. cRNA products of the Ets-1 gene but has been
After sequence excision, circular lariat species later shown to also appear in different tissues and
with a 2′ to 5′ phosphodiester bond are formed cells [60] (Fig. 4.2). Interestingly, some circRNAs
[53]. Given the strong similarity in mechanism were reported to be even more abundantly
and characteristics of group II introns and eukary- expressed than their cognate linear mRNAs, while
otic spliceosomal introns, they are thought to be also the contribution of differences in molecular
evolutionarily related [51]. stability between the linear and circular products
of a host gene cannot be excluded here [21, 41].
2.5.2 Spliceosomal-Derived Circular CircRNAs range from barely 100 to sizes over
Intronic RNAs (CiRNAs) 4000 nucleotides [42]. They are mainly found
Spliceosomal splicing is the main, highly con- within the cytoplasm of eukaryotes [61]. Circular
served, mechanism of mRNA processing within transcripts usually encompass two to five “shuf-
eukaryotic cells [54]. During intron excision a 2′ fled” or “scrambled” exons and bear the possibil-
to 5′ transester is formed and released. Circular ity to undergo alternative splicing [2]. Their
intronic RNAs (ciRNAs) are formed from such genomic loci are usually flanked by repetitive
lariats, which have been additionally degraded complementary sequences which enhance the
from their 3′ end up to the branch point but have circularization efficiency [62].
not been further degraded beyond this. Their CircRNAs are derived by a backsplicing
sizes may vary from under 200 to over 3000 mechanism, where a 3′ splice site of a down-
nucleotides [42]. The blockage of degradation stream exon is effectively spliced to end up on an
beyond the branch point depends on a consensus upstream exon or the same exon’s 5′ splice site
motif of a 7 nucleotide GU-rich element near the [21, 63]. The next chapter will give a more
5′ splice site and an 11 nucleotide C-rich element detailed insight into the different proposed mech-
near the branch point [8] (Fig. 4.1). anisms of backsplicing [61].
4 Circular RNA Splicing 47

3  plicing of circRNA from ORF

S Mechanistically, the canonical pre-mRNA
Exons splicing is a two-step transesterification reaction.
First, the 2′ branch point adenosine hydroxyl
3.1 Canonical Splicing group, which is situated within the intron to be
of Pre-mRNA spliced, nucleophilically attacks the upstream 5′
end of the intron (Fig. 4.2a). This results in a lar-
The ~3*109 base pairs large human genome only iat intermediate, which is covalently closed by a
consists of about 1% protein-coding 2′ to 5′ phosphodiester bond. The upstream
sequences referred to as exons [64, 65]. The large exon’s 3′ hydroxyl group is now free to attack the
rest of the genome, though not being translated downstream exon’s 5′ phosphate, completely
into amino acid chains, fulfill many other func- splicing out the intron in lariat form and leaving
tions, among them increasing overall genomic the exons linked together in a linear coding
stability [66], controlling the chromatin accessi- sequence manner.
bility via epigenetic modifications [67, 68], and
regulation of transcription [69–71]. Protein-­
coding genes are transcribed in the nucleus by 3.2  ircRNAs Are a Product
C
RNA polymerase II resulting in a single-stranded of an Alternative Splicing
pre-mRNA containing introns and exons in Process
genomic sequence [72]. During transcription
there is no discrimination between exonic and The generation of circular RNAs (circRNA) from
intronic regions; both are equally transcribed exons, is more similar to the formation of mRNAs
after initiation of transcription at the promoter as opposed to the generation of other circular
site; the RNA-polymerase complex is producing RNA species (see, e.g., viral RNA genomes
one continuous pre-mRNA transcript of the which are circularized by host ligases in the pre-
genomic DNA [73]. In the further process of vious chapter), requires processing by the canonic
mRNA maturation, a 5′ 7-methylguanosine cap spliceosome. CircRNAs were first identified as
and a 3′-poly-adenosine tail are added [72, “scrambled exons” and thought to be splicing
74–78]. errors or waste products of the splicing process
In order to create functional coding mRNAs [2, 82, 83], since their order of exons is inverted
that can be translated by cytoplasmic ribosomes, compared to the exonic arrangement on the
the interceding intronic regions are removed genomic open reading frame [2, 5, 7, 21]. Already
from the primary transcript by RNA splicing. The in the year 1993, early after the first description
multimeric enzyme complex made up of small of scrambled exons by the Vogelstein lab in 1991,
nuclear uracil-rich ribonucleoprotein particles circularity of exonic RNA was observed in
(U-snRNPs), referred to as the spliceosome [47], eukaryotes by the Bailleul laboratory in RNA
assembles in a stepwise manner on specific products of the Ets-1 gene as well as by Capel
splice-signaling intronic guide sequences [79]. and colleagues in RNA products of the Sry gene
The mRNA splicing can be canonical or nonca- in adult mouse testis [4, 5, 55, 84]. Cocquerelle
nonical, depending on the guiding sequence of and colleagues moreover reported in their pio-
the respective intron to be removed. Over 99% of neering research work the generation of two cir-
all pre-mRNAs feature canonical splice sites 5′ cular RNAs from the Ets-1 gene by incorporation
GU and 3′ AG on the respective ends of the of different exons into the different final RNA
intronic regions [80]. These canonical splice sites circles, thereby also reporting alternative splicing
function as recognition signals for U-snRNP of circular exonic RNAs for the very first time [4]
binding, here the 5′ end of the U1-snRNP directly (Fig. 4.2c).
binds to the 5′ splice signal sequence on the pre-­ Alternative splicing allows for mRNAs with a
mRNA, thereby promoting the splicing reaction variety of differently composed exons resulting
[81]. from a single gene locus, which can code for
48 N. Eger et al.

multiple isoforms of a protein with different ing approaches revealed that these inverted
functional properties. Alternative splicing is fur- repeats, especially Alu sequences, are two fold as
ther contributing to functional specification of likely to occur in regions bordering on circular-
proteins to certain tissue demands, as was pre- izing exons compared to non-circularizing exons
dicted early on by Walter Gilbert in response to [9, 92]. These retro-transposed genomic elements
the discovery of the fact that genes consist of have previously been demonstrated to contribute
coding as well as noncoding sequences [85]. The to alternative over canonical splicing [93, 94]. At
ryanodine receptor gene, whose alternative least in higher eukaryotes, these regions seem to
mRNA splicing in cardiomyocytes allows for play a major role in circularization. For lower
highly specialized contribution to intracellular eukaryotes like Saccharomyces, these repetitive
calcium signaling by different resulting protein regions are less common, and therefore likely
isoforms within the heart, is a good example in other mechanisms are at play [22, 95]. Studies in
this regard [86–88]. The mechanism of alterna- Drosophila however demonstrated that inverted
tive splicing, according to the combinatorial repeats are not essential for circularization, at
model, is controlled by regulatory factors that least in this organism [62]. Therefore, inverted
bind to pre-mRNA and induce certain splicing sequences flanking the exon seem to promote cir-
patterns, determining the inclusion or exclusion cularization but not be mandatory for formation
of a certain exon on the final mRNA product. It of all yet known circRNAs.
has been the proposed mechanism facilitating
circRNAs formation in the first description of
scrambled exons in 1991 [2, 89]. 3.3 Backsplicing Models
The number of exons in a single circRNA dif-
fers between one and five, with two or three The alternative splicing mechanism of backsplic-
exons being most frequently included in the cir- ing generates circRNAs by connection of the 3′
cle. Interjacent intronic regions are mostly end of a containing exon to either their own or a
excised but are sometimes fully or in part different upstream exon’s 5′ region [96]. This
included into the circularized molecule [62, 90]. leads to an inverted or scrambled order of exons
Both splice sites of the canonical splice signal are in comparison to the genomic sequence [2, 4, 5,
required for successful exon circularization 9] (Fig. 4.2).
through the process of alternative splicing [84]. Two models for the biogenesis of circRNA
This so-called “trans-splicing” process which backsplicing have been proposed: the lariat
was already reported in 1985 is independent of model, sometimes also referred to as exon skip-
the actual exon sequence but can be modulated ping model, and the direct backsplicing model
by flanking intron structures and has at least been [22, 58, 97] (Fig. 4.3a and b). The essential dif-
shown to be true for a subset of circRNAs [47, ference between these two mechanisms is the
91]. order in which splicing events do occur. In the
Repetitive inverted sequences upstream and lariat model, canonical splicing of the pre-mRNA
downstream of the circularizing exon contribute occurs first, while backspliced circRNAs are
to this circularization process. It has been pro- products of further processing of lariat intermedi-
posed that the intronic base pairing of these ates. In the direct backsplicing model on the other
repetetive inverted sequences facilitates the for- hand, circRNAs are generated first.
mation of a secondary RNA hairpin structures Mechanistically, the two models can be distin-
[92] (Figs. 4.2 and 4.3). These inverted repeat guished by the different intermediates that occur
regions have been proven mandatory for circular- during the splice processes. These can be ana-
ization of the gene Sry in mice, where a 400 lyzed and detected by their specific steric proper-
nucleotide sequence was minimally required for ties in order to identify which mechanism is
successful circle formation [9, 55]. In humans responsible for the circularization of a specific
progressing next generation sequenc- circRNA [22, 97]. One method to analyze these
4 Circular RNA Splicing 49

properties would be two-dimensional denaturing [21] (Fig. 4.3b). This brings an upstream branch
polyacrylamide gel electrophoresis [95, 98]. point close to a downstream 5′ splice acceptor
According to the lariat model circRNAs are site and therefore sterically enables the 2′
products of previous splicing, in which an alter- hydroxyl group’s nucleophilic attack, creating a
native exon has been excised from the final Y-shaped intermediate. This frees the circular-
mRNA product, ending up in an isolated lariat izing exon’s 3′ hydroxyl group to in turn attack
intermediate. This intermediate RNA molecule is its 5′ phosphate. The final products of this reac-
not permanently stable, and spliced lariats are tion are a circularized RNA molecule contain-
usually quickly degraded within the cell [99]. ing the circularizing exon or exons, as well as
Debranching endonucleases specifically recog- the pre-­ mRNA transcript whose secondary
nize the 2′ to 5′ phosphodiester bond that is char- hairpin structure now contains a 2′, 5′-phos-
acteristic for lariats and by digestion linearize phodiester linkage which can be removed by
them. Linear RNAs without protection of a poly-­ canonical splicing resulting in a linear mRNA
adenosine tail or a 5′ 7-methylguanosine cap are [22].
easy targets for nucleases. The direct backsplicing model may explain
During generation of circRNAs via the lariat the high expression of certain circRNAs, which
model splicing, a second double lariat intermedi- for some genes even exceeds their linear counter-
ate occurs (Fig. 4.3a). First, the splicing of the parts, as has been shown for circular and linear
linear pre-mRNA occurs as per usual, with the RNAs species derived from the KIAA0182 and
downstream branch point adenosine’s 2′ hydroxyl MAN1A2 genes in immune cells [41]. The
group attacking an upstream splice acceptor site majority of the RNAs expressed from these loci
beyond an alternative exon. The result of this showed a scrambled order of exons. Interestingly,
substitution reaction is a canonical lariat struc- this finding was unrelated to alternative splicing,
ture containing an exon and linked by a 2′ to 5′ thereby giving evidence for a direct backsplicing
phosphodiester bond. In a second step, this mechanism without lariat intermediates.
excised lariat undergoes internal backsplicing, Contribution of intronic pairing of inverted repeat
and an upstream branch point attacks a down- Alu sequences has been indicated to promote cir-
stream splice acceptor, resulting in a double lariat cularization even without exon skipping in
structure. The alternative exon’s 3′ hydroxyl human fibroblasts [21].
group is now able to perform a nucleophilic
attack on its own upstream splice acceptor site, Acknowledgments This work was supported by the
resulting in circularization by forming a 3′ to 5′ German Cardiac Society (Deutsche Gesellschaft für
Kardiologie (DGK)) to Jes-Niels Boeckel.
phosphodiester bond [22]. Circular RNA forma-
tion by the lariat model has been demonstrated in Competing Financial Interests The authors declare no
the yeast Schizosaccharomyces pombe gene competing financial interests.
Mrsp1. Yeast genomes rarely have repetitive
sequences, which may be an explanation for the
prevalence of lariat model circularization in these
organisms [22, 95, 98]. Interestingly, RNA References
sequencing also revealed lariat species within
human fibroblast RNA preparations [21]. 1. Hsu MT, Coca-Prados M (1979) Electron micro-
Direct backsplicing presumes that circRNA scopic evidence for the circular form of RNA in the
cytoplasm of eukaryotic cells. Nature 280:339–340
biogenesis is a process independent from prior 2. Nigro JM, Cho KR, Fearon ER et al (1991) Scrambled
exon excision and lariat formation. This model exons. Cell 64:607–613
proposes that inverted repeat regions on both 3. Zaphiropoulos PG (1997) Exon skipping and circu-
sides of the prospectively circularized exon or lar RNA formation in transcripts of the human cyto-
chrome P-450 2C18 gene in epidermis and of the rat
exons, first results in a secondary stem-loop androgen binding protein gene in testis. Mol Cell Biol
structure formation within the RNA transcript 17:2985–2993
50 N. Eger et al.

4. Cocquerelle C, Mascrez B, Hétuin D et al (1993) 22. Barrett SP, Wang PL, Salzman J (2015) Circular RNA
Mis-splicing yields circular RNA molecules. FASEB biogenesis can proceed through an exon-containing
J 7:155–160 lariat precursor. Elife 4:e07540
5. Cocquerelle C, Daubersies P, Majérus MA et al 23. Diener TO, Raymer WB (1967) Potato spindle tuber
(1992) Splicing with inverted order of exons occurs virus: a plant virus with properties of a free nucleic
proximal to large introns. EMBO J 11:1095–1098 acid. Science 158:378–381
6. Burd CE, Jeck WR, Liu Y et al (2010) Expression 24. Flores R, Hernández C, Martínez de Alba AE et al
of linear and novel circular forms of an INK4/ARF-­ (2005) Viroids and viroid-host interactions. Annu Rev
associated non-coding RNA correlates with athero- Phytopathol 43:117–139
sclerosis risk. PLoS Genet 6:e1001233 25. Bonfiglioli, McFadden, Symons RH (1994) In situ
7. Aebi M, Weissman C (1987) Precision and orderli- hybridization localizes avocado sunblotch viroid on
ness in splicing. Trends Genet 3:102–107 chloroplast thylakoid membranes and coconut cadang
8. Zhang Y, Zhang XO, Chen T et al (2013) Circular cadang viroid in the nucleus. Plant J 6(1):99–103
intronic long noncoding RNAs. Mol Cell 51:792–806 26. Przybilski R, Gräf S, Lescoute A et al (2005)
9. Dubin RA, Kazmi MA, Ostrer H (1995) Inverted Functional hammerhead ribozymes naturally encoded
repeats are necessary for circularization of the mouse in the genome of Arabidopsis thaliana. Plant Cell
testis Sry transcript. Gene 167:245–248 17:1877–1885
10. Sanger HL, Klotz G, Riesner D et al (1976) Viroids 27. Steger G, Perreault JP (2016) Structure and associ-
are single-stranded covalently closed circular RNA ated biological functions of viroids. Adv Virus Res
molecules existing as highly base-paired rod-like 94:141–172
structures. Proc Natl Acad Sci U S A 73:3852–3856 28. Flores R, Grubb D, Elleuch A et al (2011) Rolling-­
11. Schultz ES, Folsom D (1923) A “spindling-tuber dis- circle replication of viroids, viroid-like satellite RNAs
ease” of Irish potatoes. Science 57:149. https://doi. and hepatitis delta virus: variations on a theme. RNA
org/10.1126/science.57.1466.149 Biol 8(2):200–206
12. Diener TO (1971) Potato spindle tuber “virus”. IV. A 29. Flores R, Gas ME, Molina-Serrano D et al (2009)
replicating, low molecular weight RNA. Virology Viroid replication: rolling-circles, enzymes and ribo-
45:411–428 zymes. Viruses 1:317–334
13. Stoler MH, Broker TR (1986) In situ hybridization 30. Branch AD, Robertson HD (1984) A replication cycle
detection of human papillomavirus DNAs and mes- for viroids and other small infectious RNA’s. Science
senger RNAs in genital condylomas and a cervical 223:450–455
carcinoma. Hum Pathol 17:1250–1258 31. Flores (1999) Viroids with hammerhead ribozymes:
14. Summers J, Jones SE, Anderson MJ (1983) some unique structural and functional aspects with
Characterization of the genome of the agent of eryth- respect to other members of the group. Biol Chem
rocyte aplasia permits its classification as a human 380(7-8):849–854
parvovirus. J Gen Virol 64(Pt 11):2527–2532 32. Macnaughton TB, Shi ST, Modahl LE et al (2002)
15. Kos A, Dijkema R, Arnberg AC et al (1986) The Rolling circle replication of hepatitis delta virus RNA
hepatitis delta (delta) virus possesses a circular is carried out by two different cellular RNA polymer-
RNA. Nature 323:558–560 ases. J Virol 76:3920–3927
16. Soma A, Onodera A, Sugahara J et al (2007) 33. Reid CE, Lazinski DW (2000) A host-specific func-
Permuted tRNA genes expressed via a circular RNA tion is required for ligation of a wide variety of
intermediate in Cyanidioschyzon merolae. Science ribozyme-­processed RNAs. Proc Natl Acad Sci U S A
318:450–453 97:424–429
17. Lykke-Andersen J, Aagaard C, Semionenkov M et al 34. Petkovic S, Müller S (2015) RNA circularization
(1997) Archaeal introns: splicing, intercellular mobil- strategies in vivo and in vitro. Nucleic Acids Res
ity and evolution. Trends Biochem Sci 22:326–331 43:2454–2465
18. Salgia SR, Singh SK, Gurha P et al (2003) Two reac- 35. Barral PM, Sarkar D, Su ZZ et al (2009) Functions
tions of Haloferax volcanii RNA splicing enzymes: of the cytoplasmic RNA sensors RIG-I and MDA-5:
joining of exons and circularization of introns. RNA key regulators of innate immunity. Pharmacol Ther
9:319–330 124:219–234
19. Danan M, Schwartz S, Edelheit S et al (2012) 36. Yoneyama M, Onomoto K, Jogi M et al (2015) Viral
Transcriptome-wide discovery of circular RNAs in RNA detection by RIG-I-like receptors. Curr Opin
Archaea. Nucleic Acids Res 40:3131–3142 Immunol 32:48–53
20. Tang TH, Rozhdestvensky TS, d’Orval BC et al 37. Abelson J, Trotta CR, Li H (1998) tRNA splicing.
(2002) RNomics in Archaea reveals a further link J Biol Chem 273:12685–12688
between splicing of archaeal introns and rRNA pro- 38. Knapp G, Ogden RC, Peebles CL et al (1979) Splicing
cessing. Nucleic Acids Res 30:921–930 of yeast tRNA precursors: structure of the reaction
21. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular intermediates. Cell 18:37–45
RNAs are abundant, conserved, and associated with 39. Lu Z, Filonov GS, Noto JJ, Schmidt CA, Hatkevich
ALU repeats. RNA 19:141–157 TL, Wen Y, Jaffrey SR, Matera AG (2015) Metazoan
4 Circular RNA Splicing 51

tRNA introns generate stable circular RNAs in vivo. 56. Pasman Z, Been MD, Garcia-Blanco MA (1996)
RNA 21:1554–1565 Exon circularization in mammalian nuclear extracts.
40. Suzuki H, Zuo Y, Wang J et al (2006) Characterization RNA 2:603–610
of RNase R-digested cellular RNA source that con- 57. Caldas C, So CW, MacGregor A et al (1998) Exon
sists of lariat and circular RNAs from pre-mRNA scrambling of MLL transcripts occur commonly and
splicing. Nucleic Acids Res 34:e63 mimic partial genomic duplication of the gene. Gene
41. Salzman J, Gawad C, Wang PL, Lacayo N, Brown PO 208:167–176
(2012) Circular RNAs are the predominant transcript 58. Zaphiropoulos PG (1996) Circular RNAs from tran-
isoform from hundreds of human genes in diverse cell scripts of the rat cytochrome P450 2C24 gene: cor-
types. PLoS One 7:e30733 relation with exon skipping. Proc Natl Acad Sci U S
42. Lasda E, Parker R, Parker ROY (2014) Circular A 93:6536–6541
RNAs : diversity of form and function. RNA 59. Baralle FE, Giudice J (2017) Alternative splicing as a
20:1829–1842 regulator of development and tissue identity. Nat Rev
43. Grabowski PJ, Zaug AJ, Cech TR (1981) The inter- Mol Cell Biol 18:437–451
vening sequence of the ribosomal RNA precursor 60. Salzman J, Chen RE, Olsen MN et al (2013) Cell-type
is converted to a circular RNA in isolated nuclei of specific features of circular RNA expression. PLoS
Tetrahymena. Cell 23:467–476 Genet 9:e1003777
44. Durovic P, Dennis PP (1994) Separate pathways for 61. Zhang XO, Dong R, Zhang Y et al (2016) Diverse
excision and processing of 16S and 23S rRNA from alternative back-splicing and alternative splicing land-
the primary rRNA operon transcript from the hyper- scape of circular RNAs. Genome Res 26:1277–1287
thermophilic archaebacterium Sulfolobus acido- 62. Zhang XO, Wang HB, Zhang Y et al (2014)
caldarius: similarities to eukaryotic rRNA processing. Complementary sequence-mediated exon circulariza-
Mol Microbiol 13:229–242 tion. Cell 159:134–147
45. Dennis PP, Ziesche S, Mylvaganam S (1998) 63. Chen LL (2016) The biogenesis and emerging roles of
Transcription analysis of two disparate rRNA operons circular RNAs. Nat Rev Mol Cell Biol 17:205–211
in the halophilic archaeon Haloarcula marismortui. 64. 1000 Genomes Project Consortium, Auton A, Brooks
J Bacteriol 180:4804–4813 LD et al (2015) A global reference for human genetic
46. Kruger K, Grabowski PJ, Zaug AJ et al (1982) Self-­ variation. Nature 526:68–74
splicing RNA: autoexcision and autocyclization of the 65. Lynch M, Conery JS (2003) The origins of genome
ribosomal RNA intervening sequence of Tetrahymena. complexity. Science 302:1401–1404
Cell 31:147–157 66. Khanduja JS, Calvo IA, Joh RI, Hill IT, Motamedi M
47. Grabowski PJ, Seiler SR, Sharp PA (1985) A mul- (2016) Nuclear noncoding RNAs and genome stabil-
ticomponent complex is involved in the splicing of ity. Mol Cell 63:7–20
messenger RNA precursors. Cell 42:345–353 67. Böhmdorfer G, Wierzbicki AT (2015) Control of
48. Qian L, Vu MN, Carter M, Wilkinson MF (1992) A chromatin structure by long noncoding RNA. Trends
spliced intron accumulates as a lariat in the nucleus of Cell Biol 25:623–663
T cells. Nucleic Acids Res 20:5345–5350 68. Magistri M, Faghihi MA, St Laurent G, Wahlestedt
49. Tabak HF, Van der Horst G, Smit J et al (1988) C (2012) Regulation of chromatin structure by long
Discrimination between RNA circles, interlocked noncoding RNAs: focus on natural antisense tran-
RNA circles and lariats using two-dimensional poly- scripts. Trends Genet 28:389–396
acrylamide gel electrophoresis. Nucleic Acids Res 69. Vance KW, Ponting CP (2014) Transcriptional regu-
16:6597–6605 latory functions of nuclear long noncoding RNAs.
50. Hedberg A, Johansen SD (2013) Nuclear group I Trends Genet 30:348–355
introns in self-splicing and beyond. Mob DNA 4:1 70. Wilusz JE, Sunwoo H, Spector DL (2009) Long non-
51. Lambowitz AM, Zimmerly S (2011) Group II introns: coding RNAs: functional surprises from the RNA
mobile ribozymes that invade DNA. Cold Spring world. Genes Dev 23:1494–1504
Harb Perspect Biol 3:1–19 71. Bonasio R, Shiekhattar R (2014) Regulation of tran-
52. Lehmann K, Schmidt U (2003) Group II introns: scription by long noncoding RNAs. Annu Rev Genet
structure and catalytic versatility of large natural ribo- 48:433–455
zymes. Crit Rev Biochem Mol Biol 38:249–303 72. Lee TI, Young RA (2000) Transcription of eukaryotic
53. Li-Pook-Than J, Bonen L (2006) Multiple physical protein-coding genes. Annu Rev Genet 34:77–137
forms of excised group II intron RNAs in wheat mito- 73. Nogales E, Louder RK, He Y (2017) Structural
chondria. Nucleic Acids Res 34:2782–2790 insights into the eukaryotic transcription initiation
54. Soesanto W, Lin HY, Hu E et al (2009) Mammalian machinery. Annu Rev Biophys 46:59–83
target of rapamycin is a critical regulator of cardiac 74. Jurado AR, Tan D, Jiao X et al (2014) Structure and
hypertrophy in spontaneously hypertensive rats. function of pre-mRNA 5′-end capping quality control
Hypertension 54:1321–1327 and 3′-end processing. Biochemistry 53:1882–1898
55. Capel B, Swain A, Nicolis S et al (1993) Circular 75. Moore MJ, Proudfoot NJ (2009) Pre-mRNA process-
transcripts of the testis-determining gene Sry in adult ing reaches back to transcription and ahead to transla-
mouse testis. Cell 73:1019–1030 tion. Cell 136:688–700
52 N. Eger et al.

76. Furuichi Y, Shatkin AJ (2000) Viral and cellular cardiomyocyte Ca2+ signaling and susceptibility to
mRNA capping: past and prospects. Adv Virus Res apoptosis. Circ Res 100:874–883
55:135–184 88. Lanner JT, Georgiou DK, Joshi AD, Hamilton SL
77. Shandilya J, Roberts SGE (2012) The transcription (2010) Ryanodine receptors: structure, expression,
cycle in eukaryotes: from productive initiation to molecular details, and function in calcium release.
RNA polymerase II recycling. Biochim Biophys Acta Cold Spring Harb Perspect Biol 2:a003996
1819:391–400 89. Mabon SA, Misteli T (2005) Differential recruitment
78. Wahle E, Rüegsegger U (1999) 3′-End processing of pre-mRNA splicing factors to alternatively spliced
of pre-mRNA in eukaryotes. FEMS Microbiol Rev transcripts in vivo. PLoS Biol 3:1893–1901
23:277–295 90. Boeckel J-N, Jaé N, Heumüller AW et al (2015)
79. Frendewey D, Keller W (1985) Stepwise assembly Identification and characterization of hypoxia-­
of a pre-mRNA splicing complex requires U-snRNPs regulated endothelial circular RNA. Circ Res
and specific intron sequences. Cell 42:355–367 117(10):884–890
80. Burset M, Seledtsov IA, Solovyev VV (2000) 91. Konarska MM, Padgett RA, Sharp PA (1985)
Analysis of canonical and non-canonical splice Trans splicing of mRNA precursors in vitro. Cell
sites in mammalian genomes. Nucleic Acids Res 42:165–171
28:4364–4375 92. Liang D, Wilusz JE (2014) Short intronic repeat
81. Hwang DY, Cohen JB (1996) Base pairing at the sequences facilitate circular RNA production. Genes
5′ splice site with U1 small nuclear RNA promotes Dev 28:2233–2247
splicing of the upstream intron but may be dispens- 93. Lev-Maor G, Ram O, Kim E et al (2008) Intronic Alus
able for slicing of the downstream intron. Mol Cell influence alternative splicing. PLoS Genet 4:1–12
Biol 16:3012–3022 94. Hu S, Wang X, Shan G (2016) Insertion of an Alu
82. Liu J, Liu T, Wang X, He A (2017) Circles reshaping element in a lncRNA leads to primate-specific mod-
the RNA world: from waste to treasure. Mol Cancer ulation of alternative splicing. Nat Struct Mol Biol
16:58 23:1011–1019
83. Huang S, Yang B, Chen BJ et al (2017) The emerg- 95. Barrett SP, Salzman J (2016) Circular RNAs: analy-
ing role of circular RNAs in transcriptome regulation. sis, expression and potential functions. Development
Genomics 109:401–407 143:1838–1847
84. Starke S, Jost I, Rossbach O, Schneider T et al (2015) 96. Braun S, Domdey H, Wiebauer K (1996) Inverse
Exon circularization requires canonical splice signals. splicing of a discontinuous pre-mRNA intron gener-
Cell Rep 10:103–111 ates a circular exon in a HeLa cell nuclear extract.
85. Gilbert W (1978) Why genes in pieces? Nature Nucleic Acids Res 24:4152–4157
271:501 97. Chen LL, Yang L (2015) Regulation of circRNA bio-
86. Valdivia HH (2007) One gene, many proteins: alter- genesis. RNA Biol 12:381–388
native splicing of the ryanodine receptor gene adds 98. Schindewolf C, Braun S, Domdey H (1996) In vitro
novel functions to an already complex channel pro- generation of a circular exon from a linear pre-mRNA
tein. Circ Res 100:761–763 transcript. Nucleic Acids Res 24:1260–1266
87. George CH, Rogers SA, Bertrand BMA et al (2007) 99. Hesselberth JR (2013) Lives that introns lead after
Alternative splicing of ryanodine receptors modulates splicing. Wiley Interdiscip Rev RNA 4:677–691
 Circular RNAs Biogenesis
in Eukaryotes Through Self-­ 5
Cleaving Hammerhead Ribozymes

Marcos de la Peña

Abstract Abbreviations
Circular DNAs are frequent genomic mole-
cules, especially among the simplest life circRNA circular RNA
beings, whereas circular RNAs have been HHR hammerhead ribozyme
regarded as weird nucleic acids in biology. LTR long terminal repeat
Now we know that eukaryotes are able to PBS primer binding site
express circRNAs, mostly derived from back- PPT polypurine tract
splicing mechanisms, and playing different RT retrotranscriptase
biological roles such as regulation of RNA TSD target site duplication
splicing and transcription, among others.
However, a second natural and highly efficient
pathway for the expression in vivo of cir-
cRNAs has been recently reported, which 1 Introduction
allows the accumulation of abundant small
(100–1000 nt) non-coding RNA circles Genomic circular DNAs are frequent macromol-
through the participation of small self-­cleaving ecules among simple organisms, from small pro-
RNAs or ribozymes called hammerhead ribo- karyotic plasmids to the larger genomes of many
zymes. These genome-encoded circRNAs bacteriophages or viruses, bacteria, archaea and
with ribozymes seem to be a new family of plastids. On the other hand, circular RNAs have
small and nonautonomous retrotransposable been regarded as very rare nucleic acids in biol-
elements of plants and animals (so-called ret- ogy till very recently. Now we know that numer-
rozymes), which will offer functional clues to ous life beings express stable circRNAs [1], and
the biology and evolution of circular RNA among them, it is noteworthy the recent discov-
molecules as well as new biotechnological ery of a myriad of splicing-derived circRNAs in
tools in this emerging field. eukaryotes [2–4] with diverse functions in regu-
lation of splicing [5] and transcription [6], small
Keywords RNAs biology [7], RNA-mediated inheritance
Circular RNA · Retrotransposons · Ribozyme and epigenetics [8] and some others, as described
in this book. However, it has been recently
reported that eukaryotes have a second natural
M. de la Peña (*) pathway that allows the expression in vivo of
IBMCP (CSIC-UPV), Valencia, Spain abundant circular RNAs [9, 10]. This alternative
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 53

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_5
54 M. de la Peña

mechanism does not require any classical spli- 2  he Family of Small Self-­
T
ceosome reaction but the involvement of small Cleaving RNAs
self-cleaving RNAs or ribozymes called ham- and the Singular Case
merhead ribozymes (HHRs) [11]. of the Hammerhead
The finding of catalytic RNAs or ribozymes Ribozyme
more than 30 years ago [12, 13] propelled the
revolution in the RNA field and started with the Among the simplest ribozymes so far described,
uninterrupted discovery of the many differ- it can be highlighted the enigmatic group of small
ent roles and capabilities of this macromolecule (50–200 nt) self-cleaving RNAs, which all
in biology. Moreover, the ground-breaking dis- catalyse a sequence-specific intramolecular reac-
covery of ribozymes strongly supported the tion of transesterification. This reaction starts by
hypothesis of the prebiotic RNA world [14], a nucleophilic attack of the 2′ oxygen to the adja-
where RNAs carried out both informative (RNA cent 3′ phosphate, resulting in cleavage of the
genomes) and catalytic (ribozymes) roles. phosphodiester bond to form two RNA products
Somehow, it is thought that these primal RNA with a 5′-hydroxyl and a 2′,3′-cyclic phosphate
molecules would have evolved to present organ- ends each (Fig. 5.1a). The family of small self-­
isms based in DNA and proteins as the genetic cleaving ribozymes is composed so far by nine
material and catalytic machines, respectively different classes: hammerhead (HHR) [24, 25],
[15–17]. Proofs supporting this hypothesis are hairpin (HPR) [26], human hepatitis-δ (HDV)
the existence of RNA genomes among the sim- [27], Varkud satellite (VS) [28], GlmS [29],
plest organisms (such as RNA viruses and twister [30], twister sister, hatchet and pistol [31]
viroids), as well as catalytic and regulatory ribo-­ ribozymes. The HHR was the first discovered and
functions among all living beings, where RNA one of the best known members of the family of
itself is the final molecule in charge of the activ- small self-cleaving ribozymes. It is composed of
ity. A remarkable example of a catalytic RNA a conserved catalytic core of 15 nucleotides sur-
would be the central machine of life, the ribo- rounded by three double helixes (I to III), which
some [18], which is the universal ribozyme that adopt a γ-shaped fold where helix I interacts with
catalyses the peptide bond formation during pro- helix II through tertiary interactions required for
tein synthesis in all known living entities. This efficient in vivo activity (Fig. 5.1b) [32–34].
fact allows to connect DNA and proteins through There are three possible circularly permuted
a catalytic RNA, which offers a solution to the topologies for the HHR, named type I, type II or
chicken or the egg (or more precisely, the DNA type III, depending on the open-ended helix
or the protein) causality dilemma. Other key (Fig. 5.1c). The HHR were first found encoded in
ribozymes and regulatory RNAs considered the small circRNA genomes of a group of infec-
as ancient relics of the prebiotic RNA world tious subviral agents of plants, such as viral RNA
would be the autocatalytic introns [12] and small satellites and viroids [24, 25], where it catalyses
ribozymes [19], the RNase P [13], the spliceo- a self-cleavage transesterification reaction
some [20], the riboswitches [21], most non-­ required for the rolling-circle replication of these
coding RNAs (such as those small RNA guides pathogens. Surprisingly, few other examples of
found in the CRISPR [22] and the RNAi [23] HHR motifs were also found encoded in the
pathways) or even the circRNAs described in this genomes of some unrelated eukaryotes such as
book, which altogether confirm the extraordinary newts, trematodes or even some mammals,
potential of any RNA molecule present in a living among others [35–39]. In 2010, different labs
organism. reported the widespread occurrence of HHR
5 Circular RNAs Biogenesis in Eukaryotes Through Self-Cleaving Hammerhead Ribozymes 55

Fig. 5.1 RNA self-cleavage by the hammerhead ribo- the three possible hammerhead ribozyme topologies
zyme (a) Mechanism of internal transesterification in the (types I, II and III). Dotted and continuous lines refer to
RNA. The cleavage reaction starts with an attack of the 2′ non-canonical and Watson-Crick base pairs, respectively.
hydroxyl to the 3′ phosphate, followed by a bipyramidal The three topologies have been reported in the genomes of
transition state. The cleavage products are a 2′,3′-cyclic bacteriophages and prokaryotes. Type-I hammerheads are
phosphate at the 5′ RNA product and a 5′-hydroxyl at the mostly found in metazoan genomes, whereas typical type-­
3′ RNA product. (b) Classic two (left)- and three (right)- III motifs are found in the plants and their infectious cir-
dimensional diagrams of the hammerhead ribozyme motif. cRNAs (viroidal RNAs). N stands for any nucleotide,
Black boxes indicate the highly conserved nucleotides (in whereas R stands for purines (A or G), Y for pyrimidines
white letters) at the catalytic core. (c) Representation of (U or C) and H for either A, U or C

motifs in prokaryotic and eukaryotic genomes be more frequent than previously thought.
[40–43], including our own genome [44], which Although the precise biological roles of all these
confirmed that the HHR was a ubiquitous cata- genomic self-cleaving ribozymes are still under
lytic RNA motif in all life kingdoms [45, 46]. study, a direct connection with retrotransposons
Interestingly, the occurrence of genomic HHR and other mobile genetic elements has been
motifs along the tree of life seems to follow a reported for most of them [10, 48–51].
kind of structural or functional compartmental-
ization. This way, anyone of the three topologies
of the HHR motif (types I, II and III, Fig. 5.1c) 3 Hammerhead Ribozymes
can be frequently detected in the genomes of pro- in Plant Genomes Promote
karyotes and bacteriophages. However, metazoan circRNA Expression:
genomes mostly show type-I HHR motifs, The Retrozymes
whereas plant genomes, as well as their subviral
agents, almost exclusively show the presence of Two examples of type-III HHR motifs were
type-III HHR motifs. originally reported in the genome of A. thaliana
Other small self-cleaving RNA motifs have [35]. Numerous copies of this ribozyme were
been also found widespread in DNA genomes, also detected in the genomes of diverse flower-
such as HDV [47] or twister ribozymes [30], ing plants [40]. In many instances, these HHRs
which confirms that small catalytic RNAs would have been found as tandem repeats of several
56 M. de la Peña

Fig. 5.2 Genomic plant

retrozymes (a)
Schematic
representation (top) of a
full genomic retrozyme
element of plants. Target
side duplications (TSDs)
delimiting the retrozyme
are shown in grey boxes.
Long terminal repeats
(LTRs) are shown in
black boxes. The
positions of the primer
binding site (PBS), the
polypurine tract (PPT),
the hammerhead
ribozymes (HHR) and
the typical sizes
encompassed by the
ribozymes are indicated.
The resulting self-­
cleaved retrozyme RNA
after transcription
(middle) and
circularization (bottom)
is indicated. (b) An
example of a northern
blot analysis of RNA
extracts (~30 μg each)
from physic nut
(Jatropha curcas)
leaves, young seedlings
and seeds. Samples were
run on a 5% denaturing
PAGE and were detected
using a digoxigenin-­
labelled J. curcas
retrozyme fragment as a
probe, which revealed
the presence of both
circular and linear RNA
forms in each plant
tissue as indicated at the
right. Ethidium bromide
staining of the 5S rRNA
is shown at the bottom
as a loading control

motifs (usually two or three) separated by a few topology: they are delimited by 4 bp target site
hundred base pairs. These observations have duplications (TSDs), whereas HHRs are embed-
been recently extended in our lab, and we have ded in direct long terminal repeats (LTRs) of
reported the occurrence of hundreds of type-III ~350 bp. LTRs delimit a central region (~300–
HHRs in more than 40 plant species [10]. 700 bp), which begins with the primer binding
Comparative genomics revealed that sequences site (PBS, corresponding to the tRNAMet) and
flanked by tandem HHR motifs sized from 600 finishes with the polypurine tract (PPT)
to 1000 bp with almost no identity. However, sequences characteristic of LTR retrotranspo-
these genomic repetitive elements show a similar sons [52] (Fig. 5.2a). Altogether, these elements
5 Circular RNAs Biogenesis in Eukaryotes Through Self-Cleaving Hammerhead Ribozymes 57

Fig. 5.3 Minimum free energy secondary structure pre- each circRNA structure, and dotted lines indicate pre-
dictions for (a) a retrozyme circRNA of Jatropha curcas dicted tertiary interactions between HHR loops based on
(Entry KX273075.1), (b) the Nepovirus satellite RNA previous models [60, 61]. Self-cleavage sites are indicated
sTRSV (Entry M14879.1) and (c) the viroid CChMVd with arrowheads. Kissing-loop interactions described for
(Entry AJ878085.1). HHR sequences are shown in purple CChMVd [62] are shown. Numbering for each circRNA
(positive polarity) and green (negative polarity). The cor- starts at the self-cleavage site
responding structure of the HHRs motifs are shown under

were classified as a new family of nonautono- RNAs (up to 1 ng per μg of total RNA) of the
mous retrotransposons with hammerhead ribo- precise size encompassed by the HHR motifs
zymes (so-called retrozymes) similar to other (Fig. 5.2b), which strongly indicates an RNA
nonautonomous retroelements of plants like self-processing activity by the ribozymes during
TRIMs [53] and SMARTs [54]. Most likely, in vivo transcription followed by RNA circular-
autonomous retrotransposons of the Ty3-Gypsy ization. Although sequence identity between ret-
family would mobilize the small retrozymes rozymes from non-related plant species is very
based on the sequence s­ imilarities (PBS and 5′ low, secondary structure predictions for these cir-
and 3′ LTR ends) between both types of cRNAs show similar architecture and high stabil-
retroelements. ity (Fig. 5.3a). These structured circRNAs with
Northern blot analysis and RT-PCR experi- type-III hammerhead ribozymes highly resemble
ments of diverse somatic and reproductive tissues those infectious circRNAs of plants, such as viral
from several plant species, such as physic nut, satellite RNAs and viroids (Fig. 5.3b and c),
strawberry, eucalyptus or citrus plants, revealed which indicates a clear evolutionary relationship
the presence of high levels of circular and linear between all of them [9].
58 M. de la Peña

4 Tandem Copies many copies (dozens to even hundreds) of type-I

of Hammerhead Ribozymes HHR motifs (Fig. 5.4). These type-I HHRs not
in Metazoan Genomes only show a characteristic set of tertiary interac-
tions but a very short or even no helix III at all,
Previous bioinformatic searches in metazoan which indicates that these ribozymes may require
genomes have also revealed the widespread occur- in many instances the adoption of dimeric HHR
rence of the HHR motif in animals [10, 40, 45]. As conformations to self-cleave efficiently [55, 24].
observed in plants, these ribozymes are usually The minimal type-I HHRs of metazoan retrozymes
found in close tandem copies, suggesting that highly resemble those described in the pseudo-
genomic retrozymes in animals may express simi- LTRs of the autonomous Penelope-like retroele-
lar circRNA molecules with HHRs, which would ments (PLEs) of metazoans and other eukaryotes
also accumulate in metazoan transcriptomes. [48], which somehow links these two families of
However, several differences can be highlighted retrotransposons. On the other hand, animal retro-
between plant and animal retrozymes. On the one zymes are composed of smaller minimal repeats in
hand, none of the characteristic sequences of plant tandem (150–300 bp), indicating that the expected
retrozymes, such as LTRs, PBS or PPT, are present animal ­ circRNAs are also smaller than those
in their animal counterparts. Moreover, whereas described in plants. These repeats, however, are
plant retrozymes only show a few HHR motifs frequently, but not always, flanked by TSDs as
(usually, just two copies per retrozyme) of the type well, although these are slightly larger (8–12 bp)
III, ribozymes in metazoan retrozymes occur as than those found in plant retrozymes (4 bp) [9, 10].

Fig. 5.4 Metazoan retrozymes tion of three examples of circRNAs derived from meta-
Schematic representation (top) of a typical genomic retro- zoan retrozymes (rotifers, corals and arthropods) are
zyme element present in metazoan genomes. shown (bottom). HHR sequences in the circRNAs are
Target side duplications (TSDs) delimiting the retrozyme shown in purple letters, and the self-cleavage sites are
are shown in grey boxes. Tandem repeats of around 300 indicated with arrowheads. Numbering starts after the
bp are indicated with arrows. Typical type-I HHRs are HHR self-cleavage site
shown. Minimum free energy secondary structure predic-
5 Circular RNAs Biogenesis in Eukaryotes Through Self-Cleaving Hammerhead Ribozymes 59

Recent analysis done in our lab with diverse transcription depending on tissues and/or their
retrozyme-containing metazoans has confirmed genomic location. In any case, nascent RNA tran-
that, as suspected, these organisms accumulate scripts would follow co-transcriptional self-­
abundant circRNAs in most of the analysed tis- processing by tandem self-cleaving HHRs,
sues, in a similar way as described for plants (De producing linear RNAs with 5′-OH and
la Peña and Cervera, to be published). Altogether, 2′,3′-cyclic-phosphate ends. Whereas the step of
the resulting landscape offered by genomic self-cleavage is expected to occur with high effi-
HHRs (either type I or III) in eukaryotes indi- ciency for plant retrozymes carrying type-III
cates that close tandem copies of this ribozyme HHRs, in the case of metazoan retrozymes, self-­
allows the expression of small circRNAs (100– cleavage frequently requires the adoption of a
1000 nt). Although the presence of other small dimeric conformation of minimal type-I HHRs,
self-cleaving ribozymes in eukaryotic genomes which is expected to be slightly less efficient than
have been described, such as HDV and twister the monomeric version [55]. As summarized in
ribozymes [30, 47], the characteristic occurrence Fig. 5.5, covalent circularization of the resulting
in close tandem repeats seems to be exclusively self-cleaved RNAs through either the HHR itself
restricted to the case of the hammerhead or a host RNA ligase factor [57] would finish in
ribozyme. stable circRNAs. As the most plausible model,
these circRNAs are the final template for ret-
rotranscription, whereas linear retrozyme RNAs
5 A Proposed Mechanism would be intermediaries and/or by-products of
for the Expression the circRNAs. In the case of plants, circRNAs
and Spreading of Eukaryotic derived from LTR-like retrozymes could be
circRNAs with Hammerhead primed by any cellular tRNAMet through their
Ribozymes PBS motifs. Then, retrotranscriptases encoded by
Ty3-Gypsy LTR retrotransposons would produce
Retrozymes are a new and atypical group of non- cDNAs of different lengths, thanks to the circular
autonomous eukaryotic retroelements with self-­ nature of the RNA template. In the case of meta-
cleaving hammerhead ribozymes. In plants, zoan retrozymes, retrotranscriptases encoded by
genomic retrozymes resemble other small nonau- non-LTR retrotransposon (such as PLEs or
tonomous LTR retrotransposons such as TRIMs LINEs) would be responsible of carrying out this
[53] and SMARTs [54]. As nonautonomous ret- latter step of cDNA synthesis. Finally, the result-
rotransposons, retrozymes do not show protein-­ ing cDNAs would be integrated in new genomic
coding regions but self-cleaving HHR motifs in locations through the machinery of the autono-
their LTRs, which, most likely, are responsible of mous retrotransposons (Fig. 5.5). A last question
the accumulation in vivo of circular and linear to be addressed is related to the very high levels
RNAs of the precise size encompassed by the of circRNAs with HHRs detected in most organ-
HHRs. Regarding the life cycle of retrozymes, isms analysed. Genomic retrozymes are fre-
the most plausible model would start with the quently found as many copies (from dozens to
transcription of the genomic retrozyme. Similar thousands of repeats) within a given genome,
retroelements, such as TRIMs or autonomous which suggests that even low transcription activ-
LTR retrotransposons, are known to be generally ity would result in abundant levels of circRNAs.
transcribed by RNA Pol II, although examples of However, most of the obtained data indicates that
RNA Pol III-transcribed retrotransposons have only a few retrozyme copies would be
been also reported [56]. Plant and metazoan ret- ­transcriptionally active [10], which suggests that
rozymes do not seem to contain any recognizable the higher stability of these structured circRNAs
promoter, and in consequence, a feasible hypoth- with ribozymes compared with linear RNAs
esis could be that retrozymes may undergo Pol-­ would be the reason of their high levels of accu-
driven (either I, II or III) read-through mulation in vivo. Moreover, the presence of a
60 M. de la Peña

Fig. 5.5 Model for the life cycle of retrozymes cell factors (new biological roles), or retrotranscriptases
A full genomic retrozyme containing at least two HHRs in encoded by autonomous retrotransposons. In the latter
tandem is transcribed (top), and the resulting RNA would case, the resulting cDNAs from retrotranscription of a cir-
self-process through the HHRs to give a linear RNA with cular RNA template would have different lengths depend-
5′-OH and 2′,3′-cyclic phosphate ends. The linear RNA ing on the processivity of the retrotranscriptase. In a final
would be circularized through an RNA ligase activity, and step, the machinery of the retrotransposon would integrate
the resulting circRNA(+) could be recognized for either the retrozyme DNAs at new genomic loci
endogenous RNA polymerases (replication cycles), other

high sequence heterogeneity observed for a pop- major drivers of genome evolution with a role in
ulation of circRNAs in a given organism, together shaping the genomes that they inhabit. In this
with the presence of retrozyme RNAs of the neg- regard, genomic retrozymes and their associated
ative polarity, also suggests the intriguing possi- circRNAs would have similar evolutionary
bility of replication of the circRNAs through impact as any other retroelement. However, the
endogenous polymerases. atypically high accumulation levels of RNA cir-
cles encoded by genomic retrozymes in the tran-
scriptomes of most eukaryotic tissues, either
6 Functional somatic or reproductive, suggest that other bio-
and Biotechnological logical roles can be possible. In this regard, sev-
Applications of circRNAs eral genic circRNAs have been found to play a
with Ribozymes role as microRNA sponges [7], and a comparable
role for retrozyme circRNAs would be feasible.
Retrotransposons, and mobile genetic elements Moreover, the highly structured circRNAs
in general, constitute a major fraction of nuclear derived from genomic retrozymes are suitable to
genomes of most eukaryotes. Historically, these be recognized and processed by the RNAi
genomic sequences have been regarded as junk machinery of the cell, and, consequently, these
DNA, but now we know that retroelements are abundant circRNAs with ribozymes would be
5 Circular RNAs Biogenesis in Eukaryotes Through Self-Cleaving Hammerhead Ribozymes 61

potential templates for the production of miRNA/ However, most of these HHRs in the genomes of
siRNAs with specific regulatory roles in the biol- primitive plants are type-I motifs, which are more
ogy of the organisms where they are expressed. related to those found in metazoan than in angio-
At the same time, we already know many exam- sperm genomes. This observation would indicate
ples of co-option or domestication of the trans- that retrozymes in flowering plants may have a
posable elements by their hosts as adaptations to different origin, either due to a de novo origin in
diverse problems [58]. Usually, these domestica- angiosperms by chance or through horizontal
tions are performed with transposon-derived pro- transfer from other organisms containing type-III
teins, but also small ribozymes such as 3′ UTR HHRs such as bacteria [40–42]. In any case, all
HHRs [39] and intronic HHRs [44] or HDV ribo- these data suggest that genome-encoded cir-
zymes [59] seem to be examples of retroelement cRNAs with HHRs would be more frequent mol-
domestication. Consequently, circRNAs with ecules in eukaryotic transcriptomes than
HHRs in some organisms could have been spe- previously thought. Moreover, circRNAs with
cifically co-opted to play precise functions, a type-III HHRs in plants allow to propose an evo-
possibility that should be studied in the future. lutionary path for the origin of the small infec-
Regarding the biotechnological applications tious circRNAs with HHRs of plants (viroids and
of tandem small ribozymes in the expression of viral satellite RNAs), which may come by chance
circRNAs, it has to be pointed out that the mech- from the abundant reservoirs of circRNAs pres-
anism of backsplicing described for the synthesis ent in plant transcriptomes [9]. In contrast, meta-
of most genic circRNAs seems to be a complex zoan retrozymes with type-I HHRs seem to
pathway, which still requires deeper study in indicate that this HHR topology would be more
order to fully understand and use for practical efficient for circRNA expression in animals.
applications. In this regard, our current knowl- Future in vivo experiments will help us to better
edge about small ribozymes allow us to design understand this new tool in the knowledge of the
much easier approaches for the expression of cir- biology and biotechnology of eukaryotic
cRNAs, which, moreover, could reach higher circRNAs.FundingFunding for this work was
accumulation levels as observed for most eukary- provided by the Ministerio de Economía y
otic retrozymes so far analysed. Moreover, Competitividad of Spain and FEDER funds
in vitro synthesis of circRNAs through self-­ (BFU2014-56094-P and BFU2017-87370-P).
cleaving ribozymes may also offer a straightfor-
ward approach for the production and study of
specific genic circRNAs from eukaryotes. References
1. Lasda E, Parker R (2014) Circular RNAs: diversity of
form and function. RNA 20(12):1829–1842
7 Conclusions and Future 2. Salzman J, Gawad C, Wang PL et al (2012) Circular
Prospects RNAs are the predominant transcript isoform from
hundreds of human genes in diverse cell types. PloS
Genomic retrozymes are a new family of eukary- One 7(2):e30733
3. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular
otic retrotransposons, which spread through cir- RNAs are abundant, conserved, and associated with
cRNAs with hammerhead ribozymes among ALU repeats. RNA 19(2):141–157
plant and animal genomes [10]. In plants, retro- 4. Wang PL, Bao Y, Yee MC et al (2014) Circular RNA
zymes seem to be mostly restricted to eudicots, is expressed across the eukaryotic tree of life. PloS
One 9(6):e90859
although the presence of putative retrozymes 5. Ashwal-Fluss R, Meyer M, Pamudurti NR et al (2014)
with tandem HHR copies were also detected in circRNA biogenesis competes with pre-mRNA splic-
some monocots, primitive land plants (like the ing. Mol Cell 56(1):55–66
spikemoss Selaginella moellendorffii) and algae 6. Memczak S, Jens M, Elefsinioti A et al (2013)
Circular RNAs are a large class of animal RNAs with
(such as Chlamydomonas reinhardtii) [10, 40]. regulatory potency. Nature 495(7441):333–338
62 M. de la Peña

7. Hansen TB, Jensen TI, Clausen BH et al (2013) 26. Buzayan JM, Gerlach WL, Bruening G (1986) Non-­
Natural RNA circles function as efficient microRNA enzymatic cleavage and ligation of RNAs comple-
sponges. Nature 495(7441):384–388 mentary to a plant virus satellite RNA. Nature
8. Talhouarne GJ, Gall JG (2014) Lariat intronic RNAs 323:349–353
in the cytoplasm of Xenopus tropicalis oocytes. RNA 27. Kuo MY, Sharmeen L, Dinter-Gottlieb G et al (1988)
20(9):1476–1487 Characterization of self-cleaving RNA sequences on
9. De la Peña M, Cervera A (2017) Circular RNAs the genome and antigenome of human hepatitis delta
with hammerhead ribozymes encoded in eukary- virus. J Virol 62(12):4439–4444
otic genomes: the enemy at home. RNA Biol 28. Saville BJ, Collins RA (1990) A site-specific self-­
14(8):985–991 cleavage reaction performed by a novel RNA in
10. Cervera A, Urbina D, de la Peña M (2016) Retrozymes Neurospora mitochondria. Cell 61(4):685–696
are a unique family of non-autonomous retrotranspo- 29. Winkler WC, Nahvi A, Roth A et al (2004) Control
sons with hammerhead ribozymes that propagate in of gene expression by a natural metabolite-responsive
plants through circular RNAs. Genome Biol 17(1):135 ribozyme. Nature 428(6980):281–286
11. De la Peña M, Garcia-Robles I, Cervera A (2017) 30. Roth A, Weinberg Z, Chen AG et al (2014) A wide-
The hammerhead ribozyme: a long history for a short spread self-cleaving ribozyme class is revealed by
RNA. Molecules 22(1):78–89 bioinformatics. Nat Chem Biol 10(1):56–60
12. Kruger K, Grabowski PJ, Zaug AJ et al (1982) Self-­ 31. Weinberg Z, Kim PB, Chen TH et al (2015) New
splicing RNA: autoexcision and autocyclization of the classes of self-cleaving ribozymes revealed by
ribosomal RNA intervening sequence of Tetrahymena. comparative genomics analysis. Nat Chem Biol
Cell 31(1):147–157 11(8):606–610
13. Guerrier-Takada C, Gardiner K, Marsh T et al (1983) 32. De la Peña M, Gago S, Flores R (2003) Peripheral
The RNA moiety of ribonuclease P is the catalytic regions of natural hammerhead ribozymes greatly
subunit of the enzyme. Cell 35(3 Pt 2):849–857 increase their self-cleavage activity. EMBO
14. Gilbert W (1986) The RNA world. Nature 319:618 J 22(20):5561–5570
15. Crick FH (1968) The origin of the genetic code. J Mol 33. Khvorova A, Lescoute A, Westhof E et al (2003)
Biol 38(3):367–379 Sequence elements outside the hammerhead ribo-
16. Orgel LE (1968) Evolution of the genetic apparatus. zyme catalytic core enable intracellular activity. Nat
J Mol Biol 38(3):381–393 Struct Biol 10(9):708–712
17. Woese CR (1968) The fundamental nature of the 34. Martick M, Scott WG (2006) Tertiary contacts distant
genetic code: prebiotic interactions between poly- from the active site prime a ribozyme for catalysis.
nucleotides and polyamino acids or their derivatives. Cell 126(2):309–320
Proc Natl Acad Sci U S A 59(1):110–117 35. Przybilski R, Graf S, Lescoute A et al (2005)
18. Steitz TA, Moore PB (2003) RNA, the first macromo- Functional hammerhead ribozymes naturally encoded
lecular catalyst: the ribosome is a ribozyme. Trends in the genome of Arabidopsis thaliana. Plant Cell
Biochem Sci 28(8):411–418 17(7):1877–1885
19. Ferre-D’Amare AR, Scott WG (2010) Small self-­ 36. Ferbeyre G, Smith JM, Cedergren R (1998)
cleaving ribozymes. Cold Spring Harb Perspect Biol Schistosome satellite DNA encodes active hammer-
2(10):a003574 head ribozymes. Mol Cell Biol 18(7):3880–3888
20. Valadkhan S, Manley JL (2001) Splicing-­ 37. Rojas AA, Vazquez-Tello A, Ferbeyre G et al (2000)
related catalysis by protein-free snRNAs. Nature Hammerhead-mediated processing of satellite
413(6857):701–707 pDo500 family transcripts from Dolichopoda cave
21. Winkler W, Nahvi A, Breaker RR (2002) Thiamine crickets. Nucleic Acids Res 28(20):4037–4043
derivatives bind messenger RNAs directly to regulate 38. Epstein LM, Gall JG (1987) Self-cleaving transcripts
bacterial gene expression. Nature 419(6910):952–956 of satellite DNA from the newt. Cell 48(3):535–543
22. Mojica FJ, Diez-Villasenor C, Garcia-Martinez J et al 39. Martick M, Horan LH, Noller HF et al (2008) A
(2005) Intervening sequences of regularly spaced pro- discontinuous hammerhead ribozyme embed-
karyotic repeats derive from foreign genetic elements. ded in a mammalian messenger RNA. Nature
J Mol Evol 60(2):174–182 454(7206):899–902
23. Fire A, Xu S, Montgomery MK et al (1998) Potent and 40. De la Peña M, Garcia-Robles I (2010) Ubiquitous
specific genetic interference by double-stranded RNA presence of the hammerhead ribozyme motif along
in Caenorhabditis elegans. Nature 391(6669):806–811 the tree of life. RNA 16(10):1943–1950
24. Prody GA, Bakos JT, Buzayan JM et al (1986) 41. Jimenez RM, Delwart E, Luptak A (2011) Structure-­
Autolytic processing of dimeric plant virus satellite based search reveals hammerhead ribozymes in the
RNA. Science 231(4745):1577–1580 human microbiome. J Biol Chem 286(10):7737–7743
25. Hutchins CJ, Rathjen PD, Forster AC et al (1986) 42. Perreault J, Weinberg Z, Roth A et al (2011)
Self-cleavage of plus and minus RNA transcripts Identification of hammerhead ribozymes in all
of avocado sunblotch viroid. Nucleic Acids Res domains of life reveals novel structural variations.
14(9):3627–3640 PLoS Comput Biol 7(5):e1002031
5 Circular RNAs Biogenesis in Eukaryotes Through Self-Cleaving Hammerhead Ribozymes 63

43. Seehafer C, Kalweit A, Steger G et al (2011) From involved in restructuring plant genomes. Proc Natl
alpaca to zebrafish: hammerhead ribozymes wherever Acad Sci U S A 98(24):13778–13783
you look. RNA 17(1):21–26 54. Gao D, Chen J, Chen M et al (2012) A highly con-
44. De la Peña M, Garcia-Robles I (2010) Intronic ham- served, small LTR retrotransposon that preferentially
merhead ribozymes are ultraconserved in the human targets genes in grass genomes. PloS one 7(2):e32010
genome. EMBO Rep 11(9):711–716 55. Forster AC, Davies C, Sheldon CC et al (1988) Self-­
45. Hammann C, Luptak A, Perreault J et al (2012) cleaving viroid and newt RNAs may only be active as
The ubiquitous hammerhead ribozyme. RNA dimers. Nature 334(6179):265–267
18(5):871–885 56. Kalendar R, Tanskanen J, Chang W et al (2008)
46. Garcia-Robles I, Sanchez-Navarro J, De la Peña M Cassandra retrotransposons carry independently
(2012) Intronic hammerhead ribozymes in mRNA transcribed 5S RNA. Proc Natl Acad Sci U S A
biogenesis. Biol Chem 393(11):1317–1326 105(15):5833–5838
47. Webb CH, Riccitelli NJ, Ruminski DJ et al (2009) 57. Flores R, Grubb D, Elleuch A et al (2011) Rolling-­
Widespread occurrence of self-cleaving ribozymes. circle replication of viroids, viroid-like satellite RNAs
Science 326(5955):953 and hepatitis delta virus: variations on a theme. RNA
48. Cervera A, De la Peña M (2014) Eukaryotic penelope-­ Biol 8(2):200–206
like retroelements encode hammerhead ribozyme 58. Jangam D, Feschotte C, Betran E (2017) Transposable
motifs. Mol Biol Evol 31(11):2941–2947 element domestication as an adaptation to evolution-
49. Eickbush DG, Eickbush TH (2010) R2 retrotrans- ary conflicts. Trends Genet 33(11):817–831
posons encode a self-cleaving ribozyme for pro- 59. Salehi-Ashtiani K, Luptak A, Litovchick A et al
cessing from an rRNA cotranscript. Mol Cell Biol (2006) A genomewide search for ribozymes reveals
30(13):3142–3150 an HDV-like sequence in the human CPEB3 gene.
50. Ruminski DJ, Webb CH, Riccitelli NJ et al (2011) Science 313(5794):1788–1792
Processing and translation initiation of non-long ter- 60. Chi YI, Martick M, Lares M et al (2008) Capturing
minal repeat retrotransposons by hepatitis delta virus hammerhead ribozyme structures in action by modu-
(HDV)-like self-cleaving ribozymes. J Biol Chem lating general base catalysis. PLoS Biol 6(9):e234
286(48):41286–41295 61. Dufour D, de la Peña M, Gago S et al (2009) Structure-­
51. Kennell JC, Saville BJ, Mohr S et al (1995) The VS function analysis of the ribozymes of chrysanthemum
catalytic RNA replicates by reverse transcription as a chlorotic mottle viroid: a loop-loop interaction motif
satellite of a retroplasmid. Genes Dev 9(3):294–303 conserved in most natural hammerheads. Nucleic
52. Gorinsek B, Gubensek F, Kordis D (2004) Acids Res 37(2):368–381
Evolutionary genomics of chromoviruses in eukary- 62. Gago S, De la Peña M, Flores R (2005) A kissing-­
otes. Mol Biol Evol 21(5):781–798 loop interaction in a hammerhead viroid RNA criti-
53. Witte CP, Le QH, Bureau T et al (2001) Terminal-­ cal for its in vitro folding and in vivo viability. RNA
repeat retrotransposons in miniature (TRIM) are 11(7):1073–1083
 Part IV
Molecular Mechanisms and Gene
Regulation of Circular RNAs
 Circular RNAs Act as miRNA
Sponges 6
Amaresh Chandra Panda

Abstract 1 Introduction
Majority of RNAs expressed in animal cells
lack protein-coding ability. Unlike other cel- RNA molecules were conventionally believed to
lular RNAs, circular (circ)RNAs include a transfer the genetic information coded in the
large family of noncoding (nc)RNAs that lack genomic DNA into specific proteins [1]. However,
the 5′ or 3′ ends. The improvements in high-­ the protein-coding mRNAs represent only ~5% of
throughput RNA sequencing and novel bioin- the human transcriptome, while the rest of the
formatics tools have led to the identification of transcriptome is noncoding (nc)RNAs [2]. The
thousands of circRNAs in various organisms. vast majority are ribosomal (r)RNA and transfer
CircRNAs can regulate gene expression by (t)RNA, both involved in translation [1, 3]. The
influencing the transcription, the mRNA turn- other categories of ncRNAs include microRNAs
over, and translation by sponging RNA-­ (miRNAs), pseudogenes, long (l)ncRNAs, and
binding proteins and microRNAs. Given the circular (circ)RNAs [4–6]. In 1976, electron
broad impact of circRNA on miRNA activity, microscopy of plant viroid discovered covalently
there is huge interest in understanding the closed single-stranded RNA molecules for the
impact of miRNA sponging by circRNA on first time [7]. Later, the hepatitis delta virus
gene regulation. In this review, we summarize (HDV) was found to have a circRNA genome [8].
our current knowledge of the miRNA-­ Another report suggested the expression of
circRNA interaction and mechanisms that scrambled exon RNA from tumor suppressor
influence gene expression. gene DCC in human cells [9]. Due to lack of
RNA-sequencing technologies and inability to
Keywords map the circRNAs to the genome, circRNAs were
mRNA · miRNA · circRNA · Competing completely neglected for last two decades.
endogenous RNA · Translation · miRNA Traditional molecular biology techniques used for
sponge RNA analysis cannot differentiate circRNAs from
linear RNAs [10, 11]. Interestingly, the innova-
tions in next-generation sequencing associated
with new computational pipelines to map cir-
cRNAs to the genome have moved circRNA to
the forefront of RNA research [12–15]. Most of
A. C. Panda (*) the circRNAs are found to be abundant, conserved
Institute of Life Sciences, across species, and often show ­ tissue-­
specific
Bhubaneswar, Odisha, India

© Springer Nature Singapore Pte Ltd. 2018 67

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_6
68 A. C. Panda

expression pattern [13, 16]. Recent studies estab- events including cell proliferation, migration, dif-
lished that covalently closed circRNAs are gener- ferentiation, and apoptosis [29–33]. Given their
ated from the canonical splicing machinery by a importance in gene expression regulation, there is
process called backsplicing [17]. CircRNAs are enormous interest in understanding the regulatory
categorized as exonic (E), intronic (I), and exon- mechanisms that can regulate miRNA function.
intron (EI) circRNAs based on the primary tran- Accumulating evidence indicates that circRNAs
script sequence they are generated from [15, play a crucial role in gene expression regulation
17–19]. circRNAs are very stable due to lack of partly by inhibiting miRNA activity (Fig. 6.1). In
the 5′-3′ ends which makes them resistant to exo- this review, several circRNA-­miRNA interactions
nucleases [20, 21]. Recent studies reported that and their function are listed as follows (Table 6.1).
some may act as sponges for miRNAs, sponges
for RBPs, compete with linear splicing, and trans- CDR1as The CDR1as is generated from the
lated into peptides [4, 18, 22]. Growing evidence antisense transcript of CDR1 gene (also termed
indicated that circRNAs involved in various cel- as ciRS-7). CDR1as was the first miRNA sponge
lular events including proliferation, differentia- reported to negatively regulate miR-7 and found
tion, apoptosis, and metastasis [23]. This review to be expressed in brain tissue, neuroblastoma,
briefly discusses the impact of circRNAs on key astrocytoma, HeLa cells, and lung carcinoma
cellular processes by acting as a competing [14, 34–36]. The CDR1as has more than 60 bind-
endogenous RNA (ceRNA) for target miRNAs. ing sites for miR-7, and its resistance to miRNA-­
mediated RNA degradation makes this a perfect
ceRNA for miR-7 [12, 14]. Expression of
2 MiRNA Sponging by circRNA CDR1as inhibits miR-7 activity that leads to
increase in expression of miR-7 targets.
MicroRNAs are small, evolutionarily conserved Coexpression of CDR1as and miR-7 is necessary
ncRNA molecules predicted to regulate ~30% of for sponging activity of circRNA. The sponging
the protein-coding genes [24, 25]. There are of miR-7 by ciRS-7 is reported to affect the
~1800 miRNAs which have been identified in expression of ubiquitin protein ligase A (UBE2A)
humans [26]. The miRNA genes are transcribed in Alzheimer’s disease [37]; Myrip and Pax6 in
into primary miRNA (pri-miRNA) by RNA poly- insulin secretion and synthesis, respectively [38];
merase II followed by processing with Drosha and epidermal growth factor receptor (EGFR) in
and DGCR8 to generate pre-miRNA [27]. The cancer [36, 14, 39]. Upregulation of CDR1as in
pre-miRNA is exported to the cytoplasm and pro- gastric cancer suppresses miR-7 activity which
cessed by Dicer to produce ~19–22 nt mature leads to more aggressive oncogenic phenotype
miRNA. The functional miRNA is incorporated mediated by PTEN/PI3K/AKT pathway [40].
in the effector RNA-induced silencing complex Another study reported the upregulation of
(RISC) and activates the RISC complex to bind CDR1as and downregulation of miR-7 in hepato-
the target mRNA that has sequence complemen- cellular carcinoma (HCC) tissue compared with
tarity with the loaded miRNA. The miRNAs usu- the adjacent non-tumor tissues [34]. The overex-
ally target the 3′ untranslated regions (UTRs) of pression of miR-7 or silencing of CDR1as inhib-
specific mRNA targets and regulate their stability ited the HCC cell proliferation and invasion by
and/or translation [28]. The level of complemen- inhibiting the expression of target genes CCNE1
tarity between the mRNA and miRNA determines and PIK3CD. Together, CDR1as act as an onco-
the mechanism of miRNA action on target gene in HCC through sponging miR-7 [34].
mRNA. As each miRNAs can have complemen-
tarity with many mRNAs, they have the potential
to regulate multiple genes [29]. miRNAs are cir-SRY The sex-determining region Y gene
shown to be involved in posttranscriptional regu- produces a circRNA known as cir-SRY. The cir-­
lation of gene expression in nearly all cellular SRY is highly expressed in adult mouse testis
6 Circular RNAs Act as miRNA Sponges 69

Fig. 6.1 Schematic representation of circRNA biogene- Most circRNA sponges are enriched in miRNA-binding
sis and their impact on gene expression by sponging sites. Overexpression of circRNA leads to inactivation of
miRNA miRNAs, thereby upregulating miRNA target gene
(a) The canonical splicing machinery generates mature expression. (d) The decrease in circRNA expression
linear RNA from the primary transcript by removing allows higher levels of functional miRNAs to suppress the
intervening introns. (b) The circRNA is generated by the expression of mRNAs containing miRNA-binding sites
“head-to-tail” backsplicing of the circularizing exons. (c)

[41]. A recent study reported that there are 16 cir-ITCH Cir-ITCH (itchy E3 ubiquitin protein
putative miR-138-binding sites present in cir-­ ligase) is generated from the ICTH gene. Cir-­
SRY which can act as ceRNA and thus poten- ICTH was reported to enrich in miRNA regula-
tially regulate expression of miR-138 target tory elements (MREs) for miR-7, miR-17, and
genes [14, 42]. miR-214 and act as a sponge for these miRNAs,
70 A. C. Panda

Table 6.1 Potential circRNA-miRNA-mRNA regulatory networks

CircRNA name Sponged miRNA miRNA target gene Diseases/tissue References
CDR1as/ miR-7 UBE2A Alzheimer’s disease [37]
ciRS-7
CDR1as/ miR-7 Myrip and Pax6 Diabetes [38]
ciRS-7
CDR1as/ miR-7 CCNE1 and Hepatocellular carcinoma [34]
ciRS-7 PIK3CD
CDR1as/ miR-7 EGFR Glioblastoma [14, 39]
ciRS-7
CDR1as/ miR-7 PI3K Gastric cancer [40]
ciRS-7
cir-SRY miR-138 TWIST2 Colorectal cancer [14, 42]
cir-ITCH miR-7, miR-17, ITCH Esophageal squamous cell carcinoma, [43–46]
and miR-214 bladder, lung, and colorectal cancer
circHIPK3 miR-124 Aquaporin 3 Hepatocellular carcinoma [48]
circHIPK3 miR-558 Heparanase Bladder cancer [49]
circHIPK3 miR-379 IGF-1 Non-small cell lung cancer [50]
circPVT1 let-7 IGF2BP1, KRAS, Senescence [51]
and HMGA2
circPVT1 miR-125 E2F2 Gastric cancer [52]
circRNA-CER miR-136 MMP13 Osteoarthritis [53]
circRNA-­ miR-29a VEGFA Bladder carcer [54]
MYLK
circTCF25 miR-103a-3p and CDK6 Bladder cancer [55]
miR-107
circHIAT1 for CDC42 Clear cell renal cell carcinoma [56]
miR-195-5p/29a-­
3p/29c-3p
HRCR miR-223 ARC Cardiac hypertrophy [57]
cir-ZNF609 miR-150-5p AKT3 Hirschsprung disease [58]
hsa_ miR-145 BAG4, E2F5, and Colorectal cancer [59]
circ_001569 FMNL2
hsa_ miR-29c-3p Osteosarcoma [60]
circ_001564
circVMA21 miR-200c XIAP Intervertebral disc degeneration [61]
circRNA_ miR-138-5p MMP13, COX-2, Osteoarthritis [62]
Atp9b and IL-6
circDOCK1 miR-196a-5p BIRC3 Oral squamous cell carcinoma [63]
circMTO1 miR-9 P21 Hepatocellular carcinoma [64]
hsa_ miR-129-5p Notch1 Hepatocellular carcinoma [65]
circ_0005986
hsa_ miR-106b CDK6 Colorectal cancer [66]
circ_000984
hsa_ miR-138 TERT and PD-L1 Colorectal cancer [67]
circ_0020397
hsa_ miR-449a IL6R Osteosarcoma [68]
circ_0009910
circGFRA1 miR-34a GFRA1 Triple negative breast cancer [69]
hsa-­ miR-24 caspase-1 Osteosarcoma [70]
circ-­0016347
circWDR77 miR-124 FGF2 Vascular smooth muscle cells [71]
(continued)
6 Circular RNAs Act as miRNA Sponges 71

Table 6.1 (continued)

CircRNA name Sponged miRNA miRNA target gene Diseases/tissue References
circACTA2 miR-548f-5p α-SMA Vascular smooth muscle cells [72]
hsa_ miR-22 ErbB3 Lung adenocarcinoma [73]
circ_0012673
circRNA_ miR-424 LATS1 Gastric cancer [74]
LARP4
circ-ABCB10 miR-1271 Breast cancer [75]

leading to upregulation of miRNA target gene The circHIPK3 overexpression in NCI-H1299

ITCH. Cir-ITCH was reported to have antitumor promoted cell proliferation and circHIPK3
activity by suppressing the Wnt/β-catenin signal- silencing in NCI-H2170-inhibited cell prolifera-
ing by regulating ICTH expression in various tion. CircHIPK3 was found to sequester miR-379
cancers including esophageal squamous cell car- and increase the expression levels of miR-379
cinoma and bladder, lung, and colorectal cancer target IGF1, leading to increase in cell prolifera-
[43–46]. tion [50].

circHIPK3 The second exon of homeodomain-­ circPVT1 Circular RNA expression pattern in
interacting protein kinase 3 (HIPK3) gene gener- proliferating (early-passage) and senescent (late-­
ates a circRNA called circHIPK3 that can act as passage) human diploid WI-38 fibroblasts were
a sponge for nine miRNAs including miR-124. analyzed and identified hundreds of differentially
Silencing of CircHIPK3 reduced cell growth expressed senescence-associated circRNAs
through suppressing miR-124 activity which (SAC-RNAs) [51]. One of the SAC-RNA called
directly interacts with circHIPK3 [47]. circPVT1 was significantly downregulated in
CircHIPK3 was upregulated in HCC tissues [48]. senescent fibroblasts. Further, circPVT1 silenc-
CircHIPK3 acted as a sponge for miR-124 and ing in proliferating fibroblasts promoted cellular
suppressed miR-124 activity in HCC, leading to senescence. CircPVT1 selectively sponged let-7
upregulation of miR-124 target gene aquaporin 3 and promoted the expression of let-7 target genes
(AQP3). Further, increase in AQP3 expression including IGF2BP1, KRAS, and HMGA2.
promoted cell proliferation and migration in Together, these data suggested that the SAC-­
HCC cells. Together, these results suggested that RNA circPVT1, elevated in the proliferating
circHIPK3 regulated HCC growth through the cells, inhibits endogenous let-7 activity to enable
miR-124-AQP3 axis [48]. CircHIPK3 was also a proliferative phenotype [51]. Another report
found to be downregulated in bladder cancer tis- suggested that the circPVT1 is often upregulated
sues compared with normal bladder tissues, and in gastric cancer (GC) tissues compared with
the level of circHIPK3 negatively correlates with matched normal tissues [52]. The circPVT1 acted
bladder cancer grade [49]. Increase in circHIPK3 as a sponge for the members of the miR-125 fam-
level led to inhibition of migration, invasion, and ily and promoted cell proliferation. Further, cir-
angiogenesis of bladder cancer cells in vitro. cPVT1 expression level was correlated with the
CircHIPK3 acted as a ceRNA for miR-558 and survival of patients with gastric cancer. In sum,
inhibit miR-558 activity, thereby regulating the circPVT1 acts as a proliferative factor and prog-
expression of heparanase (HPSE) in bladder can- nostic marker in gastric cancer [52].
cer cells [49]. Another study reported the expres-
sion pattern of circHIPK3 in six non-small cell
lung cancer (NSCLC) cell lines [50]. The NSCLC circRNA-CER A recent study explored the cir-
cell lines NCI-H2170 and NCI-H1299 had the cRNA expression pattern and function of chon-
highest and lowest expression level, respectively. drocyte extracellular matrix (ECM)-related
72 A. C. Panda

circRNAs (circRNA-CER) in cartilage. Several miR-223 to suppress cardiac hypertrophy. ARC

circRNAs were differentially expressed in osteo- was identified to be the downstream target to
arthritis samples compared to normal cartilage. mediate the function of miR-223 in cardiac hyper-
The circRNA-CER expression was upregulated trophy. The circRNA HRCR was found to seques-
in chondrocytes upon interleukin-1 (IL-1) and ter and inhibit the function miR-223, leading to
tumor necrosis factor α (TNFα) treatment [53]. upregulation of ARC expression which is involved
The circRNA-CER was found to have five puta- in cardiac hypertrophy and heart failure [57].
tive MREs for miR-636, miR-665, miR-217,
miR-646, and miR-136. The MMP13 gene was
regulated by miR-136 which is sponged by cir-ZNF609 In Hirschsprung disease (HSCR) the
circRNA-­CER. The silencing of circRNA-CER expression of cir-ZNF609 was lower as compared
increased chondrocyte extracellular matrix for- with normal bowel tissues. The silencing of cir-
mation by inhibiting MMP13 expression [53]. ZNF609 led to suppression of proliferation and
migration of cells. Further, cir-ZNF609 was found
to regulate the expression of AKT3 by acting as a
circRNA-MYLK A recent study found that the decoy for miR-150-5p. These findings suggest that
circRNA-MYLK and VEGFA were significantly the cir-ZNF609-miR-150-5p-AKT3 axis plays a
upregulated in bladder carcinoma. The circRNA-­ critical role in the onset of HSCR [58].
MYLK directly binds to miR-29a that leads to
increase in expression of miR-29a target
VEGFA. Overexpression of circRNA-MYLK has_circ_001569 The expression of hsa_
activated VEGFA/VEGFR2 and downstream circ_001569 was upregulated in colorectal can-
Ras/ERK signaling pathway by acting as a cer tissues and predicted to sponge miR-145. The
ceRNA for miR-29a [54]. miR-145 can modulate the expression of BAG4,
E2F5, and FMNL2 transcripts in colorectal can-
cer cells. Expression of circ_001569 upregulated
circTCF25 The circRNA circTCF25 was pre- the expression of BAG4, E2F5, and FMNL2 by
dicted to sponge miR-103a-3p and miR-107. The sponging miR-145, leading to colorectal cancer
overexpression of circTCF25 sequestered miR-­ cell proliferation and invasion [59].
103a-­3p and miR-107, leading to upregulation of
their target CDK6 which in turn enhanced the pro-
liferation and migration of bladder cancer cells [55]. hsa_circ_001564 The circRNA hsa_
circ_0001564 is derived from the gene called
CANX and significantly upregulated in osteosar-
circHIAT1 The expression of circHIAT1 was coma cell lines. Silencing of hsa_circ_0001564
downregulated in clear cell renal cell carcinoma reduced the proliferation by inducing cell cycle
(ccRCC) compared to the adjacent normal tis- arrest and apoptosis in HOS and MG-63 cells.
sues. Androgen receptor (AR) downregulated the Further, hsa_circ_0001564 was found to be a
expression of circHIAT1 by suppressing tran- sponge for miR-29c-3p which could reverse the
scription of its host gene, hippocampus abundant tumorigenic effect of circ_0001564. These data
transcript 1 (HIAT1). The circHIAT1 may act as suggested that hsa_circ_0001564 plays a critical
a sponge for miR-195-5p/29a-3p/29c-3p to mod- role in osteosarcoma by acting as a ceRNA for
ulate CDC42 expression which is linked to miR-29c-3p [60].
ccRCC cell migration and invasion [56].

circVMA21 The role of circVMA21 was

HRCR A recent work suggested that a heart-­ explored in nucleus pulposus (NP) cells and
related circRNA (HRCR) could act as a sponge for degenerative NP tissues from intervertebral disc
6 Circular RNAs Act as miRNA Sponges 73

degeneration (IVDD) patients. The circVMA21 HCC tissues and positively correlated with sur-
was found to directly interact with miR-200c vival of HCC patients. Biochemical assays
which inhibits the expression of the target gene revealed the interaction of miR-9 with circ-
X-linked inhibitor-of-apoptosis protein (XIAP). MTO1in HCC cells. The circMTO1 silencing led
The NP cell function and viability was regulated to the suppression of p21 which is a target of
by miR-200c through suppression of miR-9, leading to increase in proliferation and
XIAP. Together, circVMA21 could inhibit NP invasion of HCC cells. Taken together, these data
cell apoptosis through miR-200c-XIAP axis [61]. suggest that circMTO1 inhibits HCC progression
by acting as ceRNA for miR-9 to upregulate the
expression of p21 [64].
circRNA_Atp9b In osteoarthritis, circRNA_
Atp9b was overexpressed in mouse chondrocytes
upon interleukin-1 beta (IL-1β) treatment. hsa_circ_0005986 The circRNA hsa_
Further, circRNA_Atp9b silencing upregulated circ_0005986 was found to be downregulated in
type II collagen expression and suppressed the HCC tissue samples compared with adjacent
expression of MMP13, COX-2, and IL-6. miR-­ non-tumorous tissues [65]. Furthermore, hsa_
138-­5p was found to be sponged by circRNA_ circ_0005986 expressions were significantly
Atp9b, and their expression levels are negatively downregulated in HCC cell lines, HepG2,
correlated. Moreover, the effects of circRNA_ SMMC7721, and HCCLM3 compared to normal
Atp9b on extracellular matrix catabolism and hepatic cell line L02. Interestingly, the level of
inflammation were partly reversed by inhibition hsa_circ_0005986 downregulation was found to
of miR-138-5p. In sum, these data suggested that be correlated with Barcelona clinic liver cancer
the extracellular matrix in chondrocytes is regu- (BCLC) stage, chronic hepatitis B family history,
lated by circRNA_Atp9b through sponging miR-­ and tumor diameters. miR-129-5p was one of the
138-­5p [62]. miRNA that could be sponged by hsa_
circ_0005986 and regulated the target gene
Notch1. Silencing of hsa_circ_0005986 increased
circDOCK1 A TNF-α-induced apoptotic model miR-129-5p activity and downregulated Notch1
was developed for oral squamous cell carcinoma expression, leading to increase in cell prolifera-
(OSCC) to study the impact of circDOCK1 on tion and tumorigenesis in HCC [65].
apoptosis. The silencing of circDOCK1 increased
the apoptosis in OSCC cells. Bioinformatics anal-
ysis predicted the interaction of circDOCK1 with hsa_circ_000984 The circular RNA hsa_
miR-196a-5p which targets BIRC3. Interestingly, circ_000984 generated from the CDK6 gene was
both overexpression of miR-­196a-­5p or knock- significantly overexpressed in colorectal cancer
down of circDOCK1 led to suppression of BIRC3 (CRC) tissues from patients as well as in the
which is a negative regulator of apoptosis. CRC cell lines [66]. Further, the expression level
Together, these results suggested that the apopto- of hsa_circ_000984 was positively associated
sis of OSCC cells is regulated by circDOCK1 with CRC advancement. The knockdown of hsa_
through sponging miR-196a-5p [63]. circ_000984 expression led to inhibition of cell
proliferation, migration, and invasion in CRC
cell lines. Hsa_circ_000984 could act as a sponge
circMTO1 A recent study reported the circRNA for miR-106b, leading to upregulation of miR-­
expression profile in HCC and identified circ- 106b target CDK6. Together, these data suggest
MTO1, generated from the gene mitochondrial that the hsa_circ_000984 could upregulate CDK6
translation optimization 1 (MTO1). The level of by inhibiting miR-106b activity, thereby promot-
circMTO1 was found to be downregulated in ing colon cancer growth and metastasis [66].
74 A. C. Panda

hsa_circ_0020397 Another study reported that hsa-circ-0016347 Hao J et al. reported that the
the circRNA hsa_circ_0020397 was upregulated, circ-0016347 acted as a ceRNA for miR-214 in
while its target miR-138 was downregulated in osteosarcoma cell leading to upregulation of
CRC cells. Furthermore, overexpression of hsa_ miR-24 target caspase-1. Moreover, circ-0016347
circ_0020397 could inhibit the miR-138 activity was found to induce proliferation and invasion of
indicating that hsa_circ_0020397 act as a ceRNA osteosarcoma cells. These data suggested that the
for miR-138. The hsa_circ_0020397 promoted circ-0016347 plays a crucial role in osteosar-
the expression of miR-138 targets telomerase coma progression by acting as a sponge for miR-­
reverse transcriptase (TERT) and programmed 124, which could be used as a potential target for
death-ligand 1 (PD-L1) by sponging miR-138, development of therapy for osteosarcoma [70].
thereby promoting cell viability and invasion of
CRC cells. Together, these data suggest that hsa_
circ_0020397 plays a crucial role in CRC patho- circWDR77 The circRNA expression profiling
genesis by acting as a sponge for miR-138 [67]. in glucose-induced vascular smooth muscle cells
(VSMCs) discovered hundreds of differentially
expressed circRNAs. CircWDR77 is one of the
hsa_circ_0009910 In osteosarcoma, the hsa_ upregulated circRNAs whose silencing led to
circ_0009910 expression was found to be upreg- inhibition of proliferation and migration of
ulated and silencing of circ_0009910 promoted VSMCs. Computational prediction suggested the
cell cycle arrest and apoptosis. The miR-449a interaction of circWDR77 with miR-124.
expression was found to be suppressed in osteo- Furthermore, circWDR77 inhibited miR-124
sarcoma cells and predicted to be sponged by activity and upregulated the expression of miR-­
circ_0009910. Further, miR-449a could target 124 target fibroblast growth factor 2 (FGF2) in
and downregulate the expression of IL6R in VSMCs. Together, these results indicated that the
osteosarcoma cells, thereby promoting inhibition proliferation and migration of VSMCs were reg-
of cell proliferation, cell cycle arrest, and apopto-ulated through circWDR77/miR-124/FGF2 axis
sis. The transcript level of IL6R was also found to [71].
be negatively correlated with the level of miR-­
449a in osteosarcoma cells. Taken together, these
data suggested that the carcinogenesis of osteo- circACTA2 Neuregulin-1 (NRG-1) was found
sarcoma cells was induced by the circ_0009910/ to be upregulated and cleaved in response to
miR-449a/IL6R axis [68]. transforming growth factor-β1 in VSMCs.
NRG-1 was also known to promote the expres-
sion of an extracellular epidermal growth factor-­
circGFRA1 A little is known about the role of like domain and intracellular domain
circRNA in triple negative breast cancer (TNBC). (NRG-1-ICD) which induced circular ACTA2
A recent study reported that the circGFRA1 was (alpha-actin-2; circACTA2) expression. Further,
upregulated in TNBC cell lines and tissues. circACTA2 acted as a sponge for miR-548f-5p
Kaplan-Meier survival analysis suggested that which upregulated α-SMA expression, leading to
the level of circGFRA1 was negatively correlated increase in stress fiber formation and cell con-
with survival. CircGFRA1 silencing led to the traction in VSMCs. Together, these data indicated
suppression of proliferation and induced apopto- that circACTA2 fine-tunes the α-SMA expression
sis in TNBC cells. Further, biochemical assays and VSMC contraction through the NRG-1-ICD/
found that circGFRA1 can directly bind and circACTA2/miR-548f-5p axis [72].
inhibit miR-34a activity, leading to increase in
miR-34a target gene GFRA1. In sum, circG-
FRA1 act as a sponge for miR-34a to regulate hsa_circ_0012673 The hsa_circ_0012673
GFRA1 expression in TNBC [69]. expression was upregulated in lung adenocarci-
6 Circular RNAs Act as miRNA Sponges 75

noma (LAC) tissues compared with adjacent 3 Web Tools for Analysis
non-tumor tissues. Further, the expression level of miRNA-circRNA
of circ_0012673 was positively correlated with Interaction
tumor size. Biochemical assays revealed that
hsa_circ_0012673 promoted LAC proliferation Besides in-depth studies on functional circRNAs,
by acting as a sponge for miR-22, which inhibits several circRNA databases have been developed
erb-b2 receptor tyrosine kinase 3 (ErbB3) [73]. to explore the interaction of circRNAs with miR-
NAs. The databases such as CircInteractome,
Circ2Traits, CircNet, and StarBase v2.0 provide
circRNA_LARP4 The large tumor suppressor excellent platforms to predict the miRNA-­
kinase 1 (LATS1) acts as a tumor suppressor in circRNA interactions (Table 6.2).
gastric cancer by regulating the Hippo signaling
pathway. The circRNA_ LARP4 was downregu- circInteractome The CircInteractome (http://
lated in gastric cancer and predicted to sponge circinteractome.nia.nih.gov) database is the first
miR-424 computationally. The direct interaction web tool developed to predict the miRNAs-­
between miR-424 and LATS1 or circLARP4 was circRNA interactions for circRNAs listed in
verified by various biochemical assays. CircBase [76–78]. Furthermore, this is the only
Overexpression of miR-424 suppressed the web tool available to date for designing divergent
expression of miR-424 target gene LATS1 which primer and siRNA against circRNAs. Together,
led to increase in proliferation and invasion of the CircInteractome web tool provides bioinfor-
gastric cancer cells. In sum, circLARP4 act as a matic analyses of miRNA-binding sites on cir-
novel tumor suppressive factor by sponging miR-­ cRNAs and predicts potential miRNA sponge
424-­5p which modulate the expression of circRNAs that are crucial for posttranscriptional
LATS1 in gastric cancer cells [74]. gene regulation [76].

circ-ABCB10 The circ-ABCB10 was signifi- circ2Trait The circ2Traits (http://gyanxet-beta.

cantly overexpressed in breast cancer tissue. com/circdb/) is a database of 1951 human cir-
Further experiments suggested that silencing of cRNAs and their association with 105 human dis-
circ-ABCB10 inhibited the proliferation and pro- eases [79]. The circRNAs were categorized based
moted apoptosis of breast cancer cells. on number of disease-associated SNPs, AGO inter-
Bioinformatics and biochemical analysis found action sites, and interaction with disease-­associated
that miR-1271 can be sponged by circ-ABCB10. miRNA. Circ2Trait also checks the enrichment of
Furthermore, the effect of circ-ABCB10 on sets of genes in the miRNA-­circRNA interactome
breast cancer cells was rescued by miR-1271. that is associated with ­ particular diseases.
These data indicated that circ-ABCB10 promotes Circ2Trait also provides the complete information
breast cancer pathogenesis via sponging miR-­ of miRNA-circRNA-­ mRNA-lncRNA interaction
1271 [75]. networks for the human diseases.

Table 6.2 Web tools for prediction of circRNA-miRNA interactions

Database name URL References
CircInteractome https://circinteractome.nia.nih.gov/ [76, 77]
Circ2Traits http://gyanxet-beta.com/circdb/ [79]
CircNet http://circnet.mbc.nctu.edu.tw/ [80]
StarBase v2.0 http://starbase.sysu.edu.cn/ [81]
76 A. C. Panda

CircNet The CircNet (http://circnet.mbc.nctu. transcripts, and their translatability into proteins.
edu.tw/) is a database to explore circRNA expres- With ever-increasing evidence of functional cir-
sion in specific tissues, circRNAs isoforms, cir- cRNA and discovery of novel molecular mecha-
cRNA sequences, and circRNA-miRNA nisms of action, there is no doubt that circRNAs
interactions. The CircNet database was the first will be used for disease diagnosis and treatment
database to report the expression of tissue-­ in near future.
specific circRNAs and circRNA-miRNA-gene
interaction network. In sum, the CircNet is an Acknowledgments This work was supported by the
interactive web interface for visualization and Science and Engineering Research Board, a statutory
analysis of the regulatory network of circRNA, body of the Department of Science and Technology
(DST), Government of India (SERB/F/6890/2017-18).
miRNA, and genes [80].
Conflicts of Interest The authors have no conflicts of
interest to declare.
starBase v2.0 The starBase v2.0 (http://star-
base.sysu.edu.cn/) is the first database to system-
atically identify the regulatory RNA-RNA and
protein-RNA interaction networks using the References
experimentally supported CLIP-Seq data. This
study identified ∼9000 miRNA-circRNA regula- 1. Crick F (1970) Central dogma of molecular biology.
tory interactions. Moreover, starBase v2.0 pro- Nature 227(5258):561–563
2. Derrien T, Johnson R, Bussotti G et al (2012) The
vides CLIP-supported miRNA target sites for GENCODE v7 catalog of human long noncoding
interacting circRNAs. This web server predicts RNAs: analysis of their gene structure, evolution, and
the functional interaction of miRNA-circRNA expression. Genome Res 22(9):1775–1789
and their coordinated regulatory networks [81]. 3. Noller HF (1991) Ribosomal RNA and translation.
Annu Rev Biochem 60:191–227
4. Hsiao KY, Sun HS, Tsai SJ (2017) Circular RNA –
new member of noncoding RNA with novel functions.
4 Conclusion and Future Exp Biol Med (Maywood) 242(11):1136–1141
5. Wery M, Kwapisz M, Morillon A (2011) Noncoding
Directions RNAs in gene regulation. Wiley Interdiscip Rev Syst
Biol Med 3(6):728–738
With the advancement in circRNA enrichment 6. Yoon JH, Abdelmohsen K, Gorospe M (2013)
and sequencing technology, a huge number of Posttranscriptional gene regulation by long noncod-
ing RNA. J Mol Biol 425(19):3723–3730
circRNAs have been identified in various organ- 7. Sanger HL, Klotz G, Riesner D et al (1976) Viroids are
isms, and the number will most likely increase. single-stranded covalently closed circular RNA mol-
The circRNAs are currently one of the focus ecules existing as highly base-paired rod-like struc-
areas in biological research, and the field is still tures. Proc Natl Acad Sci U S A 73(11):3852–3856
8. Kos A, Dijkema R, Arnberg AC et al (1986) The
in its early stages. There is huge interest in how hepatitis delta (delta) virus possesses a circular
circRNAs are generated and what are their bio- RNA. Nature 323(6088):558–560
logical functions. Although thousands of cir- 9. Nigro JM, Cho KR, Fearon ER et al (1991) Scrambled
cRNAs have been identified, only a handful of exons. Cell 64(3):607–613
10. Jeck WR, Sharpless NE (2014) Detecting and charac-
circRNAs has been reported to have a biological terizing circular RNAs. Nat Biotechnol 32(5):453–461
function. Almost all studies on circRNAs largely 11. Panda AC, Abdelmohsen K, Gorospe M (2017)
focused on their interactions with miRNAs and RT-qPCR detection of senescence-associated circular
RBPs. The expression of vast number and types RNAs. Methods Mol Biol 1534:79–87
12. Memczak S, Jens M, Elefsinioti A et al (2013)
of circular RNAs increases the difficulty level for Circular RNAs are a large class of animal RNAs with
understanding their regulatory mechanisms. regulatory potency. Nature 495(7441):333–338
Further intensive studies are required to under- 13. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular
stand the biogenesis of circRNAs, functional RNAs are abundant, conserved, and associated with
ALU repeats. RNA 19(2):141–157
interaction of circRNAs with genome and mRNA
6 Circular RNAs Act as miRNA Sponges 77

14. Hansen TB, Jensen TI, Clausen BH et al (2013) 33. Munk R, Panda AC, Grammatikakis I et al (2017)
Natural RNA circles function as efficient microRNA Senescence-associated microRNAs. Int Rev Cell Mol
sponges. Nature 495(7441):384–388 Biol 334:177–205
15. Panda AC, De S, Grammatikakis I et al (2017) High-­ 34. Yu L, Gong X, Sun L et al (2016) The circular RNA
purity circular RNA isolation method (RPAD) reveals Cdr1as act as an oncogene in hepatocellular carci-
vast collection of intronic circRNAs. Nucleic Acids noma through targeting miR-7 expression. PLoS One
Res 45(12):e116 11(7):e0158347
16. Xia S, Feng J, Lei L et al (2017) Comprehensive 35. Tang W, Ji M, He G et al (2017) Silencing CDR1as
characterization of tissue-specific circular RNAs inhibits colorectal cancer progression through
in the human and mouse genomes. Brief Bioinform regulating microRNA-7. Onco Targets Ther
18(6):984–992 10:2045–2056
17. Chen LL, Yang L (2015) Regulation of circRNA bio- 36. Peng L, Yuan XQ, Li GC (2015) The emerging land-
genesis. RNA Biol 12(4):381–388 scape of circular RNA ciRS-7 in cancer (review).
18. Panda AC, Grammatikakis I, Munk R et al (2017) Oncol Rep 33(6):2669–2674
Emerging roles and context of circular RNAs. Wiley 37. Lukiw WJ (2013) Circular RNA (circRNA) in
Interdiscip Rev RNA 8(2):e1413 Alzheimer’s disease (AD). Front Genet 4:307
19. Li Z, Huang C, Bao C et al (2015) Exon-intron cir- 38. Xu H, Guo S, Li W et al (2015) The circular RNA
cular RNAs regulate transcription in the nucleus. Nat Cdr1as, via miR-7 and its targets, regulates insulin
Struct Mol Biol 22(3):256–264 transcription and secretion in islet cells. Sci Rep
20. Chen LL (2016) The biogenesis and emerging 5:12453
roles of circular RNAs. Nat Rev Mol Cell Biol 39. Kefas B, Godlewski J, Comeau L et al (2008)
17(4):205–211 microRNA-7 inhibits the epidermal growth factor
21. Zhang Y, Xue W, Li X et al (2016) The biogenesis of receptor and the Akt pathway and is down-regulated
nascent circular RNAs. Cell Rep 15(3):611–624 in glioblastoma. Cancer Res 68(10):3566–3572
22. Abdelmohsen K, Panda AC, Munk R et al (2017) 40. Pan H, Li T, Jiang Y et al (2018) Overexpression of
Identification of HuR target circular RNAs uncovers circular RNA ciRS-7 abrogates the tumor suppres-
suppression of PABPN1 translation by CircPABPN1. sive effect of miR-7 on gastric cancer via PTEN/
RNA Biol 14(3):361–369 PI3K/AKT signaling pathway. J Cell Biochem
23. Haque S, Harries LW (2017) Circular RNAs (cir- 119(1):440–446
cRNAs) in health and disease. Genes (Basel) 41. Capel B, Swain A, Nicolis S et al (1993) Circular
8(12):353 transcripts of the testis-determining gene Sry in adult
24. Berezikov E, Guryev V, van de Belt J et al (2005) mouse testis. Cell 73(5):1019–1030
Phylogenetic shadowing and computational 42. Long L, Huang G, Zhu H et al (2013) Down-regulation
identification of human microRNA genes. Cell of miR-138 promotes colorectal cancer metastasis via
120(1):21–24 directly targeting TWIST2. J Transl Med 11:275
25. Rajewsky N (2006) microRNA target predictions in 43. Huang G, Zhu H, Shi Y et al (2015) cir-ITCH
animals. Nat Genet 38(Suppl):S8–S13 plays an inhibitory role in colorectal cancer by
26. Kozomara A, Griffiths-Jones S (2011) miRBase: regulating the Wnt/beta-catenin pathway. PLoS One
integrating microRNA annotation and deep-­ 10(6):e0131225
sequencing data. Nucleic Acids Res 39(Database 44. Yang C, Yuan W, Yang X et al (2018) Circular RNA
issue):D152–D157 circ-ITCH inhibits bladder cancer progression by
27. Siomi H, Siomi MC (2010) Posttranscriptional regu- sponging miR-17/miR-224 and regulating p21, PTEN
lation of microRNA biogenesis in animals. Mol Cell expression. Mol Cancer 17(1):19
38(3):323–332 45. Li F, Zhang L, Li W et al (2015) Circular RNA
28. Pillai RS (2005) MicroRNA function: multiple mech- ITCH has inhibitory effect on ESCC by suppress-
anisms for a tiny RNA? RNA 11(12):1753–1761 ing the Wnt/beta-catenin pathway. Oncotarget
29. Zhou X, Yang PC (2012) MicroRNA: a small 6(8):6001–6013
molecule with a big biological impact. Microrna 46. Wan L, Zhang L, Fan K et al (2016) Circular RNA-­
1(1):1 ITCH suppresses lung cancer proliferation via inhib-
30. Lee KP, Shin YJ, Panda AC et al (2015) miR-431 iting the Wnt/beta-catenin pathway. Biomed Res Int
promotes differentiation and regeneration of old 2016:1579490
skeletal muscle by targeting Smad4. Genes Dev 47. Zheng Q, Bao C, Guo W et al (2016) Circular RNA
29(15):1605–1617 profiling reveals an abundant circHIPK3 that regu-
31. Panda AC, Abdelmohsen K, Gorospe M (2017) SASP lates cell growth by sponging multiple miRNAs. Nat
regulation by noncoding RNA. Mech Ageing Dev Commun 7:11215
168:37–43 48. Chen G, Shi Y, Liu M et al (2018) circHIPK3 regu-
32. Panda AC, Sahu I, Kulkarni SD et al (2014) miR-­ lates cell proliferation and migration by sponging
196b-­mediated translation regulation of mouse insu- miR-124 and regulating AQP3 expression in hepato-
lin2 via the 5′UTR. PLoS One 9(7):e101084 cellular carcinoma. Cell Death Dis 9(2):175
78 A. C. Panda

49. Li Y, Zheng F, Xiao X et al (2017) Circ HIPK3 63. Wang L, Wei Y, Yan Y et al (2018) CircDOCK1 sup-
sponges mi R-558 to suppress heparanase presses cell apoptosis via inhibition of miR196a5p
expression in bladder cancer cells. EMBO Rep by targeting BIRC3 in OSCC. Oncol Rep
18(9):1646–1659 39(3):951–966
50. Tian F, Wang Y, Xiao Z et al (2017) Circular RNA 64. Han D, Li J, Wang H et al (2017) Circular RNA circ-
circHIPK3 promotes NCI-H1299 and NCI-H2170 MTO1 acts as the sponge of microRNA-9 to suppress
cell proliferation through miR-379 and its target hepatocellular carcinoma progression. Hepatology
IGF1. Zhongguo Fei Ai Za Zhi 20(7):459–467 66(4):1151–1164
51. Panda AC, Grammatikakis I, Kim KM et al (2017) 65. Fu L, Chen Q, Yao T et al (2017) Hsa_
Identification of senescence-associated circular circ_0005986 inhibits carcinogenesis by acting as
RNAs (SAC-RNAs) reveals senescence suppressor a miR-129-5p sponge and is used as a novel bio-
CircPVT1. Nucleic Acids Res 45(7):4021–4035 marker for hepatocellular carcinoma. Oncotarget
52. Chen J, Li Y, Zheng Q et al (2017) Circular RNA 8(27):43878–43888
profile identifies circPVT1 as a proliferative factor 66. Xu XW, Zheng BA, Hu ZM et al (2017) Circular
and prognostic marker in gastric cancer. Cancer Lett RNA hsa_circ_000984 promotes colon cancer growth
388:208–219 and metastasis by sponging miR-106b. Oncotarget
53. Liu Q, Zhang X, Hu X et al (2016) Circular RNA 8(53):91674–91683
related to the chondrocyte ECM regulates MMP13 67. Zhang XL, Xu LL, Wang F (2017) Hsa_circ_0020397
expression by functioning as a MiR-136 ‘Sponge’ in regulates colorectal cancer cell viability, apopto-
human cartilage degradation. Sci Rep 6:22572 sis and invasion by promoting the expression of the
54. Zhong Z, Huang M, Lv M et al (2017) Circular RNA miR-138 targets TERT and PD-L1. Cell Biol Int
MYLK as a competing endogenous RNA promotes 41(9):1056–1064
bladder cancer progression through modulating 68. Deng N, Li L, Gao J et al (2018) Hsa_circ_0009910
VEGFA/VEGFR2 signaling pathway. Cancer Lett promotes carcinogenesis by promoting the expression
403:305–317 of miR-449a target IL6R in osteosarcoma. Biochem
55. Zhong Z, Lv M, Chen J (2016) Screening differ- Biophys Res Commun 495(1):189–196
ential circular RNA expression profiles reveals the 69. He R, Liu P, Xie X et al (2017) circGFRA1 and
regulatory role of circTCF25-miR-103a-3p/miR- GFRA1 act as ceRNAs in triple negative breast can-
107-CDK6 pathway in bladder carcinoma. Sci Rep cer by regulating miR-34a. J Exp Clin Cancer Res
6:30919 36(1):145
56. Wang K, Sun Y, Tao W et al (2017) Androgen recep- 70. Jin H, Jin X, Zhang H et al (2017) Circular RNA
tor (AR) promotes clear cell renal cell carcinoma hsa-circ-0016347 promotes proliferation, invasion
(ccRCC) migration and invasion via altering the and metastasis of osteosarcoma cells. Oncotarget
circHIAT1/miR-195-5p/29a-3p/29c-3p/CDC42 sig- 8(15):25571–25581
nals. Cancer Lett 394:1–12 71. Chen J, Cui L, Yuan J et al (2017) Circular RNA
57. Wang K, Long B, Liu F et al (2016) A circular RNA WDR77 target FGF-2 to regulate vascular smooth
protects the heart from pathological hypertrophy muscle cells proliferation and migration by spong-
and heart failure by targeting miR-223. Eur Heart ing miR-124. Biochem Biophys Res Commun
J 37(33):2602–2611 494(1–2):126–132
58. Peng L, Chen G, Zhu Z et al (2017) Circular RNA 72. Sun Y, Yang Z, Zheng B et al (2017) A novel regula-
ZNF609 functions as a competitive endogenous RNA tory mechanism of smooth muscle alpha-actin expres-
to regulate AKT3 expression by sponging miR-150-5p sion by NRG-1/circACTA2/miR-548f-5p axis. Circ
in Hirschsprung’s disease. Oncotarget 8(1):808–818 Res 121(6):628–635
59. Xie H, Ren X, Xin S et al (2016) Emerging roles of 73. Wang X, Zhu X, Zhang H et al (2018) Increased
circRNA_001569 targeting miR-145 in the prolif- circular RNA hsa_circ_0012673 acts as a sponge
eration and invasion of colorectal cancer. Oncotarget of miR-­ 22 to promote lung adenocarcinoma
7(18):26680–26691 proliferation. Biochem Biophys Res Commun
60. Song YZ, Li JF (2018) Circular RNA hsa_ 496(4):1069–1075
circ_0001564 regulates osteosarcoma proliferation 74. Zhang J, Liu H, Hou L et al (2017) Circular RNA_
and apoptosis by acting miRNA sponge. Biochem LARP4 inhibits cell proliferation and invasion of gas-
Biophys Res Commun 495(3):2369–2375 tric cancer by sponging miR-424-5p and regulating
61. Cheng X, Zhang L, Zhang K et al (2018) Circular LATS1 expression. Mol Cancer 16(1):151
RNA VMA21 protects against intervertebral disc 75. Liang HF, Zhang XZ, Liu BG et al (2017) Circular
degeneration through targeting miR-200c and X RNA circ-ABCB10 promotes breast cancer prolifera-
linked inhibitor-of-apoptosis protein. Ann Rheum Dis tion and progression through sponging miR-1271. Am
77(5):770–779 J Cancer Res 7(7):1566–1576
62. Zhou ZB, Du D, Huang GX et al (2018) Circular RNA 76. Dudekula DB, Panda AC, Grammatikakis I et al
Atp9b, a competing endogenous RNA, regulates the (2016) CircInteractome: a web tool for exploring
progression of osteoarthritis by targeting miR-138-5p. circular RNAs and their interacting proteins and
Gene 646:203–209 microRNAs. RNA Biol 13(1):34–42
6 Circular RNAs Act as miRNA Sponges 79

77. Panda AC, Dudekula DB, Abdelmohsen K et al 80. Liu YC, Li JR, Sun CH et al (2016) CircNet: a
(2018) Analysis of circular RNAs using the web tool database of circular RNAs derived from tran-
CircInteractome. Methods Mol Biol 1724:43–56 scriptome sequencing data. Nucleic Acids Res
78. Glazar P, Papavasileiou P, Rajewsky N (2014) 44(D1):D209–D215
circBase: a database for circular RNAs. RNA 81. Li JH, Liu S, Zhou H et al (2014) starBase v2.0:
20(11):1666–1670 decoding miRNA-ceRNA, miRNA-ncRNA and
79. Ghosal S, Das S, Sen R et al (2013) Circ2Traits: a protein-­RNA interaction networks from large-scale
comprehensive database for circular RNA potentially CLIP-Seq data. Nucleic Acids Res 42(Database
associated with disease and traits. Front Genet 4:283 issue):D92–D97
 Regulation of Transcription
by Circular RNAs 7
Rumela Bose and Rupasri Ain

Abstract regulation on various cell transcriptome

Circular RNAs (circRNAs) are a class of non- remains largely unknown. In this chapter, we
coding RNA that are present in wide variety of will review the regulatory effects of circRNAs
cells in various tissue types across species. in the transcription of their own or other genes.
They are non-polyadenylated, single-stranded, Also, we will discuss the association of cir-
covalently closed RNAs. CircRNAs are more cRNAs with miRNAs and RNA-binding pro-
stable than other RNAs due to lack of 5′ or 3′ teins (RBPs), with special reference to
end leading to resistance to exonuclease diges- Drosophila circMbl and their role as an
tion. The length of circRNAs varies from 1 to “mRNA trap,” which might play a role in its
5 exons with retention of introns in mature cir- regulatory potential transcriptionally or
cRNAs with ~25% frequency. They are pri- posttranscriptionally.
marily found in the cytosol within the cell
although the mechanism of their nuclear Keywords
export remains elusive. However, there is a circRNAs · Noncoding RNA · Transcription
subpopulation of circRNAs that remain in the regulation · Splicing · miRNA sponge
nucleus and regulate RNA-Pol-II-mediated
transcription. Bioinformatic approaches min-
ing RNA sequencing data enabled genome-­
wide identification of circRNAs. In 1 Introduction
mammalian genome over 20% of the expressed
genes in cells and tissues can produce these With the advent of our understanding of molecu-
transcripts. Owing to their abundance, stabil- lar regulation of gene expression in last three
ity, and diverse expression profile, circRNAs decades, the layers of complexity have increased
likely play a pivotal role in regulatory path- in every possible step in gene regulation. Circular
ways controlling lineage determination, cell RNAs have emerged as a new player of gene reg-
differentiation, and function of various cell ulation intervening in transcriptional as well as
types. Yet, the impact of circRNA-mediated posttranscriptional processes. Although transport
of circRNAs is not well understood, their cellular
R. Bose · R. Ain (*) compartmentalization dictates their role in gene
Division of Cell Biology and Physiology, CSIR-­ regulation.
Indian Institute of Chemical Biology, Circular RNAs exist in diverse isoforms in
Kolkata, West Bengal, India wide variety of organisms ranging from
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 81

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_7
82 R. Bose and R. Ain

s­ elf-­replicating viroids to humans. The simplest derived from rRNA and tRNA intronic sequences
type of circular RNAs with 250–400 nucleotides [3]. The other types of circRNAs are primarily
exist as single-stranded RNA genome of viroids, found in higher eukaryotic cells. Most of the cir-
which are plant pathogens. Yet another example cular RNAs in eukaryotes are produced by a pro-
of viral genome circular RNA is that of hepatitis cess termed as backsplicing, wherein a
delta virus, which is a human pathogen, and downstream splice donor site is joined to an
length of the circle is approximately 1.7 kb. upstream splice acceptor site to form circular
Second types of circRNAs primarily comprised molecules (Fig. 7.1).
of excised intronic circles. Based on the type of This gives rise to three broadly defined iso-
introns, they can be of four different types: (a,b) forms: circular intronic RNAs (ciRNA), exonic
circRNAs containing either excised group I or circular RNAs (ecircRNA), and exon-intron cir-
group II introns formed by ribozymes in some cular RNAs (EIciRNA) which have different
eukaryotes, bacteria, and viruses as RNA pro- modes of biogenesis (Fig. 7.1). Intronic cir-
cessing by-product; (c) circular intronic RNAs cRNAs are derived from intron lariats produced
(ciRNAs) formed in eukaryotes by spliceosome during pre-mRNA splicing [4], exonic circular
mediated 2′-5′ branchpoint attack followed by RNAs are produced co-transcriptionally compro-
degradation of downstream intronic sequence; mising the linear mRNA splicing [5], while the
and (d) circRNAs containing excised tRNA exon-intron circRNAs form circles with the inter-
introns found in some archaea as tRNA process- vening intron retained between them.
ing by-product. The third type of circRNAs arise In addition, eukaryotic circRNAs may also
either as by-product of rRNA processing or as arise either by trans-splicing or exon scrambling.
intermediate during tRNA processing in some In trans-splicing event, exons from two separate
archaea and algae [1]. The first discovery of an mRNAs are joined together, followed by
intron-derived circular RNA was made from an backsplicing-­mediated rearrangements of the
intervening sequence in the pre-rRNA transcript exons in a circular orientation [6]. Scrambled
of the ciliate Tetrahymena [2]. Transcriptome-­ exons may arise due to genomic rearrangement,
wide study of the archaeon Sulfolobus solfatari- tandem duplication, trans-splicing, or backsplic-
cus P2 revealed several novel circular transcripts ing [1]. With the help of RT-PCR with divergent

Fig. 7.1 Biogenesis and diversity of circular RNAs or multiple exons (b) or retain an unspliced intron between
Circular RNAs are produced by backsplicing in which a two exons (c). Sometimes they can be entirely composed
downstream splice donor site is joined to an upstream of an intron (d). Colored bars, exons; black lines, introns;
splice acceptor site. These circular RNAs show a wide SD splice donor, SA splice acceptor
range of diverse isoforms. They may consist of a single (a)
7 Regulation of Transcription by Circular RNAs 83

primers, as well as computational algorithms, the be retained in the nucleus in a manner similar to
order of some exons in matured eukaryotic incompletely spliced mRNAs [16]. This is a very
mRNAs were found to be inconsistent with interesting yet unexplored area of circular RNA
respect to their gene sequences. Such exons are biology.
called scrambled exons. Patrick Brown and his Nuclear localized circRNAs (ciRNAs,
team statistically evaluated the origin of circular EIciRNAs) primarily function at the transcription
RNAs from scrambled exons. About 32% of the level as demonstrated by their interaction with
scrambled exon junctions indicated toward the Pol II and U1 snRNP complex. On the other
occurrence of a hypothetical circular RNA asso- hand, certain circRNAs can alternatively splice
ciated with them. They tested for six genes their own pre-mRNA to produce circular tran-
including nine scramble exon-containing iso- scripts. This circularization of the exons inhibits
forms. Consistent with the statistical findings, the production of the canonical protein from their
each of these transcripts were resistant to RNaseR locus thus regulating the expression of these
further confirming their circular nature [7]. genes [17]. Majority of the circRNAs are exported
Bioinformatic analysis by Jeck et al. showed to the cytoplasm, where they interact with RBPs
that sequences flanking exons that form circular and miRNAs sequestering them into specific cel-
RNAs are more likely to contain complementary lular locations [18–21]. These circRNAs have
Alu elements [8]. More than 12,000 and 6000 cir- several conserved binding sites for RBPs and
cRNAs have been found in Arabidopsis thaliana miRNAs, allowing them to posttranscriptionally
and Oryza sativa, respectively, which are derived regulate expression of the target genes. The list of
from intronic, exonic, as well as other sequences circRNAs acting as RBP and miRNA sponge is
[9]. CircRNAs accumulate during heat stress in given in the Table 7.1 and 7.2.
Arabidopsis, implying its potential role in heat Although circRNAs are “noncoding” RNAs,
stress [10]. Evolutionary significance of cir- there are instances where they may code protein.
cRNAs lies in the fact that several of them are One such naturally occurring protein encoding
conserved in sequences as well as function within circular RNA is found in the hepatitis delta virus,
same or different species. For example, circular where a circRNA codes for a 122 amino acid pro-
RNAs produced from the HIPK family are con- tein in infected mammalian cells [22]. Insertion
served between humans and rodents [8]; muscle- of internal ribosomal entry site (IRES) in cir-
blind circular RNA regulates alternative splicing cRNAs leads to translation by endogenous ribo-
across Drosophila species in a conserved manner some [23]. In their experiment, the ribosomes
[5]. Various databases have been designed to traversed the circles repeatedly and gave rise to
identify and for quick access to unified data about high molecular weight multimeric polypeptides.
different aspects of circRNAs. Some of the com- By introducing thrombin cleavage site upstream
monly used databases are circ2Traits [11], cir- of the IRES sequence, these high molecular
cBase [12], circNet [13], circRNADb [14], etc. weight products could be cleaved to produce sin-
Once produced, circRNAs are either exported gle polypeptides of 40kD size (as was predicted
to the cytoplasm or retained inside the nucleus. to be encoded by the circular RNA) [23].
Molecular mechanisms as well as factors govern-
ing circRNAs’ nuclear export are not well under-
stood. As the potential function of a circRNA is 2 The Transcription Machinery
much dependent on the cellular milieu to which it in Eukaryotes
is exposed, it is important to determine its export
mechanism. Emerging evidences suggest that 2.1 Eukaryotic RNA Polymerases
they may surpass the nucleus during mitosis [6]
or are exported by mRNA transporters recruited Eukaryotic nuclei contain three RNA polymer-
at the exon-exon junction during splicing [15]. ases which have different locations within the
Additionally, intron-containing EIciRNAs may nucleus and also different responses to salt and
84 R. Bose and R. Ain

Table 7.1 CircRNAs that act as RBP sponge

Name Protein it binds Mode of action Reference
circMbl MBL Regulates its own biosynthesis [5]
circPABPN1 HuR Prevents binding of HuR to PABPN1 mRNA and lowers [49, 50]
its translation
circFoxo3 Id-1, E2F1, Reduced nuclear and mitochondrial translocation of [50, 51]
HIF-α, and FAK these factors during cardiac stress causing cardiac
senescence
P21, CDK2 Facilitates inhibition of CDK2 by p21, inhibiting cell [52]
cycle progression
MDM2, p53 Facilitates MDM2-mediated ubiquitination of p53 [18]
circANRIL PES1 Competes with rRNA binding of PES1 impairing [50, 53]
ribosome biogenesis
circAmotl1 c-myc, STAT3, Translocates them to the nucleus affecting transcription [17,
PDK1, AKT1 19–21]
ciRS-7/Cdr1as AGO Inhibits miRNA-mediated degradation of target mRNAs [54]
circ CDYL, circNFATC3, IMP-3 _ [55]
circANKRD17

Table 7.2 CircRNAs that act as miRNA sponge

Name miRNA it binds Mode of action Reference
ciRS-7/ miR-7 Prevents downregulation of target genes [54]
Csr1as
circZNF609 miR-150-5p Modulates AKT expression in Hirschsprung’s disease [56]
circSry miR-138 Regulates hypoxia-induced apoptosis in cardiac myocytes [17, 57]
circZNF91 miR-23, Prevents downregulation of ZNF225, ZNF486, and ZNF85 (miR-23) [5, 58]
miR-181, tumor suppressor genes RB1 and RBAK (miR-181) ZNF20 and
miR-199 ZNF791 (miR-199)
circHRCR miR-223 Removes the translation inhibition of apoptosis inhibitor with CARD [17, 59]
domain (ARC)
circMFACR miR-652-3p Regulates mitochondrial dynamics, apoptosis of cardiac myocytes, and [17, 60]
myocardial infarction
circWDR77 miR-124 Vascular smooth muscle cell proliferation [17, 61]
hsa-­ miR-19a Decreased apoptosis in human aortic smooth muscle cells [62]
circ-­000595
hsa-­ miR-145 Positively regulates cell proliferation [63]
circ-­001569

divalent cations, and RNA polymerase I, which Pol II, have been discovered till date [25, 26].
transcribes rRNA genes, is located in the nucleo- Richard Young categorized the originally identi-
lus, while RNA polymerase II and III, which fied subunits into three broad classifications: the
mainly transcribe mRNA and tRNAs, respec- core subunits which are indispensible for
tively, are located in the nucleoplasm. These ­polymerase structure and function, the common
three enzymes are differently sensitive to the poi- subunits which are found in all three polymer-
son alpha-amanitin [24]. The subunit structures ases, and the nonessential subunits, which are
of these three polymerases from different eukary- conditionally dispensable for enzyme activity.
otes have been well studied. All these structures The three polypeptides Rpb1, Rpb2, and Rpb3
contain multiple subunits, some of which are are absolutely required for enzyme activity. The
common to all three polymerases. Twelve sub- Rpb1 subunit, identified as the functional homo-
units of yeast, Saccharomyces cerevisiae, RNA logue of the bacterial polymerase β’ subunit, is
7 Regulation of Transcription by Circular RNAs 85

involved in DNA binding. Under physiological the transcription start site. However, TATA-less
conditions, the Rpb1 subunit exists in two iso- promoters are frequently found in the cells, as in
forms IIo and IIa based upon the state of phos- the case of house-keeping genes and develop-
phorylation. The amino acid sequence of IIa mentally regulated homoeotic genes. Promoter
subunit shows a repeat string of seven amino proximal elements include CCAAT box and GC
acids with the following consensus sequence: box (GGGCGG) located at ~100 bp and ~200 bp
Tyr-Ser-Pro-Thr-Ser-Pro-Ser. Because this upstream of the transcription start site. The core
sequence is found at the carboxy terminus of the promoter drives basal level of transcription and is
IIa subunit, it is named as the carboxy terminal the binding site for TATA-binding protein (TBP)
domain (CTD). This CTD is likely to get phos- and associated factors (TAFs), whereas the pro-
phorylated at the Ser, Thr residues, transforming moter proximal elements regulate true level of
IIa to the IIo subunit. The existence of two forms transcription and are binding sites for gene-­
of Rpb1 subunit in the cells implies that they specific transcription factors. Yet another cis-­
serve different purpose in transcription. It is acting element that regulates transcription is
indeed the case wherein Pol II with Rpb1 IIa enhancer element located upstream or down-
(referred to as Pol IIA) can bind to promoter and stream of transcription start site. Activators or
thus initiate transcription, and Pol II with Rpb1 repressors of transcription bind to these elements
IIo (referred to as Pol IIo) is the species that car- and regulate rate of transcription.
ries out elongation [27, 28]. The Rpb2 subunit is The class I promoters, recognized by Pol I, are
involved in nucleotide binding at the active site of not well conserved across species. It consists of a
the enzyme in all the three polymerases. There is core element surrounding the transcription start
one 20-amino acid region of Rpb3 that shares site and an upstream control element about
great similarity to E. coli α-subunit. Also same 100 bp further upstream. The spacing between
kinds of polymerase assembly defects are seen in the two elements is very important as insertion or
RPB3 mutant yeasts as in E. coli α-subunit deletion of bases between them greatly reduces
mutants. Thus it has been predicted that Rpb3 is promoter strength [29]. The classical genes are
required for the appropriate assembly of the RNA transcribed by Pol III promoters that are located
polymerase holoenzyme. Five subunits – Rpb5, within the genes. The internal promoter of 5S
Rpb6, Rpb8, Rpb10, and Rpb12 – are common to rRNA gene is split into three regions: box A, a
all three polymerases and serve general purpose short intermediate element, and box C. One class
of transcription like processivity or fidelity. The III gene, 7SL, contains a weak internal promoter
two nonessential subunits Rpb4 and 9 are not and a sequence at the 5′-flanking region of the
absolutely required for polymerase activity under gene that is required for high level transcription.
normal conditions, but RPB4 and 9 mutants are Other class III genes (e.g., 7SK and U6 RNA)
inviable at high temperatures. completely lack internal promoters and contain
class II-like promoters that lie in the 5′-flanking
region and contains TATA box [30].
2.2 Eukaryotic Promoters

The three polymerases have different structures 2.3 Transcription Factors

and transcribe different genes and therefore rec-
ognize different promoters. The promoters recog- General transcription factors drive basal level of
nized by Pol II are termed as class II promoters transcription by binding to core promoter ele-
which have two parts: the core promoter and an ment and the RNA polymerase to form a pre-­
upstream proximal element. The core promoters initiation complex (PIC). The minimal PIC
generally constitute of the TATA box, an initiator includes RNA pol-II and six general transcription
site located approximately 25–30 bp upstream of factors that are TFIIA, TTFIIB, TFIID, TFIIE,
86 R. Bose and R. Ain

TFIIF, and TFIIH. Binding and sequential recruit- 3.1 Regulation of Transcription
ment of TFs have been demonstrated by various at Initiation Step: Role
scientists using gel mobility shift and various of Exon–Intron Circular RNAs
other assays [31, 32]. The largest general TF,
TFIID, contains various subunits that include Eukaryotic transcription can be primarily tuned
TBP and 16 TAFs. The saddle-shaped TBP binds at the initiation step, which involves formation of
to the promoter in the DNA minor groove creat- the pre-initiation complex at the promoter.
ing a bend in the DNA followed by recruitment Almost all known transcription factors assemble
of TFIIA and TFIIB, respectively. TBP mutants at this point further stabilizing the complex, stim-
are not only deficient in class II genes but also in ulating the rate of transcription. To test whether
class I and III genes suggesting that TBP is a uni- noncoding RNAs can regulate transcription, Li
versal transcription factor required by all three et al. [41] performed cross-linking followed by
RNA Pols. RNA Pol-II binds to TFIIF to form the immunoprecipitation (CLIP) using RNA Pol
Pol-II complex. TFIIB recruits the PolII complex II-specific antibody. RNA sequencing of Pol II
to the promoter and helps the complex bind cor- CLIP samples and further bioinformatics analy-
rectly. This is followed by binding of TFIIE and ses revealed as many as 111 circRNAs to be asso-
TFIIH binding to the complex resulting in forma- ciated with Pol II. Out of these 111 circRNAs, 15
tion of basal PIC. TFIIH subunits possess ATPase were EIciRNAs. Fluorescence in situ hybridiza-
and helicase activity that create negative super tion (FISH) revealed that 2 of these 15 EIciRNAs,
helical tension resulting in unwinding of one turn circEIF3J and circPAIP2, were exclusively local-
of DNA to form the transcription bubble. TFIIA, ized in the nucleus. Knocking down of these two
B, E, F, and H leave once RNA elongation begins, EIciRNAs, using either short interfering RNA
but TFIID stays till the end of elongation [33, (siRNA) or RNase H-based antisense oligonucle-
34]. otides (ASO), resulted in decrease in the parent
In addition to general transcription factors that transcripts (eif3j and paip2) in both HeLa and
drive basal level of transcription, there are about HEK293 cells, without any effect on the neigh-
2600 transcription factors coded by human boring genes’ transcripts (ctdspl2 and matr3). To
genome. These transcription factors possess understand whether this decrease was due to
DNA binding domain and transcription activa- decrease in the transcription of the respective
tion domain and interact with other proteins and mRNAs, nuclear run-on experiments were per-
increase the level of transcription as much as formed with nuclei extracted from circEIF3J and
100-fold. They drive tissue-specific and cell type-­ circPAIP2 knockdown cells. It was found that
specific expression or repression of various knockdown of circEIF3J and circPAIP2 indeed
genes. resulted in lower EIF3J and PAIP2 transcription,
whereas knockdown of EIF3J and PAIP2 with
siRNA had no effect on their transcription. Not
3  ircular RNAs Act as Potent
C only that circEIF3J and circPAIP2 were found to
Regulators of Transcription co-localize with the genomic loci of their paren-
tal genes as revealed by RNA-DNA double
CircRNAs regulate transcription at the initiation FISH. These data collectively indicates that cir-
as well as the elongation step. In addition they cEIF3J and circPAIP2 may regulate the expres-
also regulate gene expression posttranscription- sion of their parental genes in cis. However, these
ally. An elaborate interplay among diverse pro- EIciRNAs were not confined to their parental
tein coding and noncoding RNA species during gene loci only, thus leaving a possibility of trans
transcription has been described [35–40]. The effects of these circRNAs on other loci as well.
various roles played by different circular RNAs The obvious question that would arise is
at different stages of transcription have been whether these EIciRNAs directly associate with
described below. Pol II or other factors of the pre-initiation com-
7 Regulation of Transcription by Circular RNAs 87

plex to exert their effects on transcription. Pull 3.2 Regulation of Transcription

down experiments with specific oligos corre- During Elongation:
sponding to different regions of either circEIF3J Interaction of Intronic
or circPAIP2 led to coprecipitation of U1A and circRNAs with
U1C snRNPs, U1 snRNA, along with Pol II sug- Elongating Pol II
gesting these interactions to be specific.
Interestingly, sites within the promoter and also Although in most cases transcription is regulated
regions of the first exon of parent genes copre- at the initiation stage, transcriptional regulation
cipitated in these experiments. U1 snRNA is a can also take place during elongation. Circular
core-splicing component that associates specifi- RNAs, intronic circRNAs in particular, provide
cally with TFIIH which is a general transcription such example where they control transcription at
initiation factor [42]. The role of U1 snRNA in the elongation step. It is generally believed
transcription initiation is to stimulate the forma- introns are unused part of the mRNA which are
tion of the first phosphodiester bond by Pol II unstable and rapidly degraded. One way by
[42]. Each of these EIciRNAs has one U1 which intronic RNAs can prevent their degrada-
snRNA-binding site in their retained intron. tion and accumulate in the cells is by circulariza-
Sterically blocking this site not only decreased tion, known as intronic circular RNAS (ciRNAs).
interaction of U1 snRNA with EIciRNA but also These ciRNAs do not have their own promoters
with Pol II and EIciRNA and Pol II with promot- but are derived from the spliced introns of their
ers of the parental genes of circEIF3J and cir- parent transcripts, sometimes enhancing the pro-
cPAIP2. As a result, the transcription of the duction of the latter. One such ciRNA identified
parental genes of the corresponding circRNAs is the ci-ankrd52 [4]. The ci-ankrd52 is derived
also decreased. Conversely, the binding of Pol II from the second intron of the ankrd52 gene,
with specific U1 snRNPs (U1A and U1C) to their which codes for a protein with unknown func-
parent gene promoters also requires the presence tion, containing a large ankyrin repeat domain.
of the EIciRNAs [41]. Pull down with U1A- and Synthetic antisense oligodeoxynucleotides
U1C-specific antibodies, but not with any other (ASO), which target the intron-derived ci-­
snRNP (like Lsm10), or auxiliary factors ankrd52, successfully downregulated the expres-
(U2AF65 and U2AF35), led to coprecipitation of sion of these circular RNAs leading to a decrease
substantial amount of EIciRNAs. Furthermore, in the parent mRNA level as well. ASO specific
chromatin immunoprecipitation (ChIP) experi- to ci-ankrd52, being complementary to the
ments revealed that only U1A and U1C and not ankrd52 pre-mRNA intron, may bind to the pre-­
U2AF65 and U2AF35 interacted with the pro- mRNA and subsequently leads to its degradation
moter regions of some genes like EIF3J and resulting in decreased mRNA levels. Inability of
PAIP2, but not their neighboring genes. In line ASOs against introns, adjacent to the ci-ankrd52
with these findings, U1 snRNA was found to be to reduce ankrd52 mRNA, excludes this possibil-
co-localized with majority of the circEIF3J or ity. In addition, co-expression of ASOs and cor-
circPAIP2 within the nucleus using dual RNA responding intronic RNAs except ci-ankrd52
FISH. failed to reduce ankrd52 mRNA levels confirm-
Available experimental evidence therefore ing the specificity of ci-ankrd52 in regulating its
suggest that specific RNA-RNA interaction parental mRNA expression. Similar results were
between U1 snRNA and the EIciRNAs followed obtained with ASO-mediated knockdown of two
by interaction of the EIciRNA-snRNP complex other ci-RNAs, ci-mcm5, and ci-sirt7.
with the Pol II at the promoter site leads to upreg- In the quest to find out the mechanism of
ulation of transcription of their parent genes ciRNA-mediated reduction of parental mRNAs,
(Fig. 7.2). authors tested three hypotheses.
88 R. Bose and R. Ain

Fig. 7.2 Regulation of transcription initiation by act with U1-snRNA through specific RNA-RNA interac-
EIciRNAs tions. This EIciRNA-U1snRNP complex further interacts
Exon-intron circular RNAs are composed of exons and with the Pol II transcription initiation complex at the pro-
unspliced introns retained in between the exons. They are moter of parent genes and promote their transcription.
produced as by-product of gene transcription. They inter- Orange and light blue bars, exons; green bar, intron

A. CiRNAs acting as miRNA sponge: Presence II interaction with ci-mcm5 and ci-sirt7 was dem-
of only a few miRNA-binding sites on these onstrated using similar assays.
ciRNAs and their exclusive nuclear localiza- Phosphorylation of Pol II is pivotal to tran-
tion exclude the possibility of these circRNAs scription elongation process [27]. Association of
acting as miRNA sponges. ciRNAs with phosphorylated Pol II therefore
B. CiRNAs required for proper mRNA process- confirms their regulatory role in transcription
ing: Analysis of relative abundance of splic- elongation (Fig. 7.3).
ing intermediates revealed that the ciRNA
and its downstream introns are processed at a
similar rate. However, knockdown of ci-­ 3.3 Circular RNAs
ankrd52 gave rise to new isoforms of the Posttranscriptionally Regulate
mRNA with retained introns containing stop Gene Expression by Acting
codons that lead to nonsense-mediated decay as miRNA Sponges
(NMD) of the parent mRNAs. These results
indicate that ciRNAs regulate mRNA CircRNAs that are exported from the nucleus and
processing. are located in the cytoplasm have several binding
C. Transcriptional regulation by ciRNAs: DNA-­ sites for miRNAs and compete with the target
RNA dual FISH revealed presence of ciRNAs mRNAs for miRNA binding in the cytoplasm
in the elongating transcript of the parent gene. thus regulate gene expression at the posttran-
scriptional level. The best characterized circRNA
Furthermore, biotinylated ci-ankrd52 inter- that acts as miRNA sponge is the vertebrate
acted with phosphorylated RNA Pol II in in vitro ciRS-7, also known as Cdr1as that acts as sponge
pull down assay. This was further substantiated for miR-7. Produced from the vertebrate cerebel-
by co-immunoprecipitation of phosphorylated lar degeneration-related 1 (CDR1) antisense tran-
RNA Pol II with ci-ankrd52 in PA1 cell line. Pol script, Cdr1as is preferentially expressed in
7 Regulation of Transcription by Circular RNAs 89

Fig. 7.3 Regulation of

transcription elongation
by intronic circRNA
Intronic ci-ankrd52 is
produced from the
second intron of the
ankrd52 gene. After its
synthesis it accumulates
at its site of transcription
where it interacts with
phosphorylated RNA
Pol II and enhances
transcription elongation
of its parent gene.
Colored bars, exons;
black lines, introns

human and mouse brains [14]. It has over 60 a ~70% increase in Cdr1as overexpressing cells.
binding sites for miR-7 as analyzed by PAR-­ In mouse islet cells, the percentage was as high as
CLIP experiments with human AGO [43]. Cdr1as ~90%. Thus, Cdr1as affects the insulin secretion
and miR-7 are co-expressed in neuronal tissues, by upregulating its biosynthesis. Interestingly,
pancreas, and pituitary gland and also in murine Myrip and Pax6, which are involved in insulin
pancreatic tissue-derived MIN6 cell line. biosynthesis and secretion, are also potential tar-
Specifically, high level of co-expression is found gets of miR-7. Thus by acting as miR-7 sponge,
in the developing midbrain of D13.5 mouse Cdr1as upregulates levels of Myrip and Pax6 in
embryos. As expected, downregulation of Cdr1as the cell. In line with this argument, Cdr1as over-
led to downregulation of miR-7 targets along expression led to significant upregulation of
with house-keeping genes in HEK293 cells, indi- Myrip and Pax6 mRNA in MIN6 cells by up to
cating miR-7-mediated repression of targets in 70% and 50%, respectively. An even better result
absence of Cdr1as. was observed in mouse islets. As expected,
Interestingly, Cdr1as has been shown to regu- ­ectopic overexpression of miR-7 led to decrease
late insulin transcription and secretion in mouse in Myrip and Pax6 mRNA levels by 40–50%.
islets [44]. Stimulation of islet cells with either Therefore, overexpressed Cdr1as could bind and
forskolin or PMA led to increased expression of sequester miR-7 in the cytoplasm and hence
Cdr1as, but miR-7 was reduced under the same abolish its inhibitory effects on the Myrip and
condition. Overexpression of miR-7 and Cdr1as Pax6 mRNAs, which in turn elevates insulin bio-
separately, in MIN6 and isolated islet cells, led to synthesis and secretion (Fig. 7.4).
decrease and increase in insulin secretion, respec-
tively, as evaluated by glucose-stimulated insulin
secretion assay (GSIS). These results indicate 3.4 CircularMbl RNA
that miR-7 directly regulate levels of insulin tran- Posttranscriptionally
script in the cell, whereas Cdr1as binds and to Regulates Its Own Expression
miR-7 and acting as a sponge reverses the effect by Acting as an RBP Sponge
of miR-7 on insulin transcript levels. As expected,
in miR-7 overexpressing MIN6 cells, there was CircRNAs in some cases regulate its own expres-
~25% decrease in insulin content as compared to sion by tuning the posttranscriptional events.
90 R. Bose and R. Ain

Fig. 7.4 Schematic diagram showing Cdr1as as a potent increased expression of Cdr1as leads to sequestration and
regulator of insulin transcription in mouse pancreatic inhibition of miR-7. Thus Pax6 mRNA is translated to pro-
β-cells and its secretion duce a transcriptional activator, which translocates to the
In absence of any secretagogue, miR-7 binds to the 3′UTR nucleus and promotes the transcription of insulin gene. On
of its targets Pax6 and Myrip mRNA, thereby decreasing the the other hand, the product of Myrip gene is involved in
expression and secretion of insulin in mouse β-cells (left). In translocation and secretion of insulin, which is also increased
the presence of secretagogues-like forskolin or PMA, due to Cdr1as overexpression (right)

Drosophila muscleblind (Mbl) circular RNA is script which is translated to give rise to MBL pro-
one such transcript which drives its own expres- tein. As the level of the MBL protein builds up, it
sion through alternative splicing of its precursor binds to the mbl pre-mRNA and causes it to
RNA. The MUSCLEBIND protein (MBL) in backsplice into circMbl. As a consequence the
Drosophila is required for the development of level of linear mbl transcript is diminished and so
muscle and photoreceptor cells in the fly eye. It is is the MBL protein. Furthermore, the circMbl
expressed in the cells of embryonic muscle, and itself contains several MBL protein-binding sites
Mbl deficiency is embryonic lethal [45]. MBL and thus can act as an RBP sponge. The circMbl
protein promotes the splicing of the second exon binds to and sequesters MBL protein, lowering
of its own pre-mRNA into a circular transcript, its free cellular concentration so that it can no
circMbl, It thus competes with mbl mRNA pro- longer produce circMbl transcripts thereby low-
duction thereby decreasing the levels of MBL ering its own level (Fig. 7.5). Thus circMbl regu-
protein [46]. At low levels of MBL protein, the lates its own expression via MBL sequestration
Mbl mRNA is spliced to produce the linear tran- in a negative feedback mechanism.
7 Regulation of Transcription by Circular RNAs 91

Fig. 7.5 Drosophila

circMbl negatively
regulates the expression
of its own gene
(a) When MBL protein
is low, the mbl transcript
is spliced to produce a
linear mRNA which is
translated to produce
MBL protein. (b) As the
amount of MBL protein
rises, it binds to its
pre-mRNA and causes
the second exon to
circularize to form
circMbl thereby
competing with splicing
and synthesis of native
MBL protein.
Furthermore, circMbl
binds to and sequesters
MBL protein by its
several MBL-binding
sites. This gradually
lowers its free cellular
concentration and
consequently decreasing
its own expression in a
negative feedback loop

3.5  ircular RNAs Act as “mRNA

C Phenotypically, they had normal limb develop-
Trap”: A Novel Mechanism ment but had renal agenesis with incomplete
in Regulating Gene ­penetrance. The authors put forth the model of
Expression “mRNA trap” from these observations. In this
model, the circular Fmn RNAs sequesters the lin-
The previous section describes how circRNA ear fmn transcripts from being translated into
regulates the expression of a functional protein functional FORMIN protein.
thereby regulating its own expression. In this sec- The “mRNA trap” phenomenon is also
tion we will discuss about how some circRNAs observed in patients with dystrophinopathy. The
sequester the translation start site on their linear dystrophin (DMD) gene produces several circu-
transcripts to monitor their protein expression lar transcripts. In the skeletal muscle of patients,
level. The Formin (fmn) gene, which is respon- scrambled RNAs in the form of circular dystro-
sible for development of limbs and kidney in phin RNAs are produced, at the expense of linear
mouse, was reported to produce circular exonic in-frame transcripts reducing the levels of func-
RNAs (ecircRNAs) comprising exons 4 and 5 tional proteins [6, 48].
[47]. Fmn mutant mice with deletion of exon 4 In mouse and humans, HIPK2 and HIPK3 loci
and 5 did not produce the circfmn transcripts. generate exonic circular RNAs from the exon that
92 R. Bose and R. Ain

contains ATG start codon, yet they are not trans- 11. Ghoshal S, Das S, Sen R et al (2013) Circ2Traits:
a comprehensive database for circular RNA poten-
lated to yield any protein product. For HIPK3 tially associated with disease and traits. Front Genet
locus in particular, the circular isoform is more 4(283):283
abundant than its linear isoform, which does not 12. Glazar P, Papavasileiou P, Rajewsky N (2014)
translate into any protein. Thus this mechanism circBase: a database for circular RNAs. RNA
20(11):1666–1670
can be viewed as yet another example of “mRNA-­ 13. Liu YC, Li JR, Sun CH et al (2016) CircNet: a
trapping” event by circular RNAs in order to database of circular RNAs derived from tran-
modulate gene expression at posttranscriptional scriptome sequencing data. Nucleic Acids Res
level [6, 8]. 44(D1):D209–D215
14. Chen X, Han P, Zhou T et al (2016) circRNADb: a
comprehensive database for human circular RNAs
Acknowledgment Supported by CSIR-Indian Institute with protein-coding annotations. Sci Rep 6:34985
of Chemical Biology internal support grant, Rumela Bose 15. Le Hir H, Gatfield D, Izaurralde E et al (2001)
is a recipient of Shyama Prasad Mukherjee predoctoral The exon-exon junction complex provides a bind-
fellowship from the Council of Scientific and Industrial ing platform for factors involved in mRNA export
Research, India. and nonsense-mediated mRNA decay. EMBO
J 20(17):4987–4997
Conflict of Interest The authors declare that there is no 16. Ebbesen KK, Kjems J, Hansen TB (2016) Circular
conflict of interest. RNAs: identification, biogenesis and function.
Biochim Biophys Acta 1859(1):163–168
17. Han B, Chao J, Yao H (2018) Circular RNA and its
mechanisms in disease: from the bench to the clinic.
Pharmacol Ther 187:31. https://doi.org/10.1016/j.
References pharmthera.2018.01.010
18. Du WW, Fang L, Yang W et al (2017) Induction of
1. Lasda E, Parker R (2014) Circular RNAs: diversity of tumor apoptosis through a circular RNA enhancing
form and function. RNA 20(12):1829–1842 Foxo3 activity. Cell Death Differ 24(2):357–370
2. Grabowski PJ, Zuag AJ, Cech TR (1981) The inter- 19. Yang ZG, Awan FM, Du WW et al (2017) The circu-
vening sequence of the ribosomal RNA precursor lar RNA interacts with STAT3, increasing its nuclear
is converted to a circular RNA in isolated nuclei of translocation and wound repair by modulating Dnmt3a
Tetrahymena. Cell 23(2):467–476 and miR-17 function. Mol Ther 25(9):2062–2074
3. Danan M, Schwartz S, Edelheit S et al (2012) 20. Yang Q, Du WW, Wu N et al (2017) A circular RNA
Transcriptome-wide discovery of circular RNAs in promotes tumorigenesis by inducing c-myc nuclear
Archaea. Nucleic Acids Res 40(7):3131–3142 translocation. Cell Death Differ 24(9):1609–1620
4. Zhang Y, Zhang XO, Chen T et al (2013) Circular 21. Zeng Y, Du WW, Wu Y et al (2017) A circular RNA
intronic long noncoding RNAs. Mol Cell binds to and activates AKT phosphorylation and
51(6):792–806 nuclear localization reducing apoptosis and enhanc-
5. Ashwal-Fluss R, Meyer M, Pamudurti NR et al (2014) ing cardiac repair. Theranotics 7(16):3842–3855
circRNA biogenesis competes with pre-mRNA splic- 22. Kos A, Dijkema R, Arnberg AC et al (1986) The hepa-
ing. Mol Cell 56(1):55–66 titis delta (δ) virus possesses a circular RNA. Nature
6. Jeck WR, Sharpless NE (2014) Detecting and charac- 323(6088):558–560
terizing circular RNAs. Nat Biotechnol 32(5):453–461 23. Chen CY, Sarnow P (1995) Initiation of protein syn-
7. Salzman J, Gawad C, Wang PL et al (2012) Circular thesis by the eukaryotic translational apparatus on cir-
RNAs are the predominant transcript isoform from cular RNAs. Science 268(5209):415–417
hundreds of human genes in diverse cell types. PLoS 24. Weinmann R, Roeder RG (1974) Role of DNA-­
One 7(2):e30733 dependent RNA polymerase III in the transcription
8. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular of the tRNA and 5S RNA genes. Proc Nalt Acad Sci
RNAs are abundant, conserved, and associated with USA 71(5):1790–1794
ALU repeats. RNA 19(2):141–157 25. Woychik NA, Liao SM, Kolodziej PA et al (1990)
9. Ye CY, Chen L, Liu C et al (2015) Widespread Subunits shared by eukaryotic nuclear RNA polymer-
noncoding circular RNAs in plants. New Phytol ases. Genes Dev 4(3):313–323
208(1):88–95 26. Kolodziej PA, Woychik N, Liao SM et al (1990) RNA
10. Pan T, Sun X, Liu Y et al (2017) Heat stress alters polymerase II subunit composition, stoichiometry,
genome-wide profiles of circular RNA in Arabidopsis. and phosphorylation. Mol Cell Biol 10(5):1915–1920
Plant Mol Biol 96(3):217–229 27. Payne JM, Laybourn PJ, Dahmus ME (1989) The
transition of RNA polymerase II from initiation to
7 Regulation of Transcription by Circular RNAs 93

elongation is associated with phosphorylation of the lin transcription and secretion in islet cells. Sci Rep
carboxy-terminal domain of subunit IIa. J Biol Chem 5:12453
264(33):19621–19629 45. Begemann G, Paricio N, Artero R et al (1997)
28. Corden JL (1990) Tails of RNA polymerase II. Trends Muscleblind, a gene required for photoreceptor dif-
Biochem Sci 15:383–387 ferentiation in Drosophila, encodes novel nuclear
29. Haltiner MM, Smale ST, Tjian R (1986) Two distinct Cys3His-type zinc-finger-containing proteins.
promoter elements in the human rRNA gene identi- Development 124(21):4321–4331
fied by linker scanning mutagenesis. Mol Cell Biol 46. Barrette SP, Salzman J (2016) Circular RNAs: analy-
6(1):227–235 sis, expression and potential functions. Development
30. Murphy S, Di Liegro C, Melli M (1987) The in vitro 143(11):1838–1847
transcription of the 7SK RNA gene by RNA poly- 47. Chao CW, Chan DC, Kuo A et al (1998) The mouse
merase III is dependent on the presence of an upstream Formin (Fmn) gene: abundant circular RNA tran-
promoter. Cell 51(1):81–87 scripts and gene-targeted deletion analysis. Mol Med
31. Buratowski S, Hahn S, Guarente L et al (1989) Five 4(9):614–628
intermediate complexes in transcription initiation by 48. Gualandi F, Trabanelli C, Rimessi P et al (2003)
RNA polymerase II. Cell 56(4):549–561 Multiple exon skipping and RNA circularization con-
32. Flores O, Lu H, Killeen M et al (1991) The small tribute to the severe phenotypic expression of exon 5
subunit of transcription factor IIF recruits RNA poly- dystrophin deletion. J Med Genet 40(8):e100
merase II into the preinitiation complex. Proc Natl 49. Abdelmohsen K, Panda AC, Munk R et al (2017)
Acad Sci USA 88(22):9999–10003 Identification of HuR target circular RNAs uncovers
33. Goodrich JA, Tjian R (1994) Transcription factors IIE suppression of PABPN1 translation by CircPABPN1.
and IIH and ATP hydrolysis direct promoter clearance RNA Biol 14(3):361–369
by RNA polymerase II. Cell 77(1):145–156 50. Du WW, Zhang C, Yang W et al (2017) Identifying
34. Allison LA (2012) Fundamental molecular biology, and characterizing circRNA-protein interaction.
2nd edn. Wiley, Hoboken Theranostics 7(17):4183–4191
35. Chen H, Zhang J, Yuan G et al (2014) Complex inter- 51. Du WW, Yang W, Chen Y et al (2017) Foxo3 circu-
play among DNA modification, noncoding RNA lar RNA promotes cardiac senescence by modulating
expression and protein-coding RNA expression in multiple factors associated with stress and senescence
Salvia miltiorrhiza chloroplast genome. PLoS One responses. Eur Heart J 38(18):1402–1412
9(6):e99314 52. Du WW, Yang W, Liu E et al (2016) Foxo3 circular
36. Salmena L, Poliseno L, Tay Y et al (2011) A ceRNA RNA retards cell cycle progression via forming ter-
hypothesis: the Rosetta stone of a hidden RNA lan- nary complexes with p21 and CDK2. Nucleic Acids
guage? Cell 146(3):353–358 Res 44(6):2846–2858
37. Franco-Zorrilla JM, Valli A, Todesco M et al 53. Holdt LM, Stahringer A, Sass K et al (2016) Circular
(2007) Target mimicry provides a new mechanism non-coding RNA ANRIL modulates ribosomal
for regulation of microRNA activity. Nat Genet RNA maturation and atherosclerosis in humans. Nat
39(8):1033–1037 Commun 7:12429
38. Flynn RA, Martin L, Spitale RC et al (2015) 54. Hansen TB, Jensen TI, Clausen BH et al (2013)
Dissecting noncoding and pathogen RNA-protein Natural RNA circles function as efficient microRNA
interactomes. RNA 21(1):135–143 sponges. Nature 495(7441):384–388
39. Hamilton MJ, Young MD, Sauer S et al (2015) The 55. Schneider T, Hung LH, Schreiner S et al (2016)
interplay of long non-coding RNAs and MYC in can- CircRNA-protein complexes: IMP3 protein compo-
cer. AIIMS Biophys 2(4):794–809 nent defines subfamily of circRNPs. Sci Rep 6:3131
40. Tay Y, Rinn J, Pandolfi PP (2014) The multilayered 56. Peng L, Chen G, Zhu Z et al (2017) Circular RNA
complexity of ceRNA crosstalk and competition. ZNF609 functions as a competitive endogenous RNA
Nature 505(7483):344–352 to regulate AKT3 expression by sponging miR-150-5p
41. Li Z, Huang C, Bao C et al (2015) Exon-intron cir- in Hirschsprung’s disease. Oncotarget 8(1):808–818
cular RNAs regulate transcription in the nucleus. Nat 57. He S, Liu P, Jian Z et al (2013) miR-138 protects
Struct Mol Biol 22(3):256–264 cardiomyocytes from hypoxia-induced apoptosis via
42. Kwek KY, Murphy S, Furger A et al (2002) U1 MLK3/JNK/c-Jun pathway. Biochem Biophys Res
snRNA associates with TFIIH and regulates transcrip- Commun 441(4):763–769
tional initiation. Nat Struct Biol 9(11):800–805 58. Schnall-Levin M, Rissland OS, Johnston WK et al
43. Memczak S, Jens M, Elefsinioti A et al (2013) (2011) Unusually effective microRNA targeting
Circular RNAs are a large class of animal RNAs with within repeat-rich coding regions of mammalian
regulatory potency. Nature 495(7441):333–338 mRNAs. Gnome Res 21(9):1395–1403
44. Xu H, Guo S, Li W et al (2015) The circular RNA 59. Wang K, Long B, Liu F et al (2016) A circular RNA
Cdr1as, via miR-7 and its targets, regulates insu- protects the heart from pathological hypertrophy
94 R. Bose and R. Ain

and heart failure by targeting miR-223. Eur Heart ing miR-124. Biochem Biophys Res Commun
J 37(33):2602–2611 494(1–2):126–132
60. Wang K, Gan TY, Li N et al (2017) Circular RNA 62. Zheng C, Niu H, Li M et al (2015) Cyclic RNA hsa-­
mediates cardiomyocyte death via miRNA-dependent circ--000595 regulates apoptosis of aortic smooth
upregulation of MTP18 expression. Cell Death Differ muscle cells. Mol Med Rep 12(5):6656–6662
24(6):1111–1120 63. Xie H, Ren X, Xin S et al (2016) Emerging roles of
61. Chen J, Cui L, Yuang J et al (2017) Circular RNA circRNA_001569 targeting miR-145 in the prolif-
WDR77 target FGF-2 to regulate vascular smooth eration and invasion of colorectal cancer. Oncotarget
muscle cells proliferation and migration by spong- 7(18):26680–26691
 Functional Analysis of Circular
RNAs 8
Shanmugapriya, Hisham Alkatib Huda,
Soundararajan Vijayarathna, Chern Ein Oon,
Yeng Chen, Jagat R. Kanwar, Mei Li Ng,
and Sreenivasan Sasidharan

Abstract gene silencing assay, luciferase reporter

Circular RNAs characterize a class of wide- assays, circRNA gain-of-­ function investiga-
spread and diverse endogenous RNAs which tion via overexpression of circular transcript
are non-coding RNAs that are made by back- assay, RT-q-PCR quantification, and other
splicing events and have covalently closed latest applicable assays. The methods
loops with no polyadenylated tails. Various described in this chapter are demonstrated on
indications specify that circular RNAs (cir- the cellular model.
cRNAs) are plentiful in the human transcrip-
tome. However, their participation in Keywords
biological processes remains mostly unde- CircRNAs · Functional validation · Cellular
scribed. To date thousands of circRNAs have model
been revealed in organisms ranging from
Drosophila melanogaster to Homo sapiens.
Functional studies specify that these tran-
scripts control expression of protein-coding 1 Introduction
linear transcripts and thus encompass a key
component of gene expression regulation. Circular RNAs (circRNAs) are closed RNA
This chapter provide a comprehensive over- transcripts made by back-splicing of a single
view on functional validation of circRNAs. pre-­mRNA that is found in all higher eukaryotes
Furthermore, we discuss the recent modern including mammals. The first circRNA was dis-
methodologies for the functional validation of covered in the early 1990s [1] as an obviously
circRNAs such as RNA interference (RNAi) befalling family of non-coding RNAs that is

Shanmugapriya · H. A. Huda · S. Vijayarathna

C. E. Oon · S. Sasidharan (*) J. R. Kanwar
Institute for Research in Molecular Medicine Faculty of Health, Nanomedicine-Laboratory of
(INFORMM), Universiti Sains Malaysia, Immunology and Molecular Biomedical Research
Pulau Pinang, Malaysia (LIMBR), School of Medicine (SoM), Deakin
University, Geelong, VIC, Australia
Y. Chen
Faculty of Dentistry, Dental Research & Training M. L. Ng
Unit, and Oral Cancer Research and Coordinating Integrative Medicine Cluster, Advanced Medical and
Centre (OCRCC), University of Malaya, Dental Institute (AMDI), Universiti Sains Malaysia,
Kuala Lumpur, Malaysia Pulau Pinang, Malaysia

© Springer Nature Singapore Pte Ltd. 2018 95

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_8
96 Shanmugapriya et al.

vastly denoted in the eukaryotic transcriptome [2, diseases in human. Hence, in this chapter we dis-
3]. Previously these circRNAs had generally cuss the latest methodologies for the validation of
been considered to be errors or by-products of circRNA transcripts which will eventually lead to
RNA splicing [4] with little biological function. the utilization of circRNA as molecular markers
However, circRNAs have now been accepted as of complex diseases in human and gene therapy
another type of endogenous non-coding RNA agent. Various experimental validation assays for
species with the development of next-generation functional analysis of circular RNAs such as RNA
sequencing (NGS) which are plentiful and interference (RNAi) gene silencing assay, lucifer-
maintained in various biological systems. ase reporter assays, CircRNA gain-of-function
CircRNAs are extremely stable in vivo compared analysis through over expression of circular tran-
with their linear counterpart RNAs due to the scripts assay, RT-q-­PCR quantification and other
absence of a 2′ to 5′ carbon linkage and free 3′ or applicable assay were discussed in this chapter
5′ ends, respectively [2], and are mostly found in with appropriate examples in cellular model.
the cytoplasm and exosomes [5]. Interestingly, a Figure 8.1 depicts the proposed model to study
huge number of circRNAs have been effectively the role of circRNA in a cellular model by using
identified in recent time in numerous cell lines latest methodologies for the validation of cir-
and across diverse species [6, 7]. cRNA transcripts.
Various properties of circRNAs have been effi-
caciously characterized from the time of its first
discovery in the early 1990s, but a comprehensive 2 Principle of Functional
understanding of their biological function remains Analysis of Circular RNAs
unclear. Various evidences from latest research
findings also suggest a possible role of circRNAs 2.1 CircRNA Quantification
in diverse human diseases [8]. In spite of these
findings in support of circRNAs’ significant pur- Quantification of circRNA abundance required
poseful roles, their influence on biological pro- application of bioinformatics to precisely analyse
cesses is still mostly unknown. Various lab-based circRNA datasets generated by deep sequencing
functional studies specify that these circRNA [9]. Currently, RNA-seq is regarded as a high-­
transcripts control expression of protein-coding throughput sequencing technique for quantifica-
linear RNA transcripts and therefore encompass a tion and functional analysis of the transcriptome
significant constituent of gene expression regula- [9, 10]. RNA-seq analyses structural features
tion. Several circRNAs with changed expression of circRNA based on RNA-­seq-­derived datasets
patterns are recognized in numerous types of dis- including find_circ, MapSplice [11],
eases. CircRNAs thus function as healing targets CIRCexplorer [12], circRNAFinder and CIRI
for treatment because of their action on various [13]. A range of computational methods have
target genes and proteins. Validation of these cir- been utilized to recognize scrambled sequences in
cRNAs permits further study and is within the RNA-seq datasets. Such approach consuming
scope of this chapter which focuses on cellular paired-end RNA-seq reads that line up to a cus-
model. To study the role of circRNA variation in tom database of all probable exon-exon pair
cells, the most upregulated or downregulated cir- junctions [14]. Utilizing computational method to
cRNAs that were the mostly affected biological map read ends from a genomic anchor site to a
function of the cells can be selected for further “breakpoint” edged by GU/AG sequence [15], or
validation. Furthermore, the circRNAs important MapSplice algorithm that segments reads to iden-
target proteins that could assist as potential targets tify the “back-splice” events. In addition, novel
for disease prevention also can be identified. algorithms designed to detect novel splicing or
Assessing the potential of circRNAs to modulate structurally mutated transcriptome are developed.
disease could assist in the detection and develop- Recent library cloning uses oligo(dT) primer in
ment of new healing policies against various cDNA synthesis to allow detection of both
8 Functional Analysis of Circular RNAs 97

Fig. 8.1 Schematically summarizes various modern experimental validation methods for circRNAs

unpolyadenylated and polyadenylated fractions of enrichment of circRNA levels due to endogenous

long circRNAs [1, 16]. Quantification of circRNA nicking of the RNA or contaminating nucleases,
uses random priming because circRNA generally and some linear RNAs are resistant to RNase R
has lower abundance in the cells. The easiest degradation [23], and different circRNAs can
quantification of circRNA expression is to count show drastically different levels of RNase R resis-
split-back spice read and include reads that do not tance [24], or potentially due to interference with
align directly to exons [17–19]. The use of algo- reverse transcription reactions [20]. For this rea-
rithms to annotate and quantify circRNA from son, it is imperative to verify its efficacy by qRT-­
RNA-seq data can yield different results. Less PCR. Hence, quantitative reverse transcription
than 1% of split-back-spliced reads are generated PCR (qRT-PCR) is a powerful tool to quantita-
from a total RNA-seq experiment, whereas 99% tively assess the relative abundance of circRNAs.
of the reads are aligned to express transcriptome In this method, the circRNA is reversely tran-
[20, 21]. Generally, circRNAs are structurally dif- scripted to a cDNA molecule that contains an
ferent from linear RNA because circRNAs are exon-exon junctional sequence which can be spe-
uncapped and unpolyadenylated. Consequently cifically targeted and amplified by primers. These
RT-PCR reactions or RNA-seq of cap-enriched, primers are recognized as “inverse” or “outward-
poly(A) selected or oligo(dT) reverse-transcrip- facing” primers since it stops the amplification of
tion-primed samples favourably identify linear RNA species that do not comprise the exon-exon
scrambled exon RNAs. Additionally, scrambled junction when aligned to the genome. However, it
exon RNAs resulting from DNA rearrangements is imperative to verify the amplified sequence
or tandem duplications would be detectable in because amplification errors might occur during
genomic DNA PCR reactions. These methods, RT step such as template switching [25], splicing
used in combination with RNase R, a 3′ to 5′ between two separate pre-mRNA molecules [26],
exoribonuclease specifically degrades linear unexpected genomic duplication and rolling cir-
RNAs, are a useful tool to validate and enrich cir- cle RT. Compared to PCR, Northern blots allow
cRNAs in a total RNA sequencing library [20, precise monitoring of species mobility [27]. This
22]. However, RNase R may cause artifactual technique utilizes probes which are intended to
98 Shanmugapriya et al.

specifically target the circularized RNA exonic Furthermore, mammalian genes can be studied
sequence in separate blots. When RNA is run on a by overexpressing mammalian vectors comprise
polyacrylamide gel matrix, differential mobility the circularized exon(s) along with flanking
forms two bands if the RNA is circular or three splicing signals and intronic sequences, which
bands if the RNA is linear [27]. In addition, two- harbour inverted repeats to assist their linking
dimensional denaturing polyacrylamide gel elec- into a circle [10, 34, 35]. However, it’s difficult to
trophoresis can be also utilized to identify the determine whether the given phenotypes are
circRNAs. Total RNA is run on a 2D gel compris- exclusively driven by circRNA because the vec-
ing various percentages of polyacrylamide in each tor also overexpresses linear RNA. In addition, a
dimension; circRNA runs in an arc, which can be common artefact may arise from rolling circle
sequenced and enriched in next-generation transcription of the plasmid. Hypothetically, the
sequencing (NGS) [28]. Otherwise, the 2D gel expression plasmid harbours circularized exon of
can be probed through Northern blotting to quan- a gene with flanking introns. If the transcription
tity specific circRNAs. Furthermore, ribosomal termination signals in the vector are circum-
RNA (rRNA) depletion and polyA-depletion are vented, the RNA polymerase will remain to tran-
general approaches utilized to enrich for cir- scribe around the entire plasmid, generating a
cRNAs in sequencing libraries. However, neither concatemer of the RNA sequence contained in
promises that the enriched sequences are abso- the plasmid. This transcript piece can lead to off-
lutely circular since numerous forms of non-cod- target effects on the cell and spurious circRNA
ing RNA will also survive in these selections. quantification. Therefore, attention has been
“TRAP electrophoresis” is another mean of veri- given in vector design to lessen the amplification
fying circRNAs. In this method, circRNAs are of inaccurate products [10]. For additional uses,
separated based on characteristic variances in such as establishing translation or analysing
movement on a gel paralleled with their corre- function of a circRNAs, these artefacts should be
sponding linear molecules [15, 29]. Additionally, entirely removed. Furthermore, suppressing cir-
one- and two-dimensional polyacrylamide gel cRNAs function can be achieved by siRNA
electrophoresis consuming diverse percentage knockdown. In this method, siRNA is designed
gels can also distinguish characteristic movement to specifically disrupt circRNA expression with-
patterns of circular RNAs as single-hit nicking or out affecting linear protein-coding RNAs.
targeted RNase H cleavage must transform the
circle RNAs to linear species with predictable
electrophoretic mobility. On the contrary, split of 2.3 Validation Through Biological
a linear RNA will produce two products on a gel Function of CircRNAs
electrophoresis [30–32].
Expression of circRNAs and its isoforms is often
specific to cell type, tissue and developmental
2.2 Interrogate the Biological stage. While the abundance of the circRNAs can
Function of CircRNAs be assessed by biochemical methods, many
biological functions of circRNAs are unidentified.
Ectopic circRNA expression plasmid is a conve- Multiple lines of evidence have shown that
nient tool to interrogate the biological function of circRNAs function as “microRNA sponge”.
circRNA [15, 33]. CircRNAs are frequently over- Notably, the circRNAs ciRS-7/CDR1as contains
expressed on plasmid by utilizing gene fragments many highly conserved target sites for microRNA
under the control of a strong promoter. miR-7 that leads to decreased miRNA activity
Circularization is prompted by inverted repeats [15]. Likewise, Sry is another abundantly
(IR) flanking the circularized exon, which seem- expressed circRNA in mouse testis that has been
ingly brings the splice acceptor (SA) and splice previously clearly demonstrated to suppress the
donor (SD) into closeness for back-­ splicing. miRNA-138. This points to the role of circRNA
8 Functional Analysis of Circular RNAs 99

in regulatory framework of post-transcriptional HIF-1α [45, 46]. Circ-­Foxo3 has been shown to
gene expression, supporting the function of interact with cell-cycle proteins CDK2 and P21
circRNAs as miRNA decoys. Additional evidence to reduce cancer cells growth [46]. In addition,
for functional circRNAs sponge has been circRNA coined circMbl harbours binding sites
demonstrated by the downregulation of circRNA, for MBL protein itself that reduces mbl mRNA
named HRCR (mm9-circ-012559) in the mice and protein production [47]. Other potential
expressing miR-223 transgene [35], suggesting function for circRNAs has been described in sub-
that role of HRCR in preventing heart failure. cellular transportation and as stable molecule
Other circRNAs have been identified to promote scaffold for assembly of complexes [15]. In short,
cell proliferation in cancer. A circRNA from the the identification of circRNAs contributing to the
HIPK3 gene (circHIPK3) was found to sponge post-transcriptional regulation of gene expres-
miR-124 to enhance cell growth [36]. Other sion and RBP sponge highlight a good capacity
studies have linked circRNA circRNA-CER of circRNA associated functionalities, as demon-
which sponges miR-136 [37] and strated by the conserved nature of circRNA
circRNA_001569 which sponges miR-145 to expression in tissue-specific abundance.
promote cell survival [38]. Another circRNA, Unrevealing these molecule functionalities
circZNF292 was found to enhance proliferation. should be the main focus in the future directions
However, it’s unknown as whether circRNA for circRNA research field.
could be a general function of miRNA decoys.
This is probably attributed by a highly conserved
sequence in the codon of circulating exons [15, 3 Modern Experimental
26] or decreased single-nucleotide polymor- Validation Assays
phisms in microRNA target sites of circularized and Functional Analysis
exons [39].This is too a matter of competing of Circular RNAs
endogenous RNA (ceRNAs) which also binds
and destabilizes miRNAs. Regardless of this 3.1 Validation of CircRNA
endogenous competition, circRNAs remain as
more effective “miRNA sponge” than ceRNAs Validation of circRNA can be performed through
because of the circular structure of circRNAs that RTqPCR, in situ hybridization and Northern
protect from exonucleases degradation [40]. blotting with the incorporation of circRNA-­
Furthermore, the circular structure also confers specific primers and probes.
intrinsic resistance against miRNA-mediated
destabilization. This shows that circRNAs function 3.1.1 RTqPCR
as important regulator for miRNA expression CircRNA identified through RNA-seq and bioin-
[41, 42]. A novel subclass of circRNAs, named formatics can be validated by performing the
EIciRNAs, has been recently identified as positive reverse transcription quantitative polymerase
regulator for gene transcription, through an inter- chain reaction (RTqPCR). Enrichment of
action with U1 small nuclear ribonucleoprotein circRNA through RNase R treatment from the
(snRNP) and RNA polymerase II in the promoter total RNA is confirmed as real circle and
region of the host gene [33]. In this regard, nonlinear RNA products by designing the
circRNAs are able to induce gene expression in divergent primers of the circRNA [29] (Fig. 8.2).
cis or trans to regulate other genome loci [43]. Nevertheless, RNase R digestion is not an
Other functions of circRNAs include acting as essential step in qPCR since the divergent primers
protein decoys [42, 44]. CircRNAs are mostly designed is not expected to amplify the linear
localized in the cytoplasm that can sequester RNAs [51]. After that, reverse transcription (RT)
protein to prevent entry into nucleus. In this of the RNA will be conducted to synthesize
regard, Circ-Foxo3 was found to reduce nuclear complementary DNA (cDNA) with the utilization
concentration of stress-­related proteins FAK and of random hexamers for priming instead of
100 Shanmugapriya et al.

probes will then be introduced to be hybridized

after the denaturation process. Labelled probes
will be hybridized to the complementary back-­
splicing of circRNA. Then, the cells will be
washed, and the probes can be detected with the
aid of cross-linking agent. Cells will be counter
strained before acquiring the images [29, 40].

3.2 Functional Analysis

of CircRNA

Functional analysis of circRNA can be performed

by using loss-of-function, gain-of-function
investigation and luciferase reporter assay.

Fig. 8.2 Schematic depiction of amplification process of 3.2.1 RNA Interference

circular RNA from back-splicing with the aid of divergent Gene silencing method is one of the commonly
primers used techniques to determine the function of
specific circRNA. Knockdown of circRNA can
oligo(dT) priming on the ground that poly(A) tail be performed with the utilization of small
is in paucity in circRNA [48]. There are two interfering RNAs (siRNAs). The siRNA will be
popular methods to perform qPCR, namely, designed so as to be complementary with the
SYBR Green or TaqMan. Although SYBR Green back-splice sequence of the circRNA. There are
is considered to be cost-effective and several ways to transfect the siRNA into the cells
straightforward [49] as compared to TaqMan, which include chemical transfection
SYBR Green is non-specific and highly believed (Lipofectamine), mechanical transfection
to produce false positive as SYBR Green dye (electroporation) and viral-mediated delivery
binds to all double-stranded DNA (dsDNA) (expression vectors). siRNA directed circRNA
including primer-dimers [50, 51]. However, knockdown will eventually lead to gene silencing,
specific outward-facing primers and TaqMan followed by inhibition of protein translation via
probes complementary to the back-splice junction upregulation of miRNA [54] (Fig. 8.4). This is
of the circRNA will be designed to validate the because circRNAs play an important role as
specific circRNA via TaqMan RTqPCR [52]. The microRNA sponges [15].
quantification and validation of circRNA can be
analysed by the amplification plot representing 3.2.2 Enhancement CircRNA Function
the fluorescent intensity versus number of cycle. Enhancement CircRNA function can be per-
Ct value known as the threshold cycle is where formed by increasing the expression of circRNA
all the quantification data begins as the first in the cells. Recapitulating the circRNA will be
distinguishable fluorescent escalates [53]. accomplished with the incorporation of expres-
sion vectors in which unabridged intron respon-
3.1.2 In Situ Hybridization sible for circRNA production is infused together
Localization, quantification and validation of with its natural splice sites and exons [55]. The
specific circRNA can be analysed through in situ expression vectors containing the circRNA
hybridization (ISH). Specific probe expressing sequences will then be transiently
complementary to the back-splicing junction of transfected to the cells [34]. The expression of
circRNA will be designed with fluorescent label circRNA from the plasmid can be distinguished
or radioactive label (Fig. 8.3). The labelled by performing Northern blot analysis, and the
8 Functional Analysis of Circular RNAs 101

Fig. 8.3 Schematic diagram of in situ hybridization

Fig. 8.4 Schematic image illustrates (a) the role of small interfering RNA (siRNA) causes knockdown of the
CircRNA as microRNA sponges which eventually down- circRNA and releases the microRNA which will bind to
regulates or inhibits the microRNA. This consequently its mRNA target and silences the gene. As a result, protein
causes an efficient transcription of mRNA and successful translation is inhibited
translation of proteins. (b) However, the introduction of

production of bona fide circRNA can be validated 3.2.3 Luciferase Reporter Assays
through RNase R treatment as the back-splicing Luciferase reporter assays can be utilized to
reaction of circular RNA transcripts makes it investigate the regulation miRNA complemen-
RNase R resistant with increased stability, unlike tary to specific circRNA. Briefly, the luciferase
the linear RNA [29, 56, 57]. Therefore, down- vector will be constructed by inserting the
stream investigations of circRNA overexpression specific circRNA sequence in the 3’ UTR into
can be conducted. the promoter-driven luciferase reported gene
102 Shanmugapriya et al.

Fig. 8.5 Schematic diagram depicts the (a) construction sequences binds to the 3′ UTR together with the reporter
of luciferase reporter vector with the insertion of specific gene, luciferase which emits bioluminescence in the pres-
circRNA sequence in the 3′ UTR. (b) After transfection of ence of its substrate, luciferin
vectors into the cells, the circRNA complementary

(Fig. 8.5). Cells will then be cotransfected with other software category specifically circRNA_
the luciferase reporter vector and the miRNA finder, CIRCexplorer, DCC, MapSplice and
mimics in order to establish the characteristic ofsegemehl could be assigned to a subcategory
circRNA as miRNA sponges by comparing the because they invent spliced alignment algorithms
luciferase activity with the negative controls [55].
to identify and investigate the back-splicing
In addition, the relationship between circRNA events under the approach of “fragmented-based”
and miRNA can also be analysed through lucifer- or “segmented read approach” which recognized
ase reporter analysis [58]. back-splicing junctions from the mapping info of
a multiple-split read’s alignment to the genome.
However, find_circ and UROBORUS could be
3.3 Bioinformatic Analysis categorized together because both develop back-­
of Circular RNA splicing events from the mapping information of
these anchors after collecting the unmapped
To detect circRNAs from RNA-seq data, there reads. Lastly, CIRI is exclusive; it can identify
are around 11 software available. This software’s the paired chiastic clipping (PCC) signals from
package could be commonly separated into two the mapping information of reads by local
groups in line with the main approaches to detect alignment with BWA-MEM combined through
circRNA. For example, KNIFE, NCLscan and orderly filtering steps to get rid of possible false
PTESFinder require the circRNA sequences with positives [59]. Figure 8.6 explains the downstream
the gene annotation info in order to identify the bioinformatics analysis which can be performed
circRNA. This approach is named “pseudo-­ to further investigate the functional annotation of
reference based” or “candidate-based” strategy. the circRNA.
NCLScan and PTESFinder construct the assumed
circRNA sequences achieving the mapping info
of the segmented anchors found after alignment 4 Conclusion
to the genome or transcriptome. While, KNIFE
instantly builds all the possible out-of-order Functional validation of circRNAs is an impor-
exon-exon junction from gene annotation tant step for circRNA-based research in order to
information before alignment. However, the ascertain their role in various diseases. In order to
8 Functional Analysis of Circular RNAs 103

Fig. 8.6 Bioinformatics analysis flowchart

4. Cocquerelle C, Mascrez B, Hetuin D et al (1993)

validate the function of circRNAs, various meth- Mis-splicing yields circular RNA molecules. FASEB
ods were discussed in this chapter. To ensure the J 7(1):155–160
possible vital role of circRNAs in diverse human 5. Li Y, Zheng QP, Bao CY et al (2015) Circular RNA
is enriched and stable in exosomes: a promising bio-
diseases, functional validation of circRNAs is the marker for cancer diagnosis. Cell Res 25(8):981–984
first step towards establishing its identity in vari- 6. Glazar P, Papavasileiou P, Rajewsky N (2014)
ous ailments and development of novel therapeu- circBase: a database for circular RNAs. RNA
tic strategies against various diseases in human. 20(11):1666–1670
7. Rybak-Wolf A, Stottmeister C, Glazar P et al (2015)
In conclusion, this chapter discussed various reli- Circular RNAs in the mammalian brain are highly
able and modern methods to validate the function abundant, conserved, and dynamically expressed.
of circRNAs which will result in further accelera- Mol Cell 58(5):870–885
tion of research on this unique circRNAs. 8. Westholm JO, Miura P, Olson S et al (2014) Genome-­
wide analysis of drosophila circular RNAs reveals their
structural and sequence properties and age-dependent
Acknowledgements Shanmugapriya was supported by neural accumulation. Cell Rep 9(5):1966–1980
the Graduate Research Assistance Scheme from Universiti 9. Conn SJ, Pillman KA, Toubia J et al (2015) The RNA
Sains Malaysia, Malaysia. binding protein quaking regulates formation of cir-
cRNAs. Cell 160(6):1125–1134
Competing Financial Interests The authors declare no 10. Kramer MC, Liang D, Tatomer DC et al (2015)
competing financial interests. Combinatorial control of Drosophila circular RNA
expression by intronic repeats, hnRNPs, and SR pro-
teins. Genes Dev 29(20):2168–2182
11. Koopman P, Munsterberg A, Capel B et al (1990)
Expression of a candidate sex-determining
References gene during mouse testis differentiation. Nature
348(6300):450–452
1. Nigro JM, Cho KR, Fearon ER et al (1991) Scrambled 12. Cocquerelle C, Daubersies P, Majerus MA et al
exons. Cell 64(3):607–613 (1992) Splicing with inverted order of exons occurs
2. Jeck WR, Sharpless NE (2014) Detecting and charac- proximal to large introns. EMBO J 11(3):1095–1098
terizing circular RNAs. Nat Biotechnol 32(5):453–461 13. Guo J (2014) Transcription: the epicenter of gene
3. Lasda E, Parker R (2014) Circular RNAs: diversity of expression. J Zhejiang Univ Sci B 15(5):409–411
form and function. RNA 20(12):1829–1842 14. Salzman J, Gawad C, Wang PL et al (2012) Circular
RNAs are the predominant transcript isoform from
104 Shanmugapriya et al.

hundreds of human genes in diverse cell types. PLoS 31. Ruskin B, Krainer AR, Maniatis T et al (1984) Excision
One 7(2):e30733 of an intact intron as a novel lariat structure during
15. Memczak S, Jens M, Elefsinioti A et al (2013) pre-mRNA splicing in vitro. Cell 38(1):317–331
Circular RNAs are a large class of animal RNAs with 32. Tabak HF, Van der Horst G, Smit J et al (1988)
regulatory potency. Nature 495(7441):333–338 Discrimination between RNA circles, interlocked
16. Grabherr MG, Haas BJ, Yassour M et al (2011) RNA circles and lariats using two-dimensional poly-
Full-length transcriptome assembly from RNA-Seq acrylamide gel electrophoresis. Nucleic Acids Res
data without a reference genome. Nat Biotechnol 16(14A):6597–6605
29(7):644–U130 33. Li Z, Huang C, Bao C et al (2015) Exon-intron cir-
17. Lu T, Cui L, Zhou Y et al (2015) Transcriptome-­ cular RNAs regulate transcription in the nucleus. Nat
wide investigation of circular RNAs in rice. RNA Struct Mol Biol 22(3):256–264
21(12):2076–2087 34. Liang D, Wilusz JE (2014) Short intronic repeat
18. Saad FA, Vitiello L, Merlini L et al (1992) A 3′ sequences facilitate circular RNA production. Genes
consensus splice mutation in the human dystrophin Dev 28(20):2233–2247
gene detected by a screening for intra-exonic 35. Wang K, Long B, Liu F et al (2016) A circular RNA
deletions. Hum Mol Genet 1(5):345–346 protects the heart from pathological hypertrophy
19. Bailleul B (1996) During in vivo maturation of and heart failure by targeting miR-223. Eur Heart
eukaryotic nuclear mRNA, splicing yields excised J 37(33):2602–2611
exon circles. Nucleic Acids Res 24(6):1015–1019 36. Zheng QP, Bao CY, Guo WJ et al (2016) Circular
20. Hansen TB, Wiklund ED, Bramsen JB et al (2011) RNA profiling reveals an abundant circHIPK3 that
miRNA-dependent gene silencing involving regulates cell growth by sponging multiple miRNAs.
Ago2-mediated cleavage of a circular antisense Nat Commun 7:11215
RNA. EMBO J 30(21):4414–4422 37. Liu Q, Zhang X, Hu XQ et al (2016) Circular RNA
21. Zaphiropoulos PG (1997) Exon skipping and circu- related to the chondrocyte ECM regulates MMP13
lar RNA formation in transcripts of the human cyto- expression by functioning as a MiR-136 ‘Sponge’ in
chrome P-450 2C18 gene in epidermis and of the rat human cartilage degradation. Sci Rep 6:22572
androgen binding protein gene in testis. Mol Cell Biol 38. Xie H, Ren X, Xin S et al (2016) Emerging roles of
17(6):2985–2993 circRNA_001569 targeting miR-145 in the prolif-
22. Houseley JM, Garcia-Casado Z, Pascual M et al eration and invasion of colorectal cancer. Oncotarget
(2006) Noncanonical RNAs from transcripts of the 7(18):26680–26691
Drosophila muscleblind gene. J Hered 97(3):253–260 39. Thomas LF, Saetrom P (2014) Circular RNAs are
23. Li XF, Lytton J (1999) A circularized sodium-­ depleted of polymorphisms at microRNA binding
calcium exchanger exon 2 transcript. J Biol Chem sites. Bioinformatics 30(16):2243–2246
274(12):8153–8160 40. Zhang Y, Xue W, Li X et al (2016) The biogenesis of
24. Hooper JE (2014) A survey of software for genome-­ nascent circular RNAs. Cell Rep 15(3):611–624
wide discovery of differential splicing in RNA-Seq 41. Zhang XO, Wang HB, Zhang Y et al (2014)
data. Hum Genomics 8:3 Complementary sequence-mediated exon circulariza-
25. Kulpa D, Topping R, Telesnitsky A (1997) tion. Cell 159(1):134–147
Determination of the site of first strand transfer during 42. Ashwal-Fluss R, Meyer M, Pamudurti NR et al (2014)
Moloney murine leukemia virus reverse transcription circRNA biogenesis competes with pre-mRNA
and identification of strand transfer-associated reverse splicing. Mol Cell 56(1):55–66
transcriptase errors. EMBO J 16(4):856–865 43. Holt CE, Schuman EM (2013) The central dogma
26. Agabian N (1990) Trans splicing of nuclear pre-­ decentralized: new perspectives on RNA function and
mRNAs. Cell 61(7):1157–1160 local translation in neurons. Neuron 80(3):648–657
27. Capel B, Swain A, Nicolis S et al (1993) Circular 44. Denzler R, Agarwal V, Stefano J et al (2014)
transcripts of the testis-determining gene Sry in adult Assessing the ceRNA hypothesis with quantitative
mouse testis. Cell 73(5):1019–1030 measurements of miRNA and target abundance. Mol
28. Awan AR, Manfredo A, Pleiss JA (2013) Lariat Cell 54(5):766–776
sequencing in a unicellular yeast identifies regulated 45. Du WW, Yang W, Chen Y et al (2017) Foxo3 circu-
alternative splicing of exons that are evolutionarily lar RNA promotes cardiac senescence by modulating
conserved with humans. Proc Natl Acad Sci U S A multiple factors associated with stress and senescence
110(31):12762–12767 responses. Eur Heart J 38(18):1402–1412
29. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular 46. Du WW, Yang W, Liu E et al (2016) Foxo3 circu-
RNAs are abundant, conserved, and associated with lar RNA retards cell cycle progression via form-
ALU repeats. RNA 19(2):141–157 ing ternary complexes with p21 and CDK2. Nucleic
30. Sanger HL, Klotz G, Riesner D et al (1976) Viroids are Acids Res 44(6):2846–2858
single-stranded covalently closed circular RNA mol- 47. Chen CY, Sarnow P (1995) Initiation of protein
ecules existing as highly base-paired rod-like struc- synthesis by the eukaryotic translational apparatus on
tures. Proc Natl Acad Sci U S A 73(11):3852–3856 circular RNAs. Science 268(5209):415–417
8 Functional Analysis of Circular RNAs 105

48. Panda AC, Abdelmohsen K, Gorospe M (2017) 54. Hansen TB, Kjems J, Damgaard CK (2013)
RT-qPCR detection of senescence-associated circular Circular RNA and miR-7 in Cancer. Cancer Res
RNAs. Methods Mol Biol 1534:79–87 73(18):5609–5612
49. Orlando C, Pinzani P, Pazzagli M (1998) 55. Zhang Y, Zhang XO, Chen T et al (2013) Circular
Developments in quantitative PCR. Clin Chem Lab intronic long noncoding RNAs. Mol Cell
Med 36(5):255–269 51(6):792–806
50. Valasek MA, Repa JJ (2005) The power of real-time 56. Suzuki H, Tsukahara T (2014) A view of pre-mRNA
PCR. Adv Physiol Educ 29(3):151–159 splicing from RNase R resistant RNAs. Int J Mol Sci
51. Wong M, Medrano J (2005) Real-time PCR for 15(6):9331–9342
mRNA quantitation. BioTechniques 39(1):75–85 57. Suzuki H, Zuo YH, Wang JH et al (2006)
52. Bachmayr-Heyda A, Reiner AT, Auer K et al (2015) Characterization of RNase R-digested cellular RNA
Correlation of circular RNA abundance with prolifer- source that consists of lariat and circular RNAs from
ation–exemplified with colorectal and ovarian cancer, pre-mRNA splicing. Nucleic Acids Res 34(8):e63
idiopathic lung fibrosis, and normal human tissues. 58. Li Y, Chen B, Huang S (2018) Identification of
Sci Rep 5:8057 circRNAs for miRNA targets by Argonaute2 RNA
53. Tajadini M, Panjehpour M, Javanmard SH (2014) immunoprecipitation and luciferase screening assays.
Comparison of SYBR green and TaqMan methods in Methods Mol Biol 1724:209–218
quantitative real-time polymerase chain reaction anal- 59. Zeng X, Lin W, Guo M et al (2017) A comprehensive
ysis of four adenosine receptor subtypes. Adv Biomed overview and evaluation of circular RNA detection
Res 3:85 tools. PLoS Comput Biol 13(6):e1005420
 Part V
Circular RNAs as Potential Disease
Biomarkers
 Circular RNA in Exosomes
9
Daniele Fanale, Simona Taverna, Antonio Russo,
and Viviana Bazan

Abstract identifying new potential exosome-based can-

Circular RNAs (circRNAs) are a novel family cer biomarkers.
of non-coding endogenous RNAs discovered In this chapter, we briefly will describe the
in all eukaryotic cells and generated through a major features and functions of exosomal cir-
particular mechanism of alternative splicing cRNAs, discussing their potential role as
called “back-splicing”. These molecules show molecular biomarkers for diagnosis, progno-
multiple functions, by acting as modulators of sis and monitoring of complex diseases,
gene and miRNA expression, and may have a including cancer.
role in several biological processes, such as
cell proliferation and invasion with, tumour
development and progression, and in several
mechanisms underlying other diseases. Their
presence has been shown to be abundant in Keywords
several body fluids such as blood and saliva. Biomarkers · CDR1as · Circular RNAs
Based on their biogenesis mechanism, cir- (circRNAs) · Exosomes · Non-coding RNAs
cRNAs may be categorized into five classes:
exonic circRNAs, intronic circRNAs, anti-
sense circRNAs, sense overlapping circRNAs
and intergenic circRNAs. Recently, the pres- 1 Introduction
ence of circRNAs, in addition to that of miR-
NAs and long non-coding RNAs, has been In addition to non-coding RNAs such as small
detected also in small extracellular vesicles RNAs (microRNAs) and long non-coding RNAs
called exosomes. Investigating the presence (LncRNAs) [1–10], circular RNAs (circRNAs)
and expression levels of serum exosomal cir- represent a novel and large family of non-coding
cRNAs could allow us, in future, to discrimi- endogenous RNAs recently discovered in all
nate cancer patients from healthy individuals, eukaryotic cells and arising from a particular
alternative splicing mechanism of precursor
Daniele Fanale and Simona Taverna contributed equally mRNAs (pre-mRNAs) [11–14]. However, some
to this work few circRNAs have been shown to have the ability
D. Fanale · S. Taverna · A. Russo (*) · V. Bazan to be translated into proteins via insertion of an
Section of Medical Oncology, Department of internal ribosomal entry site [15, 16]. Differently
Surgical, Oncological and Oral Sciences, University from linear RNAs, circRNAs are covalently
of Palermo, Palermo, Italy

© Springer Nature Singapore Pte Ltd. 2018 109

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_9
110 D. Fanale et al.

Fig. 9.1 Biogenesis of circular RNAs

Canonical splicing (left panel) leads to linear mRNAs production, while back-­splicing (right panel) leads to circRNAs
lacking free 5′ and 3′ ends

closed single-stranded transcripts produced from circRNAs dates back some decades ago, they
exonic, intronic or intergenic regions and lacking were initially considered only as non-­functional
of the typical terminal structures (5′cap and 3′ artefacts of aberrant RNA splicing [23, 28, 29].
polyadenylated tails). The lack of these structures Only thanks to the introduction of recent bioinfor-
makes them more stable and resistant to exonu- matics and RNA deep sequencing technologies,
clease R than linear RNAs [17]. Three different several circRNAs have been detected, acquiring
biogenesis mechanisms to explain the origin of the deserved importance which they hold today
circRNAs have suggested that these RNAs are [18, 19]. CircRNAs play several and crucial func-
cyclized through a back-splicing process, in tions, as they can work as microRNA (miRNA)
which an upstream splice acceptor is linked to a sponges, negatively modulating miRNA expres-
downstream splice donor through direct splice or sion, as regulators of splicing and transcriptional
splice skipping [18–20] (Fig. 9.1). To date, cir- and posttranscriptional events, and as modifiers of
cRNAs are categorized into five classes: exonic parental gene expression [30] (Fig. 9.2). For
circRNA, intronic circRNA, antisense circRNA, example, CDR1as (also called ciRS-7) functions
sense overlapping circRNA, and intergenic cir- as a miR-7 sponge, regulating, via miR-7 targets,
cRNA [21, 22]. Most of circRNAs are amply con- the insulin transcription and secretion in pancre-
served and stable across different species and atic islet cells and thus opposing to the develop-
show specific features according to the tissue/cell ment of diabetes induced by miR-7 overexpression
type and developmental stage [23, 24]. Many cir- [31]. The human/mouse ciRS-7/CDR1as and
cRNAs have been found in numerous body fluids mouse Sry are the two most representative cir-
such as blood and saliva. Since they have been cRNAs acting as miRNA sponges [32]. The func-
shown to be tissue-specific and have hallmark tion of miRNA sponge enables circRNAs to
properties, this feature could make them potential control their activity and indirectly regulate the
and useful biomarkers for diagnosis, prognosis target mRNA stability [33]. In addition, circRNAs
and monitoring of several diseases, such as can- may have a role in the regulation of cell growth
cer, osteoarthritis, diabetes and neurodegenerative and invasion processes in several tumours, includ-
pathologies [25–27]. Although the discovery of ing gastric, colon and oesophageal cancers and
9 Circular RNA in Exosomes 111

Fig. 9.2 Schematic representation of circular RNA biological functions

CircRNAs contained in exosomes can have several functions such as transcription and translation regulation, splicing
regulation, miRNA sponge and protein inhibition

may allow to develop new approaches for cancer cancer patients from healthy individuals, identi-
detection and therapy [34]. However, the biologi- fying new potential exosome-based cancer bio-
cal functions of most circRNAs remain yet not markers [37].
totally understood. In this chapter, we will focus on the major
Most of cell types secretes nanosize-­ progress in the field of circRNA biology, report-
extracellular vesicles (EVs) of endolysosomal ing the current knowledge about their presence
origin called exosomes, containing a specific and biological role in exosomes and discussing
load of mRNAs, microRNAs, and proteins able their potential significance as molecular bio-
to affect the cell behaviour and potentially useful markers for diagnosis, prognosis, and monitoring
for diagnosis of several human diseases [35]. of complex diseases, including cancer.
Recently, RNA-seq analyses proved, for the first
time, the presence of several circRNAs with
potential biological function in exosomes. In par- 2 Exosomes
ticular, human serum exosomes have been shown
to contain more than 1000 circRNAs, probably One of the most attractive methods of cell-to-cell
arising from the entry into the bloodstream of cir- communication is mediated by EVs considered
cRNAs present in tumour [36]. Investigating the as an alternative to the paracrine and endocrine
presence and expression levels of serum exo- cellular system [38, 39]. The two better
somal circRNAs could allow us to differentiate ­characterized classes of EVs are exosomes and
112 D. Fanale et al.

microvesicles. Exosomes are the most deeply ExoCarta database [51] lists 9769 proteins, 3408
studied subpopulation of EVs [40]. Although mRNAs, and 2838 miRNAs contained in exo-
exosomes were initially considered as the “gar- somes collected from 286 published studies
bage bins” of cells [41, 42], in the last decades, (www.exocarta.org). This data reflects the num-
the attention of the researchers on the functions ber of targets that can be modulated by exosomes
of these vesicles is growing exponentially. and highlights the importance of studying them.
Exosomes are nanoscale EVs with lipid bilayer Proteomic studies have demonstrated that exo-
and are released into extracellular space after somes contained cytosolic, cytoskeletal and
fusion of multivesicular bodies (MVBs) with membrane proteins, integrins, enzymes, adhesion
plasma membrane [43]. Exosome formation is a and signalling molecules. Among exosomal pro-
mechanism consisting of four stages: initiation, teins, tetraspanins and heat shock proteins are the
endocytosis, MVBs formation, and exosome most conserved molecules [52, 53]. Exosomal
secretion [44]. In the first stage, early endosomes proteins maintain their biological activities such
mature into late endosomes or MVBs; during this as antigen presentation, protein cleavage and
process the endosomal membrane invaginates to pathway activation.
produce intraluminal vesicles (ILVs) in the lumen Moreover, RNA populations in exosomes
of MVBs [45]. In this mechanism is involved were identified using high-throughput RNA-seq,
ESCRT machinery that consists of four protein including messenger RNAs (mRNAs) and many
complexes: ESCRT-0, ESCRT-I, ESCRT-II, types of non-coding RNAs, such as circRNAs,
ESCRT-III and its associated proteins, such as miRNAs, transfer RNAs (tRNAs), ribosomal
TSG101 and Alix that are used as markers of exo- RNAs (rRNAs), lncRNAs, small nuclear RNAs
somal population [46]. (snRNAs), small nucleolar RNAs (snoRNAs)
Exosomes may be collected by several bio- and piwi-interacting RNAs (piRNAs) [54, 55].
logical fluids such as blood, urine, saliva, breast These RNAs can be shuttled from parental to tar-
milk, synovial fluid, amniotic fluid, bronchoal- get cells, where they modulate target genes or are
veolar lavage fluid, malignant ascites and semen the templates for protein synthesis. It was
[47, 48]. Exosomes can be internalized by target reported that exosomes act as nano-shuttles for
cells in the closeness or cells at significant dis- miRNAs with a dual role in cancer progression.
tance from parental cells. These vesicles can have In this context, they can have oncogenic and
different fates: they may interact with cells, by tumour-suppressor functions [56]. Recently, cir-
acting as messenger shuttles, in order to transfer cRNAs were found enriched and stable in cancer
information that can modulate the phenotype of exosomes. The transport of nucleic acids by exo-
target cells. Several mechanisms mediate exo- somes ensures the protection against degradation
somal uptake, including exosome fusion with the and dilution in the extracellular space, allowing
plasma membrane of target cells, leading to the long-distance distribution through the blood-
release of exosomal contents into the cytoplasm, stream or interstitial fluid [57].
endocytosis by phagocytosis and juxtacrine sig- Several papers reported a pleiotropic role of
nalling through receptor-ligand interactions. It tumour-derived exosomes, as they are involved in
was demonstrated that cancer cells released about tumour growth, angiogenesis, metastasis, modu-
10 folds more exosomes than normal cell to lation of the microenvironment, pre-metastatic
mediate tumour progression [49]. niche formation, immunomodulation and drug
Exosomes carry bioactive cargos, including resistance [58, 59]. Since exosomes shuttle their
common and donor cell-specific proteins, lipids typical cargo through the bloodstream and medi-
and RNA and DNA molecules that reflected cells ate the horizontal transfer of genetic material
and tissue of origin and provided a snapshot of from parental to target cells [56, 60], the idea of
cells at the time of release [50]. Exosomal com- “liquid biopsy” encouraged studies on exosomes-­
position can be different from parental cells based biomarkers especially in cancer as ­potential
thanks to the selective sorting of the cargos. cancer biomarkers and theranostic devices [61].
9 Circular RNA in Exosomes 113

Recently, a position paper by the International decreased in exosomes and improved in cells,
Society for Extracellular Vesicles (ISEV) sum- upon ectopic expression of miR-7 in both
marized the recent application and current find- HEK293T and MCF-7 cells. Exosomal CDR1as
ings on the EVs-based therapies [62]. The maintained biological activity also in exosomes
translation of these vesicles in clinical practice abrogating miR-7-induced growth suppression in
requires a classification of EVs-based therapies target cells [66, 67].
in agreement with supervisory outlines [62]. Since circRNAs are abundant in exosomes,
Substantial improvement in exosomes studies they can be collected by human blood. In order to
has directed to upgraded and standardized proto- test if exo-circRNA enters into the circulation
cols for purification and storage, as well as and is quantifiable for cancer diagnosis, Li and
methods and standards for quality analyses of colleagues [68] used a xenograft mouse model of
exosome-based cures [63]. Clinical trials propos- human MHCC-LM3 cancer cells. These cells
ing EVs as theranostic nanoparticles have been were inoculated in mice, and 7 weeks later, serum
described in the early 2000s; exosome power on from mice was harvested, and exosomal cir-
clinical research is established by numerous cRNAs were isolated and quantified by qRT-PCR
current clinical trials (https://clinicaltrials.gov/). analysis. The human CDYL circRNA was
Nowadays, 20 clinical trials investigate on EVs detected in serum from tumour-bearing mice, and
as biomarkers for diagnosis, prognosis or devices the amount of this circRNA in xenografted mice
for drug delivery. Exosomes are also used as a was correlated with tumour mass [68]. To con-
new tool for clinical evaluation and screening firm the idea that circRNAs enriched in exosomes
system in liquid biopsy approaches [64, 65]. may represent biomarkers for cancer diagnosis
Only one clinical trial with circular RNAs is and prognosis, the expression profile of serum
ongoing. The aim of this study is to develop a exosomal circRNAs was explored in cancer
slightly invasive analysis to identify pancreatic patients and healthy donors. The expression of
cancer at initial stage of neoplasia and check the circRNAs in serum from 11 colorectal cancer
response to treatment, but there are no clinical tri- patients, tested by RNA-seq analysis, was signifi-
als that investigate exosomal circRNAs. cantly different from healthy donors; in cancer
patients, 67 circRNAs were lost, and 257 new cir-
cRNA types were found compared to healthy
2.1 Circular RNA in Exosomes individuals [31].
In human serum more than 1000 exosomal
In 2015, Li et al. [62] reported, for the first time, circRNAs useful to discriminate patients with
that exosomes contain abundant circRNAs. tumour from healthy controls were identified.
Genome-wide RNA-seq analyses discovered that These data suggest that circRNAs derived from
circRNAs were enriched in exosomes compared human cancer can enter in the bloodstream and
to parental cells. It was reported that circRNA be easily quantified in serum. CircRNA expres-
sorting to exosomes may be controlled by modu- sion profiles have been also performed in both
lation of associated miRNA levels in parental cells and exosomes from three isogenic colorec-
cells and may transfer biological activity to target tal cancer cell lines that vary in KRAS mutational
cells. Considering that circRNAs sponge miR- status. Although circRNAs have a tendency to be
NAs, the correlation between circRNAs and enriched in exosomes, circRNA concentration
miRNAs, about circRNA shuttled with exo- decreased at global level in mutant-KRAS cell
somes, was investigated. The circRNA, CDR1as, lines, indicating a modulation of circRNAs dur-
is known to work as a miR-7 sponge, because ing colorectal cancer progression and a possible
miR-7 mimics were introduced into HEK293T contribution of circRNAs in oncogenesis [69].
and MCF-7 cell lines and the level of CDR1as in Furthermore, RNA-seq technique allows to
exosomes and parental cells was determined. It profile circRNA expression in EVs isolated
was described that CDR1as level was deeply from serum of patients with endometrial cancer
114 D. Fanale et al.

and healthy controls. It was found that the num- high powerful detection and monitoring strate-
ber of upregulated circRNAs was higher than gies for cancer risk indication, useful for patients
that of downregulated circRNAs in EVs from to receive the most appropriate therapy and for
patients compared to healthy subjects. Xu et al. clinicians to monitor the disease progression,
[25] reported that circRNAs may act as compet- regression and recurrence. Recent studies indi-
ing endogenous RNAs in receipt cells after cate exosomes as potential biomarkers for diag-
internalization of EVs from cancer cells. They nosis, prognosis and prediction in cancer. The
identified 209 upregulated and 66 downregu- goal of exosomes used as biomarker is the sub-
lated circRNAs in EVs from serum of patients stantial reduction of sample complexity, when
with endometrial cancer compared to those compared to whole body fluids, and the low inva-
from healthy controls. The roles of differently siveness in a liquid biopsy scenario [65, 73].
expressed circRNAs by using KEGG pathway Recently, exosomal circRNAs have been sug-
enrichment analysis were investigated. The gested as potential biomarkers in cancer for their
expression of two circRNAs, hsa circ 0109046 stability and high specificity. These new findings
and hsa circ 0002577, was confirmed by could be translated in clinical practice in order to
RT-qPCR, and the circRNA/miRNA interac- discriminate patients with cancer from healthy
tions for these two circRNAs were also pre- individuals with high accuracy.
dicted. Overall, these data indicate that exosomal
circRNAs can influence target cells contributing Competing Financial Interests The authors declare no
to the identification of new mechanisms of can- competing financial interests.
cer development [25]. Recently, the studies on
circRNAs have increasing value in the field of
genomic research and their hallmark properties References
convinced the researchers to explore the powers
1. Caruso S, Bazan V, Rolfo C, Insalaco L, Fanale D,
of these molecules as biomarkers for complex Bronte G et al (2012) MicroRNAs in colorectal
diseases such as cancer [17, 70]. cancer stem cells: new regulators of cancer stem-
ness? Oncogene 1(11):e32. https://doi.org/10.1038/
oncsis.2012.33
2. Corsini LR, Bronte G, Terrasi M, Amodeo V, Fanale
3 Conclusions and Future D, Fiorentino E et al (2012) The role of microRNAs
Perspectives in cancer: diagnostic and prognostic biomarkers
and targets of therapies. Expert Opin Ther Targets
Exosomes, considered “diamonds in the rough”, 16(sup2):S103–S1S9. https://doi.org/10.1517/14728
222.2011.650632
in combination with circRNAs for their peculiar- 3. Rolfo C, Fanale D, Hong D, Tsimberidou A, Piha-Paul
ity and high specificity can increase the potential S, Pauwels P et al (2014) Impact of microRNAs in
use of both exosomes and circRNAs as markers resistance to chemotherapy and novel targeted agents
of diagnosis and prognosis for cancer patients. in non-small cell lung cancer. Curr Pharm Biotechnol
15(5):475–485. https://doi.org/10.2174/13892010156
The biomarkers are biological molecules con- 66140519123219
tained in blood, tissues and body fluids that can 4. Amodeo V, Bazan V, Fanale D, Insalaco L, Caruso
be objectively evaluated and measured as indica- S, Cicero G et al (2013) Effects of anti-miR-182 on
tors for normal and pathological conditions [71]. TSP-1 expression in human colon cancer cells: there
is a sense in antisense? Expert Opin Ther Targets
The use of biomarkers is important for early 17(11):1249–1261. https://doi.org/10.1517/14728222
detection and diagnosis of different diseases as .2013.832206
well as for monitoring the responses to treat- 5. Bronte F, Bronte G, Fanale D, Caruso S, Bronte
ments [72]. The typical features of the biomark- E, Bavetta MG et al (2016) HepatomiRNoma: the
proposal of a new network of targets for diagnosis,
ers such as stability, sensitivity and specificity prognosis and therapy in hepatocellular carcinoma.
allow their use in clinical practice. The goal of Crit Rev Oncol Hematol 97:312–321. https://doi.
precision medicine, in particular the liquid biop- org/10.1016/j.critrevonc.2015.09.007.
sies, is the discovery of cancer biomarkers with
9 Circular RNA in Exosomes 115

6. Fanale D, Amodeo V, Bazan V, Insalaco L, Incorvaia cRNA Biogenesis competes with Pre-mRNA splic-
L, Barraco N et al (2016) Can the microRNA expres- ing. Mol Cell 56(1):55–66. https://doi.org/10.1016/j.
sion profile help to identify novel targets for zole- molcel.2014.08.019
dronic acid in breast cancer? Oncotarget 7(20). 20. Conn Simon J, Pillman Katherine A, Toubia J, Conn
https://doi.org/10.18632/oncotarget.8722 Vanessa M, Salmanidis M, Phillips Caroline A et al
7. Cabibi D, Caruso S, Bazan V, Castiglia M, Bronte G, (2015) The RNA binding protein quaking regulates
Ingrao S et al (2016) Analysis of tissue and circulating formation of circRNAs. Cell 160(6):1125–1134.
microRNA expression during metaplastic transforma- https://doi.org/10.1016/j.cell.2015.02.014
tion of the esophagus. Oncotarget 7(30). https://doi. 21. Memczak S, Jens M, Elefsinioti A, Torti F, Krueger
org/10.18632/oncotarget.10291 J, Rybak A et al (2013) Circular RNAs are a large
8. Fanale D, Barraco N, Listì A, Bazan V, Russo A class of animal RNAs with regulatory potency.
(2016) Non-coding RNAs functioning in colorectal Nature 495(7441):333–338. https://doi.org/10.1038/
cancer stem cells. Adv Exp Med Biol 937:93–108. nature11928
https://doi.org/10.1007/978-3-319-42059-2_5 22. Zheng Q, Bao C, Guo W, Li S, Chen J, Chen B et al
9. Fanale D, Castiglia M, Bazan V, Russo A (2016) (2016) Circular RNA profiling reveals an abundant
Involvement of non-coding RNAs in chemo- and circHIPK3 that regulates cell growth by sponging
radioresistance of. Colorectal Cancer 937:207–228. multiple miRNAs. Nat Commun 7:11215. https://doi.
https://doi.org/10.1007/978-3-319-42059-2_11 org/10.1038/ncomms11215
10. Rizzo S, Cangemi A, Galvano A, Fanale D, Buscemi 23. Hou L-D, Zhang J (2017) Circular RNAs:
S, Ciaccio M et al (2017) Analysis of miRNA expres- an emerging type of RNA in cancer. Int
sion profile induced by short term starvation in breast J Immunopathol Pharmacol 30(1):1–6. https://doi.
cancer cells treated with doxorubicin. Oncotarget org/10.1177/0394632016686985
8(42). https://doi.org/10.18632/oncotarget.18028 24. Moran JV, Salzman J, Chen RE, Olsen MN, Wang
11. Zhong Z, Lv M, Chen J (2016) Screening differential PL, Brown PO (2013) Cell-type specific features of
circular RNA expression profiles reveals the regula- circular RNA expression. PLoS Genet 9(9):e1003777.
tory role of circTCF25-miR-103a-3p/miR-107-CDK6 https://doi.org/10.1371/journal.pgen.1003777
pathway in bladder carcinoma. Sci Rep 6:1–12. 25. Xu Z, Yan Y, Zeng S, Dai S, Chen X, Wei J et al
https://doi.org/10.1038/srep30919 (2017) Circular RNAs: clinical relevance in can-
12. Chen L, Huang C, Wang X, Shan G (2015) Circular cer. Oncotarget 9(1). https://doi.org/10.18632/
RNAs in eukaryotic cells. Curr Genomics 16(5):312– oncotarget.22846
318. https://doi.org/10.2174/13892029166661507071 26. Kristensen LS, Hansen TB, Venø MT, Kjems J (2017)
61554 Circular RNAs in cancer: opportunities and chal-
13. Barrett SP, Salzman J (2016) Circular RNAs: analy- lenges in the field. Oncogene 37(5):555–565. https://
sis, expression and potential functions. Development doi.org/10.1038/onc.2017.361
143(11):1838–1847. https://doi.org/10.1242/ 27. Greene J, Baird A-M, Brady L, Lim M, Gray SG,
dev.128074 McDermott R et al (2017) Circular RNAs: biogenesis,
14. Gao Y, Wang J, Zheng Y, Zhang J, Chen S, Zhao F function and role in human diseases. Front Mol Biosci
(2016) Comprehensive identification of internal struc- 4. https://doi.org/10.3389/fmolb.2017.00038
ture and alternative splicing events in circular RNAs. 28. Arnberg AC, Van Ommen GJB, Grivell LA, Van
Nat Commun 7:12060. https://doi.org/10.1038/ Bruggen EFJ, Borst P (1980) Some yeast mitochon-
ncomms12060 drial RNAs are circular. Cell 19(2):313–319. https://
15. Wang Y, Wang Z (2015) Efficient backsplicing pro- doi.org/10.1016/0092-8674(80)90505-x
duces translatable circular mRNAs. RNA 21(2):172– 29. Capel B, Swain A, Nicolis S, Hacker A, Walter
179. https://doi.org/10.1261/rna.048272.114 M, Koopman P et al (1993) Circular tran-
16. Chen C, Sarnow P (1995) Initiation of protein synthe- scripts of the testis-determining gene Sry in adult
sis by the eukaryotic translational apparatus on cir- mouse testis. Cell 73(5):1019–1030. https://doi.
cular RNAs. Science 268(5209):415–417. https://doi. org/10.1016/0092-8674(93)90279-y
org/10.1126/science.7536344 30. Haque S, Circular LWH (2017) RNAs (circRNAs)
17. Huang S, Yang B, Chen BJ, Bliim N, Ueberham in health and disease. Genes 8(12):353. https://doi.
U, Arendt T et al (2017) The emerging role of cir- org/10.3390/genes8120353
cular RNAs in transcriptome regulation. Genomics 31. Xu H, Guo S, Li W, Yu P (2015) The circular RNA
109(5–6):401–407. https://doi.org/10.1016/j. Cdr1as, via miR-7 and its targets, regulates insulin
ygeno.2017.06.005 transcription and secretion in islet cells. Sci Rep 5(1).
18. Jeck WR, Sorrentino JA, Wang K, Slevin MK, Burd https://doi.org/10.1038/srep12453
CE, Liu J et al (2012) Circular RNAs are abun- 32. Chen I, Chen C-Y, Chuang T-J (2015) Biogenesis,
dant, conserved, and associated with ALU repeats. identification, and function of exonic circular RNAs.
RNA 19(2):141–157. https://doi.org/10.1261/ Wiley Interdiscip Rev RNA 6(5):563–579. https://doi.
rna.035667.112 org/10.1002/wrna.1294
19. Ashwal-Fluss R, Meyer M, Pamudurti Nagarjuna 33. Kulcheski FR, Christoff AP, Margis R (2016) Circular
R, Ivanov A, Bartok O, Hanan M et al (2014) cir- RNAs are miRNA sponges and can be used as a new
116 D. Fanale et al.

class of biomarker. J Biotechnol 238:42–51. https:// biomarkers. PROTEOMICS Clin Appl 9(3–4):358–
doi.org/10.1016/j.jbiotec.2016.09.011 367. https://doi.org/10.1002/prca.201400114
34. Zhu L-P, He Y-J, Hou J-C, Chen X, Zhou S-Y, Yang S-J 48. Rodriguez-dorantes M, Romero-Cordoba S, Peralta-­
et al (2017) The role of circRNAs in cancers. Biosci Zaragoza O, Salido-Guadarrama I, Hidalgo-Miranda
Rep 37(5):BSR20170750. https://doi.org/10.1042/ A (2014) MicroRNAs transported by exosomes in
bsr20170750 body fluids as mediators of intercellular communica-
35. Pant S, Hilton H, Burczynski ME (2012) The multifac- tion in cancer. OncoTargets Ther 7:1327. https://doi.
eted exosome: biogenesis, role in normal and aberrant org/10.2147/ott.s61562
cellular function, and frontiers for pharmacological 49. Li W, Li C, Zhou T, Liu X, Liu X, Li X et al (2017) Role
and biomarker opportunities. Biochem Pharmacol of exosomal proteins in cancer diagnosis. Mol Cancer
83(11):1484–1494. https://doi.org/10.1016/j. 16(1). https://doi.org/10.1186/s12943-017-0706-8
bcp.2011.12.037 50. Hessvik NP, Llorente A (2017) Current knowl-
36. Li S, Li Y, Chen B, Zhao J, Yu S, Tang Y et al (2018) edge on exosome biogenesis and release. Cell Mol
exoRBase: a database of circRNA, lncRNA and Life Sci 75(2):193–208. https://doi.org/10.1007/
mRNA in human blood exosomes. Nucleic Acids Res s00018-017-2595-9
46(D1):D106–DD12. https://doi.org/10.1093/nar/ 51. Keerthikumar S, Chisanga D, Ariyaratne D, Al
gkx891 Saffar H, Anand S, Zhao K et al (2016) ExoCarta:
37. Siegel RL, Sahar L, Portier KM, Ward EM, Jemal A a web-based compendium of exosomal cargo. J Mol
(2015) Cancer death rates in US congressional dis- Biol 428(4):688–692. https://doi.org/10.1016/j.
tricts. CA Cancer J Clin 65(5):339–344 jmb.2015.09.019
38. Maia J, Caja S, Strano Moraes MC, Couto N, 52. Haraszti RA, Didiot M-C, Sapp E, Leszyk J, Shaffer SA,
Costa-Silva B (2018) Exosome-based cell-cell com- Rockwell HE et al (2016) High-resolution proteomic
munication in the tumor microenvironment. Front and lipidomic analysis of exosomes and microvesi-
Cell Dev Biol 20(6):18. https://doi.org/10.3389/ cles from different cell sources. J Extracell Vesicles
fcell.2018.00018 5(1):32570. https://doi.org/10.3402/jev.v5.32570
39. Taverna S, Pucci M, Alessandro R (2017) Extracellular 53. Simona F, Laura S, Simona T, Riccardo A (2013)
vesicles: small bricks for tissue repair/regeneration. Contribution of proteomics to understanding the role
Ann Transl Med 5(4):83. https://doi.org/10.21037/ of tumor-derived exosomes in cancer progression:
atm.2017.01.53 state of the art and new perspectives. Proteomics
40. Tkach M, Théry C (2016) Communication by extra- 13(10–11):1581–1594. https://doi.org/10.1002/
cellular vesicles: where we are and where we need to pmic.201200398
go. Cell 164(6):1226–1232. https://doi.org/10.1016/j. 54. Nolte-‘t Hoen ENM, Buermans HPJ, Waasdorp M,
cell.2016.01.043 Stoorvogel W, Wauben MHM, ‘t Hoen PAC. Deep
41. Wolf P (1967) The nature and significance of platelet sequencing of RNA from immune cell-derived ves-
products in human plasma. Br J Haematol 13(3):269– icles uncovers the selective incorporation of small
288. https://doi.org/10.1111/j.1365-2141.1967. non-coding RNA biotypes with potential regulatory
tb08741.x functions. Nucleic Acids Res 2012;40(18):9272–
42. Pan B-T, Johnstone RM (1983) Fate of the trans- 9285. doi:https://doi.org/10.1093/nar/gks658.
ferrin receptor during maturation of sheep retic- 55. Kim KM, Abdelmohsen K, Mustapic M, Kapogiannis
ulocytes in vitro: selective externalization of D (2017) Gorospe M. RNA in extracellular vesicles.
the receptor. Cell 33(3):967–978. https://doi. Wiley Interdisciplinary Reviews RNA 8(4):e1413.
org/10.1016/0092-8674(83)90040-5 https://doi.org/10.1002/wrna.1413
43. Colombo M, Raposo G, Théry C (2014) Biogenesis, 56. Pucci M, Reclusa Asiáin P, Duréndez Sáez E,
secretion, and intercellular interactions of exosomes Jantus-Lewintre E, Malarani M, Khan S et al (2018)
and other extracellular vesicles. Annu Rev Cell Extracellular vesicles as miRNA nano-shuttles: dual
Dev Biol 30(1):255–289. https://doi.org/10.1146/ role in tumor progression. Target Oncol. https://doi.
annurev-cellbio-101512-122326 org/10.1007/s11523-018-0551-8
44. Kowal J, Tkach M, Théry C (2014) Biogenesis and 57. Morton DJ, Kuiper EG, Jones SK, Leung SW, Corbett
secretion of exosomes. Curr Opin Cell Biol 29:116– AH, Fasken MB (2018) The RNA exosome and RNA
125. https://doi.org/10.1016/j.ceb.2014.05.004 exosome-linked disease. RNA 24(2):127–142. https://
45. Huotari J, Helenius A (2011) Endosome maturation. doi.org/10.1261/rna.064626.117
EMBO J 30(17):3481–3500. https://doi.org/10.1038/ 58. Ruivo CF, Adem B, Silva M, Melo SA (2017) The
emboj.2011.286 biology of cancer exosomes: insights and new per-
46. Henne WM, Stenmark H, Emr SD (2013) Molecular spectives. Cancer Res 77(23):6480–6488. https://doi.
mechanisms of the membrane sculpting ESCRT path- org/10.1158/0008-5472.can-17-0994
way. Cold Spring Harb Perspect Biol 5(9):a016766. 59. Steinbichler TB, Dudás J, Riechelmann H, Skvortsova
https://doi.org/10.1101/cshperspect.a016766 I-I (2017) The role of exosomes in cancer metas-
47. Boukouris S, Mathivanan S (2015) Exosomes in tasis. Semin Cancer Biol 44:170–181. https://doi.
bodily fluids are a highly stable resource of disease org/10.1016/j.semcancer.2017.02.006
9 Circular RNA in Exosomes 117

60. Valadi H, Ekström K, Bossios A, Sjöstrand M, 66. Hansen TB, Kjems J, Damgaard CK (2013)
Lee JJ, Lötvall JO (2007) Exosome-mediated Circular RNA and miR-7 in cancer. Cancer Res
transfer of mRNAs and microRNAs is a novel 73(18):5609–5612. https://doi.org/10.1158/0008-
mechanism of genetic exchange between 5472.can-13-1568
cells. Nat Cell Biol 9(6):654–659. https://doi. 67. Peng LI, Yuan XQ, Li GC (2015) The emerging land-
org/10.1038/ncb1596 scape of circular RNA ciRS-7 in cancer (Review).
61. He C, Zheng S, Luo Y, Wang B (2018) Exosome Oncol Rep 33(6):2669–2674. https://doi.org/10.3892/
theranostics: biology and translational medicine. or.2015.3904
Theranostics 8(1):237–255. https://doi.org/10.7150/ 68. Li Y, Zheng Q, Bao C, Li S, Guo W, Zhao J et al
thno.21945 (2015) Circular RNA is enriched and stable in exo-
62. Lener T, Gimona M, Aigner L, Börger V, Buzas E, somes: a promising biomarker for cancer diagnosis.
Camussi G et al (2015) Applying extracellular ves- Cell Res 25(8):981–984. https://doi.org/10.1038/
icles based therapeutics in clinical trials – an ISEV cr.2015.82
position paper. J Extracell Vesicles 4(1):30087. 69. Dou Y, Cha DJ, Franklin JL, Higginbotham JN,
https://doi.org/10.3402/jev.v4.30087 Jeppesen DK, Weaver AM et al (2016) Circular RNAs
63. Coumans FAW, Brisson AR, Buzas EI, Dignat-­ are down-regulated in KRAS mutant colon cancer
George F, Drees EEE, El-Andaloussi S et al (2017) cells and can be transferred to exosomes. Sci Rep
Methodological Guidelines to Study Extracellular 6(1):37982. https://doi.org/10.1038/srep37982
Vesicles. Circ Res 120(10):1632–1648. https://doi. 70. Qu S, Zhong Y, Shang R, Zhang X, Song W, Kjems
org/10.1161/circresaha.117.309417 J et al (2016) The emerging landscape of circular
64. Reiner AT, Witwer KW, van Balkom BWM, de RNA in life processes. RNA Biol 14(8):992–999.
Beer J, Brodie C, Corteling RL et al (2017) Concise https://doi.org/10.1080/15476286.2016.1220473
review: developing best-practice models for the 71. Henry NL, Hayes DF (2012) Cancer biomarkers.
therapeutic use of extracellular vesicles. Stem Cells Mol Oncol 6(2):140–146. https://doi.org/10.1016/j.
Transl Med 6(8):1730–1739. https://doi.org/10.1002/ molonc.2012.01.010
sctm.17-0055 72. Mayeux R (2004) Biomarkers potential uses and
65. Taverna S, Giallombardo M, Gil-Bazo I, Carreca AP, limitations. NeuroRx 1(2):182–188. https://doi.
Castiglia M, Chacártegui J et al (2016) Exosomes org/10.1602/neurorx.1.2.182
isolation and characterization in serum is feasible in 73. Rolfo C, Castiglia M, Hong D, Alessandro R,
non-small cell lung cancer patients: critical analy- Mertens I, Baggerman G et al (2014) Liquid biop-
sis of evidence and potential role in clinical prac- sies in lung cancer: the new ambrosia of researchers.
tice. Oncotarget 7(19):28748–28760. https://doi. Biochim Biophys Acta 1846(2):539–546. https://doi.
org/10.18632/oncotarget.7638 org/10.1016/j.bbcan.2014.10.001
 Circular RNAs in Blood
10
Angela Vea, Vicenta Llorente-Cortes,
and David de Gonzalo-Calvo

Abstract also discussed. Finally, perspectives for future

Recent advances in RNA sequencing and bio- studies are proposed.
informatic analysis have allowed the develop-
ment of a new research field: circular RNAs Keywords
(circRNAs). These members of the non-­ Circular RNA · Blood · Serum · Plasma ·
coding transcriptome are generated by backs- Extracellular vesicles
plicing, which results in a covalently closed,
single-stranded RNA molecule. To date, thou-
sands of circRNAs have been identified in dif-
ferent human cell types. CircRNAs are 1 Introduction
evolutionarily conserved, highly stable, cell-/
developmental stage-specific and have longer Non-coding RNAs (ncRNAs) are a heteroge-
half-lives compared with linear RNAs. neous group of RNA molecules that do not
Interestingly, different studies have demon- encode proteins but have key regulatory and
strated that circRNAs are abundantly structural functions. During the last years, most
expressed in the bloodstream. In this chapter, studies have been focused on members of this
we review the current knowledge of circRNA family, such as microRNAs (miRNAs) and long
biology in blood cells and the cell-free com- non-coding RNAs (lncRNAs). Recent advances
partment, including extracellular vesicles. The in high-throughput RNA sequencing (RNA-Seq)
potential clinical application of blood cir- and computational analysis have drawn attention
cRNAs in the biomarker and therapy fields is to a new class of ncRNAs, circular RNAs (cir-
cRNAs), as a natural feature of the cell expres-
A. Vea sion programme.
Biomedical Research Institute Sant Pau CircRNAs are single-stranded and covalently
(IIB Sant Pau), Barcelona, Spain
closed RNA molecules that lack of free caps or
V. Llorente-Cortes · D. de Gonzalo-Calvo (*) poly(A) tails [1]. CircRNAs are generated by a
Biomedical Research Institute Sant Pau (IIB Sant
Pau), Barcelona, Spain
process called backsplicing in which a splice
donor site is joined to a splice acceptor site
Institute of Biomedical Research of Barcelona
(IIBB) – Spanish National Research Council (CSIC),
upstream in the primary transcript. These
Barcelona, Spain ncRNAs are mainly formed by exons, but they
CIBERCV, Institute of Health Carlos III,
can also be derived from intronic, non-coding,
Madrid, Spain antisense, untranslated or intergenic genomic

© Springer Nature Singapore Pte Ltd. 2018 119

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_10
120 A. Vea et al.

regions [2]. Their size ranges from hundreds to Evidence of potential functions in the regula-
thousands of nucleotides. The biogenesis of tion of the transcriptome and proteome is con-
mammalian circRNAs is not fully understood tinuously emerging. Cytoplasmic circRNAs can
[1]. CircRNAs are derived from pre-mRNAs, inhibit miRNA function, acting as miRNA
which are transcribed by RNA polymerase sponges by complementary base pairing [18],
II. The process of backsplicing requires the and participating in RNA networks, acting as
canonical spliceosomal machinery and seems to competing endogenous RNAs (ceRNAs) [19].
depend on sequencing motifs within flanking By contrast, circRNA could also serve as a
introns and RNA-binding proteins (RBPs) such miRNA reservoir by stabilizing miRNAs [20].
as quaking [3, 4]. Different circular isoforms CircRNAs regulate RNA transcription by bind-
with different expression patterns can be pro- ing to RNA polymerase II [14] or DNA [21] and
duced from a given gene [5]. Interestingly, cir- can interact with RBPs participating in their stor-
cRNA production is regulated independently of age, localization and function [22]. Additionally,
the underlying linear RNA gene [6]. In some recent studies revealed that several circRNAs
cases, the expression level of circRNAs is sig- could function as coding transcripts [23].
nificantly higher than that of their corresponding Despite the great advances achieved during
linear RNA isoforms from the same gene [7]. the last years, current knowledge of the biologi-
The identification of circRNAs supports the con- cal functions and potential clinical implications
cept that genes are complex transcriptional units of circRNAs remain limited. The modulation of
that contain multiple and overlapping informa- circRNAs on gene expression plays a significant
tion [8, 9]. role in a great variety of pathological conditions,
CircRNAs were initially discovered in plant including cancer and cardiovascular disease [24,
viroids as early as the 1970s [10]. This class of 25]. Thus, circRNAs are promising therapeutic
ncRNAs were originally considered as splicing targets for future drugs. Interestingly, circRNAs
by-products, background noise or specific to a have been consistently identified in the blood-
few pathogens such as viruses [11, 12]. stream and therefore are potential minimally
Recently, transcriptome-wide circRNA analy- invasive biomarkers. In this chapter, we provide
sis has identified and characterized thousands an overview of the presence of circRNAs in
of circRNAs in diverse human cells [2, 7, 13], human blood and their biological role (Fig. 10.1).
suggesting the relevance of circRNAs in the We also discuss their future clinical application
ncRNA family. Studies of circRNAs in differ- as biomarkers and therapeutics targets.
ent species have shown that the majority of cir-
cRNAs are evolutionarily conserved at the
sequence level [7], pointing to a key role in rel- 2 Circular RNAs in Blood
evant biological processes. CircRNAs are pre-
dominantly cytoplasmic [2], although their The RNA profile of the bloodstream can reflect the
presence has also been described in the nucleus transcriptomic changes in blood cells. Furthermore,
[14], and tend to accumulate in cells with a low the circulating cell-free RNA can be informative
proliferation rate, such as neurons [15]. of the alterations in the gene expression of differ-
Circularity confers specific properties to cir- ent non-haematopoietic and haematopoietic cells.
cRNAs. In comparison to linear RNAs, cir- Therefore, the circulating transcriptomic biosigna-
cRNAs are highly stable and less susceptible to ture is a promising tool that may adequately reflect
degradation by ribonuclease R (RNase R) and the molecular fingerprint of the subject phenotype.
have longer half-lives in cells [7]. Mounting The results from a number of publications point to
data have demonstrated that circRNAs are circRNAs as novel regulatory elements of blood
expressed in a cell-, developmental stage- and cell biology and biomarkers with potential clinical
disease-specific manner [13, 16, 17]. application.
10 Circular RNAs in Blood 121

Fig. 10.1 Summary of publications demonstrating the lets) and in cell-free compartments, including extracellu-
presence of circular RNAs in blood lar vesicles. Circular RNAs have also been detected in
The presence of circular RNAs has been described in bone marrow cells
blood cells (red blood cells, white blood cells and plate-

2.1 Circular RNA in Whole Blood The top expressed blood circRNAs and the same
number of top linear RNAs showed significant
Previous investigations proposed that the whole enrichment of different biological function anno-
blood is enriched in circRNAs. RNA-Seq analy- tations, which suggests that circRNA expression
sis of two independent human whole blood sam- levels are independent of the linear RNA isoform
ples, which were processed following standard abundance. Most functions of blood circRNAs
procedures, identified 4550 and 4105 unique cir- were related to transcription regulation. Seven
cRNA candidates in each sample, with approxi- candidates were further evaluated using alterna-
mately 2400 circRNAs reproducibly detected tive methodology, including PCR. These candi-
[26]. Most blood circRNAs were derived from dates were expressed from loci that were not
protein-coding exonic regions or 5’ UTR related to specific blood-related functions, which
sequences. CircRNA expression levels were generated new questions about the function of
comparable to the circRNA-rich tissue cerebel- circRNAs in the bloodstream. Importantly, hun-
lum and > 15-fold higher compared to the liver. dreds of circRNAs were much more highly
The predicted spliced length of the blood cir- expressed—at least 30-fold—than the cognate
cRNAs (median = 343 nt) was similar to that in linear isoforms. In contrast with the liver and cer-
the liver or cerebellum (median = 394 nt and ebellum, blood circular RNA isoforms were
448 nt, respectively). However, the number of detectable even while the corresponding linear
circRNAs per gene was higher in whole blood. gene products showed low abundances.
Blood circRNAs partially overlapped circRNAs Therefore, authors proposed that circRNA levels
expressed in the cerebellum and liver at approxi- in human blood could be informative of the
mately 30% and 10%, respectively, but also con- ­coding gene activity that could not be evaluated
tain a considerable number of specific circRNAs. using classical RNA analysis. Different studies
122 A. Vea et al.

have subsequently detected a number of cir- megakaryocytes. Thus, circRNAs seem to be pro-
cRNAs in the peripheral blood [27–31]. duced in platelets rather than being inherited
The presence of exogenous circRNAs in the from their precursor cells. Authors proposed that
bloodstream should also be taken into account. the enrichment of circRNAs in platelets was
Broadbent et al. [32] reported the expression of associated with the degradation/decay of linear
1381 circRNAs during the blood stage develop- RNAs during the lifetime of the platelets. The
ment of Plasmodium falciparum. Interestingly, analysis of platelet circRNAs in circulation may
their experimentally validated circRNA candi- thus provide insights into megakaryocyte func-
dates contained predicted human miRNA bind- tion in the bone marrow.
ing sites, which indicate a potential parasite-host Maass et al. [34] generated a circRNA resource
communication mechanism in malaria. catalogue by sequencing ribosomal RNA-­
Whole blood is composed of a plethora of dif- depleted total RNA in 20 human tissues highly
ferent cells. In addition, the cell-free compart- relevant to disease-related research, including
ment can contain a number of circRNAs from platelets isolated from whole blood from a single
cells of diverse non-haematopoietic origin. subject. According to their results, the platelets
Despite their great potential as a source of bio- expressed a total of 3324 circRNAs with 2339
markers, a more detailed analysis of the circRNA unique circRNAs. Supporting previous evidence,
biology in the blood components is necessary. the number of circRNAs observed in platelets
was more abundant than in any other evaluated
tissue, including the cortex, atrium, fat, or mus-
2.2 Circular RNA in Platelets cle, among others. Furthermore, low overlap in
the circRNA pattern was observed with other
Although platelets are anucleated, they contain blood cells (neutrophils) or in the cell-free com-
RNAs in the form of non-coding transcripts. partment (serum and plasma), which again sug-
Previous investigations have demonstrated that gested the tissue-specific expression of circRNAs.
the platelet transcriptome is significantly enriched In general, circular-to-linear RNA ratios were
for circRNAs. Using 3 publicly available RNA-­ high in the tissues with abundant circRNA
Seq datasets, Alhasan et al. [33] identified 33,829 expression. For example, the platelet circRNA in
structures consistent with circRNAs. Authors the SMARCA5 gene showed a circular-to-linear
showed that circRNAs were 17- to 188-fold ratio of 151:1 in platelets. Furthermore, the
enriched in human platelets compared to nucle- authors reported that approximately 100 genes
ated tissues and identified 3162 genes signifi- hosted more than five different circRNA iso-
cantly enriched for circRNAs. Approximately forms. In some cases, such as PTPN12 or TTN
27% of circRNAs were platelet-specific when genes, they detected 18 circRNAs isoforms in the
they compared their findings with previous RNA-­ platelets, atrium and vena cava. Since platelets
Seq datasets. The mean number of circRNAs per are translationally competent [35], the authors
gene was higher in platelets (5 circRNAs per hypothesized that platelet circRNAs could serve
gene) compared to different nucleated tissues and as templates for translation. However, their
cell lines (1–2 circRNAs per gene). The expres- experimental results using mass spectrometry
sion levels of ten selected circRNAs were higher were inconclusive.
than their corresponding linear structures, with These results are also consistent with recent
circRNA isoforms from SMARCA5, UBXN7 findings. Characterization of circRNAs using an
and PNN ranging from 50- to 1000-fold more RNA-Seq approach in human platelets revealed
abundance. In contrast, all evaluated circRNA that, compared to other haematopoietic cell
isoforms were expressed at an equivalent or types, including monocytes, macrophages, T
lower level than their linear counterparts in nucle- cells and megakaryocytes, circRNAs are
ated cells. Experimental evidence also estab- ­abundant in platelets [36]. A large set of circular
lished that circRNAs are not enriched in cultured isoforms are predominantly expressed in plate-
10 Circular RNAs in Blood 123

lets (55–70%), and 95% of these abundant cir- bled exons in HeLa cells and normal primary
cRNAs were identical in resting platelets and human cells, including peripheral blood collected
platelets activated by thrombin receptor activatorfrom the same patients in remission, and H9 ES
peptide-­6 (TRAP-6). Supporting the cell speci- cells. In addition, the authors reported evidence
ficity of circRNAs, the most abundant circRNA for scrambled transcripts comprising at least 10%
of the transcripts from more than 800 genes in
in platelets, Plt-circR4, was exclusively expressed
in platelets when compared to ten different cell specific cell populations isolated from the bone
lines. marrow of a single individual: naive B cells
Different genes implicated in blood vessel (CD19+), haematopoietic stem cells (CD34+)
relaxation and platelet aggregation express plate-and neutrophils. The presence of circRNAs has
let circRNAs [34]. Unfortunately, the function of also been recently corroborated in neutrophils
this class of ncRNAs in platelets remains unclear.isolated from peripheral whole blood from a sin-
Since the deregulation of platelet activation is gle donor, with a total of 274 circRNAs, includ-
associated with a number of relevant diseases, ing 58 unique circRNAs [34]. Differences in the
including myocardial infarction and stroke, and relative abundance of circRNAs have been
ncRNAs seems to play a key role in platelet biol- observed between different leukocyte types. For
ogy, further investigations should evaluate example, the most abundant circRNAs in CD19+,
whether circRNAs mediate relevant biological CD34+ and neutrophils samples were KIAA0182,
effects in platelets. MAN1A2 and CCDC126, respectively [13].
CircRNAs represented more than half of all tran-
scripts produced by these genes. Interestingly,
2.3 Circular RNA in White Blood authors reported the expression of scrambled iso-
Cells forms from ncRNAs. These results are supported
by later findings from the same group that dem-
The presence of circRNAs in circulating leuko- onstrated the circular/linear RNA ratio and the
cyte populations has been described in leuko- pattern of circRNA isoforms from each gene, in
cytes isolated from the blood [37] and bone addition to the repertoire of genes expressing cir-
marrow [38]. In a seminal study, Salzmann et al. cRNAs, which were cell-type specific by analys-
[13] performed RNA-Seq on ribosomal RNA-­ ing 15 different cancer and non-cancer cell lines,
depleted total RNA from the diagnostic bone including the leukaemia cell line K562 [6], and
marrow of five children between the ages of 2 the results from Memczak et al. [2] who sug-
and 6 with hyperdiploid B-precursor acute lym- gested that the expression of circRNAs were in
phoblastic leukaemia. They identified a hundred part cell- and developmental stage-specific.
genes with a permutated exon order (scrambled Indeed, these authors reported specific circRNA
exons) that were predicted to be circRNAs. More patterns with 939 exclusively expressed in
than 700 isoforms with scrambled exons were CD19+ cells, 333 in CD34+ and 194 in neutro-
estimated to comprise more than 10% of all tran- phils [2]. For example, the hsa-circRNA 2149
script isoforms produced from a comparable was detected in CD19+ leukocytes but not in
number of genes. It should be noted that, due to CD34+ leukocytes or neutrophils [2].
their experimental design, an underestimation of Recent evidence suggested that circRNAs
the prevalence of circular RNA isoforms was may play a relevant role in leukocyte biology.
expected. Using RT-qPCR, they confirmed the Using publicly available RNA-Seq data from
results of the most abundant circRNAs: ESYT2, mouse macrophages, Ng et al. [39] identified an
FBXW4, CAMSAP1, KIAA0368, CLNS1A, LPS-inducible circRNA, mcircRasGEF1B, that
FAM120A, MAP3K1, ZKSCAN1, MANBA, regulates the expression and stability of ICAM-1
ZBTB46, NUP54, RARS and MGA. CircRNAs mRNA. Several TLR pathways regulate the
were not a specific feature of leukaemic cells, expression of mcircRasGEF1B, including TLR4,
since PCR results verified the presence of scram- TLR9, TLR3 and TLR2/TLR1, in RAW264.7
124 A. Vea et al.

cells but not in MEF cells. Interestingly, this cir- ferent studies have provided exhaustive evidence
cRNA has a human homologue with similar about the presence of cell-free circRNAs in the
properties. Authors proposed that circRNAs may circulation. Koh et al. [41] detected 19 circRNAs
participate in the fine-tuning immune responses in plasma samples from pregnant women using
and protection against microbial infection. an approach based on RNA-Seq and microarrays.
Different expression circRNA profiles were Maass et al. [34] also demonstrated the expres-
observed in CD28(+)CD8(+) T cells and CD28(-) sion of 57 circRNAs in plasma and 39 circRNAs
CD8(+) T cells isolated from healthy elderly or in serum using RNA-seq. Notably, these authors
adult control subjects [40]. In silico prediction reported 51 and 37 unique circRNAs in plasma
results suggested that the circRNA 100783 may and serum, respectively, compared to other clini-
play a role in phosphoprotein-associated func- cally relevant tissues. Using a circRNA microar-
tions during CD28-related CD8(+) T-cell ageing. ray, a recent study proposed the presence of a
circRNA 100783 may therefore constitute a bio- higher number of circRNAs, more than 10,000,
marker for the longitudinal tracking of T-cell age- in each of the 21 plasma samples obtained from
ing and global immunosenescence. Zhang et al. patients with cervical cancer [42]. The presence
[38] compared the circRNA expression profiles of circRNA in circulating extracellular vesicles
of bone marrow-derived macrophages under dis- has also been described (Fig. 10.2). Using RNA-­
tinct polarizing conditions. Authors showed that Seq analysis, Li et al. [43] identified 1215 cir-
189 circRNAs were differentially expressed cRNAs in human exosomes isolated from a pool
between M1 and M2 macrophages and proposed of serum obtained from three healthy donors.
that circRNAs may be implicated in macrophage Most circRNAs (90%) were derived from protein-­
differentiation and polarization. coding exons but also consisted of introns,
lncRNAs, unannotated regions and antisense
regions. Similar to previous findings, the median
2.4  ircular RNA in Red Blood
C length was 350 nt. Three candidates selected for
Cells further analyses with RT-qPCR, circ-N4BP2L2,
circ-GSE1 and circ-SMARCA5 were detected in
Despite red blood cells (RBCs) being the most serum-derived exosomes but not in exosome-­
abundant cell type in the blood, the knowledge depleted serum, which suggested that circRNAs
about the presence of circRNAs in this cell type may be transported by specific mechanisms in
is limited. Nonetheless, the biology of circRNAs circulation. Supporting the high stability of cir-
in this cell type seems to be similar to that cRNAs, the incubation of serum at room temper-
observed in other anucleated cells such as plate- ature for up to 24 h had minimal effects on
lets. Indeed, circRNAs are also highly enriched exosomal circRNA levels. Interestingly, cir-
in mature RBCs relative to nucleated cells [33]. cRNAs originated from human MHCC-LM3
Again, the expression levels of circRNAs are cancer cells in a xenograft mouse model, such as
higher than linear RNAs [33]. Since RBCs are human circRNA CDYL, could be detected in the
not able to synthesise proteins, the circRNA pro- mouse serum and correlated with tumour weight,
file may reflect the biological processes of the which provided a relevant clue about the release
erythropoietic progenitor cells from the bone of circRNA from tissues to the circulation and
marrow. the potential of circRNAs as biomarkers. Indirect
evidence has been provided by independent stud-
ies that suggests a change in plasma levels of
2.5  ircular RNA in the Blood
C ­circRNAs in postoperative gastric cancer patients
Cell–Free Compartment compared to preoperative patients [44, 45]. The
expression of approximately 2700 plasma cir-
Although the exact number of circRNAs that can cRNAs was also significantly changed after sur-
be detected in the plasma remains unknown, dif- gical removal of cervical tumours [42]. The
10 Circular RNAs in Blood 125

Fig. 10.2 Mechanisms of circular RNA release to the microvesicles. Future studies should evaluate whether cir-
extracellular space cular RNAs are transported by proteins and/or
Similar to other non-coding RNAs, circular RNAs are lipoproteins
released in the extracellular space into exosomes or

presence of circRNAs in the serum and plasma were incorporated into exosomes more than lin-
has been corroborated by a considerable number ear RNAs. The level of circRNAs in exosomes
of biomarker-based studies using different meth- was only moderately correlated with that of cel-
odologies: RNA-Seq, microarray, RT-qPCR or lular circRNAs. Importantly, circRNAs contained
RT-ddPCR [19, 44–57]. in exosomes retained biological activity. The
Results observed in the cell-free compartment exosomes containing the circRNA CDR1 abro-
suggest the circRNA secretion to the extracellu- gate the miR-7-induced in vitro inhibition of cell
lar space/circulation, as it was shown for other proliferation in receipt SMCC-7721 cancer cells.
ncRNAs such as miRNAs [58]. These hypotheses These results indicate the participation of
have been validated by different studies. ncRNAs in cell-to-cell communication [59].
CircRNAs were detected in cell-derived exo- Nonetheless, further investigations should cor-
somes released by MHCC-LM3 liver cancer cells roborate these findings. Authors suggested that
[43]. The expression level was enriched in exo- the sorting of circRNAs into exosomes may be
somes compared to cells (at least twofold). regulated, at least in part, by changes in associ-
Additionally, circular-to-linear RNA ratios in ated intracellular miRNA levels. The selective
exosomes were approximately sixfold higher packaging and release of circRNAs within extra-
than those in cells, suggesting that circRNAs cellular vesicles (microvesicles and exosomes)
126 A. Vea et al.

have also been reported in platelets [36]. Since biomarkers. First, the presence of thousands of
whole blood has been defined as the main con- circRNAs has been described in the cell-free
tributor (∼40%) towards the cell-free RNA tran- compartment. This constitutes an advantage over
scriptome [41], and platelets have been reported other ncRNAs. Only hundreds of miRNAs, the
as a major source of ncRNAs in the circulation main ncRNA class in biomarker-based studies
[35], these results are especially relevant. [64], can be efficiently detected in plasma/serum
Interestingly, a group of selected circRNAs were samples [65]. Second, circularity confers excel-
preferentially released in exosomes (FAM13B, lent biochemical properties as biomarkers.
DYRK1A, AMD1 and TMEM30) and microves- CircRNAs are free of exonuclease-mediated deg-
icles (AMD1 and DYRK1A) compared with their radation, are cell-specific, are more stable and
corresponding linear RNA. Nonetheless, other have a longer half-life than most linear RNAs due
circRNAs were preferentially retained in plate- to the absence of 5′ or 3′ ends. Third, circRNAs
lets (ASAP1). Based on the size distributions, could be detected in clinical specimens. Fourth,
released circRNAs were smaller than preferen- the circRNA circulation patterns could be modu-
tially retained circRNAs (mean of 283 nt vs. lated by different physiological states. Circulating
459 nt for microvesicles; 286 nt vs. 435 nt for circRNAs are specifically expressed during dif-
exosomes). These results suggest that circRNA ferent trimesters of pregnancy, which suggests a
size may be an additional determinant for selec- temporal dynamic regulation of these ncRNAs
tive vesicle export. Further investigations should [41]. Additionally, the deregulation of circulating
evaluate other factors that may affect sorting, circRNA levels in pathological conditions are
such as sequence motifs, as demonstrated for supported by a number of publications that have
miRNAs [60]. An alternative hypothesis to evaluated the expression of circRNAs in whole
explain circRNA secretion has been proposed by blood, blood cells and circulating cell-free com-
Lasda et al. [61]. Authors hypothesized that, due partments in a wide array of diseases, including
to the long half-live of circRNAs in cells, the coronary artery disease [19, 30], type 2 diabetes
release of circRNAs in extracellular vesicles may mellitus [31], diabetes retinopathy [54], rheuma-
constitute one possible mechanism to clear cel- toid arthritis [37], colorectal cancer [43], gastric
lular circRNAs. Indeed, the higher circular/linear cancer [44], breast cancer [53], acute myeloid
RNA ratio in extracellular vesicles compared to leukaemia [66], systemic lupus erythaematosus
the producer cell may provide evidence in this [67], intracranial aneurysm [28], pulmonary
sense. Overall, the results suggested that the tuberculosis [68] and primary biliary cholangitis
secretion of circRNAs in extracellular vesicles [56]. Indeed, the results from different studies
seems to be a common property of many cell point to the potential clinical application of cir-
types [61]. Supporting this hypothesis, circRNAs cRNAs as biomarkers. In a recent investigation
have been detected in exosomes released from with a large sample size (N = 769), cir-
three different colon cancer lines [62] and mic- cRNA_025016 was upregulated in patients with
roparticles secreted by vascular smooth muscle new-onset AF after isolated off-pump coronary
cells [46]. artery bypass grafting with high diagnostic accu-
racy (AUC = 0.802) [57]. Vasourt et al. [29]
reported that the blood levels of MICRA
3 Clinical Application ­(myocardial infarction-associated circular RNA)
were a strong predictor of left ventricle dysfunc-
The development of blood-based biomarkers is tion 3–4 months after myocardial infarction, even
of great interest for clinical practice due to the after adjusting by potential confounding factors,
relatively simple blood withdrawal procedure, in peripheral blood samples from two indepen-
compared to the more invasive tissue biopsy, and dent cohorts totalling 642 patients. MICRA
the fast and cost-effective analysis [63]. In this showed an incremental predictive value on top of
context, circRNAs may constitute a new entity of established clinical parameters and biomarkers.
10 Circular RNAs in Blood 127

The same group has recently validated these find- ncRNAs could modulate the phenotype and gene
ings using an alternative stratification criteria expression of recipient cells [71], circRNAs
[27]. In breast cancer, compared with commonly emerged as tools with great potential application
used biomarkers for diagnosis of carcinoembry- in therapeutics.
onic antigen (CEA, AUC = 0.562) and carbohy-
drate antigen 15-3 (CA15-3, AUC = 0.629),
peripheral blood circ_0001785 had higher diag- 4 Limitations and Perspectives
nostic accuracy (AUC = 0.784) [53]. Furthermore,
circRNAs seem to not only be biomarkers them- Given the emerging role for blood circRNAs as
selves but can also be combined with other biomarkers to aid in the management of patients,
ncRNAs. The serum ratio of circRNA-284 to the development of independent and multicentre
miR-221, a miRNA for which circRNA-284 has studies with large population sizes to explore the
a binding site, was increased in patients present- real clinical application of circRNAs in diagnosis
ing with an acute carotid-related ischaemic event and prognosis seems mandatory. The effect of
and showed great performance in terms of dis- potential sources of variation on circRNA levels,
crimination (AUC = 0.820) as diagnostic bio- including age, sex, comorbidities, genetic back-
markers for carotid-related cerebrovascular ground and disease stage, among others, deserves
ischaemia [46]. These results amplify the poten- particular attention. Concerning extracellular cir-
tial of circRNAs as biomarkers of disease. Given cRNAs, the identification of cellular sources and
the inverse putative functional relationship cellular targets, in addition to their function in
between miR-221 and circR-284, these results health and disease, is an interesting research field
may also provide valuable information about the that should be addressed. Indeed, it is not known
pathological mechanism linked to the disease. whether extracellular circRNAs are casually
Supporting their role as biomarkers, unique involved in the pathophysiology of the underly-
fusion-circRNAs (f-circRNAs) derived from the ing disease. Thus, it is necessary to evaluate
exons of genes affected by cancer-associated whether circRNAs are mediators in cell-to-cell
chromosomal translocation have been detected in communication or their secretion in extracellular
particular pathological conditions such as leukae- vesicles is merely a circRNA discard pathway.
mia [69]. The evaluation of this f-circRNA in Since circRNAs interact with RNA-binding pro-
blood cells and the cell-free compartment repre- teins [72], future work should also investigate the
sents an interesting diagnostic tool. transport of circRNAs complexed with proteins,
Despite the progress in circRNAs, whether or even lipoproteins, similar to that observed for
circRNAs can be potential therapeutic targets for miRNAs (Fig. 10.2) [73, 74]. It should be noted
blood conditions remains elusive [8]. As cir- that circRNAs are relatively well-conserved in a
cRNAs may be involved in a wide range of bio- broad range of species, which facilitates the
logical processes, deregulation of circRNA investigation of their biological function in dif-
expression may affect a number of pathological ferent cellular and animal models. Overall, more
mechanisms and therefore may play a causative functional studies are needed to elucidate the
role in a number of diseases [20, 70]. Previous functions of circRNAs and their possible role in
evidence suggests the potential of circRNAs in blood disease. In addition, although it has been
novel therapeutic approaches. F-circRNAs con- proposed that the half-lives of cellular circRNAs
tribute to tumour progression by increasing cell can be longer than 48 h [7], it is not clear which
proliferation and clonogenicity and protect leu- is the real half-life in extracellular fluids [75].
kaemia cells from the cytotoxic effects of cytara- The evidence presented here suggested a long
bine, a drug used for the treatment of leukaemia stability in the cell-free compartment.
[69]. Therefore, interventions aimed to block Implementation of ncRNAs is currently not
f-circRNAs could provide novel therapeutic feasible in clinical laboratories due to technologi-
strategies. Furthermore, since extracellular cal limitations and the high variability in the pre-­
128 A. Vea et al.

analytical phases. The analysis of the RNA Acknowledgements DdG-C was a recipient of Juan de la
Cierva-Incorporación grants from the Ministerio de
family in plasma/serum has strong methodologi- Economía y Competitividad (IJCI-2016-29393). CIBER
cal limitations, mainly due to the low concentra- Cardiovascular (CB16/11/00403 to DdG-C and VL-C) is
tion. The evaluation of ncRNAs in blood cells has a project of the Instituto de Salud Carlos III.
become recognized as an interesting alternative VLl-C and DdG-C are members of the CardiolincTM
network.
[76]. Nonetheless, due to the presence of cir-
cRNAs in blood cells, particularly RBCs, the Competing Financial Interests The authors declare no
possible cross-contamination in haemolytic sam- competing financial interests.
ples should be taken into account when analysing
the cell-free compartment. There are some con-
troversies related to the methodology used for
genome-wide profiling of circRNAs. RNA-Seq is References
a common method, but its detection efficiency is
limited [77]. This may explain the differences 1. Chen LL (2016) The biogenesis and emerging roles of
observed in the number of circRNAs detected circular RNAs. Nat Rev Mol Cell Biol 17(4):205–211
2. Memczak S, Jens M, Elefsinioti A et al (2013)
from the same type of samples, especially in the Circular RNAs are a large class of animal RNAs with
cell-free compartment. Recently, it has been pro- regulatory potency. Nature 495(7441):333–338
posed that circRNA microarrays could be an 3. Conn SJ, Pillman KA, Toubia J et al (2015) The RNA
interesting alternative [42]. However, contrary to binding protein quaking regulates formation of cir-
cRNAs. Cell 160(6):1125–1134
RNA-Seq, microarrays can only detect known 4. Zhang XO, Wang HB, Zhang Y et al (2014)
circRNAs. More investigations are needed to Complementary sequence-mediated exon circulariza-
identify the most sensitive and accurate technol- tion. Cell 159(1):134–147
ogy. The development of standard operating pro- 5. Khan MA, Reckman YJ, Aufiero S et al (2016)
RBM20 regulates circular RNA production from the
cedures and guidelines for best practices in Titin gene. Circ Res 119(9):996–1003
addition to automated and standardized assays is 6. Salzman J, Chen RE, Olsen MN et al (2013) Cell-type
needed. A detailed description of the methods specific features of circular RNA expression. PLoS
used for circRNA analysis is fundamental to Genet 9(9):e1003777
7. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular
ensure the reproducibility of the results. A special RNAs are abundant, conserved, and associated with
effort should be performed to clarify the circRNA ALU repeats. RNA 19(2):141–157
nomenclature. In the current chapter, we have 8. Bonizzato A, Gaffo E, Te Kronnie G et al (2016)
used the name given by each publication. CircRNAs in hematopoiesis and hematological malig-
nancies. Blood Cancer J 6(10):e483
Nonetheless, there is strong diversity among dif- 9. Tuck AC, Tollervey D (2011) RNA in pieces. Trends
ferent publications. Genet 27(10):422–432
10. Sanger HL, Klotz G, Riesner D et al (1976) Viroids are
single-stranded covalently closed circular RNA mol-
ecules existing as highly base-paired rod-like struc-
5 Conclusions tures. Proc Natl Acad Sci U S A 73(11):3852–3856
11. Cocquerelle C, Mascrez B, Hetuin D et al (1993)
Recent advances in transcriptomics have placed Mis-splicing yields circular RNA molecules. FASEB
the focus on a new species of ncRNAs called cir- J 7(1):155–160
12. Kos A, Dijkema R, Arnberg AC et al (1986) The
cRNAs. Due to their biological properties, cir- hepatitis delta (delta) virus possesses a circular
cRNAs have emerged as a new source of RNA. Nature 323(6088):558–560
blood-based biomarkers and therapeutic 13. Salzman J, Gawad C, Wang PL et al (2012) Circular
approaches. However, more research is needed to RNAs are the predominant transcript isoform from
hundreds of human genes in diverse cell types. PLoS
elucidate the biological function of circRNAs in One 7(2):e30733
the blood as well as their real potential as clinical 14. Li Z, Huang C, Bao C et al (2015) Exon-intron cir-
indicators. cular RNAs regulate transcription in the nucleus. Nat
Struct Mol Biol 22(3):256–264
10 Circular RNAs in Blood 129

15. Rybak-Wolf A, Stottmeister C, Glazar P et al (2015) 31. Zhao Z, Li X, Jian D et al (2017) Hsa_circ_0054633 in
Circular RNAs in the mammalian brain are highly peripheral blood can be used as a diagnostic bio-
abundant, conserved, and dynamically expressed. marker of pre-diabetes and type 2 diabetes mellitus.
Mol Cell 58(5):870–885 Acta Diabetol 54(3):237–245
16. Cao S, Wei D, Li X et al (2017) Novel circular RNA 32. Broadbent KM, Park D, Wolf AR et al (2011) A global
expression profiles reflect progression of patients transcriptional analysis of Plasmodium falciparum
with hypopharyngeal squamous cell carcinoma. malaria reveals a novel family of telomere-associated
Oncotarget 8(28):45367–45379 lncRNAs. Genome Biol 12(6):R56
17. Xu T, Wu J, Han P et al (2017) Circular RNA expres- 33. Alhasan AA, Izuogu OG, Al-Balool HH et al (2016)
sion profiles and features in human tissues: a study Circular RNA enrichment in platelets is a signature of
using RNA-seq data. BMC Genomics 18(Suppl transcriptome degradation. Blood 127(9):e1–e11
6):680 34. Maass PG, Glazar P, Memczak S et al (2017) A map
18. Hansen TB, Jensen TI, Clausen BH et al (2013) of human circular RNAs in clinically relevant tissues.
Natural RNA circles function as efficient microRNA J Mol Med (Berl) 95(11):1179–1189
sponges. Nature 495(7441):384–388 35. Sunderland N, Skroblin P, Barwari T et al (2017)
19. Pan RY, Liu P, Zhou HT et al (2017) Circular RNAs MicroRNA biomarkers and platelet reactivity: the clot
promote TRPM3 expression by inhibiting hsa-­ thickens. Circ Res 120(2):418–435
miR-­ 130a-3p in coronary artery disease patients. 36. Preusser C, Hung LH, Schneider T et al (2018)
Oncotarget 8(36):60280–60290 Selective release of circRNAs in platelet-derived extra-
20. Wang K, Sun Y, Tao W et al (2017) Androgen recep- cellular vesicles. J Extracell Vesicles 7(1):1424473
tor (AR) promotes clear cell renal cell carcinoma 37. Ouyang Q, Wu J, Jiang Z et al (2017) Microarray
(ccRCC) migration and invasion via altering the expression profile of circular RNAs in peripheral
circHIAT1/miR-195-5p/29a-3p/29c-3p/CDC42 sig- blood mononuclear cells from rheumatoid arthritis
nals. Cancer Lett 394:1–12 patients. Cell Physiol Biochem 42(2):651–659
21. Conn VM, Hugouvieux V, Nayak A et al (2017) A 38. Zhang Y, Zhang Y, Li X et al (2017) Microarray analy-
circRNA from SEPALLATA3 regulates splicing of its sis of circular RNA expression patterns in polarized
cognate mRNA through R-loop formation. Nat Plants macrophages. Int J Mol Med 39(2):373–379
3:17053 39. Ng WL, Marinov GK, Liau ES et al (2016) Inducible
22. Du WW, Yang W, Liu E et al (2016) Foxo3 circular RasGEF1B circular RNA is a positive regulator
RNA retards cell cycle progression via forming ter- of ICAM-1 in the TLR4/LPS pathway. RNA Biol
nary complexes with p21 and CDK2. Nucleic Acids 13(9):861–871
Res 44(6):2846–2858 40. Wang YH, Yu XH, Luo SS et al (2015)
23. Yang Y, Fan X, Mao M et al (2017) Extensive Comprehensive circular RNA profiling reveals that
translation of circular RNAs driven by N(6)- circular RNA100783 is involved in chronic CD28-
methyladenosine. Cell Res 27(5):626–641 associated CD8(+)T cell ageing. Immun Ageing
24. Devaux Y, Creemers EE, Boon RA et al (2017) 12:17
Circular RNAs in heart failure. Eur J Heart Fail 41. Koh W, Pan W, Gawad C et al (2014) Noninvasive
19(6):701–709 in vivo monitoring of tissue-specific global gene
25. Xu Z, Yan Y, Zeng S et al (2018) Circular RNAs: clini- expression in humans. Proc Natl Acad Sci U S A
cal relevance in cancer. Oncotarget 9(1):1444–1460 111(20):7361–7366
26. Memczak S, Papavasileiou P, Peters O et al (2015) 42. Li S, Teng S, Xu J et al (2018) Microarray is an effi-
Identification and characterization of circular RNAs cient tool for circRNA profiling. Brief Bioinform.
as a new class of putative biomarkers in human blood. https://doi.org/10.1093/bib/bby006
PLoS One 10(10):e0141214 43. Li Y, Zheng Q, Bao C et al (2015) Circular RNA is
27. Salgado-Somoza A, Zhang L, Vausort M et al (2017) enriched and stable in exosomes: a promising bio-
The circular RNA MICRA for risk stratification marker for cancer diagnosis. Cell Res 25(8):981–984
after myocardial infarction. Int J Cardiol Heart Vasc 44. Li P, Chen S, Chen H et al (2015) Using circular RNA
17:33–36 as a novel type of biomarker in the screening of gas-
28. Teng L, Chen Y, Chen H et al (2017) Circular RNA tric cancer. Clin Chim Acta 444:132–136
hsa_circ_0021001 in peripheral blood: a potential 45. Li T, Shao Y, Fu L et al (2018) Plasma circular RNA
novel biomarker in the screening of intracranial aneu- profiling of patients with gastric cancer and their
rysm. Oncotarget 8(63):107125–107133 droplet digital RT-PCR detection. J Mol Med (Berl)
29. Vausort M, Salgado-Somoza A, Zhang L et al (2016) 96(1):85–96
Myocardial infarction-associated circular RNA pre- 46. Bazan HA, Hatfield SA, Brug A et al (2017) Carotid
dicting left ventricular dysfunction. J Am Coll Cardiol plaque rupture is accompanied by an increase in the
68(11):1247–1248 ratio of serum circR-284 to miR-221 levels. Circ
30. Zhao Z, Li X, Gao C et al (2017) Peripheral blood cir- Cardiovasc Genet 10(4):e001720
cular RNA hsa_circ_0124644 can be used as a diag- 47. Chen S, Li T, Zhao Q et al (2017) Using circular RNA
nostic biomarker of coronary artery disease. Sci Rep hsa_circ_0000190 as a new biomarker in the diagno-
7:39918 sis of gastric cancer. Clin Chim Acta 466:167–171
130 A. Vea et al.

48. Huang M, He YR, Liang LC et al (2017) Circular cancer cells and can be transferred to exosomes. Sci
RNA hsa_circ_0000745 may serve as a diagnos- Rep 6:37982
tic marker for gastric cancer. World J Gastroenterol 63. de Gonzalo-Calvo D, Quezada M, Campuzano O et al
23(34):6330–6338 (2017) Familial dilated cardiomyopathy: a multidisci-
49. Ji W, Qiu C, Wang M et al (2018) Hsa_circ_0001649: plinary entity, from basic screening to novel circulat-
a circular RNA and potential novel biomarker for ing biomarkers. Int J Cardiol 228:870–880
colorectal cancer. Biochem Biophys Res Commun 64. de Gonzalo-Calvo D, Iglesias-Gutierrez E, Llorente-­
497(1):122–126 Cortes V (2017) Epigenetic biomarkers and cardio-
50. Li WH, Song YC, Zhang H et al (2017) Decreased vascular disease: circulating MicroRNAs. Rev Esp
expression of Hsa_circ_00001649 in gastric can- Cardiol (Engl Ed) 70(9):763–769
cer and its clinical significance. Dis Markers 65. Umu SU, Langseth H, Bucher-Johannessen C et al
2017:4587698 (2018) A comprehensive profile of circulating RNAs
51. Lu R, Shao Y, Ye G et al (2017) Low expression of hsa_ in human serum. RNA Biol 15(2):242–250
circ_0006633 in human gastric cancer and its clinical 66. Li W, Zhong C, Jiao J et al (2017) Characterization
significances. Tumour Biol 39(6):1010428317704175 of hsa_circ_0004277 as a new biomarker for acute
52. Sun H, Tang W, Rong D et al (2018) Hsa_ myeloid leukemia via circular RNA profile and bioin-
circ_0000520, a potential new circular RNA bio- formatics analysis. Int J Mol Sci 18(3):597
marker, is involved in gastric carcinoma. Cancer 67. Li H, Li K, Lai W et al (2018) Comprehensive circu-
Biomark 21(2):299–306 lar RNA profiles in plasma reveals that circular RNAs
53. Yin WB, Yan MG, Fang X et al (2017) Circulating can be used as novel biomarkers for systemic lupus
circular RNA hsa_circ_0001785 acts as a diagnostic erythematosus. Clin Chim Acta 480:17–25
biomarker for breast cancer detection. Clin Chim Acta 68. Zhuang ZG, Zhang JA, Luo HL et al (2017) The cir-
S0009–8981(17):30407–30402 cular RNA of peripheral blood mononuclear cells:
54. Zhang SJ, Chen X, Li CP et al (2017) Identification Hsa_circ_0005836 as a new diagnostic biomarker and
and characterization of circular RNAs as a new class therapeutic target of active pulmonary tuberculosis.
of putative biomarkers in diabetes retinopathy. Invest Mol Immunol 90:264–272
Ophthalmol Vis Sci 58(14):6500–6509 69. Guarnerio J, Bezzi M, Jeong JC et al (2016)
55. Zhao Q, Chen S, Li T et al (2018) Clinical values of Oncogenic role of fusion-circRNAs derived from
circular RNA 0000181 in the screening of gastric can- cancer-associated chromosomal translocations. Cell
cer. J Clin Lab Anal 32(4): e22333 165(2):289–302
56. Zheng J, Li Z, Wang T et al (2017) Microarray expres- 70. Gupta SK, Garg A, Bar C et al (2018) Quaking inhib-
sion profile of circular RNAs in plasma from primary its doxorubicin-mediated cardiotoxicity through regu-
biliary cholangitis patients. Cell Physiol Biochem lation of cardiac circular RNA expression. Circ Res
44(4):1271–1281 122(2):246–254
57. Zhang J, Xu Y, Xu S et al (2018) Plasma circular 71. Bang C, Batkai S, Dangwal S et al (2014) Cardiac
RNAs, Hsa_circRNA_025016, predict postoperative fibroblast-derived microRNA passenger strand-­
atrial fibrillation after isolated off-pump coronary enriched exosomes mediate cardiomyocyte hypertro-
artery bypass grafting. J Am Heart Assoc 7:e006642 phy. J Clin Invest 124(5):2136–2146
58. de Gonzalo-Calvo D, Cenarro A, Garlaschelli K 72. Schneider T, Hung LH, Schreiner S et al (2016)
et al (2017) Translating the microRNA signature of CircRNA-protein complexes: IMP3 protein compo-
microvesicles derived from human coronary artery nent defines subfamily of circRNPs. Sci Rep 6:31313
smooth muscle cells in patients with familial hyper- 73. Arroyo JD, Chevillet JR, Kroh EM et al (2011)
cholesterolemia and coronary artery disease. J Mol Argonaute2 complexes carry a population of circu-
Cell Cardiol 106:55–67 lating microRNAs independent of vesicles in human
59. Hergenreider E, Heydt S, Treguer K et al (2012) plasma. Proc Natl Acad Sci U S A 108(12):5003–5008
Atheroprotective communication between endothelial 74. Vickers KC, Palmisano BT, Shoucri BM et al (2011)
cells and smooth muscle cells through miRNAs. Nat MicroRNAs are transported in plasma and delivered
Cell Biol 14(3):249–256 to recipient cells by high-density lipoproteins. Nat
60. Villarroya-Beltri C, Gutierrez-Vazquez C, Sanchez-­ Cell Biol 13(4):423–433
Cabo F et al (2013) Sumoylated hnRNPA2B1 controls 75. Jeck WR, Sharpless NE (2014) Detecting and charac-
the sorting of miRNAs into exosomes through binding terizing circular RNAs. Nat Biotechnol 32(5):453–461
to specific motifs. Nat Commun 4:2980 76. Devaux Y (2017) Transcriptome of blood cells as a
61. Lasda E, Parker R (2016) Circular RNAs co-­precipitate reservoir of cardiovascular biomarkers. Biochim
with extracellular vesicles: a possible mechanism for Biophys Acta 1864(1):209–216
circRNA clearance. PLoS One 11(2):e0148407 77. Szabo L, Salzman J (2016) Detecting circular RNAs:
62. Dou Y, Cha DJ, Franklin JL et al (2016) Circular bioinformatic and experimental challenges. Nat Rev
RNAs are down-regulated in KRAS mutant colon Genet 17(11):679–692
 Circular RNA in Saliva
11
Farinaz Jafari Ghods

Abstract 1 Introduction
Although the type and amount of salivary
components are influenced by many factors, In the middle of the eighteenth century, Langley
due to easy, quick, cheap, and noninvasive argued that according to Nuck’s belief, the effect
sampling method alongside with the existence of the brain through the nerves on salivary glands
of the vast majority of the substances found in causes the flow of saliva [1]. Saliva has wide
peripheral blood and urine in it, in recent years range of functions such as lubrication, speech
saliva has been considered as an ideal biofluid facilitation, preliminary food digestion, control-
for disease research. Salivary circular RNA ling of dental/oral infections by balancing demin-
(circRNA), as an endogenous RNA molecule eralization/remineralization, oral tissue repair,
with a great variety of regulatory potency, is and antimicrobial peptides [2–5]. Saliva has been
becoming a novel focus for detecting wide considered as a research material in recent years
range of local or systemic diseases. due to its potential to detect bacterial, viral, and
Expectantly, with characterization of many systemic diseases.
more circRNAs in saliva, their motifs, and tar- While 99% of the total volume of saliva is
get sites, they can be used routinely in person- water, the remaining 1% consists of organic and
alized medicine. inorganic compounds. The major salivary glands
secrete 93% of saliva, and 7% is salivated by the
Keywords minor salivary glands [6, 7]. Being an acidic
Circular RNA · Saliva · Noninvasive sam- (pH = 6–7) multi-constituent body fluid, miner-
pling · Biofluid als, electrolytes, buffers, enzymes, enzyme inhib-
itors, growth factors, cytokines, IgM, IgG, sIgA,
and a group of glycoproteins all make key com-
ponents of the human saliva [8–10]. Most of
these components are added to saliva after filter-
ing, processing and secreting from the vascula-
ture that nourish salivary glands [11–13]. Saliva
is initially sterile, but as soon as it releases into
the oral cavity, it would be exposed to oral micro-
F. Jafari Ghods (*) organisms (bacteria, viruses, and fungi), micro-
Department of Molecular Biology and Genetics, bial products, leukocytes, erythrocytes,
Faculty of Science, Istanbul University, desquamated oral epithelial cells and cellular
Istanbul, Turkey

© Springer Nature Singapore Pte Ltd. 2018 131

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_11
132 F. Jafari Ghods

products, nucleic acids, food debris, upper respi- 2  xtracellular RNAs in Human
E
ratory tract secretions, oral mucous, and gingival Saliva
crevicular fluid forming whole saliva (WS) [6,
14–17]. The type and amount of salivary compo- 2.1 Origin of Salivary RNAs
nents, regardless of whether they are transcripts,
proteins, metabolites, or oral microbes, are influ- RNA molecules found in saliva may originate
enced by factors such as age, gender, salivary from a variety of sources, for example, from the
gland development, microbial colonization pat- cell lysis that occurs in the salivary ducts, gingi-
tern, nutritional status, and tooth development val pockets, or desquamated epithelial cells of the
status [18–21]. oral cavity. Gingival crevice fluid (GCF) harbors
Cell-free and exosomal DNA fragments and various cell types such as blood leukocytes and
different kinds of RNAs (coding and noncoding) erythrocytes and their cell contents [35]. Products
also exist in saliva [22–24]. Applying high-­ derived from the lysis of these cells are consid-
throughput RNA-Seq, the first global character- ered as important sources of salivary RNAs.
ization of human saliva transcriptome was done in RNAs that are actively secreted can also be con-
2012, and it was shown that saliva encompasses sidered as another source of RNAs in saliva.
more than 4000 RNAs belonging to variety of RNAs that have been produced in secretary cells
RNA species [25]. In saliva an intricate composi- or anywhere else in the body enter the circulatory
tion of extracellular RNA has been transpired, system and are secreted through the filtering and
including mostly mRNAs, long ncRNAs (≥200 processing into the saliva [12, 36, 37]. Since the
nucleotide-long), and small ncRNAs (<200-nucle- mouth of any healthy person contains approxi-
otide-long) such as miRNAs (19–23-nucleotide- mately 500 million bacterial cells belonging to
long), piRNAs (24–30-­nucleotide-long), 700 different colony species, it can be concluded
snoRNAs (60–300-nucleotide-long), circRNAs that the genome and RNA contents of the oral
(less than 100- to over 4000-nucleotide-long), etc. microorganisms, including bacteria, viruses, and
[26–28]. They may be originated from the apopto- fungi, can be a major source for variety of sali-
sis or necrosis. Interestingly, their degradation vary DNA and RNA molecules [38–40]. When
occurs much more slowly than exogenous the purpose of the study is to evaluate and mea-
species. sure salivary RNAs for the diagnosis of oral or
Special characteristics of saliva such as easy, systemic diseases, due to the presence of a high
quick, cheap, and noninvasive sampling, straight- fraction of microbial RNAs, a significant differ-
forward storage and transportation, lack of clot- ence in the amount of RNA composition will be
ting, high security for both the patient and the detected in whole saliva (WS) compared with the
health personnel, convenient analysis, and, most cell-free saliva (CFS) [25]. This in turn reduces
importantly, existence of the vast majority of the the sensitivity of the study performed in whole
substances found in peripheral blood and urine in saliva (WS). To overcome this problem, adding
it have made saliva as an ideal biofluid for per- subsequent steps such as low-speed centrifuga-
forming investigations. In the same vein, in vari- tion would be helpful to subtract microbial RNAs
ous studies conducted and ongoing, changes in and cell debris.
the expression levels of certain RNA molecules,
coding and noncoding (microRNAs, snoRNAs,
piRNAs, circular RNAs), have been associated 2.2 Salivary RNAs Stability
with the susceptibility or the development of cer-
tain diseases [29–34]. One of the major impe- In saliva, the most RNA molecules are degraded,
tuses in ongoing researches is to use salivary and only a percent of RNAs remain intact. In
ncRNAs as diagnostic biomarkers either for local each person, the percentage of RNAs that have
or systemic diseases. been broken is associated with the types of
11 Circular RNA in Saliva 133

­ icroorganisms present in their oral cavity and to

m RNA-induced silencing complex (RISC) binds to
the amount and type of the endonucleases and and stabilize the miRNAs in exosomes [46, 52].
exonucleases contents [41]. A more in-depth
study of fragmentated RNAs suggests that endo-
nuclease enzymes overcome exonucleases in 3 Importance of Emerging
fragmentation process. Because the exonucleases Technologies in Salivary
are much more progressive, in most cases, the Diagnostics
RNA molecule is completely disintegrated [42,
43]. On the other hand, it has been seen that deg- The rapid and growing development of knowl-
radation of salivary RNAs occurs much more edge and information on the “omics” constituents
slowly than exogenous species, and this suggests of saliva increased the hopes for the development
mechanisms to protect salivary RNA [44]. One of of biomarkers and personalized medicine, and,
these mechanisms is the association of these sali- for the first time in 2008, the term “salivaomics”
vary RNAs with macromolecules. For example, was introduced, which represents the study of the
salivary mucus contains oligomeric structures of five main salivary diagnostic components as
MUC5B and MUC7proteins [45]. The associa- genome/epigenome, transcriptome, proteome,
tion of salivary RNAs with glycosylated oligo- metabolome, and microbiome [53].
mer is a protective mechanism for the preservation With the development of the aforementioned
of salivary RNAs [44]. In a study by Turchinovich branches and the application of new technologies
et al., it was shown that large part of the miRNAs such as automated extraction, purification, the
in the extracellular environment was associated whole-genome sequencing, the profiling of DNA
with Ago2 proteins [46]. The stability of the sali- and various types of RNAs using microarray
vary RNAs has also been attributed to placement analysis, quantitative PCR (qPCR), 2-D gel elec-
of them into the extracellular vesicles (EVs), also trophoresis, mass spectrometry, blotting meth-
known as exosomes, which are small membrane ods, data and bioinformatics (ranging from rapid
vesicles (30–100 nm in diameter) and have poten- short read aligners to detailed examination of
tial to carry diverse biomolecules. In 2007, Valadi RNA expression patterns), etc., personalized
et al. showed that mRNA and miRNA molecules medicine was raised more strongly [25, 54–61].
can be entrapped into the EVs, transferred As in this new approach, instead of considering
between cells, and be functional in that new envi- the term “one size fits all,” the health status of
ronment [47]. This finding was later confirmed individuals is determined based on inherited dif-
by other investigations [48, 49]. In another study ferences, environmental conditions, and lifestyle.
comparing EVs fraction to EV-depleted salivary Discovery of novel biomarkers is of increasing
supernatant in order to investigate whether body importance for personalized medicine in which
fluid miRNAs are circulating freely or via exo- the ultimate goals are to match the right molecu-
somes, predominant existence of miRNAs in lar marker to the underlying processes involved
EVs was proven [50]. Since miRNAs, piRNAs, in the disease pathology and to design experi-
and snoRNAs had been detected only in WS, it mental assays that provide valuable information
was conceived that small RNAs are not associ- about diagnosis, prognosis, and response to the
ated with exosomes. Using next-generation therapy on drug discovery.
sequencing (NGS) of small RNAs in salivary Saliva was called “the mirror of the body”
exosomes expression of known and novel miR- because it contains most of the compounds in the
NAs alongside with piRNAs, snoRNAs, and blood and urine and has potential to reflect the
other small RNAs in exosomes was defined for current physiological state of an individual [62,
the first time [51]. Formerly conducted studies 63]. However, the presence of some of these sub-
found that GW182 protein, a component of the stances at generally lower concentrations in
134 F. Jafari Ghods

saliva makes their detection problematic [64]. through varying mechanisms such as miRNA
But by advances in highly sensitive technologies, sponges, RNA-binding proteins sponges, or scaf-
detection of minute quantities of these substances folding molecules, thus expanding the complex-
in saliva has become possible [3]. ity of downstream gene expression [65]. In a
study by Hansen et al. (2013), two circRNAs
were identified acting as miRNA sponges and
4 CircRNAs Serve play role on miRNAs targeting [74]. Besides,
as a Fingerprint in Various although these molecules are classified in the
Human Diseases noncoding RNAs group, a study in 2015 showed
that some circRNAs are likely code for proteins
After discovering of salivary exRNAs in approxi- [75]. All of these findings along with their cell-­
mately 10 years ago, myriad of studies have been type-­specific and tissue-specific expression led to
pursued to provide insights on potential use of the logical assumption that presence of altered
these molecules to detect wide range of diseases expression patterns of circRNAs in different
such as oral cancer, ovarian cancer, breast cancer, body fluids and tissues might be the reason or the
Sjögren syndrome, etc. [29, 65–68]. Recently the consequence of various human disease. At pres-
circular RNA (circRNA), as an endogenous RNA ent, several studies are under initial discovery of
molecule with a great variety of regulatory circRNAs to clarify their role in various human
potency, is becoming a novel focus in this field diseases and to introduce them as diagnostic, pre-
[69]. dictive, or therapeutic targets biomarkers
Although circular RNAs were first detected (Table 11.1).
about 30 years ago, due to their closed loop struc- Since sampling of saliva is painless and stress-­
ture, their direct mapping on the genome was not free and negates the need for trained medical
possible through traditional methods of analyz- staff, application of saliva as diagnostic medium
ing RNAs. This, in turn, has led to a delay in the is most practical for the population of pediatric
discovery of such relatively new RNA species patients. However, as the diversities in biomarker
and limited information about them [70]. In fact, levels in individuals are affected by factors such
these species of RNA molecules originally deci- as age and diet, as well as the absence of deter-
phered in ribosomal-depleted RNA-Seq data, and mining thresholds that differentiate health and
circRNAs tracing was possible in excised exons disease status, there will be limitations in appli-
or introns using deep RNA sequencing and after cability and translatability of assays in pediatric
rRNA depletion, size selection, unique mapping, population [21, 89]. So, investigators should be
and classification of them according to their rela- aware of these biological changes when defining
tive abundance using BPKM (bases per kilobase normative values and designing salivary assays.
of gene model per million mapped bases) [71]. Assays’ interpretation also should reflect these
After the development of these technologies, an findings appropriately.
increasing number of circRNAs from different
tissues, such as serum and plasma, semen, saliva,
exosomes, cancer tumors, etc., were detected 5  alivary CircRNAs, Pros,
S
[72]. By determining the ratio of these circRNAs Cons, and Potential
to the linear counterparts, it was found that circu- Solutions
lar types are expressed at higher levels. On the
other hand, studies have shown that the half-life In 2015, Bahn et al., for the first time, conducted
of circRNAs varies considerably and is, on aver- and validated the existence of circRNAs using
age, about 2.5-fold longer than the median half-­ high-throughput RNA sequencing and an in-­
life of their linear host transcripts and even can be depth bioinformatics analysis following con-
up to 50 h [73]. CircRNAs, titer miRNAs, inter- struction of circRNA libraries in cell-free fraction
fere with splicing and regulate transcription of saliva [90]. Since desquamated epithelial cells
11 Circular RNA in Saliva 135

Table 11.1 A selection of recent studies in which circRNAs were reported to be related with human diseases
Disease circRNA(s) Tissue Type Potential value References
Colorectal Cancer hsa_circ_001569 Human CRC cell Biomarker [76]
lines, CRC tissue
samples
circ-BANP CRC tissue samples Prognostic and [77]
therapeutic
biomarker
Pre-eclampsia circ_101222 Blood corpuscles Prognostic [78]
biomarker
Ischemic heart disease hsa_circ_0124644 Peripheral blood Diagnostic [79]
samples biomarker
Alzheimer’s disease CircPVT1 WI-38 human Biomarker [80]
fibroblasts
Pancreatic ductal 209 circRNAs up- or PDAC tissue circRNAs expression [81]
adenocarcinoma downregulated samples profiling
Gastric cancer hsa_circ_002059 Gastric cancer Diagnostic [82]
patients tissue, biomarker
plasma samples
Cardiomyopathy circFOX-3 Heart samples of Therapeutic, target [83]
aged patients and
mice
MI with or without left MIRC Peripheral blood Biomarker [84]
ventricular dysfunction samples
Hypopharyngeal 2392 circRNAs up- or HSCC tumor tissues circRNAs expression [85]
squamous cell downregulated profiling
carcinoma
Rheumatoid arthritis circRNA_104871, Peripheral blood circRNAs expression [86]
circRNA_003524, mononuclear cells profiling, diagnostic
circRNA_101873, biomarkers
circRNA_103047
Systemic lupus 207 circRNAs up- or Plasma samples circRNAs expression [87]
erythematosus downregulated profiling
ChronicCD28- circRNA100783 CD28(−)CD8(+)T Biomarker [88]
associated CD8(+) cells and CD28(+)
T-cell aging CD8(+)T cell

and the rest of the cell types mentioned earlier from which 327 circRNAs were noncanonical
may release their RNA content into the extracel- ones. To validate the existence of circRNAs,
lular space interrupting exRNAs profiling, in RT-PCR and TOPO-cloned PCR [93] were done
order to enriching for physiological extracellular on a selective number of them. Low degree of
RNAs from environmentally originated RNAs, overlap of these 422 circRNAs was observed
cell-free saliva was used. To gain better results among the different individuals, which suggested
circRNAs enrichment procedure was done on their high individual specificity. Finally, with
extracted total RNA samples and total RNA purpose of exploring the functional relation of
treated with RNase H for 20 min at 37 ° the circRNAs, gene ontology analysis of genes
C. Thereinafter RNA extraction was done by acid overlapping putative circRNAs was conducted.
phenol/chloroform (pH = 4.5) [91, 92]. In this According to the results, it was suggested that
study, according to the presence of rich oral these salivary circRNAs probably contribute to
microorganisms in the saliva, despite the assess- intercellular signaling pathways and inflamma-
ing of cell-free saliva, 58.5% of the reads tory responses [90].
belonged to the microbial RNA sequences. Application of saliva-based circular RNAs in
However, 422 putative circRNAs were identified personalized medicine has myriad benefits
136 F. Jafari Ghods

including (1) they can be traced in saliva due to target sites, they can be used routinely in person-
their stability, longer half-live, and relatively alized medicine.
abundance in saliva and exosomes; (2) the evolu-
tion and utilization of saliva-based circRNA bio- Acknowledgments The authors thank Prof. Dr. Ahad
markers offer easy, inexpensive, reliable, Jafari Ghods, MD., FACP, (Iran University of Medical
Sciences) for critical reading and comments on the
noninvasive, and safer approach for patients, manuscript.
physicians, and medical staff, making them much
more favorable; and (3) they exhibit tremendous Competing Financial Interests The author declares no
potential to discriminate between patients and competing financial interests.
healthy individuals according to different levels
of their expression. As a result, they can be har-
nessed as sensitive and specific biomarkers for
identification of high-risk populations, cancer References
susceptibility evaluations, prognosis or diagnosis
of diseases, developing targeted therapies, assess- 1. Garrett JR (1975) Changing attitudes on salivary
secretion—a short history on spit. Proc R Soc Med
ing responses to treatment, detecting of treatment 68:553–560
resistance, and preliminary tracing of disease 2. Schipper RG, Silletti E, Vingerhoeds MH (2007)
recurrence [94, 95]. Along with the benefits men- Saliva as research material: biochemical, physi-
tioned above, this rapidly expanding field of cochemical and practical aspects. Arch Oral Biol
52:1114–1135
research has challenges and ambiguities that 3. Farnaud SJ, Kosti O, Getting SJ et al (2010) Saliva:
need to be addressed including: physiology and diagnostic potential in health and dis-
ease. Sci World J 10(16):434–456
1. Still little is known about the genetic networks 4. Malathi N, Mythili S, Vasanthi HR (2014) Salivary
diagnostics: a brief review. ISRN Dent 2014:158786
and pathways regulated by salivary circRNAs, 5. Khurshid Z, Naseem M, Sheikh Z et al (2016) Oral
so that deeper understanding of them may antimicrobial peptides: types and role in the oral cav-
allow investigators to knock out circRNAs by ity. Saudi Pharm J 24(5):515–524
genome-editing tools and diminish their 6. Hofman LF (2001) Human saliva as a diagnostic
specimen. J Nutr 131:1621S–1625S
destructive activities or by applying functional 7. Llena-Puy C (2006) The rôle of saliva in maintaining
knockout rescue experiments and design and oral health and as an aid to diagnosis. Med Oral Patol
produce ectopic circRNA expression plasmids Oral Cir Bucal 11(5):E449–E455
in order to induce their overexpression [96]. 8. Lawrence HP (2002) Salivary markers of systemic
disease: noninvasive diagnosis of disease and
2. It has shown that circular species of RNA monitoring of general health. J Can Dent Assoc
molecules form as by-products through mis-­ 68(3):170–174
splicing events; circularization of RNAs and 9. Almstahl A, Wikstrom M (2003) Electrolytes in stim-
artifacts results from use of reverse transcrip- ulated whole saliva in individuals with hyposalivation
of different origins. Arch Oral Biol 48:337–344
tase enzyme in sequencing technique [97]. 10. Aps JK, Martens LC (2005) Review: the physiology
of saliva and transfer of drugs into saliva. Forensic Sci
In a recent study five common algorithms com- Int 150:119–131
pared and it was indicated that up to 40% of pre- 11. Haeckel R, Hanecke P (1996) Application of saliva for
drug monitoring: an in vivo model for transmembrane
dicted circRNAs were only flagged by one transport. Eur J Clin Chem Clin Biochem 34:171–191
algorithm while less than 20% of the circRNAs 12. Forde M, Koka S, Ecker S et al (2006) Systemic
included into this study were identified by all five assessments utilizing saliva, part I: general consid-
[98]. Thus, the necessity of establishment of sta- erations and current assessments. Int J Prosthodont
19:43–52
tistical methods in order to overcome these tech- 13. Lee YH, Wong DT (2009) Saliva: an emerging bio-
nological barriers to estimate false detection rates fluid for early detection of diseases. Am J Dent
and discriminate true circ-RNAs and by-products 22(4):241–248
is evident. Hopefully, by solving aforementioned 14. Humphrey SP, Williamson RT (2001) A review
of saliva: normal composition, flow, and function.
problems alongside with characterization of J Prosthet Dent 85:162–169
many more circRNAs in saliva, their motifs and
11 Circular RNA in Saliva 137

15. Aps JK, Van Den Maagdenberg K, Delanghe JR et al target genes in cases with and without renal involve-
(2002) Flow cytometry as a new method to quantify ment. Kidney Blood Press Res 42(6):1322–1337
the cellular content of human saliva and its relation to 35. Delima AJ, Van Dyke TE (2003) Origin and function
gingivitis. Clin Chim Acta 321:35–41 of the cellular components in gingival crevice fluid.
16. Falcão DP, da Mota LM, Pires AL et al (2013) Periodontol 2000 31:55–76
Sialometry: aspects of clinical interest. Rev Bras 36. Kaufman E, Lamster IB (2002) The diagnostic appli-
Reumatol 53(6):525–531 cations of saliva—a review. Crit Rev Oral Biol Med
17. Carpenter GH (2013) The secretion, components, 13:197–212
and properties of saliva. Annu Rev Food Sci Technol 37. Streckfus CF, Bigler L, Dellinger T et al (2004)
4:267–276 The expression of the c-erbB-2 receptor protein in
18. Ben-Aryeh H, Fisher M, Szargel R et al (1990) glandular salivary secretions. J Oral Pathol Med
Composition of whole unstimulated saliva of 33:595–600
healthy children: changes with age. Arch Oral Biol 38. Fabian TK, Fejerdy P, Csermely P (2008) Salivary
35:929–931 genomics, transcriptomics, and proteomics: the
19. Salvolini E, Mazzanti L, Martarelli D et al (1999) emerging concept of the oral ecosystem and their use
Changes in the composition of human unstimulated in the early diagnosis of cancer and other diseases.
whole saliva with age. Aging 11:119–122 Curr Genomics 9:11–21
20. Nuriel-Ohayon M, Neuman H, Koren O (2016) 39. Parahitiyawa NB, Scully C, Leung WK et al (2010)
Microbial changes during pregnancy, birth, and Exploring the oral bacterial flora: current status and
infancy. Front Microbiol 7:1031 future directions. Oral Dis 16:136–145
21. Hassaneen M, Maron JL (2017) Salivary diagnostics 40. Lif Holgerson P, Öhman C, Rönnlund A et al (2015)
in pediatrics: applicability, translatability, and limita- Maturation of oral microbiota in children with or
tions. Front Public Health 20(5):83 without dental caries. PLoS One 10:e0128534
22. Vandewoestyne M, Van Hoofstat D, Franssen A et al 41. Li Y, Zhou X, St John MA et al (2004) RNA profil-
(2013) Presence and potential of cell free DNA in ing of cell-free saliva using microarray technology.
different types of forensic samples. Forensic Sci Int J Dent Res 83:199–203
Genet 7:316–320 42. Meyer S, Temme C, Wahle E (2004) Messenger RNA
23. Tu M, Wei F, Yang J et al (2015) Detection of exo- turnover in eukaryotes: pathways and enzymes. Crit
somal biomarker by electric field-induced release and Rev Biochem Mol Biol 39:197–216
measurement (ERIFM). J Vis Exp 95:52439 43. Wong BC, Chiu RW, Tsui NB et al (2005) Circulating
24. Yao W, Mei C, Nan X et al (2016) Evaluation and placental RNA in maternal plasma is associated with
comparison of in vitro degradation kinetics of DNA a preponderance of 5mRNA fragments: implications
in serum, urine and saliva: a qualitative study. Gene for noninvasive prenatal diagnosis and monitoring.
15:142–148 Clin Chem 51:1786–1795
25. Spielmann N, Ilsley D, Gu J et al (2012) The human 44. Park NJ, Li Y, Yu T et al (2006) Characterization of
salivary RNA transcriptome revealed by massively RNA in saliva. Clin Chem 52(6):988–994
parallel sequencing. Clin Chem 58:1314–1321 45. Lamblin G, Degroote S, Perini JM et al (2001) Human
26. Brosnan CA, Voinnet O (2009) The long and the short airway mucin glycosylation: a combinatory of carbo-
of noncoding RNAs. Curr Opin Cell Biol 21:416–425 hydrate determinants which vary in cystic fibrosis.
27. Esteller M (2011) Non-coding RNAs in human dis- Glycoconj J 18:661–684
ease. Nat Rev Genet 12:861–874 46. Turchinovich A, Weiz L, Langheinz A et al (2011)
28. Lasda E, Parker R (2014) Circular RNAs: diversity of Characterization of extracellular circulating
form and function. RNA 20:1829–1842 microRNA. Nucleic Acids Res 39:7223–7233
29. Li Y, St John MA, Zhou X et al (2004) Salivary tran- 47. Valadi H, Ekström K, Bossios A et al (2007) Exosome-­
scriptome diagnostics for oral cancer detection. Clin mediated transfer of mRNAs and microRNAs is a
Cancer Res 10:8442–8450 novel mechanism of genetic exchange between cells.
30. Dalmay T (2008) MicroRNAs and cancer. J Intern Nat Cell Biol 9:654–659
Med 263:366–375 48. Michael A, Bajracharya SD, Yuen PS et al (2010)
31. Mercer TR, Dinger ME, Mattick JS (2009) Long non-­ Exosomes from human saliva as a source of
coding RNAs: insights into functions. Nat Rev Genet microRNA biomarkers. Oral Dis 16:34–38
10:155–159 49. Koga Y, Yasunaga M, Moriya Y et al (2011) Exosome
32. Reis EM, Verjovski-Almeida S (2012) Perspectives of can prevent RNase from degrading microRNA in
long non-coding RNAs in cancer diagnostics. Front feces. J Gastrointest Oncol 2:215–222
Genet 3:32 50. Gallo A, Tandon M, Alevizos I et al (2012) The major-
33. Tang H, Wu Z, Zhang J et al (2013) Salivary lncRNA ity of microRNAs detectable in serum and saliva is
as a potential marker for oral squamous cell carci- concentrated in exosomes. PLoS One 7:e30679
noma diagnosis. Mol Med Rep 7:761–766 51. Ogawa Y, Taketomi Y, Murakami M et al (2013)
34. Jafari Ghods F, Topal Sarikaya A, Arda N et al (2017) Small RNA transcriptomes of two types of exosomes
MiRNA and mRNA profiling in systemic lupus reveals in human whole saliva determined by next generation
a novel set of cytokine – related miRNAs and their sequencing. Biol Pharm Bull 36:66–75
138 F. Jafari Ghods

52. Ogawa Y, Miura Y, Harazono A et al (2011) Proteomic 70. Jeck WR, Sharpless NE (2014) Detecting and charac-
analysis of two types of exosomes in human whole terizing circular RNAs. Nat Biotechnol 32(5):453–461
saliva. Biol Pharm Bull 34:13–23 71. Yang L, Duff MO, Graveley BR et al (2011)
53. Wong DT (2012) Salivaomics. J Am Dent Assoc Genomewide characterization of non-polyadenylated
143:19S–24S RNAs. Genome Biol 12(2):R16
54. Ramachandran P, Boontheung P, Xie Y et al (2006) 72. Memczak S, Papavasileiou P, Peters O et al (2015)
Identification of N-linked glycoproteins in human Identification and characterization of circular RNAs
saliva by glycoprotein capture and mass spectrometry. as a new class of putative biomarkers in human blood.
J. Proteome Res 5:1493–1503 PLoS One 10(10):e0141214
55. Park NJ, Zhou X, Yu T et al (2007) Characterization 73. Enuka Y, Lauriola M, Feldman ME et al (2016)
of salivary RNA by cDNA library analysis. Arch Oral Circular RNAs are long-lived and display only mini-
Biol 52:30–35 mal early alterations in response to a growth factor.
56. Whitelegge JP, Zabrouskov V, Halgand F et al Nucleic Acids Res 44(3):1370–1383
(2007) Protein-sequence polymorphisms and post-­ 74. Hansen TB, Jensen TI, Clausen BH et al (2013)
translational modifications in proteins from human Natural RNA circles function as efficient microRNA
saliva using top–down fourier-transform ion cyclotron sponges. Nature 495:384–388
resonance mass spectrometry. Int J Mass Spectrom 75. Wang Y, Wang Z (2015) Efficient back splicing pro-
268:190–197 duces translatable circular mRNAs. RNA 21:172–179
57. Hu Z, Zimmermann BG, Zhou H et al (2008) 76. Xie H, Ren X, Xin S et al (2016) Emerging roles of
Exon-­level expression profiling: a comprehensive circRNA_001569 targeting miR-145 in the prolif-
transcriptome analysis of oral fluids. Clin Chem eration and invasion of colorectal cancer. Oncotarget
54:824–832 7(18):26680–26691
58. Wei F, Wang J, Liao W et al (2008) Electrochemical 77. Zhu M, Xu Y, Chen Y et al (2017) Circular BANP,
detection of low-copy number salivary RNA based an up-regulated circular RNA that modulates cell pro-
on specific signal amplification with a hairpin probe. liferation in colorectal cancer. Biomed Pharmacother
Nucleic Acids Res 36:e65 88:138–144
59. Sondej M, Denny PA, Xie Y et al (2009) Glycoprofiling 78. Zhang YG, Yang HL, Long Y, Li WL (2016) Circular
of the human salivary proteome. Clin Proteomics RNA in blood corpuscles combined with plasma pro-
5:52–68 tein factor for early prediction of pre-eclampsia. BJOG
60. Hu S, Jiang J, Wong DT (2010) Proteomic analysis of 123:2113. https://doi.org/10.1111/1471-0528.13897
saliva: 2D gel electrophoresis, LC–MS/MS, and west- 79. Zhao Z, Li X, Gao C et al (2017) Peripheral blood cir-
ern blotting. Methods Mol Biol 666:31–41 cular RNA hsa_circ_0124644 can be used as a diag-
61. Oshlack A, Robinson MD, Young MD (2010) From nostic biomarker of coronary artery disease. Sci Rep
RNA-seq reads to differential expression results. 7:39918
Genome Biol 11:220 80. Panda AC, Grammatikakis I, Kim KM et al (2017)
62. Mandel ID (1993) Salivary diagnosis: promises, Identification of senescence-associated circular
promises. Ann N Y Acad Sci 694:1–10 RNAs (SAC-RNAs) reveals senescence suppressor
63. Kaczor-Urbanowicz KE, Martin Carreras-Presas CircPVT1. Nucleic Acids Res 45(7):4021–4035
C, Aro K et al (2017) Saliva diagnostics – current 81. Li H, Hao X, Wang H et al (2016) Circular RNA
views and directions. Exp Biol Med (Maywood) expression profile of pancreatic ductal adenocarci-
242(5):459–472 noma revealed by microarray. Cell Physiol Biochem
64. Miller SM (1994) Saliva testing--a non-traditional 40(6):1334–1344
diagnostic tool. Clin Lab Sci 7(1):39–44 82. Li P, Chen S, Chen H et al (2015) Using circular RNA
65. Majem B, Rigau M, Reventós J, et al (2015) Non-­ as a novel type of biomarker in the screening of gas-
coding RNAs in saliva: emerging biomarkers for tric cancer. Clin Chim Acta 444:132–136
molecular diagnostics. Farah CS, ed. Int J Mol Sci 83. Du WW, Yang W, Chen Y et al (2017) Foxo3 circu-
16(4):8676–8698 lar RNA promotes cardiac senescence by modulating
66. Lee YH, Kim JH, Zhou H et al (2012) Salivary tran- multiple factors associated with stress and senescence
scriptomic biomarkers for detection of ovarian can- responses. Eur Heart J 38:1402–1412
cer: for serous papillary adenocarcinoma. J Mol Med 84. Vausort M, Salgado-Somoza A, Zhang L et al (2016)
(Berl) 90:427–434 Myocardial infarction-associated circular RNA
67. Zhang L, Xiao H, Karlan S et al (2010) Discovery and predicting left ventricular dysfunction. J Am Coll
­
preclinical validation of salivary transcriptomic and Cardiol 68:1247–1248
proteomic biomarkers for the noninvasive detection 85. Cao S, Wei D, Li X et al (2017) Novel circular RNA
of breast cancer. PLoS One 5:e15573 expression profiles reflect progression of patients
68. Hu S, Wang J, Meijer J et al (2007) Salivary pro- with hypopharyngeal squamous cell carcinoma.
teomic and genomic biomarkers for primary Sjogren’s Oncotarget 8(28):45367–45379
syndrome. Arthritis Rheum 56:3588–3600 86. Ouyang Q, Wu J, Jiang Z et al (2017) Microarray
69. Li Y, Zheng Q, Bao C et al (2015) Circular RNA is expression profile of circular RNAs in peripheral
enriched and stable in exosomes: a promising bio- blood mononuclear cells from rheumatoid arthritis
marker for cancer diagnosis. Cell Res 25(8):981–984 patients. Cell Physiol Biochem 42(2):651–659
11 Circular RNA in Saliva 139

87. Li H, Li K, Lai W et al (2018) Comprehensive circu- 92. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular
lar RNA profiles in plasma reveals that circular RNAs RNAs are abundant, conserved, and associated with
can be used as novel biomarkers for systemic lupus ALU repeats. RNA 19(2):141–157
erythematosus. Clin Chim Acta 480:17–25 93. Cortés-López M, Gruner MR, Cooper DA et al
88. Wang YH, Yu XH, Luo SH et al (2015) (2018) Global accumulation of circRNAs during
Comprehensive circular RNA profiling reveals that aging in Caenorhabditis elegans. BMC Genomics
circularRNA100783 is involved in chronic CD28- 19(1):8
associated CD8(+) T cell ageing. Immun Ageing 94. Abu N, Jamal R (2016) Circular RNAs as promising
12:17 biomarkers: a mini-review. Front Physiol 7:355
89. Turner-Cobb JM, Rixon L, Jessop DS (2008) A pro- 95. Greene J, Baird AM, Brady L et al (2017) Circular
spective study of diurnal cortisol responses to the RNAs: biogenesis, function and role in human dis-
social experience of school transition in four-year-old eases. Front Mol Biosci 4:38
children: anticipation, exposure, and adaptation. Dev 96. Barrett SP, Salzman J (2016) Circular RNAs: analy-
Psychobiol 50:377–389 sis, expression and potential functions. Development
90. Bahn JH, Zhang Q, Li F et al (2015) The landscape of 143(11):1838–1847
microRNA, Piwi-interacting RNA, and circular RNA 97. Chen I, Chen CY, Chuang TJ (2015) Biogenesis,
in human saliva. Clin Chem 61(1):221–230 identification, and function of exonic circular RNAs.
91. Suzuki H, Zuo Y, Wang J et al (2006) Characterization WIREs RNA 6:563–579
of RNase R-digested cellular RNA source that con- 98. Hansen TB, Venø MT, Damgaard CK et al (2016)
sists of lariat and circular RNAs from pre-mRNA Comparison of circular RNA prediction tools. Nucleic
splicing. Nucleic Acids Res 34:e63 Acids Res 44(6):e58
 Emerging Role of Circular RNAs
as Potential Biomarkers 12
for the Diagnosis of Human
Diseases

Rupal Ojha, Raj Nandani, Nina Chatterjee,

and Vijay Kumar Prajapati

Abstract depicted to get upregulated in breast cancer

In the eukaryotic transcriptome, the evolution- which is associated with disease progression.
ary conserved circular RNAs naturally occur Furthermore, it has been observed that cir-
from the family of noncoding RNAs. Circular cRNAs are abundantly present within the
RNAs possess a unique feature to interact with mammalian brain tissues. In epileptic condi-
nucleic acids and ribonucleoproteins and are tion, Circ-EFCAB2 was observed to get nota-
establishing themselves as an obligatory com- bly upregulated within patients. Taking the
position for the regulatory messages which are above conditions into consideration, circular
encoded by the genome. The back-splicing RNAs have proven themselves as promising
mechanism leads to the formation of circular- noninvasive biomarker for the detection of
ized RNA, and because of this they become human diseases.
resistant to exonuclease-mediated degrada-
tion. The differential and aberrant expression Keywords
of circular RNAs can be detected with the help Circular RNA · Biomarker · Diagnosis ·
of various profiling methods by using serum, Human diseases
plasma, and tissue samples. In this chapter, we
have highlighted the role of circular RNAs as
putative biomarker for the detection of various
human diseases along with its profiling meth- 1 Introduction
ods. Here we have discussed the differentially
expressed circular RNAs in neurological dis- The noncoding single-stranded circular RNAs
orders and infectious diseases along with can- are stable, bountiful, and evolutionary conserved.
cer diseases. For instance, in case of pulmonary Like other forms of RNA, the circular RNAs also
tuberculosis, hsa_circRNA_001937 was originated in the nucleus and are then exported to
upregulated, while hsa_circRNA_102101 got cytoplasm with the help of nuclear pore complex.
downregulated; Hsa_circ_000178 was The presence of circular RNA in cytoplasm is ten
times greater in comparison to the linear RNAs.
The role of circular RNA is still unclear, but some
R. Ojha · R. Nandani · N. Chatterjee · V. K. Prajapati
(*) circular RNA play a major character in gene reg-
Department of Biochemistry, School of Life ulation as well as in pathophysiological processes
Sciences, Central University of Rajasthan, by conducting itself as miRNA sponge, exporter
Bandarsindri, Ajmer, Rajasthan, India of RNA, and binding protein molecules. Apart
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 141

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_12
142 R. Ojha et al.

from gene regulation, they also play a key role in In 1979, circRNAs were primarily observed in
the normal homeostasis, for example, in regula- mammalian cells with the help of electron
tion and control of cell cycle, development of microscopy [6]. Further, the presence of circRNA
embryo, regulation of metabolic activities, and was reported in yeast and viroid viruses, but till
stress condition of cell [1]. Maintenance of con- date, there are very few circular RNAs reported
ventional homeostasis is very important for nor- for humans. In the 1990s, the circRNAs were first
mal functioning of living organisms. The identified from DCC transcript study in human
circumstances under which circular RNAs are cells [7]. The circular RNAs were first experien-
formed are quite different from those of the linear tial in Sry gene of mouse adult testis. This gene
ones because circular RNAs form a covalent plays a vital role in the sex determination of
closed-loop for circularization, and it is possible embryo. Many experimental assays were per-
because they lack 5′ (cap) and 3′ (polyadenyl- formed including RT-PCR and RNAase protec-
ation) end. Due to the circularization process, tion assays to confirm the presence of circular
they possess resistance to exonuclease-mediated RNA, and it was concluded that the circular
degradation. The unique feature makes the circu- RNAs were highly profuse in testis. These circu-
lar RNA highly stable and allows its interaction lars are RNAs differentially expressed at the time
to many molecules including miRNAs (miRNA of infection or disease. This distinctive character-
sponges) and spliceosomes complex (RNA-­ istic of circular RNA makes them unique and can
protein complex) which further helps in the tran- be used as potential biomarker for the evaluation
scription process [2]. The splicing mechanism of of human diseases or infections. The high stabil-
circular RNA is mediated by alternative and ity and expression of circular RNAs in blood or
back-splicing of primary mRNA which leads to other body fluids make them unique biomarkers.
the formation of circular RNA in which the exon Many clinical trials have been conducted to iden-
part gets shuffled and forms different protein tify the role of remarkable circular RNAs in vari-
products. In back-splicing, the donor splice site is ous disease forms, from the clinical serum
joined to acceptor splice site which exists at the samples. In this chapter, we will emphasize the
upstream region and forms circRNAs as product, differential upregulation and downregulation of
whereas in case of normal splicing, the donor circular RNAs at the juncture of different dis-
splice site is usually joined to downstream accep- eases along with molecular identification
tor splice site [3]. Most of the circRNAs can be methods.
procured from exonic or intronic forms, as well
as from 5′-3′ untranslated regions. The difference
between the exonic and intronic circRNAs is that 2  ole of circRNA as Potential
R
the former one connects the exons and forms Diagnostic and Prognostic
truncated but functional protein products, Biomarker
whereas the intronic circRNAs formed by the
joining of introns is meagre in eukaryotes [2]. CircRNAs are highly stable and resistant to deb-
During the intronic splicing, the lariat structure ranching and exonuclease-mediated degradation
formed is different from exonic splicing because in comparison to linear form of RNAs. They arise
of the formation of 2′-5′ carbon linkage at the in nucleus and exported to the cytoplasm via exo-
splicing junction. Later on, it was observed that somes, where they differentially expressed and
the lariat structure formed from intronic RNAs is function as potential biomarker for the diagnosis
very stable and possesses the properties to of diseased condition. Their presence in mamma-
degrade the 3′ appendages and leave the remains lian plasma, serum samples, and body fluids (e.g.,
behind [4]. This devised lariat products are saliva) secreted from various types of tissues
known as circular intronic RNAs, while the makes them remarkable. In comparison to linear
RNAs which exhibit 3′-5′ junctions are known as RNAs, the level of circular RNAs is higher in the
circular exonic RNAs [5] (Fig. 12.1). serum sample, as the canonical s­ plicing-­mediated
12 Emerging Role of Circular RNAs as Potential Biomarkers for the Diagnosis of Human Diseases 143

Fig. 12.1 RNA splicing mechanism noncanonical splicing also known as “back-splicing.”
Formation of CircRNA isoform diversity. General over- Three forms of circular RNAs are formed during splic-
view of types of splicing that includes canonical linear ing – exonic circRNA, intronic RNA, and intronic +
splicing and noncanonical splicing. A great diversity of exonic circRNA. Colored boxes represent exons; black
circRNAs is generated from a single genomic locus by lines represent introns

linear RNAs got degraded after a short span of comparison to healthy individuals. In case of gas-
time, so, this groovy feature of circular RNA tric cancer condition, has_circ_0001649 gets
makes them potential diagnostic and prognostic downregulated in the serum and tumor tissue
biomarker because of their high stability. With the samples and is responsible for the development
help of body fluids and serum sampling, the upreg- of metastasis and consequently can be used as
ulating and downregulating circular RNAs can be noninvasive prognostic biomarker [8]. Further, in
easily identified. Even at room temperature, cir- hepatocellular carcinoma circRNA, hsa_
cRNAs are more stable than linear ones, at least circ_0001649 is downregulated as the tumori-
for 24 h. In many diseased conditions including genic condition intensifies. The infectious
cardiovascular diseases, neurological disorders, hepatitis B and cancerous hepatocellular carci-
cancer, and infectious diseases, the differential noma (HCC) are interrelated to each other. The
expression of circRNAs has been observed. For circRNA_100338 regulates the expression levels
instance, ciRS-7 in Alzheimer’s disease was seen of miR-141-3p which is a disease-relevant
to be upregulated which helps in the degradation miRNA and proliferates the metastasis condition.
of amyloid peptides. Secondly, it was observed Thus, the overall differential and tissue-specific
that the circular RNAs are lavishly present in the expression of circular RNAs in diseased as well
heart tissues. During the clinical trials, it was iden- as normal condition makes them conventional
tified that the patients who suffered from ischemia biomarkers for the diagnosis and prognosis pur-
disease had upregulated hsa_circ_0124644 in pose (Fig. 12.2).
144 R. Ojha et al.

Fig. 12.2 Overview of the differentially expressed cir- The upregulated circRNAs are indicated by green color
cRNAs in various diseases pyramids, while the downregulated ones are shown in red
color pyramids

3  ole of circRNA as miRNA

R For instance, the circular RNA ciRS-7 or CDR1
Sponge has multiple binding sites for miR-7 and func-
tions as a miRNA sponge which is tremendously
As it has already been conversed that circular expressed during the development of nervous
RNAs interact with many molecules, among system in mammals [10]. Thereby, it would be
them, miRNAs are very usual. When the circular more useful to identify the consequences over
RNA interacts with the miRNA, they form neuronal development in absence of ciRS-7. So,
miRNA sponges which further suppress the tran- to decipher the role of ciRS-7 in brain develop-
script and lead to gene silencing. MiRNAs are a ment, people chose zebra fish as animal model
class of small noncoding RNAs which play a because it has lost cdr1 locus, while miR-7
vital role in transcriptional silencing. They bind expresses highly in embryonic brain and con-
to the target mRNA, due to complementarity served as well. Furthermore, the researchers
among the bases, and perform the process of identified a tool named Morpholino, which is
transcriptional silencing in order to control the widely known as Morpholino oligomer. This
gene regulation. Sponges can be either artificial oligomer can be used for the regulation of gene
or natural and have multiple miRNA binding expression and works as a sponge by base pairing
sites according to miRNA gene. The two natu- with the target molecules. When this Morpholino
rally occurring circRNA sponges include SRY type of oligomer synthetically is given as treat-
and CDR1/ciRS-7, aiming miR-138 and miR-7, ment in zebra fish, they observed defects in brain
respectively. The multiple binding sites present development process. These findings demon-
on the sponge are particularly specific for the strate that ciRS-7 efficiently interacts with miR-7
miRNA and act as a competitive inhibitor, hence and causes loss of function of miR-7 which
overwhelming the binding of miRNA with its tar- results in defective midbrain development [11,
get mRNA for posttranscriptional silencing [9]. 12]. This study connotes a noteworthy ­relationship
12 Emerging Role of Circular RNAs as Potential Biomarkers for the Diagnosis of Human Diseases 145

Fig. 12.3 Putative role of circRNAs as sponge transcription of the mRNA. But, in the diseased condition,
CircS-7 acts as sponge or decoy of miR-7. Binding of circRNAs are not able to totally sponge the miR-7 that in
miRNA-7 to circRNA-­7 during the normal condition pre- turn binds to mRNA leading to silencing of that gene
vents its binding with mRNA, and this leads to normal

between miR-7 and ciRS-7. Due to the higher important for the diagnosis of various diseases
stability, nowadays, synthetically developed including cancers, neurological disorders, infec-
circular RNAs are being used as sponge for the tious diseases, and many more. The early detec-
regulation of gene expression. During the hepati- tion of diseased condition will be advantageous in
tis C virus infection, miRNA-122 highly the treatment of ailment. As it has already been
expresses and supports in dissemination of hepa- mentioned that circRNAs were initially detected
titis infection. Recently, an anti-miRNA drug via electron microscopy, this technique was
named Miravirsen has been identified, which unable to discriminate between linear and circular
plays a vital role in sequestration of miRNA-122 forms of RNA. After that several analytical meth-
and its activity as circRNA does. Because of the ods were recognized for the detection and quanti-
sequestration process, miRNA is unable to bind fication of circRNAs including microarray
to the target molecule and hence does not per- technique, RT-PCR followed by PCR (polymerase
form the posttranscriptional silencing of gene chain reaction), and Northern blot [14]. But due to
[13]. As circular RNAs possess multiple binding low sensitivity and specificity, these techniques
sites for miRNAs, the artificial circular RNAs as are incapable to examine the full sequence of cir-
miRNA sponges can be widely used for the cRNAs. Hence, due to certain drawbacks, these
sequestration of disease-pertinent miRNAs abovementioned techniques cannot be that much
(Fig. 12.3). reliable. Next method includes RNA-Seq, stan-
dard high-throughput RNA sequencing analytical
technique, which is widely being used for the
detection of circular RNAs, but due to lack of dis-
4 Techniques for the Analysis crimination property between the linear and cir-
of circRNAs Along with cular RNA, there will be generation of
Computational Approaches indeterminate results which is not conventional
[15]. To overcome all the complications associ-
The efficient experimental analysis of circular ated with the aforementioned analytical methods,
RNA can be achieved by the help of various circu- recently developed RAPD and high-throughput
lar RNA profiling methods. Their detection is RNA sequencing united with computational
146 R. Ojha et al.

approach are used for the high-purity separation 25% of new case [17]. Mycobacterium is aerobic
of circRNAs on the basis of quantitative and qual- and resides inside the human body easily and can
itative analyses. The circular RNAs lack 5′ cap flourish there happily to cause infection. TB is
and 3′ polyadenylation due to which they become categorized into two types based on the forms of
highly stable. The interaction of circRNAs with bacteria, the latent pulmonary TB which is con-
various ribosome-­binding proteins and miRNAs sidered to be asymptomatic and the active TB
affects the gene regulation and expression. When which is symptomatic. About one quarter of the
the RNA sample is treated with exoribonucleases, world’s population is affected from latent form,
all the linear forms of RNAs get degraded, and and this form could develop into symptomatic
only the circular RNA remnants are left. This form as the immune system becomes compro-
RAPD technique determines the circRNA, origi- mised. Initial symptoms of the disease include
nated from the back-splicing of transcriptome. extreme cough, fever, night sweats, blood in spu-
Many bioinformatics tools nowadays have been tum, excessive body weight loss, loss of appetite,
utilizing for the identification of novel back- difficulty in breathing, anxiety, and severe chest
spliced exonic, intronic, and exo-intronic circular pains and ultimately lead to lung cancer and
RNAs. The most popular tool associated with the death. TB is contagious and can spread through
RNA-Seq is find_circ; this analytical tool detects air, sneeze, or spit of the infected patients [17].
the back-­spliced exon reads produced from the Due to the decreasing survival rates, there is
RNA-Seq technique. It has ability to read the urgent need to diagnose the disease condition
sequences of more than 100 nucleotides. Next, ­primarily. This can be achieved with the help of
CIRCexplorer, the solitary annotating tool which small noncoding circRNAs which can be used as
is able to identify the back-spliced exonic and diagnostic and prognostic biomarkers. These
intronic circRNAs sequence reads, is obtained ­circRNAs play important role in gene regulation
from RNA-seq [16]. Subsequently, circRNA_ and contribute to the development of many human
finder is used for determining the proximal splice disorders. During the diseased (pulmonary)
sites by using RNA-Seq data. These aforesaid condition, there is a dysregulation in the
­
profiling methods help in the identification of cir- ­expression of these circRNAs; some show down-
cular RNAs whose expression levels are associ- regulation, while some show upregulation. A
ated with the various diseases. total of 37 circRNAs have shown dysregulated
expression in the PBMC cells of infected patients
when compared with the healthy cohorts. Among
5  ircular RNAs Role
C them, 13 circRNAs were upregulated which
in the Diagnosis of Infectious include 10 exonic, 2 intronic, and 1 sense overlap-
Diseases ping, while among 24 downregulated circRNAs,
there were 15 exonic, 4 intronic, 3 overlapping, 1
5.1 Tuberculosis intergenic, and 1 antisense seen. The upregulated
circRNAs were hsa_circRNA_001937, hsa_
Tuberculosis (TB) is an infectious disease caused circ_0000414, hsa_circ_0000681, hsa_circ_
by bacteria Mycobacterium tuberculosis that most 0002113, hsa_circ_0002362, hsa_circ_0002908,
often affects the lungs (pulmonary) and some- hsa_circ_000879, hsa_circ_0063179hsa_cir-
times the bones. According to WHO, tuberculosis cRNA_009024, hsa_circRNA_005086, hsa_
(TB) is considered to be one of the top ten causes circRNA_103948, hsa_circRNA_003524,
of morbidity and mortality worldwide. They gave hsa_circRNA_015879, hsa_circRNA_009377,
an estimate that in the year 2015, approximately hsa_circRNA_103285, hsa_circRNA_406505,
10.4 million people were suffering from TB and and hsa_circRNA_005232. The downregulated
1.8 million people died due to it. Almost every circRNAs were hsa_circRNA_102101, hsa_
part of world is affected by TB, but in 2016 the circRNA_104964, hsa_circRNA_104296,
largest number of new TB cases occurred in Asia, hsa_circRNA_003416, hsa_circRNA_002971,
with 45% of new cases, followed by Africa, with hsa_circRNA_007738, hsa_circRNA_000686,
12 Emerging Role of Circular RNAs as Potential Biomarkers for the Diagnosis of Human Diseases 147

hsa_circRNA_048148, hsa_circRNA_092458, majorly affects males and is responsible for

and hsa_circRNA_002465. These circRNAs reg- approximately 600,000 deaths annually across
ulate several physiological processes inside the the world. HCV infection is a dominant risk fac-
human body such as autophagy, apoptosis, cell tor in the majority of areas of Asia and sub-­
cycle regulation, and proliferation [18–21]. Saharan Africa that have a high incidence rate of
Dysregulation of the circRNAs leads to malfunc- HCC [24]. The etiological agents that lead to
tioning of the physiological processes and ulti- development of HCC are elevated viral loads,
mately results in several pathological conditions. having hepatitis virus envelope and surface anti-
The pathologies develop due to interruption of gens. Several other factors like obesity, excessive
many signaling and biological pathways in human alcohol abuse, metabolic syndrome, diabetes,
PBMCs such as “chemokine signaling pathway,” hereditary hemochromatosis with cirrhosis, etc.
“Fc gamma R-mediated phagocytosis,” “neuro- are also associated with development of HCC
trophin signaling pathway,” “cytokine-cytokine [25, 26]. The pathogenesis of HCC is due to dif-
receptor interaction,” and “bacterial invasion of ferent genetic/epigenetic aberrations and altera-
epithelial cells” [22, 23] ultimately lead to devel- tions with many signaling pathways that lead to
opment of TB. So, increase or decrease in the the disease. Various symptoms of the disease
number circRNAs during the disease can serve as include weight loss, pain, emotional distress,
clinical diagnostic markers and therapeutic tar- sleep disturbance, anemia, nutritional deficien-
gets. Detection of TB is hard to be early diag- cies, ascites, nausea vomiting, esophageal reflux,
nosed; several methods for diagnosis include pain, peripheral edema, and constipation. Many
smear microscopy and mycobacterial culture. research studies have been done for analyzing the
These are simple but have poor sensitivity, are dysregulated expression circular RNAs during
costly, and are time-consuming. The other method the diseased conditions. Research investigation
used for the identification of acid fast bacteria by Shichang Cui et al. revealed that a total of 24
Mycobacterium tuberculosis is Lowenstein- circRNAs were upregulated, while 23 circRNAs
Jensen culture method, but again it is very time were downregulated. Among these the top five
consuming process, as it takes an average of 4–5 upregulated circRNAs were hsa_cir-
weeks to yield the results. The circRNAs are cRNA_104351, hsa_circRNA_102814, hsa_cir-
highly nuclease-resistant, more stable than linear cRNA_103489, hsa_circRNA_102109, and
transcripts, and may be released into the extracel- hsa_circRNA_100381. Furthermore, the top five
lular space via the exosomes; these peculiar char- downregulated circRNAs were hsa_cir-
acteristics make them more potent to be used as a cRNA_100327, hsa_circRNA_101764, hsa_cir-
biomarker. Therefore, use of these circRNAs as cRNA_101092, hsa_circRNA_001225, and
biomarker can be a help in the diagnosis process hsa_circRNA_102904. Whereas, other research-
as they are rapid, sensitive, and cost-effective. ers have detected the dysregulated expression of
some other circRNAs. The four circRNAs- circ-
MTO1, hsa_circ_0001649, circZKSCAN1, and
5.2 Hepatitis hsa_circ_0004018 were downregulated, while
three circRNAs named, hsa_circ_0005075, ciRS-­
An inflammatory ailment of liver leads to hepati- 7(Cdr1as), and circRNA_100338 got upregulated
tis, which is caused by viral infection. There are a during HCC condition. CircMTO1 is responsible
total of five types of hepatitis A, B, C, D, and for repression of HCC progression by sponging
E. Among them hepatitis B virus (HBV) and hep- miR-9 that in turn leads to increased expression
atitis C virus (HCV) are associated with hepato- of p21 gene. CiRS-7 targets miR-7 and leads to
cellular carcinoma. Hepatocellular carcinoma enhancement of cell proliferation and invasion by
(HCC) is the cancer of liver tissues; it is consid- promoting CCNE1, and PIK3CD expression
ered to be the sixth most common type of cancer CircZKSCAN1 is responsible for inhibiting the
worldwide. The major number of HCC has been HCC cell growth, migration, and invasion by
linked to be caused by chronic hepatitis B. It regulating cancer cell signaling pathways.
148 R. Ojha et al.

CircRNAs, hsa_circ_0004018, circRNA_100338, lung cancer caused approximately 1.69 million

and circRNA_000839, play roles in HCC devel- deaths worldwide. The lung cancer is classified
opment [21, 27, 28]. CircRNAs regulate cancer into two categories on the basis of prognosis con-
development through a number of mechanisms, dition non-small cell lung cancer (NSCLC),
including miRNA sponges, modulating which is the major one and small cell lung cancer
epithelial-­mesenchymal transition, Wnt signaling (SCLC). Approximately 85% of the lung cancers
pathway, the p53 PIK3CA, and β-catenin gene are of non-small cell lung cancer type. Lung ade-
mutation [29]. Literature survey has suggested nocarcinoma is a subtype of NSCLC that forms in
that circular RNAs are initially present in viruses mucus-secreting glands throughout the lung and
as unique noncoding RNA molecules. During the is seen to be spreading at very fast rate among the
course of infection, these circRNAs are trans- nonsmokers [30, 31]. The various etiological
ferred to the host. These CircRNAs are stable agents that are considered to cause lung cancer
structure and have tissue-specific expression and include smoking cigarettes, chewing tobacco,
are widely present in the cytoplasm of eukaryotic snuffing powdered form of tobacco, passive
organisms, in the circular form. They are also smoking, and occupational exposure to chemicals
responsible for developing the disease by regulat- like arsenic, cadmium, and pollutants, etc.
ing gene expression by competing with the Symptoms include coughing up blood, ache dur-
endogenous RNAs of the cells. It modulates the ing breathing, persistent breathlessness, loss of
function of miRNAs, by terminating the suppres- appetite, unexplained weight loss, feeling of tired-
sion from their targets, which in turn leads to ness, persistent coughing, etc. Differential expres-
modulated expression levels of other associated sion of circRNAs has been reported by some
RNA molecules. The interaction between cir- researchers when they compared the tissue sam-
cRNAs and disease-associated miRNAs indicates ple from NSCLC patients and healthy individuals.
that circRNAs are important for disease regula- Zhang et al. reported a total of ten dysregulated
tion. The different available diagnostics methods circRNAs, out of which five were upregulated
are Liver Imaging Reporting and Data System and five were downregulated. The upregulated
(LIRADS), CT, MRI, contrast imaging, ultra- circRNAs were hsa_circ_0000735, hsa_
sound, etc. These aforesaid diagnosis procedures circ_0016760, hsa_circ_0003645, hsa_
are somewhat time-consuming, less sensitive, circ_0087862, and hsa_circ_0026134, while the
require validation, and expensive. Because of the downregulated circRNAs were hsa_
drawbacks of above-discussed diagnostic meth- circ_0005730, hsa_circ_0091000, hsa_
ods, the circular RNAs as biomarkers come into circ_0014130, hsa_circ_0071989, and hsa_
light with unique feature to diagnose the diseased circ_0092368. Out of these circRNAs, nine were
condition primarily. exonic and only one was intronic. H ­sa_
circ_0014130 was responsible for negatively reg-
ulating the KEGG pathway that leads to cancer
6  ircular RNAs Role
C development [32]. Another group of researchers
in Diagnosis of Cancers found that cir-ITCH was downregulated in lung
cancer tissues and is responsible for cell prolifera-
6.1 Lung Cancer tion and metastasis, whereas its overexpression
inhibits lung cancer cell proliferation. The
Lung cancer is the cancer of lung tissues; it is upregulated cir-ITCH acts as tumor suppressor
characterized by the abnormal growth of the cells by acting as sponge of miR-7 and miR-214 as
and tissues, alveoli, bronchi, and the surrounding well as it also inhibits the Wnt pathway, leading
epithelial tissues of the lung. It is the most com- to reduced lung cancer cell proliferation [33].
mon cancer among men and ranks third among CircRNA_100876 upregulated level is related to
women in terms of its occurrence and mortality. lymph node metastasis and tumor progression
According to WHO report, in the year 2015 alone, [34]. Luo et al. reported the upregulated level of
12 Emerging Role of Circular RNAs as Potential Biomarkers for the Diagnosis of Human Diseases 149

hsa_circ_0000064 in lung cancer tissues [35]. reported the differential expression of circular
Circ_0013958/miR-134/cyclin D1 was associated RNAs by profiling of peripheral blood samples
with the development of lung adenocarcinoma in breast cancer patients. They have reported that
[36]. CircRNAFOXO3 (circRNAforkhead box during breast cancer, 19 circRNAs were upregu-
O3 class) has antioncogenic role; it acts as a tumor lated, while 22 circRNAs were showing the
suppressor by sponging miR-155 during the nor- downregulation. Further, it was observed that
mal conditions, but during the diseased condi- among the aforementioned number of circular
tions, it was found to be downregulated [37]. RNAs, hsa_circ_0001785, hsa_circ_00680333,
CircPRKC1 was overexpressed in lung adenocar- and hsa_circ_0108942 circRNAs were expressed
cinoma and hence responsible for cell prolifera- aberrantly. These findings suggested the abnor-
tion and tumorigenesis. It works as sponge for mal expression of hsa_circ_000178 during
both miR-545 and miR-589 and abrogates the breast cancer and hence proved the prognostic
suppression of the pro-tumorigenic transcription and diagnostic value of specific circular RNA
factor E2F7 [38]. Circular RNAs (circRNAs) are [39]. Additionally, hsa_circ_103110, hsa_
a class of endogenous noncoding RNAs that have circ_104821, and hsa_circ_104689 were found
been demonstrated to be potential regulators in to be upregulated at the time of cancerous condi-
the development and progression of lung cancers. tion [18]. Further, it was identified that these cir-
These novel circRNAs are differentially expressed cular RNAs are also associated with different
during the diseased conditions, and utilization of signaling pathways. For instance, an oncogenic
these circRNAs as biomarkers provides new protein encoded by circular RNA hsa_
insights in diagnostic processes. circ_103110 is seen to be involved in other can-
cer development such as ovarian cancer,
colorectal cancer, and squamous cancer. The cir-
6.2 Breast Cancer cRNAs expressed during the breast and colon
cancer are interlinked, for example, has_
Breast cancer is a fatal malignant cancerous dis- circ_104821 associated with PI3K and FAK sig-
ease among women, especially in the naling pathways of breast cancer as well as colon
USA. According to 2011 WHO report 50,8000 cancer. This was also concluded that in tumori-
women suffered and died because of breast can- genic tissues, the expression of hsa_circ_006054,
cer proliferation (http://www.who.int/cancer/ hsa_circ_406697, and hsa_circ_100219 was
detection/breastcancer/en/). The occurrence of downregulated. CircFOXO3, a tumor suppres-
breast cancer is very common in developed sor, is found to be downregulated in tumor tis-
regions as compared to developing regions. sues, which is another potential biomarker for
Several studies have been performed on the breast cancer diagnosis [40]. Thus, the promis-
breast cancer and it was concluded that due to ing feature of circRNAs as biomarker for the
modern lifestyle, various risks are allied with the diagnosis of cancer condition is remarkable.
disease which includes frequent consumption of
alcohol, delayed age pregnancy, and extreme
usage of estrogen and progesterone hormones. 6.3 Gastric Cancer
Apart from this, many epigenetic changes such
as methylation, acetylation, and phosphorylation Gastric cancer or stomach cancer arises due to
are also responsible for the progression of the uncontrolled cell growth and is the fifth most
tumorigenic breast cancer. The differential common cancerous condition. According to
expression of circular RNAs during the breast WHO, every year around 723,000 deaths are
cancer condition is more concentrating. Circular caused due to stomach cancer, globally. The most
RNAs came into picture because of their disease common type of gastric cancer condition is ade-
diagnosis property, so that the disease condition nocarcinoma, which develops from the mucosal
can be cured in an early stage. A recent study cells, produces mucous, and functions as cover-
150 R. Ojha et al.

ing or layer of internal organs, especially the cases occur in people aged 65 years or older, and
stomach. The symptoms of the disease include relatively men are affected more than women
dysphagia, gastric problems like heartburn and [41]. There are several etiological agents like
bloating, indigestion, and weight loss. The sur- bacteria, viruses, smoking, and red meat con-
vival rate is decreasing day by day among people sumption, increased use of aspirin, and physical
who have developed gastric cancer. So, for early inactivity. People with inflammatory bowel dis-
detection of disease, circular RNAs, the potential ease and infected with bacterium Helicobacter
biomarkers, are used for diagnosis purposes. pylori and Streptococcus gallolyticus are at
Circular RNAs are most stable and hence play a higher risk of developing colorectal cancers in
vital role in cell cycle regulation. They have the the future [42]. The colon and rectum are parts of
property to act as putative biomarker for the diag- the large intestine, which is the lower part of the
nosis of proliferative cancerous condition. By digestive system. Symptoms are not very particu-
profiling methods, researchers have identified lar; they are common as in case of other diseases.
that in gastric cancer condition, the number of They include diarrhea, constipation, narrowing
downregulated circRNAs is greater than the of the stool, rectal bleeding, blood in the stool,
upregulated circRNAs. Recently, a study was cramping or abdominal pain, weakness and
conducted in which it was depicted that hsa_ fatigue, and unexpected weight loss. The circular
circ_0000520 is notably downregulated in the RNAs are expressed during the diseased condi-
patients affected from gastric cancer as compared tions and found to regulate a variety of essential
to healthy controls. This hsa_circ_0000520 have biological functions such as cell proliferation,
showed to be involved in disease progression as it development, apoptosis, and pathologies. The
is negatively assisted with TNM stage. TNM circRNAs as miRNA sponge are believed to reg-
staging is a cancer staging system through which ulate the functions of miRNAs, as they are
the stage of any cancerous condition can be deter- involved in suppressing translation of the mRNA
mined. circPVT1, a tumor suppressor, is identi- and sometimes they are even capable of degrad-
fied which indicated as good biomarker for the ing the mRNAs if there is perfect complementar-
prediction of survival among the gastric cancer ity between them. Bachmayr-Heyda et al.
patients. Lastly, one more study was conducted reported the differential expression of several
between healthy and diseased patients, but it was genes between normal colon mucosa and CRC
identified that CNIH4, HIAT1, and KIAA0907 mucosa samples. These genes were responsible
circRNAs were downregulated during gastric for encoding certain mRNAs when further ana-
cancer condition. The all aforementioned circular lyzed. These mRNAs had a peculiar characteris-
RNAs play a dynamic role in the diagnosis of tic of being circular. The total count revealed that
gastric carcinoma. 39 circRNAs are significantly differentially
expressed, out of which 11 of them are upregu-
lated and 28 downregulated. Among the
6.4 Colorectal Cancer upregulated ones, the top circRNAs were
­
circCCDC66, circ-BANP, ciRS-7 (Cdr1as), hsa_
Colorectal cancer (CRC) can be linked to the list circ_0000069, hsa_circ_001569, and hsa_
of cancers caused by lifestyle factors. It is the circ_0020397, while the top downregulated ones
cancer that affects the colon or rectum. The third were hsa_circRNA_103809, hsa_cir-
most common type of cancer causes majority of cRNA_10470, hsa_circ_001988, and cir-
cancer-related deaths globally [21]. As per geo- cRNA0003906. According to Hsiao et al.,
graphical distribution it was found that the circCCDC66 and Circ-BANP play an important
colorectal cancer seems to be highest in south role in cell proliferation, migration, and invasion
and Midwest part while lowest in the western that leads to progression of tissues toward can-
parts of the world, among them, USA ranked first cerous condition. Hsa_circ_ 001569 is expressed
to encounter the majority of cases. Majority of at upregulated level and acts as a sponge of miR-
12 Emerging Role of Circular RNAs as Potential Biomarkers for the Diagnosis of Human Diseases 151

145 that increases colorectal cancer cell prolifer- suffering from AD only in the USA. Furthermore,
ation by targeting E2F5, BAG4, and FMNL [43]. there are high chances of these values to undergo
Circ-7 (Cdr1as) acts as miR-7 sponge to regulate a hike of 13.8 million by mid of this century. It is
expression of CDR1 gene. Studies have shown unfortunate enough that there are no effective
that CDR1 is expressed at high levels in CRC tis- therapeutics that are victorious in preventing the
sues. CDR1 is associated with tumor size, TNM gradual progression of AD till date [49]. In prior
stage, lymph node metastasis, and poor overall studies, it has been observed that the neurons are
survival [44]. The downregulated expression of enriched with circRNAs, and they get concen-
hsa_-circ_001988 was associated with tumor dif- trated in high amounts when the brain reaches
ferentiation and perineural invasion during senescence, which provide hints that there is a
colorectal cancer [45]. The downregulated possible role of circRNAs in age-related neuro-
expression of hsa_circRNA_103809 and hsa_cir- degenerative diseases. In AD neuropathy the
cRNA_104700 in colorectal cancer tissues was accumulation of amyloid-β (Aβ) peptide due to
associated with cancer metastasis [46]. Hsa_ the amyloid precursor protein (APP) proteolytic
circ_0020397 promotes TERT and PD-L1 processes is strongly believed since ages to be
expression by acting as sponge of miR-138 to one of the key steps. Tau phosphorylation and
regulate CRC cell viability, apoptosis, and inva- GSK3β levels are upregulated by Aβ circRNA-­
sion [47]. One of the few circRNA that have the derived peptide, which is an Aβ circRNA-­
dual behavior is cir-ITCH; its downregulation encoded expression of an AB-related peptide.
inhibits the Wnt/β-catenin pathway and promotes Both GSK3β and tau phosphorylation are hall-
ITCH expression including p63, p73, and Notch1 marks for the AD progression. It has been
and is usually associated with tumor formation, observed that the mutations that are caused in
whereas upregulated expression of cir-ITCH genes like APP and presenilin which actively par-
reduces cell proliferation [48]. These circRNAs ticipate in APP protein proteolytic processing
are present inside the exosomes in the cancer tis- speed up the accumulation of Aβ peptides rap-
sues, plasma, and serum samples and are specifi- idly. The accumulation of protein causes the pep-
cally expressed during colorectal cancers. Thus, tides to polymerize into pieces of insoluble
they can be used as putative biomarkers for the amyloid plaques, and toxic forms and drives
diagnosis of various oncogenic conditions. ahead to tau protein hyperphosphorylation via
GSK3β activation, followed by the formation of
subsequent tangles of neurofibrils which form
7  ircular RNAs Role
C inside ­neurons, ultimately triggering a relay of
in Diagnosis of Neurological events that result into neuron death. Alzheimer’s
Disorders associated with this is known as familial
Alzheimer’s disease (FAD). In the case of spo-
7.1 Alzheimer’s Disease radic Alzheimer’s disease (SAD), Aβ circRNA
and Aβ circRNA-DP play a vital role and found
One of the most common dementia is Alzheimer’s within the entire human population [50].
disease (AD) which involves rigorous loss of Naturally immune to the exonucleolytically
memory, behavior capacities, and thinking with RNA degradation, the circRNAs are abundantly
progression in senescence. It is an age-associated distributed in the brain tissues of mammals.
neurodegenerative disorder. The neuropathologi- Some noncoding RNAs which are evolutionary
cal characterization is marked with the presence conserved such as miRNA-7 are not just over-
of neurofibrillary tangles and neurotic plaques, populated in human brain but are associated with
while clinical characteristics involve the gradual circRNA (ciRS-7), within the same tissues.
Impairment in cognitive skills. As far as the ciRS-7 comprises of multiple, tandem anti-
recent statistical records are concerned, 476,000 miRNA-7 sequences [51]. ciRS-7 behaves as a
people within the age group of 65 and above are competing, endogenous, anticomplementary
152 R. Ojha et al.

miRNA “sponge” which plays a role in adsorp- numbers. Furthermore, synaptic fractions include
tion and thereby quenches normal miRNA-7 synaptosomes and neuropil [54]. According to
functions. When a deficiency in the ciRS-7 and the histological findings, it has been confirmed
ciRS-7 “sponge activities” takes place, it can be that the synaptic reorganization of the limbic
expected to elevate ambient miRNA-7 levels in structures gradually enhances the hyperexcitabil-
the brain cells of patients affected with AD which ity of the cortices that may have contributions to
finally results in the downregulation of selective the onset of epilepsy. Recent studies have por-
miRNA-7-sensitive messenger RNA (mRNA) trayed the significant role of circRNAs in tempo-
goals [52]. ral lobe sclerosis (TLS), which is a pathological
entity that comes along with temporal lobe epi-
lepsy (TLE). TLS is a type of an abnormal corti-
7.2 Epilepsy cal findings associated with epilepsy which
serves to study the alterations within the temporal
At an outlook, the epilepsy comprises of a series neocortex genes in victims suffering from TLE
of brain malfunctioning which ranges from one patients. After a rigorous study, it was found that
of those that are much benign to the ones which a significant dysregulation was observed among
are life-challenging, severe, and disabling. some of the below-mentioned circRNAs – circ-­
Epilepsy is well known for disturbing the normal EFCAB2, circ-STK24, and circ-VPS37C were
neuronal activity pattern that ultimately results in observed to get notably upregulated within the
causing weird sensations, strange emotions and patients suffering from TLE as in comparison to
behavior or even muscle spasms, convulsions, those with the non-sufferers. On the other hand,
and loss of consciousness. The seizures, which circ-DROSHA, circ-CCT4, circ-UBQLN1, circ-­
are the most consistent emblem for the disease, USP9Y, and circ-STK17A were all downregu-
result due to brain damage, abnormal brain devel- lated. The most significant dysregulation was
opment, incorrect brain wiring, or a dispropor- observed in circ-EFCAB2 and circ-­
tionate nerve signaling chemicals widely known DROSHA. Furthermore, the top five miRNAs of
as neurotransmitters (https://www.ncbi.nlm.nih. circ-EFCAB2 were predicted to be miR-3929,
gov/pubmedhealth/PMHT0023036/). Circular miR-6780a-5p, miR-6884-5p, miR-4739, and
RNAs are a category of long noncoding RNAs miR5787 while for circ-DROSHA were miR-­
which have a closed-loop-like structure that par- 651-­3p, miR-1252-5p, miR-548ab, miR-­
ticipate in regulation of gene expression as 4762-­ 3p, and miR-4446-5p. The interaction
microRNA sponges and are highly stable in between these circRNAs and miRNAs resulted in
nature [53]. The circRNAs are the resultant of the formation “miRNA sponges” [55].
loop splicing mechanism, are critical regulatory
RNA molecules, and are novel. The communica-
tion of circRNAs with microRNAs (miRNAs) 7.3 Moyamoya Disease
behaves as miRNA sponge for the regulation of
gene expression through circRNAs-miRNA-­ The moyamoya disease (MMD) is an infrequent
messenger RNA (mRNA) axes in case of neuro- cerebrovascular disease that is characterized by
logical disorders. Again, circRNAs are well gradual occlusion of the internal intracranial
known for being evolutionary conserved and for carotid artery accompanied with the formation of
their expression in time, in cell type, and also in collateral vessels abnormally at the bottom area
gene-specific manners. Initially, it was observed of the brain called the basal ganglia [56]. The
that most circRNAs are normally expressed at name “moyamoya” means “puff of smoke” in
very small amounts, but further detailed investi- Japanese language, and it demonstrates the look
gation gradually reported that some circRNAs of tiny vessels tangled together and formed in
show tissue type-specific and cell type-specific order to compensate for the blockage caused
expression with elevated transcript and copy (https://www.ninds.nih.gov/Disorders/All-
12 Emerging Role of Circular RNAs as Potential Biomarkers for the Diagnosis of Human Diseases 153

Disorders/Moyamoya-Disease-Information- ers are men within the age group of 50s to early
Page). The disease was first identified in Japan in 60s [59]. It has been indicated in few researchers
the 1960s, and then it has been detected in several that a prion form of alpha-synuclein protein prob-
other countries across the globe. Most of the vic- ably causes this disorder [60]. MSA, on a molec-
tims are notably found in Asian countries in com- ular level, is classified as a distinct member from
parison to North America or Europe. The disease the clan of neurodegenerative disorders named as
primarily manifests children but can also make α-synucleinopathies. Α-synuclein is a fibrillary
the adults suffer transient ischemic attack, sei- protein that gets fragmented into oligodendro-
zures, hemorrhage, and ischemic stroke. In recent cytes and provides insulation and support to the
studies, it has been revealed that regulatory RNAs neuronal cells in the brain [61]. Often there’s a
such as long noncoding RNAs or microRNA tendency to get confused with the symptoms of
(miRNA) have strong contributions toward the idiopathic Parkinson’s disease and Parkinsonism-­
development of MMD. CircRNAs have more predominant MSA. Thus, it’s essential to identify
number of binding sites and higher levels of potential biomarkers which would facilitate an
expression when compared to other kinds of early and differential diagnosis of multiple sys-
miRNA sponges and are hence considered much tem atrophy. CircRNAs are the resultants of
more efficient in terms of sequestering miRNAs back-splicing which is the reciprocal of the
and regulating gene expression than linear RNAs canonical splicing of linear RNAs. CircRNAs are
[57]. CircRNAs have been linked up with various immune to hydrolysis by RNA exonucleases as
disorders that involve various neurological dis- they are covalently closed-loop structures.
ease, and they are correlated with the expression Furthermore, circRNAs have a half-life time of
of the miRNAs [58]. It has been found for the 48 h, which is very less in case of linear RNAs
first time that 146 circRNAs have been expressed [62]. Analysis of transcriptome data, which was
in victims suffering from MMD, and these likely retrieved from the frontal cortex of the multi-­
have enough contribution toward the develop- system atrophy brains, identified a significant
ment of this disease. Among these 146 circRNAs, upregulation of five circular RNAs within the
the level of expression varied; 29 of these were MSA tissue – hsa_circ_0038374, hsa_
upregulated, whereas 117 of them were down- circ_0005540, hsa_circ_0032891, hsa_
regulated. Hsa_circRNA_067130, hsa_cir- circ_0108308, and hsa_circ_00054211.
cRNA_067209, and hsa_circRNA_062557 were Henceforth, the perceived changes in circRNAs
upregulated, while hsa_circRNA_089763, hsa_ level can be attributed to the premature
circRNA_089761, and hsa_circRNA_100914 ­oligodendritic malfunction which is due to the
were downregulated with highest fold variations, α-synuclein aggregation [63].
which gave enough evidence to state that these From a bird’s eye view, it can be concluded
circRNAs are potential biomarkers for the diag- that circRNAs are diverse endogenous noncoding
nosis of moyamoya disease [56]. RNAs that occur naturally and regulate gene
expressions. CircRNAs are extremely stable mol-
ecules and are densely populated within the brain
7.4  ultiple System Atrophy
M and exosomes. The abundance of circRNAs tran-
in the Brain scripts in peripheral blood and brain, in compari-
son to other tissues, has registered these molecules
Multiple system atrophy is an infrequent spo- as an attractive proxy for the successful detection
radic neurodegenerative disease which comes and monitoring of neuronal diseases. They are
along with symptoms of slow movement, trem- considered as an ideal biomarker because they
ors, postural instability, and muscle rigidity due can successfully transverse the blood-brain bar-
to malfunctioning of the ataxia and autonomic rier, thereby facilitating the studies related to
nervous system. The onset of this disease takes neurological disorders [64]. In particular
place in adulthood. Around 55% of MSA suffer-
154 R. Ojha et al.

circRNAs-­based monitoring is suitable option for portrayed significant dysregulation in pregnant

the detection of neurodegenerative diseases. women affected with PE in comparison with nor-
mal pregnant mothers. Hsa_circRNA_101222,
hsa_circRNA_101151, hsa_circRNA_104018,
8  ircular RNAs Role
C hsa_circRNA_101862, hsa_circRNA_100385,
in Diagnosis of Preeclampsia hsa_circRNA_105016, hsa_circRNA_103842, and
Condition hsa_circRNA_001535 were upregulated; on the
other hand, hsa_circRNA_101160, hsa_cir-
8.1 Preeclampsia cRNA_400084, hsa_circRNA_101676, and hsa_
circRNA_101218 were downregulated. Hence, it
Preeclampsia (PE) is a pregnancy disorder which can be used as an early prognosis of pregnant
is characterized with high blood pressure and fre- women who are at the verge of developing PE.
quent presence of protein in the urine of the
mother. The disease usually begins manifesting
after 5 months of pregnancy. It has been reported 9 Conclusion
as one of the major causes of fetal and maternal
morbidity and mortality. Approximately 10% of In a nutshell, here we illuminated the role of
the pregnant women are victimized with this in circular RNAs associated with different age-old
developing countries [65, 66]. The route of patho- and recently discovered disorders those are vic-
genesis for this hasn’t been unwrapped, yet an toriously manifesting human life. The areas
early determination is necessary. A combination covered in this chapter include infectious dis-
of biomarkers is required for a prior diagnosis as eases, oncogenic diseases, and neurological
because PE is accompanied with heterogeneous disorders. Firstly, in case of pulmonary tuber-
manifestations. According to the evidences accu- culosis, the expression of several circRNAs was
mulated so far, it has been demonstrated that non- imbalanced. It was studied that a total of 37
coding RNA have an association with circRNAs were aberrantly expressed in dis-
preeclampsia [67]. CircRNAs are a type of non- eased individuals, where 13 were upregulated
coding RNA which have the potential to regulate and 24 were downregulated. Infectious disease
other RNA transcripts by eventually competing further leads to the development of cancer in
for shared miRNAs. This phenomenon has drawn several cases; one such is hepatitis, widely
attention toward the fact that they can act as known as hepatocellular carcinoma which
remarkable diagnostic markers. CircRNAs are a develops due to genetic/epigenetic aberrations
unique category of RNA molecules that can be and alterations in many signaling pathways. A
categorized as intronic, intergenic, or exonic. total of 24 circRNAs were overexpressed, while
CircRNAs do not possess free 3′ or 5′ ends and 23 circRNAs were observed to undergo a lower
can hence bypass the action of RNA exonuclease, expression in victims. Among them, circMTO1,
thus making it highly stable in nature. It has been hsa_circ_0001649, circZKSCAN1, and hsa_
found that circRNAs were enriched with miRNA circ_0004018 were found to get downregulated,
binding sites, thereby behaving as a miRNA while three circRNAs hsa_circ_0005075, ciRS-
sponge within cells, relieving the inhibition of 7(Cdr1as), and circRNA_100338 were upregu-
miRNA target genes, and henceforth elevating lated. Cancer which is characterized by the
the level of expression of the targeted gene [12]. uncontrolled growth of cells can also be
The phenomenon is termed as competitive endog- detected with the help of circular RNAs. In
enous RNA mechanism. Keeping this context tumorigenic cancerous conditions such as lung
into consideration, it was studied that circRNAs cancer, gastric cancer, breast cancer, and
act as an important element in preeclampsia caused colorectal cancer, circular RNA has portrayed
during pregnancy [68]. After rigorous studies, it itself as a remarkable diagnostic marker. In the
was discovered that 12 circRNAs in blood cells colorectal cancer, 39 circRNAs were signifi-
12 Emerging Role of Circular RNAs as Potential Biomarkers for the Diagnosis of Human Diseases 155

cantly expressed as compared to the normal being most stable in nature are used as biomark-
adjacent tissues, in which, circCCDC66, circ- ers, and their dysregulation can significantly
BANP, ciRS-7 (Cdr1as), hsa_circ_0000069, help in the detection of the disease. Aβ circRNA
hsa_circ_001569, and hsa_circ_0020397 and ciRS-7 are the two circRNAs which get
ranked top in the list of the upregulated ones, upregulated and downregulated, respectively,
while hsa_circRNA_103809, hsa_cir- symbolizing the onset of Alzheimer’s disease.
cRNA_10470, hsa_circ_001988, and cir- Epilepsy is another disorder accompanied with
cRNA0003906 were occupying the initial weird sensation and strange emotions. Circ-
position in the list of downregulation. Lung EFCAB2 along with few others gets upregu-
cancers are the most common type of cancer lated, while circ-DROSHA accompanied with
worldwide. A total of ten circRNAs were imbal- others is observed to get downregulated which
anced, out of which five were upregulated and behaves like a biomarker for disease detection. A
five were downregulated. The upregulated cir- total of 146 circRNAs are imbalanced during
cRNAs were hsa_circ_0000735 and hsa_ moyamoya disease, caused due to a blockage in
circ_0016760 with few others in the list, while the carotid artery, for example, Hsa_cir-
the downregulated ones were hsa_circ_0005730, cRNA_067130 is upregulated, while hsa_cir-
hsa_circ_0091000, hsa_circ_0014130, etc. cRNA_089763 is downregulated, indicating the
CircRNA has again proved its potential in prevalence of this disease. A rare sporadic neu-
determining gastric cancer which is one of the rodegenerative disease is a multiple system atro-
root causes of death globally. Hsa_circ_0000520 phy which has an overexpression of five
was found to get notably downregulated in significant circRNAs that depict that a person
patients suffering from gastric cancer espe- showing such upregulation of circRNA, for
cially during the TNM stage. Another fatal example, hsa_circ_0038374, hsa_circ_0108308,
malignant cancer which is prevalent among and so on, is victimized with MSD. Preeclampsia
women is breast cancer. It was found that a total is an emerging disorder reportedly manifesting
of 19 circRNAs (e.g., hsa_circ_0001785, etc.) pregnant mothers. Circular RNAs have played a
were overexpressed, whereas 22 (e.g., hsa_ vital role for the detection of this disease. In
circ_006054, etc.) of them had a lower recent researches, it was found that 12 circRNAs,
expression. namely, hsa_circRNA_101222, hsa_cir-
Lastly, one of the most debilitating and cRNA_105016, and few others, had significant
untreatable condition which occurs due to degen- upregulation, strengthening the fact that the level
eration of the neurons is labeled as neurodegen- of circRNAs can be used as both prognostic and
erative disorders. The condition prevails due to diagnostic measures.
the gradual degeneration or eventual death of the In each of this, it has been closely scrutinized
nerve cells. This ultimately results into difficul- that circRNAs are playing a significant role as a
ties associated with movement, memory, sensa- diagnostic measure to identify the disorder and
tions, tremors, postural defects, seizure, and can also serve as a prognostic measure for the
other mental dysfunction (known as dementias). early determination of the disease. With all the
There are numerous reported neuronal disorders, evidences accumulated, it can be claimed that
among which few of them are covered here – circRNAs are functioning as potential biomark-
Alzheimer’s, epilepsy, multiple system atrophy, ers for all the different fields of diseases that
and moyamoya. Alzheimer is a well-known dis- manifest human life.
ease that is associated with rigorous memory
loss with age. A huge percentage of the popula- Acknowledgments RO is thankful to the Central
tion worldwide is suffering from this disease, University of Rajasthan for providing UGC fellowship.
and no such remarkable diagnostic or prognostic
Disclosure Statement No potential conflict of interest
approaches have been yet developed. This is was reported by authors.
where circRNAs come into action. CircRNAs
156 R. Ojha et al.

References 19. Du WW, Fang L, Yang W et al (2017) Induction of

tumor apoptosis through a circular RNA enhancing
Foxo3 activity. Cell Death Differ 24(2):357
1. Zhi-Chun Z, Xiao-Long G, Li X (2017) The novel
20. Du WW, Yang W, Liu E et al (2016) Foxo3 circular
roles of circular RNAs in metabolic organs. Genes Dis
RNA retards cell cycle progression via forming ter-
5(1):16–23
nary complexes with p21 and CDK2. Nucleic Acids
2. Chen Y, Li C, Tan C et al (2016) Circular RNAs: a
Res 44(6):2846–2858
new frontier in the study of human diseases. J Med
21. Bachmayr-Heyda A, Reiner AT, Auer K et al (2015)
Genet 53(6):359–365
Correlation of circular RNA abundance with prolifer-
3. Memczak S, Papavasileiou P, Peters O et al (2015)
ation–exemplified with colorectal and ovarian cancer,
Identification and characterization of circular RNAs
idiopathic lung fibrosis, and normal human tissues.
as a new class of putative biomarkers in human blood.
Sci Rep 5:8057
PLoS One 10(10):e0141214
22. Qian Z, Liu H, Li M et al (2018) Potential diagnostic
4. Carriero S, Damha MJ (2003) Template-mediated
power of blood circular RNA expression in active pul-
synthesis of lariat RNA and DNA. J Org Chem
monary tuberculosis. EBioMedicine 27:18–26
68(22):8328–8338
23. Huang Z-K, Yao F-Y, Xu J-Q et al (2018) Microarray
5. Li J, Yang J, Zhou P et al (2015) Circular RNAs in
expression profile of circular RNAs in peripheral
cancer: novel insights into origins, properties, func-
blood mononuclear cells from active tuberculosis
tions and implications. Am J Cancer Res 5(2):472
patients. Cell Physiol Biochem 45(3):1230–1240
6. Lasda E, Parker R (2016) Circular RNAs co-­precipitate 24. Zhu RX, Seto W-K, Lai C-L et al (2016) Epidemiology
with extracellular vesicles: a possible mechanism for of hepatocellular carcinoma in the Asia-Pacific region.
circRNA clearance. PLoS One 11(2):e0148407 Gut Liver 10(3):332–339
7. Jeck WR, Sharpless NE (2014) Detecting and charac- 25. Beasley RP (1988) Hepatitis B virus. The major
terizing circular RNAs. Nat Biotechnol 32(5):453 etiology of hepatocellular carcinoma. Cancer
8. Li WH, Song YC, Zhang H (2017) Decreased expres- 61(10):1942–1956
sion of Hsa_circ_00001649 in gastric cancer and its 26. Custer B, Sullivan SD, Hazlet TK et al (2004) Global
clinical significance. Dis Markers 2017:4587698 epidemiology of hepatitis B virus. J Clin Gastroenterol
9. Ebert MS, Sharp PA (2010) MicroRNA sponges: 38(10):S158–S168
progress and possibilities. RNA 16(11):2043–2050 27. Yu L, Gong X, Sun L et al (2016) The circular RNA
10. Rybak-Wolf A, Stottmeister C, Glažar P et al (2015) Cdr1as act as an oncogene in hepatocellular carci-
Circular RNAs in the mammalian brain are highly noma through targeting miR-7 expression. PLoS One
abundant, conserved, and dynamically expressed. 11(7):e0158347
Mol Cell 58(5):870–885 28. Fu L, Yao T, Chen Q et al (2017) Screening dif-
11. Barrett SP, Salzman J (2016) Circular RNAs: analy- ferential circular RNA expression profiles reveals
sis, expression and potential functions. Development hsa_circ_0004018 is associated with hepatocellular
143(11):1838–1847 carcinoma. Oncotarget 8(35):58405
12. Memczak S, Jens M, Elefsinioti A et al (2013) 29. McKillop IH, Moran DM, Jin X et al (2006) Molecular
Circular RNAs are a large class of animal RNAs with pathogenesis of hepatocellular carcinoma. J Surg Res
regulatory potency. Nature 495(7441):333 136(1):125–135
13. Jost I, Shalamova LA, Gerresheim GK et al (2018) 30. Wu K, House L, Liu W et al (2012) Personalized
Functional sequestration of microRNA-122 from targeted therapy for lung cancer. Int J Mol Sci
Hepatitis C Virus by circular RNA sponges. RNA Biol 13(9):11471–11496
28:1–8 31. Ma PC (2012) Personalized targeted therapy in
14. Panda AC, De S, Grammatikakis I et al (2017) High-­ advanced non-small cell lung cancer. Cleve Clin
purity circular RNA isolation method (RPAD) reveals J Med 79:eS56–eS60
vast collection of intronic circRNAs. Nucleic Acids 32. Zhang S, Zeng X, Ding T et al (2018) Microarray pro-
Res 45(12):e116–e116 file of circular RNAs identifies hsa_circ_0014130 as
15. Voelkerding KV, Dames SA, Durtschi JD (2009) a new circular RNA biomarker in non-small cell lung
Next-generation sequencing: from basic research to cancer. Sci Rep 8(1):2878
diagnostics. Clin Chem 55(4):641–658 33. Wan L, Zhang L, Fan K et al (2016) Circular RNA-­
16. Chen L, Yu Y, Zhang X et al (2016) PcircRNA_ ITCH suppresses lung cancer proliferation via inhib-
finder: a software for circRNA prediction in plants. iting the Wnt/β-catenin pathway. Biomed Res Int
Bioinformatics 32(22):3528–3529 2016(1):1579490
17. World Health Organization (2016) Global tuberculo- 34. Yao J-T, Zhao S-H, Liu Q-P et al (2017) Over-­
sis report 2016. World Health Organization, Geneva expression of CircRNA_100876 in non-small cell
18. Lü L, Sun J, Shi P et al (2017) Identification of cir- lung cancer and its prognostic value. Pathol Res Pract
cular RNAs as a promising new class of diagnostic 213(5):453–456
biomarkers for human breast cancer. Oncotarget 35. Luo Y-H, Zhu X-Z, Huang K-W et al (2017) Emerging
8(27):44096 roles of circular RNA hsa_circ_0000064 in the
12 Emerging Role of Circular RNAs as Potential Biomarkers for the Diagnosis of Human Diseases 157

proliferation and metastasis of lung cancer. Biomed 52. Cogswell JP, Ward J, Taylor IA et al (2008) Identification
Pharmacother 96:892–898 of miRNA changes in Alzheimer’s disease brain and
36. Zhu X, Wang X, Wei S et al (2017) hsa_circ_0013958: CSF yields putative biomarkers and insights into dis-
a circular RNA and potential novel biomarker for lung ease pathways. J Alzheimers Dis 14(1):27–41
adenocarcinoma. FEBS J 284(14):2170–2182 53. Chen L-L, Yang L (2015) Regulation of circRNA bio-
37. Zhang Y, Zhao H, Zhang L (2018) Identification of genesis. RNA Biol 12(4):381–388
the tumor-suppressive function of circular RNA 54. You X, Vlatkovic I, Babic A et al (2015) Neural
FOXO3 in non-small cell lung cancer through spong- circular RNAs are derived from synaptic genes and
ing miR-155. Mol Med Rep 17(6):7692–7700 regulated by development and plasticity. Nat Neurosci
38. Ji W, Qiu C, Wang M et al (2018) Hsa_circ_0001649: 18(4):603
a circular RNA and potential novel biomarker for 55. Li J, Lin H, Sun Z et al (2018) High-throughput data
colorectal cancer. Biochem Biophys Res Commun of circular RNA profiles in human temporal cortex tis-
497(1):122–126 sue reveals novel insights into temporal lobe epilepsy.
39. Yin WB, Yan MG, Fang X, et al (2017) Circulating Cell Physiol Biochem 45(2):677–691
circular RNA hsa_circ_0001785 acts as a diagnos- 56. Zhao M, Gao F, Zhang D et al (2017) Altered expres-
tic biomarker for breast cancer detection. Clin Chim sion of circular RNAs in Moyamoya disease. J Neurol
Acta. https://doi.org/10.1016/j.cca.2017.10.011 Sci 381:25–31
40. Cocquerelle C, Mascrez B, Hetuin D et al (1993) 57. Wilusz JE, Sharp PA (2013) A circuitous route to non-
Mis-splicing yields circular RNA molecules. FASEB coding RNA. Science 340(6131):440–441
J 7(1):155–160 58. Lin S-P, Ye S, Long Y et al (2016) Circular RNA
41. Siegel RL, Sahar L, Portier KM et al (2015) Cancer expression alterations are involved in OGD/R-induced
death rates in US congressional districts. CA Cancer neuron injury. Biochem Biophys Res Commun
J Clin 65(5):339–344 471(1):52–56
42. Boleij A, van Gelder MM, Swinkels DW et al (2011) 59. Bensimon G, Ludolph A, Agid Y et al (2008)
Clinical importance of Streptococcus gallolyticus infec- Riluzole treatment, survival and diagnostic criteria
tion among colorectal cancer patients: systematic review in Parkinson plus disorders: the NNIPPS study. Brain
and meta-analysis. Clin Infect Dis 53(9):870–878 132(1):156–171
43. Xie H, Ren X, Xin S et al (2016) Emerging roles of 60. Prusiner SB, Woerman AL, Mordes DA et al (2015)
circRNA_001569 targeting miR-145 in the prolif- Evidence for α-synuclein prions causing multiple sys-
eration and invasion of colorectal cancer. Oncotarget tem atrophy in humans with parkinsonism. Proc Natl
7(18):26680 Acad Sci U S A 112(38):E5308–E5317
44. Weng W, Wei Q, Toden S et al (2017) Circular RNA 61. McCann H, Stevens CH, Cartwright H et al (2014)
ciRS-7—a promising prognostic biomarker and a α-Synucleinopathy phenotypes. Parkinsonism Relat
potential therapeutic target in colorectal cancer. Clin Disord 20:S62–S67
Cancer Res 23(14):3918–3928 62. Enuka Y, Lauriola M, Feldman ME et al (2015)
45. Wang X, Zhang Y, Huang L et al (2015) Decreased Circular RNAs are long-lived and display only mini-
expression of hsa_circ_001988 in colorectal cancer mal early alterations in response to a growth factor.
and its clinical significances. Int J Clin Exp Pathol Nucleic Acids Res 44(3):1370–1383
8(12):16020 63. Chen BJ, Mills JD, Takenaka K et al (2016)
46. Zhang P, Zuo Z, Shang W et al (2017) Characterization of circular RNAs landscape
Identification of differentially expressed circu- in multiple system atrophy brain. J Neurochem
lar RNAs in human colorectal cancer. Tumor Biol 139(3):485–496
39(3):1010428317694546 64. Li Y, Zheng Q, Bao C et al (2015) Circular RNA is
47. Zhang Y, Zhang Y, Li X et al (2017) Microarray analy- enriched and stable in exosomes: a promising bio-
sis of circular RNA expression patterns in polarized marker for cancer diagnosis. Cell Res 25(8):981
macrophages. Int J Mol Med 39(2):373–379 65. Selvaggi L, Loverro G, Schena F et al (1988) Long
48. Huang G, Zhu H, Shi Y et al (2015) cir-ITCH plays an term follow-up of women with hypertension in preg-
inhibitory role in colorectal cancer by regulating the nancy. Int J Gynaecol Obstet 27(1):45–49
Wnt/β-catenin pathway. PLoS One 10(6):e0131225 66. Díaz LM, Díaz MPN, Serrano ND et al (2011) The
49. Huang J-L, Qin M-C, Zhou Y et al (2018) prognosis for children of mothers with preeclamp-
Comprehensive analysis of differentially expressed sia. Part 2: long-term effects. Arch Argent Pediatr
profiles of Alzheimer’s disease associated circular 109(6):519–524
RNAs in an Alzheimer’s disease mouse model. Aging 67. Ji L, Brkić J, Liu M et al (2013) Placental tropho-
(Albany NY) 10(2):253 blast cell differentiation: physiological regulation and
50. Mo D (2018) The role of Aβ circRNA in Alzheimer’s pathological relevance to preeclampsia. Mol Asp Med
disease. bioRxiv:260968 34(5):981–1023
51. Burmistrova O, Goltsov A, Abramova L et al (2007) 68. Burd CE, Jeck WR, Liu Y et al (2010) Expression
MicroRNA in schizophrenia: genetic and expres- of linear and novel circular forms of an INK4/ARF-­
sion analysis of miR-130b (22q11). Biochem Mosc associated non-coding RNA correlates with athero-
72(5):578–582 sclerosis risk. PLoS Genet 6(12):e1001233
 Circular RNAs as Novel Biomarkers
for Cardiovascular Diseases 13
Qiulian Zhou, Zhongrong Zhang, Yihua Bei,
Guoping Li, and Tianhui Wang

Abstract digestive diseases. Here we summarize recent

Cardiovascular diseases are among the most studies on circRNAs and compare the charac-
serious diseases, which are a leading cause of teristics of circRNAs with traditional bio-
death across the world. Early diagnosis and markers. Finally, we highlight the value of
prognosis prediction are keys for treatment circRNAs as potential biomarkers for cardio-
and reduction of death rates. Circular RNAs vascular diseases, including acute myocardial
(circRNAs) play a critical role in the physiol- infarction, heart failure, coronary artery dis-
ogy and pathology of biological system and ease, and hypertension. In conclusion, cir-
participate in the development of diseases. In cRNAs may be promising biomarkers for
addition, circRNAs are relative stable and cardiovascular diseases.
abundant. Therefore, many studies have sug-
gested that circRNAs could be used as bio- Keywords
markers for diseases, such as neurological Circular RNA · Biomarkers · Cardiovascular
diseases, cancers, immune diseases, and diseases

Q. Zhou Z. Zhang · Y. Bei · T. Wang (*)

Shanghai Applied Radiation Institute, School of Cardiac Regeneration and Ageing Lab, Institute of
Environmental and Chemical Engineering, Shanghai Cardiovascular Sciences, School of Life Science,
University, Shanghai, China Shanghai University, Shanghai, China
Cardiac Regeneration and Ageing Lab, Institute of Shanghai Key Laboratory of Bio-Energy Crops,
Cardiovascular Sciences, School of Life Science, School of Life Sciences, Shanghai University,
Shanghai University, Shanghai, China Shanghai, China
e-mail: [email protected]
Shanghai Key Laboratory of Bio-Energy Crops,
School of Life Sciences, Shanghai University, G. Li
Shanghai, China Cardiovascular Division of the Massachusetts
General Hospital, Harvard Medical School,
Boston, MA, USA

© Springer Nature Singapore Pte Ltd. 2018 159

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_13
160 Q. Zhou et al.

1 Introduction of Circular RNA consist of intronic sequences. Exon-intron circu-

lar RNAs consist of both exonic and intronic
Circular RNAs (circRNAs) are currently a sequences. They predominantly localize in the
research focus in the field of noncoding RNAs, nucleus and have association with the RNA poly-
including microRNAs (miRNAs), long noncod- merase II. Therefore, exon-intron circular RNAs
ing RNAs (lncRNAs), etc. [1]. Noncoding RNAs are regarded as transcriptional regulators. Three
are involved in the development of many diseases types of circRNAs share the same characteristics,
[2], including cardiovascular diseases. They are such as stable, abundant, and conservative.
recognized as regulators of gene expression.
However, in the past, circRNAs were ignored and
considered as transcriptional noise [3] or errant 3 Characteristics of circRNAs
by-products of splicing [4], owing to their unique
structure and low abundance [4]. Based on Desirable biomarkers are required to be stable
advances in bioinformatics and biotechnology, a against harsh conditions, for example, tempera-
large number of abundant and significant cir- ture variation, extreme pH, and thawing cycles
cRNAs were discovered and identified. The evi- [11, 12]. Biomarkers should have high stability in
dence is mounting that circRNAs play essential blood and other body fluids. The biggest natural
roles in various physiological and pathological enemy of RNA is RNA exonucleases (or RNase
pathways. R). CircRNAs are closed circular molecules con-
sist of covalently closed loop structure and there-
fore resistant to RNA exonucleases that solely
2 Biogenesis of circRNAs digests linear transcripts [13–17]. CircRNAs are
more stable compared to linear RNAs [18].
CircRNAs are closed circular molecules, which Furthermore, based on their stability in blood and
are distinguished from other linear RNA mole- other body fluids [19, 20], circRNAs are suitable
cules. First circRNA was found by electron biomarkers for disease diagnosis.
microscopy from RNA viruses [5]. Nowadays, Clinical trials must be properly based upon
more and more circRNAs are being identified by reliable earlier trials and sufficient preclinical
RNA sequencing and bioinformatics. Due to the studies. CircRNAs are conserved across mam-
closed loop characteristic of circRNAs, they can mals [21], including human, mouse, and rat [22,
be extracted from the ribosomal RNA and linear 23]. In hearts, RNA-Seq analysis found that
RNA molecules in total RNA. RNA sequencing plenty of circRNAs could be detected in mammal
can obtain more circRNAs from specific accumu- species: human (16427), mouse (9953), and rat
lation of circRNAs samples. The covalently (13086). Among them, about 30% (3171) were
closed loop structure of circRNAs, also called conserved between mouse and rat, and about
“back-splicing,” is formed by joining of 3′ splice 10% (1288) were conserved in all three species
site to 5′ splice site [6, 7]. There are three types of [24]. These circRNAs conserved in different spe-
circRNAs: exonic circular RNAs (ecircRNAs) cies become abundant resources of promising
[8], intronic circular RNAs (ciRNAs) [9], and candidates for biomarkers. In addition to the ben-
exon-intron circular RNAs (EIciRNAs) [10]. efits mentioned above, circRNAs have been dem-
Exonic circular RNAs have three formation onstrating specificity in expression [25].
models: lariat-driven circularization, intron Emerging evidence have confirmed thousands of
pairing-­driven circularization, and resplicing-­ the tissue-specific expression of circRNAs [26],
driven circularization. They form from exon such as the heart, brain, skin, prostate, thyroid,
back-splicing and only consist of exonic blood, or blood cell. Through biclustering algo-
sequences. The formation of circular intronic rithm, studies have already reported thousands of
RNAs mainly depends on consensus RNA motifs tissue-specific expressed circRNAs in human
and branchpoint. Circular intronic RNAs only [26]. Except for tissue-specific expression,
13 Circular RNAs as Novel Biomarkers for Cardiovascular Diseases 161

c­ircRNAs have specific expression in different as RNA-binding protein (RBP) sponges [33, 34].
diseases. In cardiac development and diseases, Thirdly, circRNAs can be a regulator of tran-
the expression of circRNAs is regulated [27, 28]. scription and alternative splicing that directly
Through whole transcriptome sequencing, 826 influence gene expression [35].
putative circRNAs have been identified in human For all those reasons and more, circRNAs
heart with dilated cardiomyopathy patient and have been reported to play important roles in
hypertrophic cardiomyopathy [27]. Through physiological and pathological processes [36–
microarray expression profile assay, 63 circRNAs 38], such as proliferation, apoptosis, and autoph-
were differentially expressed in mouse between agy. CircRNA circHIPK3 regulates the viability
heart failure and sham group [28]. and proliferation of endothelial cell viability by
In particular, circRNAs (3841) are found inhibiting the activity of miR-30a-3p [36].
unique in different composition and cell types of CircRNA TTBK2 regulates the cell proliferation
blood [14], such as plasma (51), serum (37), neu- and apoptosis of human glioma cells by regulat-
trophils (58), and platelets (2339). Circulating ing miR-217/HNF1β/Derlin-1 pathway [37].
blood markers have enormous potential to be In addition, circRNAs can play regulatory
applied to clinical practice. Human peripheral function in diseases, such as neurological dis-
whole blood is an easily accessible body fluid. eases [39, 40], cancers [41], immune diseases
Thousands of circRNAs have been detected in [42], aging [43], diabetes [44], digestive diseases
blood. Except for the blood, distinct circRNAs [45], reproductive system diseases, skin, muscles
have also been founded in other body fluids, such and bones diseases, and especially cardiovascular
as saliva [19]. On the other hand, circRNAs have diseases [46, 47]. CircRNA HRCR can function
been found enriched in exosomes [29, 30]. as an endogenous miR-223 sponge through
Exosomes are currently regarded as specific upregulation of ARC, which attenuate hypertro-
secretory membrane vesicles and are involved in phic responses [48]. CircRNA Cdr1as can inhibit
intercellular communication. Exosomes are also miR-7a through upregulating PARP and SP1,
thought to be the way of transferring miRNA in which induces myocardial apoptosis and aggra-
the circulatory system, which indirectly reflects vates myocardial infarction injuries [49].
the changes of miRNAs in diseases. Serum exo-­ Furthermore, circRNAs constitute promising
circRNAs may be involved in intercellular com- new biomarkers in multiple diseases [25, 50].
munication and can be used as indicators of CircRNA hsa_circ_0014130 is significantly
diseases development. Therefore, exo-circRNAs upregulated in non-small cell lung cancer
in the blood of patients are expected to be mark- (NSCLC) tissues that may be a new biomarker of
ers for early diagnosis. According to the charac- NSCLC [51]. CircRNA hsa_circ_0004277 is
teristics talked above, circRNAs are promising downregulated in mononuclear cells from bone
new biomarkers. marrow of acute myeloid leukemia (AML)
patients [52]. CircRNA hsa_circ_0001649 is
downregulated in hepatocellular carcinoma
4 Functions of circRNAs (HCC) tissues and is correlated with tumor size
and tumor embolus’s occurrence [53].
As most other noncoding RNAs, most of the cir- In addition to the above examples, circRNAs
cRNAs do not encode proteins. But circRNAs are have the potential to become biomarkers in many
involved in regulating the expression of genes cardiovascular diseases. Here we will briefly intro-
[31]. Firstly, circRNAs can adsorb some specific duce the characteristics of circRNAs by comparison
miRNAs that prevent mRNA translation, func- with the existing biomarkers of cardiovascular dis-
tioning as miRNA sponges [25, 32]. Secondly, eases, which expound the value of circRNAs. Then
circRNAs can bind to RNA-associated proteins we will briefly summarize recent findings on cir-
and form RNA-protein complexes that function cRNAs as biomarkers in cardiovascular diseases.
162 Q. Zhou et al.

5 Biomarkers miR-133, miR-423-5p, miR-126) [66–68].

of Cardiovascular Diseases However, only a few of them have been validated,
and they are necessary to prove by large-scale
Cardiovascular diseases threaten human health clinical studies [69]. Traditional biomarkers are
and have become one of the leading diseases also limited by low stability [70]. Therefore, cir-
which causes death in the world [54]. Nowadays, cRNAs, which are stable, conservative, and spe-
more and more risk factors are aggravating the cific [71, 72], might open a novel avenue.
development and progression of cardiovascular
diseases. Early diagnosis and prognosis predic-
tion are keys for treatment and reducing mortal- 6  ircular RNAs as Biomarkers
C
ity. There are multiple diagnostic techniques for in Cardiovascular Disease
cardiovascular diseases, such as blood pressure
monitoring, dynamic electrocardiogram, cardio- 6.1 Acute Myocardial Infarction
vascular ultrasound, magnetic resonance imag-
ing, electronic computed tomography, and so on. Acute myocardial infarction is a myocardial
These diagnostic techniques require lots of necrosis caused by acute and persistent ischemia.
resources and precise instruments. What’s more, Acute myocardial infarction can be complicated
these diagnostics still cannot ensure the accurate with arrhythmia, shock, or heart failure, which is
diagnosis of the occurrence and development of often life-threatening. It tends to occur in patients
cardiovascular diseases. with coronary atherosclerosis stenosis and coro-
Some blood markers are commonly regarded nary spasm. Acute myocardial infarction carries a
as sensitive biomarkers for cardiovascular dis- high mortality rate and become a global chal-
eases, such as atrial natriuretic peptide (ANP), lenge [73, 74]. Reperfusion is the most common
creatine kinase (CK), myocardial band isoen- treatment strategy for myocardial infarction
zyme (CK-MB), and cardiac troponin (cTn) [55]. patient. This treatment strategy will cause myo-
Brain natriuretic peptide (BNP) and N-terminal cardial ischemia-reperfusion injury leading to
portion of its prohormone (NT-proBNP) are also left ventricular remodeling and even heart failure.
used as biomarkers for cardiovascular diseases There are some biomarkers for clinical assess-
[56]. But there are problems difficult to solve. ment of the severity, course, and prognosis of
Firstly, the detection time is uncertain, since acute myocardial infarction. Creatine kinase(CK)
markers keep changing over time and exhibit a and cardiac troponin(cTn) are markers for myo-
relative “delayed” release time [57]. Secondly, cardial injury [75–77]. In skeletal muscles and
disease degree and patient’s background are var- heart muscles, creatinine kinase facilitates the
ied, while these biomarkers are low in specificity transfer of energy phosphates between outside
[58]. Thirdly, there are still many kinds of cardio- and inside mitochondria. The increased concen-
vascular diseases without proper marker, such as tration of creatine kinase in serum indicates myo-
coronary artery disease. As a result, there remain cardial damage, such as myocardial infarction or
many challenges in the field of cardiovascular ischemia. But it can also be caused by vigorous
diseases diagnosis. exercise or other kinds of tissue damage [78].
There is growing evidence that noncoding Cardiac troponin I and T untimely are released
RNAs have the potential to become biomarkers into plasma when actin and myosin filaments
in many cardiovascular diseases. Circulating occur in pathologic degeneration in the heart.
miRNAs are suggested as suitable biomarkers for Thus, cardiac troponin can be a marker in myo-
cardiovascular diseases [59–61], including acute cardial infarction. But the abnormal release of
myocardial infarction (miR-1, miR-133, miR-­ cardiac troponin can also occur in myocarditis
208, and miR-499) [62, 63]; acute coronary syn- [75, 76], pulmonary embolism with acute right
drome (miR-208a, miR-34a, miR-133a, heart overload [79], and even heart failure [80].
miR-499) [64, 65]; and heart failure (miR-499, Therefore, new biomarkers are required.
13 Circular RNAs as Novel Biomarkers for Cardiovascular Diseases 163

Recently, a study observed that a circRNA myocardial structure and function. Heart failure
called MICRA (myocardial infarction-associated is a major cause of morbidity and mortality
circular RNA) is absent in peripheral blood cells, worldwide [91–93], even though its molecular
plasma, and serum samples [81, 82]. The particu- mechanisms become more and more clear and a
lar circRNA (MICRA) is formed by 874 nucleo- large amount of pharmacological interventions
tides, which is mainly derived from zinc finger come to light. In current guidelines, the plasma
protein 609 (ZNF609) gene exon 1 [82, 83]. levels of a natriuretic peptide (such as brain natri-
MICRA is associated with development of heart uretic peptide, BNP, and N-terminal pro brain
failure after myocardial infarction and can be natriuretic peptide, NT-proBNP) are important
used to predict left ventricular remodeling after diagnostic indicators of acute heart failure [94].
acute myocardial infarction by multivariable anal- Most of biomarkers for the assessment of heart
yses (0.83[0.69–1.01]; p = 0.066) [81]. Lower failure are widely acknowledged and used. But
level of MICRA classifies patients into decreased there is still limitation that circulating levels of
ejection fraction groups; and high level of MICRA BNP will significantly change with the develop-
is able to predict a preserved ejection fraction at ment of acute myocardial infarction [95].
4 months. When MICRA is included, a multivari- Therefore, these biomarkers cannot be used in
able clinical model, Akaike Information Criteria, the management of heart failure patients, who
becomes more accurate in prediction of preserved have a complicated disease history.
ejection fraction at 4 months. CircRNA MICRA An increasing number of RNA including miR-
is potential to become a novel biomarker in the NAs [19, 96] and lncRNAs [83, 97, 98] have been
future prognostication strategies of acute myocar- found with enormous potential as biomarkers for
dial infarction [81]. heart failure. LncRNA LIPCAR was found
In another study, circRNA MICRA was found downregulated early after myocardial infarction
lower in 642 myocardial infarction patients com- and meanwhile upregulated in plasma of 788
pared to 86 healthy volunteers [82]. Besides, heart failure patients [97]. LncRNA NRON and
MICRA was significantly lower in patients who LncRNA MHRT were found upregulated in
have a low ejection fraction. In particular, both plasma of 104 heart failure patients [99].
univariate and multivariate analyses proved that CircRNAs have great potential to act as bio-
circRNA MICRA may be a strong predictor for marker for myocardial infarction and heart fail-
prognoses of myocardial infarction after ure. By comparing the transcriptome sequencing
3–4 months. Collectively, circRNA MICRA is through bioinformatic analysis, 1163 circRNAs
suggested to be able to inform the risk of devel- were found changed in hypertrophied myocardial
oping left ventricular dysfunction for myocardial tissues grouping with or without heart failure
infarction patients. [28]. Through analyzing by quantitative PCR, 29
upregulated and 34 downregulated circRNAs
was validated [28]. Thus, these circRNAs may be
6.2 Heart Failure potential biomarkers for heart failure. Besides, a
circRNA MICRA was downregulated in
Heart failure is a disorder of the systolic function decreased ejection fraction groups (heart failure).
and/or diastolic function of the heart [84–86]. This circRNA might become a biomarker for
Heart failure is not an independent disease, but a both myocardial infarction and heart failure.
terminal stage in the development of heart dis-
eases [87, 88]. Almost all kinds of cardiovascular
diseases can eventually lead to the occurrence of 6.3 Coronary Artery Disease
heart failure, such as myocardial infarction, car-
diomyopathy, hemodynamic overload, myocar- Coronary artery disease, a big threat to heart
dial injury (any cause), and inflammation [89, health, is the main cause of coronary heart
90]. All these diseases can lead to alterations of ­disease, cerebral infarction, and peripheral vas-
164 Q. Zhou et al.

cular disease. Coronary artery disease with high peripheral blood, including 12 upregulated and
morbidity and mortality pose a threat to human 10 downregulated. Hsa_circ_0124644 was sig-
health and cause burden to economies [98]. There nificantly upregulated in coronary artery disease
are some treatments for coronary artery disease, group, and this circRNA was of high sensitivity
such as medications, percutaneous coronary (0.867) and specificity (0.767) for coronary
intervention (PCI), and coronary artery bypass artery disease [108]. Therefore, circRNAs are
graft surgery (CABG) [100–102]. Minimizing prospective diagnostic biomarkers in coronary
the time between plaque rupture and treatment is artery disease.
critical for reducing the morbidity and mortality
[103]. But the diagnosis and prognosis of coro-
nary artery disease is almost blank at present 6.4 Hypertension
[104, 105].
Early detection of coronary artery disease has High blood pressure (hypertension) refers to sys-
the potential to increase survival rates and to temic arterial blood pressure (systolic pressure)
improve the life quality of patients. MiR-221 and and/or diastolic blood pressure increase as the
miR-222 have been reported reduced in carotid main characteristics (systolic blood pressure,
plaques after an acute ischemic cerebrovascular 140 mmHg or higher; diastolic blood pressure,
syndrome [106]. CircR-284 possesses a miR-221 90 mmHg or higher). Hypertension can be asso-
and miR-222 binding site. Therefore, circR-284 ciated with the function of the heart, brain, kid-
may act as a regulator on miR-221/miR-222 ney, and other organs or physical damage to the
activity. A validation study found that miR-221 is clinical syndrome. Hypertension is the most
significantly decreased in the urgent carotid ath- common chronic disease and the most important
erosclerotic plaques, compared with the asymp- risk factor for cardiovascular disease [109–111].
tomatic control. But perhaps even more Profiling and bioinformatics identified 13
interesting is that circR-284 has been reported to downregulated and 46 upregulated circRNAs in
be significantly increased in the serum of urgent hypertension patients [112]. Among them, four
carotid atherosclerotic plaque patient [103]. circRNAs (hsa-circ-0000437, hsa-circ-0008139,
Combining the expression of serum circR-284 hsa-circ-0005870, and hsa-circ-0040809) showed
and miR-221, the ratio of circR-284 and miR-221 significant difference in hypertensive patients. By
was uniquely increased in the urgent group further validation, hsa-circ-0005870 was con-
(P < 0.001); and this ratio was sensitive (0.76) firmed significantly downregulated in plasma
and specific (0.88) for detecting plaque rupture between 49 hypertension patients (systolic blood
and stroke [103]. Based on these results, the ratio pressure over 150/90 mmHg) and 49 healthy (sys-
of circR-284 and miR-221 has the potential to tolic blood pressure lower than 140/90 mmHg). In
become a diagnostic biomarker in cardiovascular addition, hsa-circ-0005870 has also been found
disease for prediction of the risk of plaque related to many biological processes, such as cel-
rupture. lular response to stress. Therefore, this circRNA
Identified by microarray analysis, there were may be a novel biomarker for hypertension.
2036 circRNAs in blood differentially expressed In human umbilical vein endothelial cells,
between coronary artery disease and control, 7388 circRNAs have been identified. Under
including 376 upregulated and 1660 downregu- hypoxia in endothelial cells, circRNAs cZNF292,
lated ones [107]. Among them, a circRNA (hsa-­ cAFF1, and cDENND4C were found upregu-
circRNA11783-­ 2) was significantly lated, whereas cTHSD1 was reduced. Among
downregulated in coronary artery disease group them, cZNF292 is the highest expressed with
but also in type 2 diabetes group [107]. In a sepa- proangiogenic activities. These circRNA may be
rate set of experiments, RNA microarray showed new therapeutic targets and endogenous bio-
22 circRNAs were differentially expressed in markers [113].
13 Circular RNAs as Novel Biomarkers for Cardiovascular Diseases 165

6.5 Chronic Thromboembolic early stages of cardiovascular diseases, the diag-

Pulmonary Hypertension nosis of diseases is important for treatment and
prognosis. Traditional nonspecific diagnosis can-
Chronic thromboembolic pulmonary hyperten- not fulfill the needs. A novel biomarker demands
sion (CTEPH) is harmful to human health. (1) noninvasiveness and safety, (2) high sensitiv-
Pulmonary hypertension is caused by obstruction ity and specificity, and (3) stability and
of pulmonary arteries and subsequent vascular abundance.
remodeling [114–116]. Early detection of pulmo- CircRNAs are suitable biomarkers for disease
nary hypertension has the potential to increase diagnosis. Firstly, circRNAs are closed circular
physiological function of heart and improve molecules that are resistant to RNA exonucleases
prognosis [117, 118]. In the blood samples of or RNase R and hence have stability in blood and
chronic thromboembolic pulmonary hyperten- other body fluids. Secondly, many circRNAs are
sion patient, 351 circRNAs showed significant reported to be conserved across mammals.
difference, including 122 upregulations and 229 Especially, circRNAs have high sensitivity and
downregulations [119]. Among them, circRNA specificity. CircRNAs can be source for search-
hsa_circ_0002062 can regulate 761 miRNAs, ing new therapeutic targets and candidates in dis-
and circRNA hsa_circ_0022342 can regulate 453 ease diagnosis [123]. Multiple studies have
miRNAs. Through miRNAs regulation, these cir- suggested that circRNAs may become potential
cRNAs may be involved in regulating the path- biomarkers for many cardiovascular diseases,
ways in chronic thromboembolic pulmonary such as heart failure, acute myocardial infarction,
hypertension [119–121], such as ErbB signaling coronary artery disease, and hypertension
pathway. Therefore, these unique circRNAs (Table 13.1).
which are changed and involved in chronic There needs to be more research on explora-
thromboembolic pulmonary hypertension may tion of circRNAs as potential biomarkers. In
be novel biomarkers for this disease. experimental design, patients’ physiological con-
ditions need to be more explicit, including crite-
ria such as ages, genders, colors, diet habits, and
7 Future Perspectives so on. There are some details we should consider:
(1) What kind of sample can be collected, blood
Cardiovascular diseases are not only manifold sample, serum, or plasma? (2) What can be used
but involute. Due to the influence of heredity for normalization, Cel-mirR-39, GAPDH, or
[122], diet, living habit, and external environ- SF3a1? (3) Which test method can be use, SYBR
ment, different people have different age of onset, or TaqMan real-time quantitative?
and some people do not have obvious symptoms It is important to make strict screening crite-
in their whole life. Nowadays, many people live a ria. Although many studies assume that circRNAs
stressful, chronically overloaded, unbalanced life have high stability, there are some conditions that
that cause rising morbidity and mortality. In the need to be investigated, such as sample collection

Table 13.1 Circulating circular RNAs as biomarkers in cardiovascular diseases

Diseases CircRNAs Regulation Normalization Sample Ref
Acute myocardial infarction MICRA Down SF3a1 Blood [81, 82]
Heart failure MICRA Down SF3a1 Blood [81, 82]
Coronary artery disease CircR-284/miR-221 Up – Serum [103]
hsa_circ_0124644 Up GAPDH Blood [108]
Hypertension Hsa-­circ-­0005870 Down GAPDH Plasma [112]
Chronic thromboembolic pulmonary hsa_circ_0002062 Down – Blood [119]
hypertension hsa_circ_0022342 Down – Blood [119]
166 Q. Zhou et al.

and treatment, equipment and procedures, factic- 12. Chen X, Ba Y, Ma L et al (2008) Characterization of
microRNAs in serum: a novel class of biomarkers
ity, and reproducible analytics method. It is time for diagnosis of cancer and other diseases. Cell Res
to make a universally accepted protocol on iden- 18(10):997–1006
tification and quantification of circRNAs, which 13. Jeck WR, Sharpless NE (2014) Detecting and
affects how popular circRNAs will be used as characterizing circular RNAs. Nat Biotechnol
32(5):453–461
biomarkers for cardiovascular diseases in the 14. Maass PG, Glazar P, Memczak S et al (2017) A map
future. of human circular RNAs in clinically relevant tis-
sues. J Mol Med (Berl) 95(11):1179–1189
Acknowledgments This work was supported by the 15. Alhasan AA, Izuogu OG, Al-Balool HH et al (2016)
grants from National Natural Science Foundation of Circular RNA enrichment in platelets is a signature
China (81770401 to Y Bei) and National Key Research of transcriptome degradation. Blood 127(9):e1–e11
and Development Program of China (2017YFC1700401 16. Panda AC, De S, Grammatikakis I et al (2017)
to Y Bei). High-purity circular RNA isolation method (RPAD)
reveals vast collection of intronic circRNAs. Nucleic
Acids Res 45(12):e116
Competing Financial Interests The authors declare no
17. Suzuki H, Zuo Y, Wang J et al (2006) Characterization
competing financial interests.
of RNase R-digested cellular RNA source that con-
sists of lariat and circular RNAs from pre-mRNA
splicing. Nucleic Acids Res 34(8):e63
18. Jeck WR, Sorrentino JA, Wang K et al (2013)
Circular RNAs are abundant, conserved, and associ-
References ated with ALU repeats. RNA 19(2):141–157
19. Bahn JH, Zhang Q, Li F et al (2015) The landscape
1. Han B, Chao J, Yao H (2018) Circular RNA and its of microRNA, Piwi-interacting RNA, and circular
mechanisms in disease: from the bench to the clinic. RNA in human saliva. Clin Chem 61(1):221–230
Pharmacol Ther 187:31. https://doi.org/10.1016/j. 20. Preusser C, Hung LH, Schneider T et al (2018)
pharmthera.2018.01.010 Selective release of circRNAs in platelet-derived
2. Esteller M (2011) Non-coding RNAs in human dis- extracellular vesicles. J Extracell Vesicles
ease. Nat Rev Genet 12(12):861–874 7(1):1424473
3. Salzman J (2016) Circular RNA expression: its 21. Rybak-Wolf A, Stottmeister C, Glazar P et al (2015)
potential regulation and function. Trends Genet Circular RNAs in the mammalian brain are highly
32(5):309–316 abundant, conserved, and dynamically expressed.
4. Nigro JM, Cho KR, Fearon ER et al (1991) Mol Cell 58(5):870–885
Scrambled exons. Cell 64(3):607–613 22. Guo JU, Agarwal V, Guo H et al (2014) Expanded
5. Sanger HL, Klotz G, Riesner D et al (1976) identification and characterization of mammalian
Viroids are single-stranded covalently closed cir- circular RNAs. Genome Biol 15(7):409
cular RNA molecules existing as highly base-­ 23. Salzman J, Chen RE, Olsen MN et al (2013) Cell-­
paired rod-like structures. Proc Natl Acad Sci USA type specific features of circular RNA expression.
73(11):3852–3856 PLoS Genet 9(9):e1003777
6. Tan WL, Lim BT, Anene-Nzelu CG et al (2017) A 24. Werfel S, Nothjunge S, Schwarzmayr T et al (2016)
landscape of circular RNA expression in the human Characterization of circular RNAs in human, mouse
heart. Cardiovasc Res 113(3):298–309 and rat hearts. J Mol Cell Cardiol 98:103–107
7. Chen LL, Yang L (2015) Regulation of circRNA bio- 25. Kulcheski FR, Christoff AP, Margis R (2016)
genesis. RNA Biol 12(4):381–388 Circular RNAs are miRNA sponges and can be used
8. Chen I, Chen CY, Chuang TJ (2015) Biogenesis, as a new class of biomarker. J Biotechnol 238:42–51
identification, and function of exonic circular RNAs. 26. Liu YC, Chiu YJ, Li JR et al (2018) Biclustering of
Wiley Interdiscip Rev RNA 6(5):563–579 transcriptome sequencing data reveals human tissue-­
9. Zhang Y, Zhang XO, Chen T et al (2013) Circular specific circular RNAs. BMC Genomics 19(Suppl
intronic long noncoding RNAs. Mol Cell 1):958
51(6):792–806 27. Khan MA, Reckman YJ, Aufiero S et al (2016)
10. Li Z, Huang C, Bao C et al (2015) Exon-intron cir- RBM20 regulates circular RNA production from the
cular RNAs regulate transcription in the nucleus. Nat titin gene. Circ Res 119(9):996–1003
Struct Mol Biol 22(3):256–264 28. Wu HJ, Zhang CY, Zhang S et al (2016) Microarray
11. Arita T, Ichikawa D, Konishi H et al (2013) expression profile of circular RNAs in heart tissue of
Circulating long non-coding RNAs in plasma mice with myocardial infarction-induced heart fail-
of patients with gastric cancer. Anticancer Res ure. Cell Physiol Biochem 39(1):205–216
33(8):3185–3193 29. Dou Y, Cha DJ, Franklin JL et al (2016) Circular
RNAs are down-regulated in KRAS mutant colon
13 Circular RNAs as Novel Biomarkers for Cardiovascular Diseases 167

cancer cells and can be transferred to exosomes. Sci hepatic stellate cell induced by irradiation. Gene
Rep 6:37982 629:35–42
30. Li Y, Zheng Q, Bao C et al (2015) Circular RNA is 46. Zeng Y, Du WW, Wu Y et al (2017) A circular RNA
enriched and stable in exosomes: a promising bio- binds to and activates AKT phosphorylation and
marker for cancer diagnosis. Cell Res 25(8):981–984 nuclear localization reducing apoptosis and enhanc-
31. Memczak S, Jens M, Elefsinioti A et al (2013) ing cardiac repair. Theranostics 7(16):3842–3855
Circular RNAs are a large class of animal RNAs 47. Holdt LM, Stahringer A, Sass K et al (2016) Circular
with regulatory potency. Nature 495(7441):333–338 non-coding RNA ANRIL modulates ribosomal RNA
32. Thomas LF, Saetrom P (2014) Circular RNAs are maturation and atherosclerosis in humans. Nat
depleted of polymorphisms at microRNA binding Commun 7:12429
sites. Bioinformatics 30(16):2243–2246 48. Wang K, Long B, Liu F et al (2016) A circular RNA
33. Abdelmohsen K, Panda AC, Munk R et al (2017) protects the heart from pathological hypertrophy
Identification of HuR target circular RNAs uncovers and heart failure by targeting miR-223. Eur Heart
suppression of PABPN1 translation by CircPABPN1. J 37(33):2602–2611
RNA Biol 14(3):361–369 49. Geng HH, Li R, Su YM et al (2016) The circular
34. Barbagallo D, Caponnetto A, Cirnigliaro M et al RNA Cdr1as promotes myocardial infarction by
(2018) CircSMARCA5 inhibits migration of glio- mediating the regulation of miR-7a on its target
blastoma Multiforme cells by regulating a molecu- genes expression. PLoS One 11(3):e0151753
lar Axis involving splicing factors SRSF1/SRSF3/ 50. Meng S, Zhou H, Feng Z et al (2017) CircRNA:
PTB. Int J Mol Sci 19(2) functions and properties of a novel potential bio-
35. Ashwal-Fluss R, Meyer M, Pamudurti NR et al marker for cancer. Mol Cancer 16(1):94
(2014) circRNA biogenesis competes with pre-­ 51. Zhang S, Zeng X, Ding T et al (2018) Microarray
mRNA splicing. Mol Cell 56(1):55–66 profile of circular RNAs identifies hsa_circ_0014130
36. Shan K, Liu C, Liu BH et al (2017) Circular as a new circular RNA biomarker in non-small cell
noncoding RNA HIPK3 mediates retinal vascu- lung cancer. Sci Rep 8(1):2878
lar dysfunction in diabetes mellitus. Circulation 52. Li W, Zhong C, Jiao J et al (2017) Characterization
136(17):1629–1642 of hsa_circ_0004277 as a new biomarker for acute
37. Zheng J, Liu X, Xue Y et al (2017) TTBK2 circu- myeloid leukemia via circular RNA profile and bio-
lar RNA promotes glioma malignancy by regulating informatics analysis. Int J Mol Sci 18(3)
miR-217/HNF1beta/Derlin-1 pathway. J Hematol 53. Qin M, Liu G, Huo X et al (2016) Hsa_circ_0001649:
Oncol 10(1):52 a circular RNA and potential novel biomarker
38. Du WW, Yang W, Liu E et al (2016) Foxo3 circular for hepatocellular carcinoma. Cancer Biomark
RNA retards cell cycle progression via forming ter- 16(1):161–169
nary complexes with p21 and CDK2. Nucleic Acids 54. Ronaldson KJ (2017) Cardiovascular disease in
Res 44(6):2846–2858 clozapine-­ treated patients: evidence, mechanisms
39. Lukiw WJ (2013) Circular RNA (circRNA) in and management. CNS Drugs 31(9):777–795
Alzheimer’s disease (AD). Front Genet 4:307 55. Heil B, Tang WH (2015) Biomarkers: their potential
40. You X, Vlatkovic I, Babic A et al (2015) Neural cir- in the diagnosis and treatment of heart failure. Cleve
cular RNAs are derived from synaptic genes and reg- Clin J Med 82(12 Suppl 2):S28–S35
ulated by development and plasticity. Nat Neurosci 56. Calzetta L, Orlandi A, Page C et al (2016) Brain
18(4):603–610 natriuretic peptide: much more than a biomarker. Int
41. Tian F, Yu CT, Ye WD et al (2017) Cinnamaldehyde J Cardiol 221:1031–1038
induces cell apoptosis mediated by a novel cir- 57. Kondkar AA, Abu-Amero KK (2015) Utility of cir-
cular RNA hsa_circ_0043256 in non-small cell culating microRNAs as clinical biomarkers for car-
lung cancer. Biochem Biophys Res Commun diovascular diseases. Biomed Res Int 2015:821823
493(3):1260–1266 58. Enroth S, Johansson A, Enroth SB et al (2014)
42. Wang YH, Yu XH, Luo SS et al (2015) Strong effects of genetic and lifestyle factors on bio-
Comprehensive circular RNA profiling reveals that marker variation and use of personalized cutoffs. Nat
circular RNA100783 is involved in chronic CD28-­ Commun 5:4684
associated CD8(+)T cell ageing. Immun Ageing 59. Mari-Alexandre J, Sanchez-Izquierdo D, Gilabert-­
12:17 Estelles J et al (2016) miRNAs regulation and its
43. Gruner H, Cortes-Lopez M, Cooper DA et al (2016) role as biomarkers in endometriosis. Int J Mol Sci
CircRNA accumulation in the aging mouse brain. 17(1)
Sci Rep 6:38907 60. Van Roosbroeck K, Pollet J, Calin GA (2013)
44. Sena CM, Pereira AM, Seica R (2013) Endothelial miRNAs and long noncoding RNAs as biomark-
dysfunction – a major mediator of diabetic vascular ers in human diseases. Expert Rev Mol Diagn
disease. Biochim Biophys Acta 1832(12):2216–2231 13(2):183–204
45. Chen Y, Yuan B, Wu Z et al (2017) Microarray pro- 61. Yang Y, Yu T, Jiang S et al (2017) miRNAs as
filing of circular RNAs and the potential regulatory potential therapeutic targets and diagnostic bio-
role of hsa_circ_0071410 in the activated human markers for cardiovascular disease with a particular
168 Q. Zhou et al.

focus on WO2010091204. Expert Opin Ther Pat spinal fluids to investigate myocardial and skeletal
27(9):1021–1029 muscle injury in medicolegal autopsy cases. Leg
62. Zhu H, Fan GC (2013) Whether circulating miRNAs Med (Tokyo) 13(5):226–232
or miRNA-carriers serve as biomarkers for acute 78. Heavens KR, Szivak TK, Hooper DR et al (2014)
myocardial infarction. J Biomark Drug Dev 1(1) The effects of high intensity short rest resistance
63. Wang F, Long G, Zhao C et al (2014) Atherosclerosis-­ exercise on muscle damage markers in men and
related circulating miRNAs as novel and sensitive women. J Strength Cond Res 28(4):1041–1049
predictors for acute myocardial infarction. PLoS 79. Tulevski II, ten Wolde M, van Veldhuisen DJ et al
One 9(9):e105734 (2007) Combined utility of brain natriuretic peptide
64. Pleister A, Selemon H, Elton SM et al (2013) and cardiac troponin T may improve rapid triage and
Circulating miRNAs: novel biomarkers of acute risk stratification in normotensive patients with pul-
coronary syndrome? Biomark Med 7(2):287–305 monary embolism. Int J Cardiol 116(2):161–166
65. Rognoni A, Cavallino C, Lupi A et al (2014) Novel 80. Stelzle D, Shah ASV, Anand A et al (2018) High-­
biomarkers in the diagnosis of acute coronary syn- sensitivity cardiac troponin I and risk of heart failure
dromes: the role of circulating miRNAs. Expert Rev in patients with suspected acute coronary syndrome:
Cardiovasc Ther 12(9):1119–1124 a cohort study. Eur Heart J Qual Care Clin Outcomes
66. Ali Sheikh MS, Salma U, Zhang B et al (2016) 4(1):36–42
Diagnostic, prognostic, and therapeutic value of 81. Salgado-Somoza A, Zhang L, Vausort M et al (2017)
­circulating miRNAs in heart failure patients asso- The circular RNA MICRA for risk stratification
ciated with oxidative stress. Oxidative Med Cell after myocardial infarction. Int J Cardiol Heart Vasc
Longev 2016:5893064 17:33–36
67. Luscher TF (2016) From heart failure to transplan- 82. Vausort M, Salgado-Somoza A, Zhang L et al (2016)
tation: genes, miRNAs, and biomarkers. Eur Heart Myocardial infarction-associated circular RNA
J 37(33):2561–2563 predicting left ventricular dysfunction. J Am Coll
68. Yan H, Ma F, Zhang Y et al (2017) miRNAs as bio- Cardiol 68(11):1247–1248
markers for diagnosis of heart failure: a systematic 83. Vausort M, Wagner DR, Devaux Y (2014) Long
review and meta-analysis. Medicine (Baltimore) noncoding RNAs in patients with acute myocardial
96(22):e6825 infarction. Circ Res 115(7):668–677
69. Viereck J, Thum T (2017) Circulating noncoding 84. Kasner M, Sinning D, Lober J et al (2015)
RNAs as biomarkers of cardiovascular disease and Heterogeneous responses of systolic and diastolic
injury. Circ Res 120(2):381–399 left ventricular function to exercise in patients with
70. Devaux Y, Creemers EE, Boon RA et al (2017) heart failure and preserved ejection fraction. ESC
Circular RNAs in heart failure. Eur J Heart Fail Heart Fail 2(3):121–132
19(6):701–709 85. Rapezzi C, Milandri A, Lorenzini M (2017) The
71. Danan M, Schwartz S, Edelheit S et al (2012) complex interplay between systolic and diastolic
Transcriptome-wide discovery of circular RNAs in function at rest and during exercise in heart failure:
archaea. Nucleic Acids Res 40(7):3131–3142 the case of cardiac amyloidosis. Eur J Heart Fail
72. Hansen TB, Jensen TI, Clausen BH et al (2013) 19(11):1466–1467
Natural RNA circles function as efficient microRNA 86. Rommel KP, von Roeder M, Oberueck C et al
sponges. Nature 495(7441):384–388 (2018) Load-independent systolic and diastolic right
73. Townsend N, Wilson L, Bhatnagar P, ventricular function in heart failure with preserved
Wickramasinghe K, Rayner M, Nichols M (2016) ejection fraction as assessed by resting and hand-
Cardiovascular disease in Europe 2016: an epide- grip exercise pressure-volume loops. Circ Heart Fail
miological update. Eur Heart J 37(42):3182–3183 11(2):e004121
74. Benjamin EJ, Blaha MJ, Chiuve SE et al (2017) 87. Beltran-Alvarez P, Tarradas A, Chiva C et al (2014)
Heart disease and stroke Statistics-2017 update: Identification of N-terminal protein acetylation and
a report from the American Heart Association. arginine methylation of the voltage-gated sodium
Circulation 135(10):e146–e603 channel in end-stage heart failure human heart.
75. Haggart PC, Adam DJ, Ludman PF et al (2001) J Mol Cell Cardiol 76:126–129
Comparison of cardiac troponin I and creatine 88. Szczurek W, Szygula-Jurkiewicz B, Zakliczynski M
kinase ratios in the detection of myocardial injury et al (2018) Prognostic utility of the N terminal pro-
after aortic surgery. Br J Surg 88(9):1196–1200 hormone of brain natriuretic peptide and the modi-
76. Shyu KG, Lin JL, Chen JJ et al (1996) Use of cardiac fied model for end stage liver disease in patients with
troponin T, creatine kinase and its isoform to moni- end stage heart failure. Pol Arch Intern Med
tor myocardial injury during radiofrequency abla- 89. Callender T, Rahimi K (2015) Heart failure and
tion for supraventricular tachycardia. Cardiology iron deficiency anaemia: a complex dance. Heart
87(5):392–395 101(8):579–580
77. Wang Q, Michiue T, Ishikawa T et al (2011) 90. Choi KH, Lee GY, Choi JO et al (2018) Outcomes
Combined analyses of creatine kinase MB, cardiac of de novo and acute decompensated heart fail-
troponin I and myoglobin in pericardial and cerebro-
13 Circular RNAs as Novel Biomarkers for Cardiovascular Diseases 169

ure patients according to ejection fraction. Heart agement of patients with Non-ST-elevation acute
104(6):525–532 coronary syndromes, and 2014 ACC/AHA guideline
91. Bayes-Genis A, Lanfear DE, de Ronde MWJ et al on perioperative cardiovascular evaluation and man-
(2018) Prognostic value of circulating microRNAs agement of patients undergoing noncardiac surgery.
on heart failure-related morbidity and mortality in Circulation 134(10):e123–e155
two large diverse cohorts of general heart failure 101. McGrath BM, Norris CM, Hardwicke-Brown E et al
patients. Eur J Heart Fail 20(1):67–75 (2017) Quality of life following coronary artery
92. Bhatt AS, Cooper LB, Ambrosy AP et al (2018) bypass graft surgery vs. percutaneous coronary
Interaction of body mass index on the association intervention in diabetics with multivessel disease: a
between N-terminal-pro-b-type natriuretic pep- five-year registry study. Eur Heart J Qual Care Clin
tide and morbidity and mortality in patients with Outcomes 3(3):216–223
acute heart failure: findings from ASCEND-HF 102. Zimmermann FM, De Bruyne B, Pijls NH, et al
(acute study of clinical effectiveness of Nesiritide (2015) Rationale and design of the fractional flow
in decompensated heart failure). J Am Heart Assoc reserve versus angiography for multivessel evalua-
7(3) tion (FAME) 3 trial: a comparison of fractional flow
93. Moser DK, Robinson S, Biddle MJ et al (2015) reserve-guided percutaneous coronary intervention
Health literacy predicts morbidity and ­ mortality and coronary artery bypass graft surgery in patients
in rural patients with heart failure. J Card Fail with multivessel coronary artery disease. Am Heart
21(8):612–618 J 170 (4):619–626 e612
94. Ponikowski P, Voors AA, Anker SD et al (2016) 103. Bazan HA, Hatfield SA, Brug A et al (2017) Carotid
2016 ESC guidelines for the diagnosis and treatment plaque rupture is accompanied by an increase in the
of acute and chronic heart failure: the task force for ratio of serum circR-284 to miR-221 levels. Circ
the diagnosis and treatment of acute and chronic Cardiovasc Genet 10(4):e001720
heart failure of the European Society of Cardiology 104. Sarkisian L, Saaby L, Poulsen TS, et al (2016)
(ESC). Developed with the special contribution of Clinical characteristics and outcomes of patients
the heart failure association (HFA) of the ESC. Eur with myocardial infarction, myocardial injury, and
J Heart Fail 18(8):891–975 nonelevated troponins. Am J Med 129(4):446 e445–
95. Talwar S, Squire IB, Downie PF et al (2000) Profile 446 e421
of plasma N-terminal proBNP following acute myo- 105. Lee PH, Lee JY, Lee CW et al (2018) Comparison
cardial infarction; correlation with left ventricular of a simple angiographic approach with a synergy
systolic dysfunction. Eur Heart J 21(18):1514–1521 between percutaneous coronary intervention with
96. Lin X, Lo HC, Wong DT et al (2015) Noncoding Taxus and cardiac surgery score-based approach for
RNAs in human saliva as potential disease biomark- left main coronary artery stenting: a pooled analy-
ers. Front Genet 6:175 sis of serial PRECOMBAT (premier of randomized
97. Kumarswamy R, Bauters C, Volkmann I et al (2014) comparison of bypass surgery versus angioplasty
Circulating long noncoding RNA, LIPCAR, pre- using Sirolimus-eluting stent in patients with
dicts survival in patients with heart failure. Circ Res left main coronary artery disease) studies. Circ
114(10):1569–1575 Cardiovasc Interv 11(1):e005374
98. Burd CE, Jeck WR, Liu Y et al (2010) Expression 106. Bazan HA, Hatfield SA, O'Malley CB et al (2015)
of linear and novel circular forms of an INK4/ARF-­ Acute loss of miR-221 and miR-222 in the athero-
associated non-coding RNA correlates with athero- sclerotic plaque shoulder accompanies plaque rup-
sclerosis risk. PLoS Genet 6(12):e1001233 ture. Stroke 46(11):3285–3287
99. Xuan L, Sun L, Zhang Y et al (2017) Circulating 107. Li X, Zhao Z, Jian D et al (2017) Hsa-­
long non-coding RNAs NRON and MHRT as novel circRNA11783-­2 in peripheral blood is correlated
predictive biomarkers of heart failure. J Cell Mol with coronary artery disease and type 2 diabetes
Med 21(9):1803–1814 mellitus. Diab Vasc Dis Res 14(6):510–515
100. Levine GN, Bates ER, Bittl JA et al (2016) 2016 108. Zhao Z, Li X, Gao C et al (2017) Peripheral blood
ACC/AHA Guideline focused update on duration circular RNA hsa_circ_0124644 can be used as a
of dual antiplatelet therapy in patients with coronary diagnostic biomarker of coronary artery disease. Sci
artery disease: a report of the American College of Rep 7:39918
Cardiology/American Heart Association Task Force 109. Hamamura K, Yanagida M, Ishikawa H et al
on clinical practice guidelines: an update of the 2011 (2018) Quantitative measurement of a candidate
ACCF/AHA/SCAI guideline for percutaneous coro- serum biomarker peptide derived from alpha2-
nary intervention, 2011 ACCF/AHA guideline for HS-­ glycoprotein, and a preliminary trial of mul-
coronary artery bypass graft surgery, 2012 ACC/ tidimensional peptide analysis in females with
AHA/ACP/AATS/PCNA/SCAI/STS guideline for pregnancy-induced hypertension. Ann Clin Biochem
the diagnosis and management of patients with sta- 55(2):287–295
ble ischemic heart disease, 2013 ACCF/AHA guide- 110. Richter MJ, Schermuly R, Seeger W et al (2016)
line for the Management of ST-Elevation myocardial Relevance of angiopoietin-2 and soluble P-selectin
infarction, 2014 AHA/ACC guideline for the man- levels in patients with pulmonary arterial hyperten-
170 Q. Zhou et al.

sion receiving combination therapy with oral trepro- 117. Ghofrani HA, D'Armini AM, Grimminger F et al
stinil: a FREEDOM-C2 biomarker substudy. Pulm (2013) Riociguat for the treatment of chronic throm-
Circ 6(4):516–523 boembolic pulmonary hypertension. N Engl J Med
111. Smukowska-Gorynia A, Tomaszewska I, 369(4):319–329
Malaczynska-Rajpold K et al (2017) Red blood 118. Sulica R, Fenton R, Cefali F (2015) Early observa-
cells distribution width as a potential prognostic tions on the use of Riociguat in a large, metropolitan
biomarker in patients with pulmonary arterial hyper- pulmonary arterial hypertension/chronic thrombo-
tension and chronic thromboembolic pulmonary embolic pulmonary hypertension treatment center.
hypertension. Heart Lung Circ doi 27:842. https:// Cardiol Ther 4(2):209–218
doi.org/10.1016/j.hlc.2017.08.007 119. Miao R, Wang Y, Wan J et al (2017) Microarray
112. Wu N, Jin L, Cai J (2017) Profiling and bioinformat- expression profile of circular RNAs in chronic
ics analyses reveal differential circular RNA expres- thromboembolic pulmonary hypertension. Medicine
sion in hypertensive patients. Clin Exp Hypertens (Baltimore) 96(27):e7354
39(5):454–459 120. Kamath-Rayne BD, Saal H, Lang S et al (2013)
113. Boeckel JN, Jae N, Heumuller AW et al Recurrent severe oligohydramnios and fetal pulmo-
(2015) Identification and characterization of nary hypoplasia associated with ErbB4 mutation.
hypoxia-­
­ regulated endothelial circular RNA. Circ Obstet Gynecol 121(2 Pt 2 Suppl 1):499–501
Res 117(10):884–890 121. Kao DP, Stevens LM, Hinterberg MA et al
114. Ghofrani HA, Simonneau G, D'Armini AM et al (2017) Phenotype-specific Association of Single-
(2017) Macitentan for the treatment of inoperable Nucleotide Polymorphisms with heart failure and
chronic thromboembolic pulmonary hypertension preserved ejection fraction: a genome-wide asso-
(MERIT-1): results from the multicentre, phase 2, ciation analysis of the cardiovascular health study.
randomised, double-blind, placebo-controlled study. J Cardiovasc Transl Res 10(3):285–294
Lancet Respir Med 5(10):785–794 122. Mitchell BD, McArdle PF, Shen H et al (2008) The
115. Hoeper MM, Madani MM, Nakanishi N et al (2014) genetic response to short-term interventions affect-
Chronic thromboembolic pulmonary hypertension. ing cardiovascular function: rationale and design
Lancet Respir Med 2(7):573–582 of the heredity and phenotype intervention (HAPI)
116. Miwa H, Tanabe N, Jujo T et al (2018) Long-term heart study. Am Heart J 155(5):823–828
outcome of chronic thromboembolic pulmonary 123. Gorlach A, Holdenrieder S (2017) Circular RNA
hypertension at a single Japanese pulmonary endar- maps paving the road to biomarker development?
terectomy center. Circ J 82(5):1428–1436 J Mol Med (Berl) 95(11):1137–1141
 Circular RNAs as Biomarkers
for Cancer 14
Lu Xia, Meiyi Song, Mengxue Sun, Fei Wang,
and Changqing Yang

Abstract that circRNAs are a class of promising future

As a type of novel noncoding RNAs, circular biomarkers for cancer diagnosis and progno-
RNAs (circRNAs) have attracted great interest sis. Here we summarize the researches on cir-
due to its different characteristics from linear cRNAs and cancer over the past few years. We
RNAs. They are abundantly and stably present expect this summary to be a stepping stone to
in the transcriptome of eukaryotic cells, with further exploration of possible circRNAs as
development stage specificity and high con- cancer biomarkers.
servatism. Because circRNAs are not easily
degraded by exonuclease RNase R, they can Keywords
exist more stably in body fluids than linear Circular RNAs (circRNAs) · Cancer ·
RNAs. Based on these unique conditions, cir- Biomarkers · Diagnosis · Prognosis
cRNAs have great potential value as clinical
diagnostic and prognostic markers. As the
research deepens, more and more evidences
suggest that circRNAs may be closely associ- 1 Introduction
ated with many diseases, especially cancer.
Numerous studies have demonstrated the With the development of biotechnology and
abnormal expression of circRNAs in cancer, computer technology, more and more “invisible
and they can regulate the occurrence and pro- substances” are exposed in the organism, such as
gression of cancer by targeting key genes. circular RNAs (circRNAs). Although do not
Abundant circRNAs in tissues and cells can be encode proteins, they are indispensable in regula-
released into saliva and blood. It is undeniable tory processes. As the latest research hotspot, cir-
cRNAs are a group of newly validated noncoding
RNA molecules which form a covalent ring
These authors contributed equally to this work. Lu Xia structure instead of having a cap structure at the
and Meiyi Song
5′ end and poly (A) tail at the 3′ end. It was found
L. Xia · M. Song · M. Sun · F. Wang (*) in RNA viruses as early as the 1970s and once
C. Yang (*) thought to be by-products of aberrant RNA splic-
Division of Gastroenterology and Hepatology,
Digestive Disease Institute, Shanghai Tongji Hospital, ing due to low expression levels [1, 2]. With the
Tongji University School of Medicine, rapid development of bioinformatics and
Shanghai, China sequencing technologies, large-scale analysis of
e-mail: [email protected]; transcriptome data becomes possible, and the
[email protected]

© Springer Nature Singapore Pte Ltd. 2018 171

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_14
172 L. Xia et al.

characteristics and functions of circRNAs, a sort dation of the target mRNA [18, 19]. Multiple
of ancient and conservative molecules, are being researches have shown that miRNAs play an
gradually unveiled. It is well known that cir- important regulatory role in the development of
cRNAs are richly expressed in various tissues tumors [20–22]. Recently independent experi-
and cells throughout the human body and play ments have demonstrated that circRNAs can
vital roles in regulating physiological and patho- serve as sponges for miRNAs, which means that
logical processes [3, 4]. CircRNAs are abundant they can bind with miRNAs and thus suppress the
in the transcriptome of eukaryotic cells and con- function of miRNAs [3, 23]. These findings trig-
servative in species. They are also stable in ger the exploration of the potential regulatory
expression and exhibit tissue and development effect of circRNAs on cancer. Technologies such
stage specificity [5–7]. In 2013, professor as circRNAs chip and qRT-PCR have been widely
William R. Jeck’s team detected more than used in the studies related to circRNAs and can-
25,000 circRNAs in human fibroblasts, while in cer. So far, many circRNAs have been found to be
the same year, Memczak et al. identified 1950 closely related to cancer and bring new dawn of
human circRNAs and 1903 mice circRNAs the early diagnosis and treatment of cancer [24].
through RNA-seq data combined with the human
leukocyte database [3]. Unlike linear RNAs, cir-
cRNAs are not easily degraded by exonuclease 2 The Characteristics
RNase R in body fluids; therefore they have the of CircRNAs
potential application value as clinical diagnostic
and prognostic markers [8]. There are numerous types of circRNAs, which
Cancer is characterized by high mortality, and are huge in number and widely distributed in the
its incidence is increasing in recent years [9]. It is organisms. The circRNAs currently found are
one of the major diseases threatening human life mainly divided into three categories according to
and social economic development and therefore their origin: exonic circRNAs (ecRNAs), intronic
brings the focus of attention [10, 11]. A survey on circRNAs (ciRNAs), and exon-intron circular
the incidence of global cancer in 2012 showed RNAs (EIciRNAs) composed of exons and
that there are approximately 14.1 million new introns. They are characterized by extremely
cancer cases and 8.2 million cancer deaths world- high stability, strong evolutionary conservation,
wide each year. Lung cancer and breast cancer and unique temporal and spatial expression.
are the most common types of cancer [9, 12]. The Studies have found that the expression levels of
essence of tumor formation is a kind of genetic circRNAs are ten times higher than their linear
disease, of which tissues infiltration and metasta- isomers [25]. The half-life of circRNAs with the
sis are the major features [13–15]. Its pathogen- peculiar loop structure is more than 48 h, which
esis refers to multiple steps and mechanisms. is far much longer than the linear mRNAs (10 h
New approaches for early diagnosis and treat- in average); this is why the circRNAs can be
ment of cancer have been sought for a long time. more stable in tissues, cells, and body fluids [25,
Biomarkers for cancer have emerged as a class of 26]. However, the expression of the circRNAs is
molecules closely related to the development and not dependent on the linear mRNA expression of
progression of tumors [16]. They play key roles its parental genes but changes with the life activi-
in many aspects such as early diagnosis, thera- ties such as growth and senescence. Moreover,
peutic monitoring, and prognostic evaluation the type and content of circRNAs in different
[17]. cells and tissues are also distinct [27, 28]. When
It is universally known that microRNAs (miR- the abundance of the circRNAs with conservative
NAs) control a large number of biological pro- nucleotide sequences increases, they can com-
cesses by direct interaction with their target pete with other RNA or miRNAs by competi-
mRNAs. This regulation can be achieved by tively binding to the RNA-binding protein (RBP)
inhibiting translation or by triggering the degra- [3]. These properties give circRNAs a unique
14 Circular RNAs as Biomarkers for Cancer 173

advantage as diagnostic and prognostic markers Since the first discovery of circRNAs in 1976,
for clinical diseases. CircRNAs are widely more and more mature technologies such as high-­
involved in the processes of physiological and throughput sequencing and gene chip have been
pathological regulation of human beings. Large applied to explore their biological functions. On
amounts of circRNAs in organisms influence the other hand, the establishment and enrichment
basic life activities such as cell proliferation, of the database also provide extremely conve-
cycle progression, cell senescence, and apoptosis nient conditions for the in-depth study of the cir-
by regulating gene expression. cRNAs. The current databases of circRNAs
The mechanisms of the discovered circRNAs include circRNABase, circBase, Circ2Traits, and
mainly include the following three types: CircNet. Through these databases, we can query
circRNAs sequences, genome annotations,
1. As molecular sponges of miRNAs: this is a expression profiles, miRNA-circRNA interaction
relatively common function of circRNAs and networks, and related disease information. With
the most promising direction in circRNAs these supports we have made tremendous prog-
researches. CircRNAs themselves contain at ress in the study of circRNAs.
least one miRNA binding site and therefore
can serve as miRNA sponges to regulate the
expression of target genes which are inhibited 3 CircRNAs in Cancer
by miRNAs through ceRNAs [3]. The ceR-
NAs network in the living body is complex, 3.1 Lung Cancer
and any minor changes can affect gene expres-
sion and induce tumorigenesis. Therefore, the The changes in lifestyle, dietary pattern, and dete-
competitive combination of circRNAs is riorating environment have led to a rising trend in
essential to maintain the balance of the ceR- the prevalence of cancer patients. Worldwide, the
NAs network. incidence and mortality of lung cancer account
2. Regulate gene transcription and cleavage: for the third most severe human tumor. Meanwhile
the mechanism of this process is diverse. It it is one of the most malignant tumors in the world
can not only enhance the expression of [33]. According to the World Cancer Report 2014
parental gene through miRNAs but also exert released by WHO, there are 1.82 million new lung
a positive feedback on its parental genes cancer cases in 2012 and 1.59 million deaths
through interaction with RNA polymerase caused by lung cancer, accounting for 13% and
II. It is also possible to regulate parental gene 19.4% of the incidence and deaths of all malig-
expression through competitive splicing fac- nant tumors [34]. The epidemiological data mani-
tors. CircRNAs can regulate parental genes fest that the morbidity and mortality rates of lung
through positive feedback and negative feed- cancer are both the highest among global malig-
back control at different expression levels [5, nant tumors. More than 220,000 cases were diag-
29, 30]. nosed with lung cancer in 2015, and the number
3. Interact with RNA-binding proteins: cir- of patients who died of lung cancer in the same
cRNAs can bind stably with RNA-binding year exceeded 158,000 [35]. Moreover, about
proteins such as AGO (argonaute), RNA poly- 85% of lung cancers are non-small cell lung can-
merase II, muscleblind protein (MBL), vari- cer, and 25–30% of non-small cell lung cancers
able factor QKI (Quaking), and eukaryotic are squamous cell carcinomas [36]. At present,
translation initiation factor 4A3 (EIF4A3). the common clinical treatment methods, includ-
They can store and transport RBP to compete ing surgical resection, radiation therapy, and che-
with RBP substrate for its binding site and motherapy, cannot reduce the incidence of lung
thereby regulate the activity of RBP and inter- cancer. The 5-year survival rate of lung cancer
fere with the normal function of a protein in a patients is only 15.9% [37]. Therefore, innovative
direct or indirect way [3, 5, 30–32]. new treatments are urgently needed.
174 L. Xia et al.

A variety of circRNAs with high expression in sion of ITCH is found positively correlated with
lung cancer have been found. Hsa_circ_0000064 hsa_circ_0043256. As a miR-1252 sponge, hsa_
is upregulated in lung cancer tissues and lung circ_0043256 is significantly downregulated,
cancer cell lines A549 and H1229. The abnormal weakening the inhibitory effect of cinnamalde-
expression of hsa_circ_0000064 is closely related hyde, when cinnamaldehyde is used to block the
to the clinical features such as tumor lymph node Wnt/β-catenin signaling pathway [45].
metastasis and TNM staging [38]. The high
expression of circRNA-100876 in non-small cell
lung cancer (NSCLC) is closely related to the 3.2 Gastrointestinal Cancer
lymph node metastasis and tumor stage of lung
cancer. Additionally, the overall survival time of Gastrointestinal cancer is one of the most common
NSCLC patients with high circ_100876 expres- malignancies which seriously threatens the life
sion is significantly shorter [39]. Circ-HIPK3 can and health of human beings [46]. The gastric and
be detected in the NSCLC cell lines H1299, colorectal cancers are the most common in this
H827, H1975, H2170, H520, and H1650. kind of diseases. Studies have shown that the inci-
Experiments have shown that circ-HIPK3 can dence and mortality of gastric cancer rank second
regulate the expression level of insulin-like in China, which is second only to lung cancer,
growth factor1(IGF1) and promote cell prolifera- while colorectal cancer ranks fifth among all
tion by binding to miR-379 in NCI-H1299 and malignancies [47]. The 5-year survival rate of
NCI-H2170 cells [40]. Hsa_circ_0013958 is sig- advanced gastric cancer is less than 30% [48].
nificantly upregulated in histiocytes and plasma Tumor markers including AFP, CEA, CA19-9, and
of lung adenocarcinoma patients and shows sta- CA50 are commonly used to assist the clinical
tistically significant correlation with lymph node diagnosis and prognosis of gastrointestinal tumors.
metastasis and tumor staging [41]. By applying In order to compensate for the lack of tumor detec-
high-throughput sequencing on tumor and adja- tion markers, attempts are made to find circular
cent tissue from four cases with nonsmoking RNAs with diagnostic and prognostic values.
early lung adenocarcinoma, Zhao et al. detected
more than 300 circRNAs differentially expressed 3.2.1 Gastric Cancer
in tumor tissues and later verified 5 of them by A study of circRNAs in gastric cancer (GC) tis-
RT-qPCR. Consistent with the results of the chip, sues and paracancerous tissues identified 467 dif-
it provided potential targets for early diagnosis ferentially expressed circRNAs, among which
and treatment of early-stage lung adenocarci- expression of 214 circRNAs were significantly
noma [42]. Cdr1as (ciRS7) contains more than increased and expression of 253 were signifi-
70 selective binding sites for miRNAs, which can cantly decreased. Most of the circRNAs could be
strongly inhibit the activity of miR-7 and hence detected with corresponding miRNA binding
activate the miR-7-regulated genes. It has a regu- sites [49]. Circ-PVT1 is a highly expressed cir-
latory effect on many diseases [23]. Importantly, cRNA in GC and a potential independent index
similar to liver cancer, Cdr1as can also inhibit for evaluating the prognosis of GC. By analyzing
tumors in lung cancer by binding to miR-7 [43]. clinical data and tumor tissues of 187 patients
The circ-ITCH is significantly reduced and can with GC, it was found that the survival rate of
act as miR-7 and miR-214 sponge in lung cancer. patients with high circ-PVT1 expression was
It can inhibit the activation of Wnt/β-catenin sig- markedly higher. The promotive effect of circ-­
naling pathway by enhancing the expression of PVT1 on GC may act through attenuating the
the ITCH gene and thereby suppresses the prolif- inhibitory effect of miR-125 on cell proliferation
eration of lung cancer cells [44]. A new mecha- by combining with it. The experiment also
nism of cinnamaldehyde intervention on NSCLC demonstrated that reducing the expression of
­
through hsa_circ_0043256/miR-1252/ITCH axis circ-­PVT1 could inhibit the proliferation of GC
was proposed in another work. The gene expres- cells [50]. In contrast, hsa_circ_0000096 was
14 Circular RNAs as Biomarkers for Cancer 175

found obviously downregulated in GC cell lines fibrosis and ovarian tumor cells and normal ovar-
and tissues compared with normal gastric epithe- ian epithelial cells and confirmed that circRNA
lial cells and paired adjacent non-tumor tissues. It abundance was negatively correlated with cell
is supposed that hsa_circ_0000096 might interact proliferation. The expression of circRNAs in
with miRNAs through endogenous competition colorectal cancer (CRC) tissues were signifi-
and thereby affect GC cell growth and migration cantly lower than that in normal tissues, and the
by interfering with cell cycle and expression of expression level were also lower in CRC cell
migration-related protein [51]. Screened by data- lines. The investigators tested four circRNAs
base and verified by qRT-PCR, hsa_circ_002059 (circ_0817, circ_3204, circ_6229, circ_7374)
was found significantly downregulated in GC tis- low-expressed in clinical specimens in colon can-
sue compared with the adjacent non-tumor tissue cer cell lines together with their corresponding
by Li et al. The authors noticed that 10 days after linear RNAs. The expression ratio of the four cir-
GC tissue resection, the expression of plasma cRNAs to their parental linear RNAs was lower.
has_circ_002059 was detected to be higher than Finally, the researchers verified the relationship
that before surgery. Low expression level of has_ between the expression of circRNAs/correspond-
circ_002059 was significantly associated with ing linear RNAs and cell proliferation, confirm-
distant metastases, TNM staging, gender, and age ing that the cell proliferation rate was negatively
[52]. The combination of hsa_circ_0000096 and correlated with the expression circRNAs [58]. As
hsa_circ_002509 can considerably improve the a circRNA generated from the exon 5–11 of
diagnosis of gastric cancer [51]. Zhang et al. BANP gene, circ-BANP is highly expressed in
established a prediction system of early recur- CRC. Knocking down of circ-BANP can signifi-
rence for patients with stage III GC based on four cantly reduce the proliferation of CRC cells.
selected circRNAs: hsa_circ_101308, has_ Circ-BANP may play an important regulatory
circ_104423, hsa_circ_104916, and hsa_ role in CRC cells and may serve as a marker for
circ_100269. The area under the curve (AUC) prognosis and treatment of CRC [59]. Recently,
could reach 0.763 and 0.711 by the test of two Guo et al. discovered a novel abnormal circRNA,
centers and could rise to 0.866 and 0.818 by join- hsa_circ_0000069, by employing unsupervised
ing TNM staging, which indicated that this hierarchical clustering analysis. They determined
circRNA-­based predictive model is an effective that hsa_circ_0000069 is highly expressed in
assessment of the risk of early recurrence after CRC through quantitative PCR analysis of 30
radical gastrectomy for GC. However, this sys- paired CRC tissues and adjacent noncancerous
tem still needs to be further verified by prospec- tissues and is closely related to the patient’s age,
tive and multicenter research [53]. Other tumor size, lymph node metastasis, and TNM
circRNAs with abnormal expression profiles in staging. Functional analysis by using specifically
GC include has_circ_0001649 (the expression designed siRNA in CRC cells confirms that
level is significantly downregulated in GC tissue knocking down hsa_circ_0000069 markedly
and upregulated in serum after operation) [54], inhibits cell proliferation, migration, and inva-
has_circ_0044516 (the expression level is signifi- sion and induces G0/G1 arrest in vitro [60].
cantly upregulated in GC tissue) [55], has_ Another study has confirmed that hsa_
circ_0014717, and hsa_circ_0000190 (the circ_001569 plays a positive regulatory role in
expression level is significantly downregulated in cell proliferation and invasion of CRC. The
GC tissues and considered related to distant authors found that the expression of has_
metastasis, tumor staging, CA19-9) [56, 57]. circ_001569 is elevated in CRC tissues and cor-
related with tumor volume, TNM staging, and
3.2.2 Colorectal Cancer prognosis. It was further confirmed that hsa_
Dietmar Pils’s team compared circRNAs expres- circ_001569 can also be used as a “sponge” to
sion levels between fibrotic lung and normal lung adsorb miR-145. This is why the expression of
tissue in patients with idiopathic pulmonary has_circ_001569 and miR-145 are negatively
176 L. Xia et al.

correlated in CRC tissue. Upregulated analysis, in order to reveal the possible circRNAs
circ_001569 increases the expression of miR-145 involved in the generation of radiation resistance
target genes E2F5, BAG4, and FMNL2 and during treatment of esophageal cancer. They ini-
hence enhances the proliferation and invasion of tially identified 57 remarkable upregulated and
CRC cells, which in turn promotes the progres- 17 remarkable downregulated circRNAs in 3752
sion of colorectal cancer [61]. Circ-CCDC66 is a candidate circRNAs (fold change ≥ 2.0 and
newly discovered circRNA which is encoded by P < 0.05), of which 9 were validated by real-time
the CCDC66 gene. Hsiao et al. found that the qPCR. Supplemented by gene ontology analysis,
expression of circ-CCDC66 was increased in the authors confirmed a large number of target
polyps and CRC and was negatively correlated genes (including most miRNAs) participating in
with the prognosis. Inhibition of circ-CCDC66 this biological process. Among them, more than
expression in vitro significantly reduces the 400 target genes are enriched in the Wnt signal-
tumor volume in nude mice, suggesting that circ-­ ing pathway. Circ_001059 and circRNA_000167
CCDC66 can regulate multiple pathological pro- are the two largest nodes of the co-expression
cesses including cell proliferation, invasion, network of circRNA/microRNA [64]. Based on
migration, and anchorage-independent growth. the detection of circRNAs expression in 684
This function is achieved by acting as the sponge cases of esophageal squamous cell carcinoma
of miR-33b and miR-93. Experiments showed (ESCC) patients and their matched paracancer-
that knocking down circ-CCDC66 can impede ous tissues, circ-ITCH was also found to play a
tumor proliferation and invasion in vivo. All of role as molecular sponge of miR-7, miR-17, and
these findings indicate that circ-CCDC66 plays miR-214 in esophageal cancer. Consistent with
an important role in progression and metastasis the mechanism of action in colorectal cancer,
of CRC [62]. Other circRNAs such as circ-ITCH increased ITCH promotes ubiquitin-mediated
can also exert the function of miRNA sponge to degradation of Dvl2 and reduces the expression
suppress tumor proliferation. A research based on of the oncogene c-myc that inhibits the Wnt sig-
45 specimens of CRC found that circ-ITCH naling pathway and ultimately suppresses the
expression is abnormally lower than that of para- tumor growth [65]. In 51 cases of ESCC patients
cancerous tissue. The bioinformatics analysis with different staging, the expression of hsa_
predicted that circ-ITCH and its parent gene circ_0067934 in the tumor tissue is obviously
ITCH possess the same miR-214, miR-7, and higher than that in paired paracancerous tissues;
miR-20a binding sites, and the firefly luciferase and the high expression level of hsa_
reporter assay in CRC cell lines HCT116 and circ_0067934 is associated with the tumor stage
SW480 confirmed that circ-ITCH can obstruct (P = 0.025). ] Higher stage of the tumor tissue is
the Wnt signaling pathway to inhibit cancer cell companioned with higher expression of hsa_
proliferation by competitively absorbing miR-7 circ_0067934. Interference of hsa_circ_0067934
and miR-20a [63]. expression by siRNA in vitro suppresses the pro-
liferation and migration of ESCC cells and
3.2.3 Esophageal Cancer blocks cell cycle progression. Cell component
Researches on the regulation mechanism of cir- analysis and fluorescence in situ hybridization
cRNAs in esophageal cancer suggest that many confirm that the circRNAs are mainly located in
of them can function as miRNAs sponges. The the cytoplasm [66].
radiation resistance obtained during radiother-
apy is considered to be the most important factor 3.2.4 Pancreatic Cancer
that affects the therapeutic effect and stimulates Pancreatic cancer (PC) is a common cancer, but
local tumor recurrence. Based on this, Su et al. there is still lack of reliable biological markers
explored the differentially expressed circRNAs for early diagnosis and evaluation of prognosis.
in radioresistant esophageal cancer cells by Current studies on the relationship between cir-
using expression profiling and bioinformatics cRNAs and PC are insufficient. Recently,
14 Circular RNAs as Biomarkers for Cancer 177

researchers have explored the expression profiles downregulation of hsa_circ_0001649 was found
of circRNAs in four pancreatic ductal adenocar- associated with inhibition of cell proliferation
cinoma (PDAC) samples and matched adjacent [74]. In contrast, the upregulated expression of
normal tissues. The result revealed that a large hsa_circ_0005075 and Cdr1as promotes cell
number of circRNAs are abnormally expressed in adhesion and proliferation, respectively [75,
PDAC, suggesting that they may be involved in 76]. As the inhibitor and sponge of miR-7,
the initiation and progression of PDAC. This dis- Cdr1as can indirectly interrupt the PI3K/Akt/
covery provides potential biological targets for mTOR signaling pathway by targeting miR-7
the diagnosis and treatment of PDAC [67]. Li [77]. It has been suggested that Cdr1as may be
et al. analyzed the tissue samples from six employed as a prognostic biomarker for hepatic
patients with PDAC by microarray technology carcinoma and a therapeutic target for microvas-
and found that compared to normal pancreatic cular infiltration [77]. Huang et al. detected
tissues, abnormal expression of circRNAs cluster another potential marker hsa_circ_100338 by
was gathered in pancreatic cancer tissues. 209 applying a circRNA microarray and verified
upregulated and 142 downregulated circRNAs miR-141-3p as its direct downstream target
were screened out. Subsequently, 7 circRNAs through computer calculation and experimental
were analyzed with quantitative reverse tran- analysis. High expression level of hsa_
scriptase polymerase chain reaction (qRT-PCR) circ_100338 indicates the metastasis progres-
in 20 groups of PC and paracancerous tissues to sion and meanwhile makes an influence on the
confirm that the results were consistent with the cumulative survival rate [78]. A circRNA-­
microarray. GO analysis and pathway analysis miRNA-­mRNA network is constructed with five
suggested that some of the dysregulated cir- upregulation circRNAs in hepatic cancer. This
cRNAs are involved in molecular biological pro- network proposes that high levels of circFUT8
cesses in pancreatic cancer, including influence (hsa_circ_101368, hsa_circ_0003028), circ-
on endocytosis and meditation of abnormal ZFR (hsa_circ_103809, hsa_circ_10072088),
expression of VEGF pathway. These results and circ-IOP11 (hsa_circ_103847, hsa_
reveal the indispensable role of circRNAs in the circ_0007915) expression are probably associ-
malignant biological behavior of PC [68]. ated with the progression of hepatic cancer [79].
Researchers focusing on the genetics and epide-
miology of hepatocellular carcinoma (HCC)
3.3 Hepatic Cancer discovered that the expression of cir-ITCH in
HCC tissues is significantly lower than that in
Hepatic cancer generally refers to malignant matched paracancerous tissues; and suggested
tumor which originates in hepatocytes or intrahe- that HCC patients with a relatively high cir-
patic bile duct epithelium [47]. As one of the ITCH expression have a better prognosis.
common malignancies of the digestive system, Collectively, the results revealed that cir-ITCH
hepatic cancer is the fifth most common cancer has an inhibitory action on HCC and may be a
with the third most common mortality in the hopeful biomarker for the assessment of suscep-
world [69]. About 250,000 people worldwide are tibility and prognosis in patients with HCC [80].
estimated to lose their lives every year due to Modulating function of hsa_circ_0015756 on
hepatic cancer, among which China accounts for hepatoblastoma cell by acting as a miR-1250-3p
about 45% [70, 71]. The 5-year survival rate of sponge has also been found based on circRNAs
advanced hepatic cancer is about 10%. Due to the microarray analysis. Silencing hsa_
lack of effective early diagnostic tools, only circ_0015756 reduces viability, proliferation,
30–40% of the patients can be diagnosed and and invasiveness of hepatoblastoma cells [81].
take appropriate treatment in early stage [72, 73]. Fu et al. conducted a series of experiments to
A series of circRNAs associated with hepatic prove the point that the expression level of hsa_
cancer are gradually discovered. A noteworthy circ_0005986 is related to the tumor size, BCLC
178 L. Xia et al.

staging, and microvascular infiltration. The advanced or recurrent cases does not exceed 1
underlying mechanism of hsa_circ_0005986-­ year [86]. Clinical diagnosis of gynecologic
regulated carcinogenesis of HCC is by modulat- malignancies is mainly by physical signs, tumor
ing Notch1 expression through interaction with markers, B-ultrasonography, computed tomog-
miR-129-5p [82]. They further showed the low raphy (CT), and magnetic resonance imaging
expression of hsa_circ_0004018 is associated (MRI). Common gynecological tumor markers
with AFP level in serum, diameter, and differen- include cancer markers such as cancer antigen
tiation of tumor and Barcelona clinical stage in 125 and 19-9 (CA-125 and CA19-­9), but their
HCC. Remarkably, hsa_circ_0004018 has the specificity is doubtful. Researches have been try-
specific expression characteristics indicative of ing to find high-specificity diagnostic markers
different HCC stage in various chronic liver dis- among circRNAs, which is of great significance
eases [83]. Based on all these findings, there are for timely treatment of gynecological cancer,
reasons to believe that circRNAs combined with reduction of metastasis, and improvement of
traditional biomarkers can make a more accu- prognosis.
rate diagnostic method for hepatic cancer.
3.4.1 Breast Cancer
As previously mentioned, Cdr1as can indirectly
3.4 Gynecologic Cancer regulate the expression of miR-7 target gene,
while Cdr1as/miR-7 can affect tumorigenesis and
Gynecologic cancer is closely related to women development of tumor through multiple pathways.
mainly including breast cancer, endometrial can- Early studies have showed that the expression
cer, cervical cancer, and ovarian cancer. The level of endogenous miR-7 is negatively corre-
incidence of gynecological tumors has increased lated with Pak1 and is positively correlated with
year by year, which causes serious harm to wom- homeodomain transcription factor HOXD10. In
en’s physical and mental health. Breast cancer the transformation process of breast cancer from
has the highest incidence of women malignancy high invasion phenotype to low invasion pheno-
tumors in the world, which is also the biggest type, Pak1 protein levels gradually increase, while
factor leading to the death of female cancer miR-7 and HOXD10 gradually decrease. In
patients [84]. Breast cancer accounts for 23% of highly invasive breast cancer cells, miR-7 can
the global new female cancer cases and accounts inhibit their proliferation activity, invasiveness,
for 14% of the total number of global cancer and tumorigenic potential. These indicate that the
deaths [84, 85]. Ovarian cancer represents the miR-7/Pak1 pathway may play an important role
highest mortality rate among gynecologic cancer in the development of breast cancer. Therefore,
with no less than 204,000 new cases and 125,000 Cdr1as is considered also involved in the regula-
deaths cases per year [86]. Only a small propor- tion of breast cancer [89]. A total of 1155 differ-
tion of patients can be diagnosed in the early entially expressed circRNAs were screened out in
stage due to its concealed characteristics and 51 breast cancer patients by using the whole
lack of effective early diagnosis. In fact, more genome transcript profile technique, of which 715
than 70% of the patients are diagnosed in were upregulated and 440 were downregulated.
advanced stage; and their 5-year survival rate is The expression levels of hsa_circ_103110, hsa_
lower than 30% [87]. Endometrial cancer origi- circ_104689, and hsa_circ_104821 were elevated
nates from the endometrial epithelium. Its mor- in breast cancer tissues, while the expression lev-
tality is second only to ovarian cancer. It accounts els of hsa_circ_006054, hsa_circ_100219, and
for 7% of all malignant tumors in women, and hsa_circ_406697 were downregulated among the
the incidence of malignant tumors in female selected circRNAs. Further investigation of the
reproductive system is as high as 20–30% [88]. circRNAs targeting complementary miRNAs
Although early-stage endometrial cancer has a response elements revealed that progesterone
good prognosis, the median survival time for receptor (PR)-negative was related to the upregu-
14 Circular RNAs as Biomarkers for Cancer 179

lation of hsa_circ_104689 and hsa_circ_104821 while the cell apoptosis accelerates. The miR-­
and the downregulation of hsa_circ_406697. The 449a was identified as a related miRNA to
diagnostic accuracy of hsa_circ_100219 was the circ_000911 by using a biotin-labeled probe
highest with the AUC of 0.78 (95% CI: 0.69– method. Overexpression of cir_000911 in breast
0.88). Combined hsa_circ_006054, hsa_ cancer could increase the expression of Notch1,
circ_100219, and hsa_circ_406697 had a higher which is a functional target of miR-449a. Signal
judgment value for judging breast cancer (AUC: transduction reporter array and western blot anal-
0.82, 95% CI: 0.73–0.90) [90]. With the purpose ysis confirmed that NF-κB signaling transduction
to investigate the expression profile and possible is a functional target of the circ_000911/miR-­
regulatory mechanisms of oncogenic circRNAs in 449a pathway [93].
breast cancer, Liang et al. used circRNA microar-
ray to screen abnormally expressed circRNAs in 3.4.2 Reproductive System Tumors
breast cancer tissues and found that circ-ABCB10 By RNA sequencing of primary ovarian cancer,
was highly expressed in breast cancer tissues. The peritoneal metastases, and lymph node metasta-
authors then verify the result of the chip by using ses in three patients with ovarian cancer, cir-
a large amount of samples. The loss-of-­function cRNAs with significant differences in expression
experiments in vitro demonstrated that subtrac- in epithelial ovarian cancer were found, includ-
tion of circ-ABCB10 level in breast cancer cells ing many new genes such as HIPK2/3 and
can inhibit cell proliferation and promote apopto- ZKSCAN1. The number of differentially
sis. The bioinformatics technique was used to pre- expressed circRNAs is much higher than that of
dict the existence of complementary sequences in the corresponding linear mRNA in metastatic
circ-ABCB10 and miR-1271, which was then lesions. In addition, various cancer-associated
verified by luciferase reporter assays. Finally, it signaling pathways including NF-κB, PI3K/Akt,
was confirmed in breast cancer cells that the inhi- and TGF-β have the opposite expression trends in
bition of miR-1271 can restore function of circ- circRNAs and linear mRNAs. Consensus of cir-
ABCB10, demonstrating the spongy effect of cRNAs expression provides new candidates for
circ-ABCB10 on miR-1271 [91]. In a study inves- cancer treatment and prognosis [94]. Endometrial
tigating whether hypoxia regulates proliferation cancer (EC) and cervical cancer also belong to
through circRNAs, a hypoxic model in breast can- the female reproductive system malignancy.
cer cells was established. The increased expres- Lately, researchers used RNA sequencing tech-
sion of circ-­ DENND4C was detected under nology to identify EC-specific circular transcrip-
hypoxic conditions, whereas knockdown of tomes. By comparison, the overall abundance of
HIF1α could reduce the expression of circ- circRNAs in EC (14,707) was lower than that in
DENND4C. This confirmed the correlation normal endometrium (21,340). On this basis the
between circ-­ DENND4C and HIF1α in the researchers identified 120 differentially expressed
hypoxic model. It was further found that knock- circRNAs between EC tissues and normal endo-
down circ-­DENND4C can inhibit the abnormal metrial tissues by collecting and analyzing sam-
proliferation of breast cancer cells in anoxic envi- ples from 6 EC patients, in which unique hotspot
ronment, indicating that circ-DENND4C has the genes, such as cancer-specific ESR1 circular iso-
function of promoting breast cancer cell prolifera- forms, were regarded with the value of EC diag-
tion under hypoxic conditions. Finally, the expres- nosis and progress detection. The circular isoform
sion level of circ-DENND4C was found related to of DNAH14 may be involved in the regulation of
the tumor volume, and the larger tumors con- tumor-associated pathways. The DMD and
tained more circ-DENND4C [92]. As a notewor- DMBT1 genes undergone significant changes
thy downregulated circRNA in breast cancer cells, during generation of the circular transcript, sug-
when the expression of circ_000911 is enhanced, gesting that they may be involved in the patho-
the proliferation, migration, and invasion ability logical changes of EC [95]. The work of
of breast cancer cells are all inhibited, and mean- Abdelmohsen’s team demonstrated an example
180 L. Xia et al.

of a functional model of a protein regulation 4 Conclusion and Prospect

through circRNAs endogenous binding in cervi-
cal cancer. They used the RIP assay to identify CircRNAs is a type of closed-circular RNA mole-
circRNAs interacting with HuR in Hela cells; and cule widely distributed in the transcriptome; they
the most obviously changed candidate was participate in the regulation of numerous biologi-
circPABPN1(hsa_circ_0031288). Excess circ-­ cal activities in organisms. CircRNAs are not eas-
PABPN1 can prevent the binding of HuR to lin- ily degraded by nucleases and are more stable than
ear PABPN1 mRNA and hence inhibits the linear RNAs, which provide a foundation for them
translation of HuR [96]. to be novel biomarkers for tumor diagnosis. The
regulatory expression mechanisms of circRNAs
are diverse. They can serve as “miRNA sponges”
3.5 Other Cancer to perform posttranscriptional regulation by com-
petitively binding to miRNAs, interact with snRNP
Some circRNAs also have the potential of bio- or RNA polymerase II in the nucleus to regulate
logical markers in other tumors. For instance, the transcription, or bind to transcription factors and
expression level of has_circ_104912 is signifi- competitively regulate classic RNA splicing.
cantly downregulated in laryngeal squamous cell CircRNAs accumulate in cells and release into
carcinoma, while the expression level of has_ exosomes and plasma. The amount of circRNAs
circ_100855 is significantly increased [97]. released into exosomes from tumor tissue is three
circ_100290 can function in oral squamous cell times more than that in tumor tissues [24].
carcinoma as a molecular sponge of miR-29 fam- CircRNAs play an irreplaceable regulatory
ily [98]; high expression of circ-TTBK2 in gli- role in the complex life process. Their abnormal
oma tissue may promote the development of expression can induce or impede the occurrence
glioma [99]. and development of cancer. They are promising
Acute pregranulocyte leukemia chromosomal biomarkers and even therapeutic targets for can-
translocation t(15;17)(q24;q21) leads to the for- cer. With the continuous development of high-­
mation of a key oncogenic fusion protein PML-­ throughput sequencing and bioinformatics
RARα, and the circRNA, f-circRNA, forms after technologies, the formation and function of cir-
the translocation of this chromosome can also be cRNAs and their relationship with cancer have
carcinogenic [100]. Circ-TRIM24 (hsa_ gradually attracted widespread attention in the
circ_0082582) and circ-FAM169A (hsa_ scientific community and become a hotspot for
circ_0007158) are significantly downregulated in researches of clinical disease after miRNAs and
bladder cancer tissues, while circ-BC048201 long noncoding RNAs (LncRNAs). For the
(hsa_circ_0061265), circ-PTK2 (hsa_ moment, Circbase, Circ2Traits, CircNet, and
circ_0005273), circ-ZFR (hsa_circ_0072088), other databases have included information of
and circ-TCF25 (hsa_circ_0041103) are signifi- more than 100,000 circRNAs, which can assist us
cantly upregulated, and the circTCF25-miR-­ to predict the regulatory relationship of circRNA-­
103a-3p/miR-107-CDK6 pathway is suggested miRNA-­mRNA. This tool is extremely conve-
as an important regulatory axis in bladder cancer nient for us to study circRNA systematically. At
[101]. 23 high- and 48 low-expression circRNAs the same time, methods for constructing or
and their corresponding binding sites for 354 ­ interfering with circRNAs have emerged and
bindable miRNA sequences are identified in matured, making it possible to artificially regu-
basal cell carcinoma [102]. The phenomenon of late the expression of intracellular circRNAs and
abnormal expression of these circRNAs in vari- being helpful to further explore the role of cir-
ous tumor tissues indicates that circRNAs have cRNAs in tumor cells.
an inseparable relationship with the occurrence By summarizing the researches on the rela-
and development of tumors. tionship between circRNAs and cancer in recent
years, we find that many circRNAs have abnormal
14 Circular RNAs as Biomarkers for Cancer 181

expression in the tumor tissue/blood/exosome of circ_002509 exhibit a high diagnostic value for
tumor patients, among which “star molecules” gastric cancer. Circ-BANP is a potential diagnos-
such as circ-ITCH and Cdr1as have differential tic marker for colorectal cancer, and the elevation
expression and play a regulatory role in different of has_circ_001569 in colorectal cancer is proven
types of cancer. Cdr1as has the potential for being helpful to the diagnosis and staging of disease.
a biomarker for gastric cancer, hepatic cancer, The rise of hsa_circ_0067934 in esophageal can-
colorectal cancer, cervical cancer, and so on. Hsa_ cer suggests the progression of tumor staging. The
circ_0000064 and hsa_circ_0013958 are related same phenomenon is found in other cancers. In
to tumor lymph node metastasis and TNM staging certain kind of cancer, there are many circRNAs
in lung cancer. Circ-­100876 is associated with the with different expression changes; on the other
prognosis of lung cancer patients. The reduction hand, the same circRNA can correspond to differ-
of has_circ_002059 in gastric cancer has a predic- ent types of cancer. The summary of cancer-­
tive role in distant metastasis and TNM staging. related circRNAs that have been uncovered is
The combination of hsa_circ_0000096 and hsa_ shown in Table 14.1.

Table 14.1 Cancer-associated circRNAs with its characteristics and related genes
Change in
Cancer type The name of circRNAs cancer Features, related molecules, and pathways
Lung cancer hsa_circ_0000064 Up Promote cancer cell proliferation and
invasion
circRNA-100876 Up High expression level suggests shorter
survival
circ-HIPK3 Up miR-379
hsa_circ_0013958 Up miR-314
Cdr1as (ciRS-7) / miR-7
circ-ITCH Up miR-7 and miR-124
hsa_circ_0043256 Down miR-1252
ciR-Sry [103] / miR-138
circ-ZEB1.5 Down miR-200a-3p [104]
circ-ZEB1.19 Down
circ-ZEB1.17 Down
circ-ZEB1.33 Down
Gastric cancer circPVT1 Up Expression level positively correlated with
survival rate; miRNA-125
hsa_circ_0000096 Down Promote cancer cell proliferation, cycle,
and migration
miR-224 and miR-200a
has_circ_002059 Down Potential for predictive treatment and
associated with distant metastases, TNM
staging, gender, and age
has_circ_0001649 Down Potential for predictive treatment and
correlated with pathological differentiation
has_circ_0044516 Up 191 miRNAs, COL1A1
has_circ_0014717 Down Related to distant metastasis, tumor
staging, CA19-9
hsa_circ_0000190 Down Related to distant metastasis, tumor
staging, CA19-9
hsa_circ_0076305 Down PGC
hsa_circ_0037362 Down C16orf73
hsa_circ_0035431 Down CGNL1
hsa_circ_0000140 [105] Down KIAA0907
(continued)
182 L. Xia et al.

Table 14.1 (continued)

Change in
Cancer type The name of circRNAs cancer Features, related molecules, and pathways
Colorectal cancer circ-BANP Up BANP
hsa_circ_0000069 Up Related to age, tumor size, lymph node
metastasis and TNM staging
hsa_circ_001569 Up Have a positive regulatory role in cancer
cell proliferation and invasion; miR-145
circCCDC66 Up miR-33b, miR-93, CCDC66
circ-ITCH Down Overexpression can inhibit cancer cell
proliferation; miR-7, miR-20a,ITCH
hsa_circ_001988 [106] Down Associated with colon cancer cell
differentiation and neurotrophic invasion
hsa_circ_0001946 [43] / CDR1
hsa_circ_0001141 [107] Down ITCH
hsa_circ_0006229 Down TNS3
Esophageal cancer circRNA_001059 Up LIN52
circRNA_100385 Up PRRX1
circRNA_104983 Up NHS
circRNA_101877 Down RFWD3
circRNA_102913 Down ATIC
circRNA_000167 Down RPPH1
circRNA_000695 Down EEFSEC
cir-ITCH Down miR-7, miR-17, miR214; Wnt/β-catenin
signaling pathways
hsa_circ_0067934 Up High expression level is associated with
the tumor stage
Pancreatic cancer Ci-sirt7 [108] / sirt7
hsa_circ_0001946 [68] Up hsa_circ_0005785-miR-181a/
miR-181b-VEGF
Hepatic cancer hsa_circ_0001649 Down Suppress cancer cell proliferation
hsa_circ_0005075 Up Promote cancer cell adhesion
Cdr1as Up miR-7; PI3K/Akt/mTOR
hsa_circ_100338 Up Associated with cumulative survival rate;
miR-141-3p
circFUT8 (hsa_circRNA_101368, Up Associated with the progression of hepatic
hsa_circ_0003028) cancer
circZFR Up Associated with the progression of hepatic
(hsa_circRNA_103809,hsa_ cancer
circ_10072088)
circIOP11 (hsa_circRNA_103847, Up Associated with the progression of hepatic
hsa_circ_0007915) cancer
cir-ITCH Down High expression level is associated with
good prognosis
hsa_circ_0015756 Up miR-1250-3p
hsa_circ_0005986 Down Expression level is related to the tumor
size, BCLC staging, and microvascular
infiltration miR-129-5p, Notch1
hsa_circ_0004018 Down Has the specific prompting for different
HCC stage
(continued)
14 Circular RNAs as Biomarkers for Cancer 183

Table 14.1 (continued)

Change in
Cancer type The name of circRNAs cancer Features, related molecules, and pathways
Breast cancer hsa_circ_103110 Up hsa_miR_339_5p
hsa_circ_104689 Up Relates to progesterone receptor negative
hsa_circ_104821 Up Relates to progesterone receptor negative
hsa_circ_100219 Down The AUC for diagnostic accuracy is 0.78
(95% CI: 0.69-0.88).
hsa_circ_406697 Down Relates to progesterone receptor negative
circ-ABCB10 Up miR-1271
circDENND4C Up Promotes cancer cell proliferation under
hypoxic conditions; HIF1α
circRNA-000911 Down miR-449a, Notch1
Ovarian cancer ciR-Sry [109] / miR-138
Endometrial cancer hsa_circ_0031288(circPABPN1) Up PABPN1
Laryngeal has_circ_104912 Down /
squamous cell has_circ_100855 Up /
carcinoma
Oral squamous cell circRNA_100290 Up miR-29 family, CDK6
carcinoma ci-mcm5 [110] / mcm5
Skin squamous cell Hsa_circ_103736 [111] Up miR-876-5p,miR-192-3p,miR-34b-­
carcinoma 3p,miR-34c-3p,miR-181b-3p
Hsa_circ_103737 [111] Up miR-877-3p,miR-876-5p,miR-181b-2-­
3p,miR-181b-3p,miR-627-3p
Hsa_circ_101555 [111] Up miR-644a,miR-485-5p,miR-889-5p,miR-­
329-­5p,miR-148a-5p
Glioma circ-TTBK2 Up Promotes the development of glioma
Acute f-circRNA / Forms after the translocation of
pregranulocyte l chromosome, has carcinogenic effects
eukemia
Bladder cancer circ-TRIM24 (hsa_circ_0082582) Down /
circFAM169A (hsa_circ_0007158) Down /
circBC048201 (hsa_circ_0061265) Up /
circPTK2 (hsa_circ_0005273) Up /
circZFR (hsa_circ_0072088) Up /
circTCF25 (hsa_circ_0041103) Up circTCF25-miR-103a-3p/miR-107-CDK6
and PI3K-Akt pathway
Cholangio ciR-Sry [112] / miR-138, RhoC
carcinoma
Basal cell Hsa_circ_0008732 [102] Up OncomiR-1, miR-19-92 family
carcinoma
Osteosarcoma Has_circ_0016347 [113] Up miR-214
Clear cell carcinoma circHIAT1 [114] up miR-195-5p,miR-29a-3p,miR-29c-3p
of kidney

With the continuous expansion of existing cRNAs function. We have reason to believe that
researches, the relationship between circRNAs with the discovery of more and more circRNAs,
and cancer networks have become increasingly the diagnosis, progression, and prognosis evalua-
clear. However, our knowledge on circRNAs is tion system of various cancer based on circRNAs
only the tip of the iceberg. There is still a long will be mature, which is of great significance for
way to go to uncover the mysterious veil of cir- cancer treatment.
184 L. Xia et al.

Acknowledgments This work was supported by the ing to new data of the Czech National Cancer
grants from National Natural Science Foundation of Registry. Klin Onkol 27(Suppl 2):19–39
China (81670571 and 81370559 to C. Yang; 81400635 to 11. Rosenberg AR, Kroon L, Chen L et al (2015)
F. Wang), Joint Projects in Major Diseases funding from Insurance status and risk of cancer mortality among
Shanghai Municipal Commission of Health and Family adolescents and young adults. Cancer 121(8):1279
Planning (2014ZYJB0201 to C. Yang), Joint Projects for 12. Torre LA, Bray F, Siegel RL et al (2015) Global can-
Novel Frontier Technology in Shanghai Municipal cer statistics, 2012. CA Cancer J Clin 65(2):87–108
Hospital from Shanghai Municipal Commission of Health 13. Kirwan A, Utratna M, O'Dwyer ME et al (2015)
and Family Planning (SHDC12014122 to C. Yang), Glycosylation-based serum biomarkers for can-
Shanghai Medical Guide Project from Shanghai Science cer diagnostics and prognostics. Biomed Res Int
and Technology Committee (14411971500 to F. Wang), 2015:490531
grants from Chinese Foundation for Hepatitis Prevention 14. Huang G, Li S, Yang N et al (2017) Recent prog-
and Control (TQGB20140141 to F. Wang), and funds ress in circular RNAs in human cancers. Cancer Lett
from Shanghai Innovation Program (12431901002 to 404:8
C. Yang). 15. Xin Z, Ma Q, Ren S et al (2017) The understanding
of circular RNAs as special triggers in carcinogen-
Competing Financial Interests The authors declare no esis. Brief Funct Genomics 16(2):80
competing financial interests. 16. Boutros PC (2015) The path to routine use of
genomic biomarkers in the cancer clinic. Genome
Res 25(10):1508–1513
17. Kalia M (2015) Biomarkers for personalized
oncology: recent advances and future challenges.
References Metabolism 64(3 Suppl 1):S16–S21
18. Hutvagner G, Simard MJ (2008) Argonaute proteins:
key players in RNA silencing. Nat Rev Mol Cell
1. Sanger HL, Klotz G, Riesner D et al (1976)
Biol 9(9):22–32
Viroids are single-stranded covalently closed cir-
19. Ebert MS, Neilson JR, Sharp PA (2007) MicroRNA
cular RNA molecules existing as highly base-­
sponges: competitive inhibitors of small RNAs in
paired rod-like structures. Proc Natl Acad Sci USA
mammalian cells. Nat Methods 4(9):721
73(11):3852–3856
20. Rupaimoole R, Calin GA, Lopez-Berestein G
2. Sorber K, Dimon MT, Derisi JL (2011) RNA-Seq
et al (2016) miRNA deregulation in cancer cells
analysis of splicing in Plasmodium falciparum
and the tumor microenvironment. Cancer Discov
uncovers new splice junctions, alternative splicing
6(3):235–246
and splicing of antisense transcripts. Nucleic Acids
21. Leva GD, Garofalo M, Croce CM (2014) microR-
Res 39(9):3820–3835
NAs in cancer. Annu Rev Pathol 9(1):287
3. Memczak S, Jens M, Elefsinioti A et al (2013)
22. Rupaimoole R, Slack FJ (2017) MicroRNA thera-
Circular RNAs are a large class of animal RNAs
peutics: towards a new era for the management of
with regulatory potency. Nature 495(7441):333
cancer and other diseases. Nat Rev Drug Discov
4. Hsiao KY, Sun HS, Tsai SJ (2017) Circular RNA –
16(3):203
New member of noncoding RNA with novel func-
23. Hansen TB, Jensen TI, Clausen BH et al (2013)
tions. Exp Biol Med 242(11):1136
Natural RNA circles function as efficient microRNA
5. Zhang Y, Zhang XO, Chen T et al (2013) Circular
sponges. Nature 495(7441):384–388
intronic long noncoding RNAs. Mol Cell
24. Li Y, Zheng Q, Bao C et al (2015) Circular RNA is
51(6):792–806
enriched and stable in exosomes: a promising bio-
6. Li Z, Huang C, Bao C et al (2015) Exon-intron cir-
marker for cancer diagnosis. Cell Res 25(8):981
cular RNAs regulate transcription in the nucleus. Nat
25. Jeck WR, Sorrentino JA, Wang K et al (2013)
Struct Mol Biol 22(3):256
Circular RNAs are abundant, conserved, and asso-
7. Rybakwolf A, Stottmeister C, Glažar P et al (2015)
ciated with ALU repeats. Rna-a Publication of the
Circular RNAs in the mammalian brain are highly
Rna. Society 19(2):141–157
abundant, conserved, and dynamically expressed.
26. Schwanhäusser B, Busse D, Li N et al (2013)
Mol Cell 58(5):870
Corrigendum: global quantification of mammalian
8. Guo JU, Agarwal V, Guo H et al (2014) Expanded
gene expression control. Nature 495(7439):126–127
identification and characterization of mammalian
27. Szabo L, Morey R, Palpant NJ et al (2015)
circular RNAs. Genome Biol 15(7):409
Statistically based splicing detection reveals neural
9. Ferlay J, Soerjomataram I, Dikshit R et al (2015)
enrichment and tissue-specific induction of circular
Cancer incidence and mortality worldwide: sources,
RNA during human fetal development. Genome Biol
methods and major patterns in GLOBOCAN 2012.
16(1):1–26
Int J Cancer 136(5):E359–E386
28. Rybak-Wolf A, Stottmeister C, Glažar P et al (2015)
10. Dusek L, Muzik J, Maluskova D et al (2014)
Circular RNAs in the mammalian brain are highly
Epidemiology of screening-targeted cancers accord-
14 Circular RNAs as Biomarkers for Cancer 185

abundant, conserved, and dynamically expressed. RNA hsa_circ_0043256 in non-small cell lung can-
Mol Cell 58(5):870 cer. Biochem Biophys Res Commun 493(3):1260
29. Li Z, Huang C, Bao C et al (2015) Exon-intron cir- 46. Aoyama T, Nishikawa K, Fujitani K et al (2016)
cular RNAs regulate transcription in the nucleus. Nat Early results of a randomized two-by-two factorial
Struct Mol Biol 22(3):256–264 phase II trial comparing neoadjuvant chemother-
30. Ashwal-Fluss R, Meyer M, Pamudurti NR et al apy with two and four courses of cisplatin/S-1 and
(2014) circRNA biogenesis competes with Pre-­ docetaxel/cisplatin/S-1 as neoadjuvant chemother-
mRNA Splicing. Mol Cell 56(1):55–66 apy for locally advanced gastric cancer. Ann Oncol
31. Conn S, Pillman K, Toubia J et al (2015) The RNA 28(8):1876–1881
binding protein quaking regulates formation of cir- 47. Chen W, Zheng R, Baade PD et al (2016) Cancer sta-
cRNAs. Cell 160(6):1125–1134 tistics in China, 2015. CA Cancer J Clin 66(2):115
32. Dudekula DB, Panda AC, Grammatikakis I et al 48. Acebron SP, Niehrs C (2016) β-Catenin-independent
(2016) CircInteractome: a web tool for exploring roles of Wnt/LRP6 signaling. Trends Cell Biol
circular RNAs and their interacting proteins and 26(12):956–967
microRNAs. RNA Biol 13(1):34–42 49. Sui W, Shi Z, Xue W et al (2017) Circular RNA
33. Howlader N, Noone AM, Krapcho M, et al and gene expression profiles in gastric cancer
(2012) Previous version: SEER cancer statistics based on microarray chip technology. Oncol Rep
review,1975–2009(Vintage 2009 Populations). 37(3):1804–1814
National Cancer Institute. http://seer.cancer.gov/ 50. Chen J (2017) 250P Circular RNA profile identifies
csr/1975_2009_pops09/ circPVT1 as a proliferative factor and prognostic
34. Ferlay J, Soerjomataram I, Dikshit R et al (2015) marker in gastric cancer. Cancer Lett 388:208–219
Cancer incidence and mortality worldwide: sources, 51. Li P, Chen H, Chen S et al (2017) Circular RNA
methods and major patterns in GLOBOCAN 2012. 0000096 affects cell growth and migration in gastric
Int J Cancer 136(5):E359 cancer. Br J Cancer 116(5):626–633
35. Institute NC (2015) Cancer of the pancreas – SEER 52. Peifei L, Shengcan C, Huilin C et al (2015) Using
stat fact sheets. http://seer.cancer.gov/statfacts/html circular RNA as a novel type of biomarker in the
36. Maemondo M, Inoue A, Kobayashi K et al (2010) screening of gastric cancer. Clin Chim Acta 444:132
Gefitinib or chemotherapy for non-small-cell 53. Zhang Y, Li J, Yu J et al (2017) Circular RNAs
lung cancer with mutated EGFR. N Engl J Med signature predicts the early recurrence of stage III
362(25):2380 gastric cancer after radical surgery. Oncotarget
37. Wang BY, Huang JY, Cheng CY et al (2013) Lung 8(14):22936–22943
cancer and prognosis in taiwan: a population-based 54. Li WH, Song YC, Zhang H et al (2017) Decreased
cancer registry. J Thorac Oncol 8(9):1128–1135 expression of Hsa_circ_00001649 in gastric can-
38. Luo YH, Zhu XZ, Huang KW et al (2017) Emerging cer and its clinical significance. Dis Markers
roles of circular RNA hsa_circ_0000064 in the pro- 2017:4587698
liferation and metastasis of lung cancer. Biomed 55. Huang YS, Jie N, Zou KJ et al (2017) Expression
Pharmacother 96:892–898 profile of circular RNAs in human gastric cancer tis-
39. Yao JT, Zhao SH, Liu QP et al (2017) Over-­ sues. Mol Med Rep 16(3):2469–2476
expression of CircRNA_100876 in non-small cell 56. Shao Y, Li J, Lu R et al (2017) Global circular RNA
lung cancer and its prognostic value. Pathol Res expression profile of human gastric cancer and its
Pract 213(5):453–456 clinical significance. Cancer Med 6(6):1173–1180
40. Fang T, Yun W, Zhe X et al (2017) Circular RNA 57. Chen S, Li T, Zhao Q et al (2017) Using circular
CircHIPK3 promotes NCI-H1299 and NCI-H2170 RNA hsa_circ_0000190 as a new biomarker in
cell proliferation through miR-379 and its target the diagnosis of gastric cancer. Clin Chim Acta
IGF1. Chin J Lung Cancer 20(7):459 466:167–171
41. Zhu X, Wang X, Wei S et al (2017) hsa_circ_0013958: 58. Bachmayrheyda A, Reiner AT, Auer K et al (2015)
a circular RNA and potential novel biomarker for Correlation of circular RNA abundance with pro-
lung adenocarcinoma. FEBS J 284(14):2170 liferation – exemplified with colorectal and ovarian
42. Zhao J, Li L, Wang Q et al (2017) CircRNA expres- cancer, idiopathic lung fibrosis, and normal human
sion profile in early-stage lung adenocarcinoma tissues. Sci Rep 5(8057):8057
patients. Cell Physiol Biochem 44(6):2138–2146 59. Zhu M, Xu Y, Chen Y et al (2017) Circular BANP, an
43. Hansen TB, Kjems J, Damgaard CK (2013) Circular upregulated circular RNA that modulates cell prolif-
RNA and miR-7 in cancer. Cancer Res 73(18):5609 eration in colorectal cancer. Biomed Pharmacother
44. Wan L, Zhang L, Fan K et al (2016) Circular RNA-­ 88:138
ITCH suppresses lung cancer proliferation via inhib- 60. Guo J, Jin L, Zhu C et al (2016) Comprehensive pro-
iting the Wnt/β-Catenin pathway. Biomed Res Int file of differentially expressed circular RNAs reveals
2016(1):1579490 that hsa_circ_0000069 is upregulated and promotes
45. Tian F, Yu CT, Ye WD et al (2017) Cinnamaldehyde cell proliferation, migration, and invasion in colorec-
induces cell apoptosis mediated by a novel circular tal cancer. Onco Targets Ther 9:7451–7458
186 L. Xia et al.

61. Xie H, Ren X, Xin S et al (2016) Emerging roles 77. Xu L, Zhang M, Zheng X et al (2017) The circular
of circRNA_001569 targeting miR-145 in the prolif- RNA ciRS-7 (Cdr1as) acts as a risk factor of hepatic
eration and invasion of colorectal cancer. Oncotarget microvascular invasion in hepatocellular carcinoma.
7(18):26680 J Cancer Res Clin Oncol 143(1):1–11
62. Hsiao KY, Lin YC, Gupta SK et al (2017) Noncoding 78. Huang XY, Huang ZL, Xu YH et al (2017)
effects of circular RNA CCDC66 promote colon can- Comprehensive circular RNA profiling reveals the
cer growth and metastasis. Cancer Res 77(9):2339 regulatory role of the circRNA-100338/miR-141-3p
63. Huang G, Zhu H, Shi Y et al (2015) cir-ITCHPlays an pathway in hepatitis B-related hepatocellular carci-
inhibitory role in colorectal cancer by regulating the noma. Sci Rep 7(1):5428
Wnt/β-catenin pathway. PLoS One 10(6):e0131225 79. Ren S, Xin Z, Xu Y et al (2017) Construction and
64. Su H, Lin F, Xia D et al (2016) Profiling and bioin- analysis of circular RNA molecular regulatory net-
formatics analyses reveal differential circular RNA works in liver cancer. Cell Cycle 16(22):2204–2211
expression in radioresistant esophageal cancer cells. 80. Guo W, Zhang J, Zhang D et al (2017)
J Transl Med 14(1):225 Polymorphisms and expression pattern of circu-
65. Li F, Zhang L, Li W et al (2015) Circular RNA ITCH lar RNA circ-ITCH contributes to the carcino-
has inhibitory effect on ESCC by suppressing the genesis of hepatocellular carcinoma. Oncotarget
Wnt/β-catenin pathway. Oncotarget 6(8):6001 8(29):48169–48177
66. Xia W, Qiu M, Chen R et al (2016) Circular RNA 81. Liu BH, Zhang BB, Liu XQ et al (2018) Expression
has_circ_0067934 is upregulated in esophageal profiling identifies circular RNA signature in hepa-
squamous cell carcinoma and promoted prolifera- toblastoma. Cell Physiol Biochem 45(2):706
tion. Sci Rep 6:35576 82. Fu L, Chen Q, Yao T et al (2017) Hsa_circ_0005986
67. Qu S, Song W, Yang X et al (2015) Microarray inhibits carcinogenesis by acting as a miR-129-5p
expression profile of circular RNAs in human pan- sponge and is used as a novel biomarker for hepato-
creatic ductal adenocarcinoma. Genomics Data cellular carcinoma. Oncotarget 8(27):43878–43888
5(C):385 83. Fu L, Yao T, Chen Q et al (2017) Screening dif-
68. Li H, Hao X, Wang H et al (2016) Circular RNA ferential circular RNA expression profiles reveals
expression profile of pancreatic ductal adenocarci- hsa_circ_0004018 is associated with hepatocellular
noma revealed by microarray. Cell Physiol Biochem carcinoma. Oncotarget 8(35):58405–58416
40(6):1334 84. Center MM (2011) Global cancer statistics. CA
69. Wang FS, Fan JG, Zhang Z et al (2014) The global Cancer J Clin 61(2):69
burden of liver disease: the major impact of China. 85. Goto N, Hiyoshi H, Ito I et al (2014) Identification
Hepatology 60(6):2099 of a novel compound that suppresses breast can-
70. Nguyen VTT, Law MG, Dore GJ (2009) Hepatitis cer invasiveness by inhibiting transforming growth
B-related hepatocellular carcinoma: epidemiologi- factor-β signaling via estrogen receptor α. J Cancer
cal characteristics and disease burden. J Viral Hepat 5(5):336–343
16(7):453–463 86. Grushko TA, Filiaci VL, Mundt AJ et al (2008) An
71. Zhang JY (2007) Mini-array of multiple tumor-­ exploratory analysis of HER-2 amplification and
associated antigens to enhance autoantibody overexpression in advanced endometrial carcinoma:
detection for immunodiagnosis of hepatocellular a Gynecologic Oncology Group study. Gynecol
carcinoma. Autoimmun Rev 6(3):143 Oncol 108(1):3–9
72. Samuel M, Chow PK, Chan SYE et al (2009) 87. Wang H, Mu X, Zhou S et al (2014) NEDD9 overex-
Neoadjuvant and adjuvant therapy for surgical pression is associated with the progression of and an
resection of hepatocellular carcinoma. Cochrane unfavorable prognosis in epithelial ovarian cancer.
Database Syst Rev 1(1):CD001199 Hum Pathol 45(2):401–408
73. Crissien AM, Frenette C (2014) Current manage- 88. Tangjitgamol S, Khunnarong J, Srijaipracharoen S
ment of hepatocellular carcinoma. Gastroenterol (2014) Medical morbidities in endometrial cancer
Hepatol 10(3):153 patients. Int J Gynecol Cancer 24(9):1623–1627
74. Qin M, Liu G, Huo X et al (2016) Hsa_circ_0001649: 89. Reddy SD, Ohshiro K, Rayala SK et al (2008)
a circular RNA and potential novel biomarker MicroRNA-7, a homeobox D10 target, inhibits
for hepatocellular carcinoma. Cancer Biomark p21-activated kinase 1 and regulates its functions.
16(1):161–169 Cancer Res 68(20):8195
75. Shang X, Li G, Liu H et al (2016) Comprehensive cir- 90. Lü L, Jian S, Shi P et al (2017) Identification of cir-
cular RNA profiling reveals that hsa_circ_0005075, cular RNAs as a promising new class of diagnostic
a new circular RNA biomarker, is involved in biomarkers for human breast cancer. Oncotarget
hepatocellular crcinoma development. Medicine 8(27):44096–44107
95(22):e3811 91. Liang HF, Zhang XZ, Liu BG et al (2017) Circular
76. Lei Y, Gong X, Lei S et al (2016) The circular RNA RNA circ-ABCB10 promotes breast cancer prolifer-
Cdr1as Act as an oncogene in hepatocellular carci- ation and progression through sponging miR-1271.
noma through targeting miR-7 expression. PLoS Am J Cancer Res 7(7):1566–1576
One 11(7):e0158347
14 Circular RNAs as Biomarkers for Cancer 187

92. Liang G, Liu Z, Tan L et al (2017) HIF1α-associated 104. Liu YC, Li JR, Sun CH et al (2016) CircNet: a data-
circDENND4C promotes proliferation of breast can- base of circular RNAs derived from transcriptome
cer cells in hypoxic environment. Anticancer Res sequencing data. Nucleic Acids Res 44(D1):D209
37(8):4337–4343 105. Peifei L, Shengcan C, Huilin C et al (2015) Using
93. Wang H, Xiao Y, Wu L et al (2018) Comprehensive circular RNA as a novel type of biomarker in
circular RNA profiling reveals the regulatory role of the screening of gastric cancer. Clin Chim Acta
the circRNA-000911/miR-449a pathway in breast 444:132–136
carcinogenesis. Int J Oncol 52(3):743–754 106. Wang X, Zhang Y, Huang L et al (2015) Decreased
94. Ahmed I, Karedath T, Andrews SS et al (2016) expression of hsa_circ_001988 in colorectal cancer
Altered expression pattern of circular RNAs in pri- and its clinical significances. Int J Clin Exp Pathol
mary and metastatic sites of epithelial ovarian carci- 8(12):16020–16025
noma. Oncotarget 7(24):36366–36381 107. Huang G, Hua Z, Shi Y et al (2015) cir-ITCH plays an
95. Chen BJ, Byrne FL, Takenaka K et al (2018) inhibitory role in colorectal cancer by regulating the
Analysis of the circular RNA transcriptome in endo- Wnt/β-catenin pathway. PLoS One 10(6):e0131225
metrial cancer. Oncotarget 9(5):5786–5796 108. Mcglynn LM, Mccluney S, Jamieson NB et al
96. Abdelmohsen K, Panda AC, Munk R et al (2017) (2015) SIRT3 & SIRT7: potential novel biomark-
Identification of HuR target circular RNAs uncovers ers for determining outcome in pancreatic cancer
suppression of PABPN1 translation by CircPABPN1. patients. PLoS One 10(6):e0131344
RNA Biol 14(3):361–369 109. Yeh YM, Chuang CM, Chao KC et al (2013)
97. Xuan L, Qu L, Han Z et al (2016) Circular RNA: MicroRNA-138 suppresses ovarian cancer cell inva-
a novel biomarker for progressive laryngeal cancer. sion and metastasis by targeting SOX4 and HIF-1α.
Am J Transl Res 8(2):932 Int J Cancer 133(4):867–878
98. Chen L, Zhang S, Wu J et al (2017) circRNA_100290 110. Yu SY, Wang YP, Chang JY et al (2014) Increased
plays a role in oral cancer by functioning as a sponge expression of MCM5 is significantly associated
of the miR-29 family. Oncogene 36(32):4551–4561 with aggressive progression anda poor prognosis of
99. Zheng J, Liu X, Xue Y et al (2017) TTBK2 circu- oral squamous cell carcinoma. J Oral Pathol Med
lar RNA promotes glioma malignancy by regulat- 43(5):344–349
ing miR-217/HNF1β/Derlin-1 pathway. J Hematol 111. Sand M, Bechara FG, Gambichler T et al (2016)
Oncol 10(1):52 Circular RNA expression in cutaneous squamous
100. Guarnerio J, Bezzi M, Jeong JC et al (2016) cell carcinoma. J Dermatol Sci 83(3):210–218
Oncogenic role of fusion-circRNAs derived from 112. Wang Q, Tang H, Yin S et al (2013) Downregulation
cancer-associated chromosomal translocations. Cell of microRNA-138 enhances the proliferation,
165(2):289–302 migration and invasion of cholangiocarcinoma cells
101. Zhong Z, Lv M, Chen J (2016) Screening differential through the upregulation of RhoC/p-ERK/MMP-2/
circular RNA expression profiles reveals the regula- MMP-9. Oncol Rep 29(5):2046–2052
tory role of circTCF25-miR-103a-3p/miR-­107-­CDK6 113. Jin H, Jin X, Zhang H et al (2017) Circular RNA
pathway in bladder carcinoma. Sci Rep 6:30919 hsa-circ-0016347 promotes proliferation, invasion
102. Sand M, Bechara FG, Sand D et al (2016) and metastasis of osteosarcoma cells. Oncotarget
Circular RNA expression in basal cell carcinoma. 8(15):25571–25581
Epigenomics 8(5):619–632 114. Wang K, Sun Y, Tao W et al (2017) Androgen recep-
103. Yang H, Luo J, Liu Z et al (2015) MicroRNA-138 tor (AR) promotes clear cell renal cell carcinoma
regulates DNA damage response in small cell lung (ccRCC) migration and invasion via altering the
cancer cells by directly targeting H2AX. Cancer circHIAT1/miR-195-5p/29a-3p/29c-3p/CDC42 sig-
Investig 33(4):126 nals. Cancer Lett 394:1–12
 Part VI
Circular RNAs and Human Diseases
 Circular RNAs in Cardiovascular
Diseases 15
Lijun Wang, Xiangmin Meng, Guoping Li,
Qiulian Zhou, and Junjie Xiao

Abstract cardiovascular system regulatory effects,

Circular RNAs (circRNAs), a group of circu- including act as miRNA sponges, interaction
lar RNA molecules with a 3′,5′-phosphodies- with RNA-binding proteins, regulated by
ter bond at the junction site, are generated by RNA-binding proteins and serve as biomark-
back-splicing of precursor mRNAs. Most of ers. In addition, potential mechanisms under-
the circular RNAs originate from the exon lying the regulatory role of circRNAs in
region of the encoded protein, and some are cardiovascular diseases will be discussed.
derived from intron regions, antisense tran-
scripts, or long noncoding RNAs. Circular Keywords
RNAs are abundantly in eukaryotic transcrip- Circular RNA · microRNA · RNA-binding
tome and participate in various biological pro- protein · Cardiovascular diseases
cesses. It is closely associated with various
diseases such as tumors, diabetes, nervous
system diseases, and cardiovascular diseases. 1 Introduction
In cardiovascular system, numerous circRNAs
have been identified and involved in important Circular RNAs (circRNAs) were first identified
processes of cardiovascular development and as by-products of abnormal splicing with limited
diseases. Here we will review the latest functional potential [1–3]. Until 2012, with the
research progress of circular RNA in cardio-
vascular diseases. Also, we will outline the
specific examples of circRNAs involved in Q. Zhou
Shanghai Applied Radiation Institute, School of
Environmental and Chemical Engineering, Shanghai
University, Shanghai, China
L. Wang · X. Meng
Cardiac Regeneration and Ageing Lab, Institute of Cardiac Regeneration and Ageing Lab, Institute of
Cardiovascular Sciences, School of Life Science, Cardiovascular Sciences, School of Life Science,
Shanghai University, Shanghai, China Shanghai University, Shanghai, China
Shanghai Key Laboratory of Bio-Energy Crops, Shanghai Key Laboratory of Bio-Energy Crops,
School of Life Sciences, Shanghai University, School of Life Sciences, Shanghai University,
Shanghai, China Shanghai, China
G. Li J. Xiao (*)
Cardiovascular Division of the Massachusetts School of Life Science, Institute of Cardiovascular
General Hospital, Harvard Medical School, Sciences, Shanghai University, Shanghai, China
Boston, MA, USA e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 191

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_15
192 L. Wang et al.

Fig. 15.1 Biogenesis of circular RNA and its regulation. regulated by RNA-binding proteins (RBPs). (d) Exon-­
(a) Circular RNA is catalyzed by spliceosomal machinery. intron circRNAs (EIciRNAs) contain both exons and
(b) Exon skipping leads to a lariat-driven circularization, introns generated by back-splicing and regulate transcrip-
and introns contain reverse complementary sequence-­ tion of its host gene. (e) Circular intronic RNAs (ciRNAs)
driven circularization. (c) Circular RNA production is derived from lariat introns during canonical splicing

development of RNA-sequencing technology, spliceosome at back-splicing sites to catalyze the

widespread and abundant circRNAs have been ligation of 5′ donor sites and 3′ acceptor sites
discovered [4–7]. CircRNAs are produced by the (Fig. 15.1a) [9]. Two models have been proposed
formation of 3′,5′-phosphodiester bond at the for the backsplice formation of circRNAs: Exon
junction site [8, 9]. CircRNAs are expressed in skipping leads to a lariat-driven circularization,
various eukaryotic organisms, especially in mam- and introns contain reverse complementary
mals, and involved in regulating many biological sequence-driven circularization (Fig. 15.1b) [8,
processes. CircRNAs are derived from precursor 12–14]. Besides, trans-acting factors RNA-­
mRNA back-splicing [10, 11]. Classically, binding proteins (RBPs) have been reported to
­circularization of RNAs requires the assemble of regulate circRNA biogenesis (Fig. 15.1c) [15–
15 Circular RNAs in Cardiovascular Diseases 193

17]. Some circRNAs contain both exons and numerous circRNAs have been discovered and
introns (exon-intron circRNAs, EIciRNAs) involved in important processes of cardiovascular
(Fig. 15.1d) [18]. Another class of circular RNAs development and diseases [42, 43]. In this fol-
derived from lariat introns during canonical lowing section, we will give a brief review about
splicing could also lead to the biogenesis of specific examples of circRNAs involved in the
circular intronic RNAs (ciRNAs) (Fig. 15.1e)
­ currently reported ways of regulating in the car-
[19]. Many studies have shown that circRNAs diovascular system (Table 15.1).
participate in many human diseases, including
cancer, diabetes, nervous system diseases, and
cardiovascular diseases [20–26]. Numerous 3  ircular RNA Acts as miRNA
C
excellent reviews have summarized the biogene- Sponge
sis of circular RNAs and the underlying mecha-
nisms [27–36]. Here we will outline the specific 3.1  ircular RNA HRCR Acts
C
examples and discuss potential mechanisms as miR-223 Sponge that
underlying the regulatory role of circRNAs in Regulates Pathological
cardiovascular diseases. Cardiac Hypertrophy

Pathological cardiac hypertrophy is a common

2 Circular RNAs reaction of the heart against increased hemody-
and Cardiovascular Diseases namics, myocardial damage, and neurohormonal
stress, manifested as increased myocardial cell
Cardiovascular diseases are the leading cause of volume, increased cardiac fibrosis, and loss of
death in human globally, higher than cancer and myocardial cells (necrosis or apoptosis) [38, 41].
other diseases [37, 38]. In addition, the increas- When the stimulation persists, pathological myo-
ing direct economic burden of cardiovascular dis- cardial hypertrophy is decompensated, resulting
ease has also become a major public health in ventricular remodeling and heart failure. Heart
problem. However, the underlying molecular failure is the end stage of many cardiovascular
mechanisms of cardiovascular diseases are still diseases and has very poor prognosis. It is the
less known and remain to be well studied [39– main cause of disability and death. In addition to
41]. In recent years, in the cardiovascular system, cardiac transplantation, currently, heart failure

Table 15.1 Circular RNAs in cardiovascular diseases

Actions CircRNA Diseases Regulation References
Act as miRNA sponges HRCR Pathological hypertrophy ↓ [44]
Cdr1as Myocardial infarction ↑ [58]
MFACR Cardiomyocytes apoptosis ↑ [61]
CircRNA_000203 Cardiac fibrosis ↑ [63]
CircRNA_010567 Cardiac fibrosis ↑ [64]
Interaction with RBPs CircFOXO3 Cardiac senescence ↑ [66]
CircAmotl1 Dox-induced cardiomyopathy ↓ [69]
CircANRIL Atherosclerosis ↓ [71]
Regulated by RBPs CircTitin Hypertrophic cardiomyopathy Unknown [78]
CircTitin Heart failure Unknown [77]
CircFhod3 CircStrn3
Serve as biomarkers MICRA Myocardial infarction Unknown [99]
CircRNA_081881 Myocardial infarction Unknown [100]
has_circ_0124644 Coronary artery disease Unknown [101]
Others cZNF292 Angiogenesis ↑ [80]
CircSLC8a1 Heart development Unknown [43]
194 L. Wang et al.

phy through targeting miR-223 and apoptosis

repressor with CARD domain. Overexpression of
HRCR could attenuate cardiomyocyte hypertro-
phy induced by ISO treatment and represses car-
diac hypertrophy as well as heart failure in vivo.
This is the first functional research of circRNAs
in pathological cardiac hypertrophy.

3.2  ircular RNA Cdr1as Acts

C
as miR-7 Sponge that
Promotes Myocardial
Infarction
Fig. 15.2 Circular RNAs act as miRNA sponges
For example, circular RNA HRCR binds to miR-223 and Acute myocardial infarction (MI) remains one of
acts as an endogenous miR-223 sponge to inhibit miR- the leading causes of morbidity and mortality
223 activity
globally. Most MI occurs due to coronary artery
diseases [50]. During acute MI, cardiomyocyte
lacks an effective treatment in clinical practice. apoptosis and necrosis trigger the early inflam-
Therefore, exploring ways to prevent the progres- matory response that clear the wound from dead
sion of pathological cardiac hypertrophy to heart cells and activate the following reparative
failure remains an important issue in the t­ reatment response [51]. Early reperfusion is currently the
of cardiovascular diseases and heart failure. most effective strategy to improve survival rates
A recent study about circular RNA HRCR in patients suffering acute myocardial infarction
(heart-related circRNA) demonstrates that HRCR [50, 52]. However, this reperfusion therapy
directly binds to miR-223 and acts as an endoge- always leads to myocardial ischemia reperfusion
nous miR-223 sponge to inhibit miR-223 activity injury (I/RI) and at risk of developing heart fail-
(Fig. 15.2) [44]. MiR-223 was first identified in ure [52, 53]. Therefore, efforts made to investi-
myeloid cells in the bone marrow and has been gate the molecular mechanisms of cardiomyocytes
reported to have a prominent role in monocyte/ apoptosis are crucial for developing new thera-
macrophage differentiation and granulocytic dif- peutic strategies. Cdr1as (antisense to the cere-
ferentiation as well as different types of cancers bellar degeneration-related protein 1 transcript),
[45–49]. In cardiomyocytes, miR-223 acts as a also termed as ciRS-7 (circular RNA sponge for
positive regulator by targeting ARC. Through miR-7), has been reported to contain more than
random screening of 100 circRNAs from cir- 70 conserved miRNA target sites and acts as a
cRNA databases as well as the use of the bioin- miR-7 sponge [54]. In mouse brain, miR-7 inhi-
formatics program RNAhybrid, HRCR was bition and Cdr1as overexpression demonstrate
found to have six target sites of miR-223. similar developmental defects [55–57]. In cardio-
Furthermore, using RNA pulldown and AGO2 myocytes, overexpression of miR-7 could reverse
immunoprecipitation assays, HRCR was proved Cdr1as-induced apoptosis by targeting PARP and
to be directly bound to miR-223. The co-­ SP1 [58]. In vivo, Cdr1as overexpression aggra-
localization of HRCR and miR-223 was verified vates MI injury evidenced by increased cardiac
by fluorescence in situ hybridization (FISH). infarct size and upregulation of PARP and SP1,
HRCR also regulates cardiomyocytes hypertro- while miR-7 overexpression has opposite effects.
15 Circular RNAs in Cardiovascular Diseases 195

3.3  ircular RNA MFACR Acts

C remodeling and heart failure. The strategy of
as miR-652-3p Sponge that anti-fibrosis for heart failure patients has been
Regulates Mitochondrial raised; however, no effective therapy for cardiac
Dynamics and Apoptosis fibrosis was reported. Therefore, understanding
in the Heart underlying mechanism of cardiac fibrosis is
important for prevention and treatment of cardiac
Numerous studies have demonstrated that mito- fibrosis. As newly discovered class of noncoding
chondrial fission dysfunction often exists in car- RNAs, the role of circular RNAs in cardiac fibro-
diovascular disease. Mitochondrial fission sis remains unknown. Recently, circRNA_000203
process protein 1 (MTFP1) or mitochondrial was reported to upregulate in diabetic mouse
18kDa protein (MTP18) is involved in the myocardium and in AngII-induced mouse car-
­mitochondrial division by regulating membrane diac fibroblasts [63]. RNA pulldown and
fission [59]. In PC-3 and HaCaT cells, MTP18 RT-qPCR assay revealed that circRNA_000203
knockdown would induce the release of cyto- could bind to miR-26b-5p. And duel luciferase
chrome c, which activates the caspase cascade assay further showed that miR-26b-5p could spe-
and leads to apoptosis [60]. The exact role of cifically interact with the 3′-UTRs of CTGF and
MTP18 in heart remains unknown. Interestingly, Col1a2. Interestingly, this interaction could be
Wang et al. reported that MTP18 deficiency abolished by overexpression of circRNA_000203
reduces mitochondrial fission and suppresses car- evidenced by elevated expression of fibrosis-­
diomyocytes apoptosis [61]. Using the bioinfor- associated gene, such as Col1a2, Col3a1, α-SMA,
matics program RNAhybrid, mitochondrial and CTGF. Therefore, circRNA_000203 could
fission, and apoptosis-related circRNA, MFACR inhibit the anti-fibrotic effect of miR-26b in
was identified and found to contain 15 binding mouse cardiac fibroblasts. In addition, another
sites on miR-652-3p. Biotin-coupled RNA pull- circRNA, circRNA_010567, was reported to pro-
down assay in cardiomyocytes was performed to mote myocardial fibrosis via suppressing miR-­
verify this interaction of miR-652-3p and 141 function [64].
MFACR. AGO2 immunoprecipitation and
reverse pulldown assay were further conducted to
confirm that MFACR could directly bind to miR-­ 4  ircular RNA Can Interact
C
652-­3p in vivo. Consistently, overexpression of with RNA-Binding Proteins
MFACR increased the protein level of MTP18, (RBPs)
while knockdown MFACR resulted in inhibition
of MTP18. Besides, inhibition of miR-652-3p 4.1  ircular RNA FOXO3 Promotes
C
would abolish the effect of MFACR knockdown Cardiac Senescence
on MTP18. In mice heart with I/R injury, knock- Through Sequestering RBPs
down of MFACR increased miR-652-3p expres-
sion and decreased MTP18 protein level. MFACR CircFOXO3 is derived from transcription factor
acts as a miR-652-3p sponge to promote mito- FOXO3 and acts as a scaffold to modulated P21
chondrial fission and cardiomyocyte apoptosis. and CDK2 interaction [65]. In NIH3T3 cell line,
circFOXO3 interacts with both P21 and CDK2,
repressing cell progression. In the heart, this
3.4  ircular RNA Acts as miRNA
C same circFOXO3 is highly expressed.
Sponge that Contributes Interestingly, Du et al. found that the expression
to Cardiac Fibrosis levels of circFOXO3 in older hearts are signifi-
cantly higher than young hearts (Fig. 15.3a) [66].
Cardiac fibrosis is the result of excess deposition Therefore, they explored the effects of cir-
of extracellular matrix in the cardiac muscle [62]. cFOXO3 on senescence. In mouse embryonic
It is a main component of adverse ventricular fibroblasts (MEFs), mouse cardiac fibroblasts
196 L. Wang et al.

Fig. 15.3 Circular RNA interaction with RNA-binding HIF-1α and FAK in the cytoplasm decreased HIF-1α in
proteins (RBPs) the nucleus and FAK in mitochondria. (b) CircAmotl1
(a) CircFOXO3 binds to transcription factors E2F1 and binds AKT and PDK simultaneously, which increased
Id1 and inhibits their translocation into the nucleus. Also, AKT phosphorylation and pAKT nuclear translocation
under stress conditions, circFOXO3 interaction with

(MCFs), NIH3T3 fibroblasts, B16 cells, and pri- Besides, the expression level of circFOXO3
mary cardiomyocytes, circFOXO3 was highly (whether overexpression or knockdown) did not
expressed in cell lines that underwent cellular change the expression of ID1, E2F1, HIF-1α, or
senescence. In vivo, in a mouse model of Dox-­ FAK. Overexpressed circFOXO3 facilitated the
induced cardiomyopathy, circFOXO3 expression localization of most transcription factors ID1 and
level was found to be associated with the level of E2F1 and anti-stress proteins HIF-1α and FAK in
tissue apoptosis as evidenced by TUNEL stain- cytoplasm, decreasing the protein levels of ID1
ing. Furthermore, immunoprecipitation and real-­ and E2F1, HIF-1α in the nucleus, and FAK in
time PCR demonstrate that circFOXO3 was mitochondria. In summary, circFOXO3 seques-
pulled down by antibodies against senescence-­ tering ID1, E2F1, HIF1α, and FAK in the cyto-
associated proteins ID1, E2F1, as well as HIF-1α plasm decreases those protein levels in the
and FAK, but linear FOXO3 mRNA was not. nucleus and promotes senescence.
15 Circular RNAs in Cardiovascular Diseases 197

4.2  ircular RNA Amotl1

C translocation, leads to activation of AKT signal-
Facilitates Cardioprotective ing pathway, and protects Dox-induced
Nuclear Translocation of pAKT cardiomyopathy.

The Ser and Thr kinase AKT plays essential roles

in diverse cellular processes. AKT signaling 4.3  ircular RNA ANRIL Regulates
C
pathways are activated by activation of PI3K and Atherosclerosis
phosphorylation of AKT [67]. In cardiovascular
disease, AKT function as a cardioprotective mol- ANRIL is a long noncoding RNA called anti-
ecule. For the activation of AKT signaling, sense noncoding RNA in the INK4 locus. This
phosphoinositide-­ dependent protein kinase 1 locus on chromosome 9p21.3 has been reported
(PDK1) was discovered for phosphorylation of to have strong association with atherosclerotic
AKT1 at T308. During this process, the localiza- vascular disease (ASVD). A class of novel circu-
tion of AKT and PDK1 to membrane sites of lar RNA products from ANRIL locus and their
PIP3 is required for the access of PDK1 to AKT expression are correlated with INK4/ARF tran-
for phosphorylation [68]. By microarray, Zeng scription and ASVD [70]. Among them, one iso-
et al. analyzed the expression levels of different form named circANRIL is well studied [71].
circular RNAs in neonatal and mature postnatal CircANRIL was identified in association with
human cardiac tissue samples, and circAmotl1 atheroprotection at human 9p21 locus.
was found to be preferentially expressed in neo- Overexpression of circANRIL in HEK-293 cells
nate cardiac tissue (Fig. 15.3b) [69]. revealed increased apoptosis and decreased pro-
Overexpression of circAmotl1 could attenuate liferation. Using qPCR and RNA immunopre-
the Dox-induced cardiomyopathy in mice model. cipitation of AGO2 analysis demonstrates that
Higher protein levels of pAKT, AKT, PDK1, and circANRIL does not regulate 9p21 protein-­
pPDK1 were detected in cells transfected with coding genes and lacks miRNA sponging activ-
circAmotl1. In the nuclei, pAKT level was also ity. Further proteomic screen shows that
elevated. The authors further investigated the circANRIL potentially binds to proteins. Through
underlying mechanism about how circAmotl1 RNA pulldown assay and label-free mass spec-
facilitate nuclear translocation of AKT and trometric analysis, 32 proteins were detected
PDK1. By conducting AGO2 immunoprecipita- with significant enrichment in circANRIL over-
tion assays, they exclude the possibility that cir- expression cells compared to controls. Next,
cAmotl1 may function as a sponge for binding RNA immunoprecipitation was performed to
miRNAs. In further studies, using computer algo- analyze circANRIL-binding proteins, and PES1
rithm, NPDock, and immunoprecipitation, cir- was identified to be the strongest binding protein.
cAmotl1 was found able to bind to PDK1 and PES1 is the component of PeBoW complex,
AKT simultaneously. The interaction between which is required for maturation of ribosomal
PDK1 and AKT could be abolished by RNase A RNAs and formation of the 60S ribosome [72–
indicating that this interaction is mediated by 74]. Northern blot and immunofluorescent stain-
RNA. Based on the docked structure of AKT-­ ing revealed that overexpression of circANRIL in
circAmotl1-­PDK1, antisense oligos complemen- cells leads to 32S and 36S pre-rRNA accumula-
tary to the binding site on circAmotl1 were tion, nucleolar stress, and P53 activation. Taken
designed to specifically block interaction between together, circANRIL overexpression could
circAmotl1 and AKT or circAmotl1 and PDK1. impair the ribosome biogenesis and P53 activa-
Interestingly, the expression of AKT, PDK1, and tion and further lead to increase in cell apoptosis
pPDK1 was not affected, but the pAKT and and decrease in proliferation. Therefore, circAN-
pPKD1 nuclear translocation was reduced. Thus, RIL was proposed to promote atheroprotection
circAmotl1 binds to AKT and PDK1 directly, by suppressing overproliferating cell types in ath-
induces AKT phosphorylation and pAKT nuclear erosclerotic plaques.
198 L. Wang et al.

5  BPs Involved in Regulating

R gain function studies demonstrated that cZNF292
the Expression of Circular regulated angiogenic sprouting. Overexpression
RNAs of circular RNA cZNF292 exhibited enhanced
proliferation, while silencing of cZNF292 inhib-
RNA-binding proteins (RBPs) play crucial roles ited spheroid sprouting. The underlying mecha-
in various cellular processes [75, 76]. It has been nism of cZNF292 was investigated further, and
reported that RNA-binding protein Quaking reg- the possibility that cZNF292 acts as miRNA
ulates circular RNA biogenesis during epithelial sponge and cis-regulator for its host gene ZNF292
to mesenchymal transition [16]. Another splicing was excluded. As a result, hypoxia-induced endo-
factor muscleblind (MBL) could specifically bind thelial circular RNA cZNF292 is indeed existing
to circMbl and affect circMbl biosynthesis [15]. and has biological functions.
In cardiovascular system, several RBPs have
been reported as key regulators of cellular func-
tion in various cardiac diseases. In doxorubicin-­ 7  ircular RNA Serves
C
induced cardiotoxicity, Quaking was also found as a Biomarker
to be downregulated, and overexpression of of Cardiovascular Disease
Quaking could inhibit doxorubicin-induced
apoptosis by regulating the biogenesis of cardiacCircular RNAs are widely present in human path-
circular RNAs [77]. In addition to Quaking, ological tissues, blood, saliva, exosomes, and
another report explored circRNAs in cardiac tis- semen [81–88]. Because of their closed loop
sue from patients with hypertrophic cardiomy- structure, they have better stability than linear
opathy (HCM) or dilated cardiomyopathy (DCM) RNAs. In addition, the expression of circular
and from non-diseased individuals [78]. RNA-­ RNAs has distinct tissue specificity and timing
binding protein 20 (RBM20) is identified as a specificity. Recently, circular RNA is expected to
regulator of mRNA splicing of a subset of genes be a novel biomarker for human diseases [89–
94]. Many studies have been reported that cir-
involved in cardiac development [79]. In this arti-
cle, RBM20 was found to be essential for the for-cRNAs acted as biomarkers in some diseases,
mation of a subset of circRNAs that originate including cancer, neurological disorders, etc.
from the titin gene. Therefore, RBPs might play [95–98]. In cardiovascular diseases, Vausort et al.
important roles in the formation of circular RNAsconducted a study to evaluate circRNA molecules
and the regulation of circRNA expression. in the peripheral blood with acute myocardial
infarction to predict postischemic reperfusion
conditions [99]. This study selected myocardial
6 Other Circular RNAs infarction with left ventricular dysfunction and
Associated finally identified a circRNA myocardial
with Cardiovascular Disease infarction-­associated circular RNA (MICRA),
which is an 874-nucleotide-long circRNA formed
Circular RNAs are widely expressed and involved mainly from exon 1 of the zinc finger protein 609
in many biological processes in cardiovascular (ZNF609) gene located on chromosome 15q22
system. Boeckel et al. investigated the circRNAs that could predict left ventricular dysfunction
in hypoxia-induced endothelial cells [80]. 3–4 months after reperfusion. Another study per-
Through bioinformatics analysis and next-­ formed by Deng et al. showed that cir-
generation sequencing, circular RNA cZNF292 cRNA_081881 can serve as a competitive
was identified. Circular RNA cZNF292 was one endogenous RNA molecule of miR-549, thereby
of the highest expressed and significantly regulating PPARγ expression [100].
hypoxia-regulated circRNAs. The characteriza- CircRNA_081881 can be detected in plasma and
tion of ZNF292 was confirmed by northern blot therefore might serve as a potential marker and
and RNase R treatment assay. Further loss-and-­ therapeutic target for acute myocardial i­ nfarction.
15 Circular RNAs in Cardiovascular Diseases 199

Another circular RNA associated with the patho- circular RNA can act as IRESs [108]. The m6A-­
physiological process of coronary artery disease, driven translation could be reduced by m6A
has_circ_0124644, has been detected in the demethylase FTO while promoted by adenosine
peripheral blood and can be used as a potential methyltransferase METTL3/14. Circular RNA
biomarker [101]. When has_circ_0124644 was SHPRH can encode a new protein that inhibits
introduced as a test marker, both specificity and the development of gliomas, and circ-ZNF609
sensitivity of the diagnoses of cardiovascular dis-
can be translated into proteins that regulate myo-
ease were significantly increased. Taken together,genesis [109, 110]. However, no translatable cir-
with the deepening of research on circular RNA, cular RNA has been found to be involved in the
there will be more circular RNAs that are regulation of cardiovascular disease. IRES-driven
expected to become new biomarkers. It will be translation is favored under stress conditions. In
greatly helpful for the diagnosis and precise treat-
the event of cardiovascular disease, it is highly
ment of cardiovascular diseases. likely that the translation of circular RNA into
proteins or peptides is involved in the regulation
of the occurrence and development of the dis-
8 Conclusions and Future ease. Due to the strong tissue and spatial specific-
Perspectives ity of circular RNAs, these circular RNA-derived
proteins or peptides may serve as better therapeu-
With the advent of deep sequencing techniques, tic targets. Therefore, it is of great significance to
in recent years, thousands of circular RNA iso- discover the function of these translatable circu-
forms from numerous tissues and organisms are lar RNAs in the heart.
characterized. A large amount of circular RNAs At present, the study of cardiovascular RNAs
have been found in the heart, and the expression in circular RNAs has mainly focused on the regu-
of many circular RNAs is closely related to the lation of circular RNAs in diseases such as myo-
occurrence and development of cardiovascular cardial infarction, atherosclerosis, and heart
diseases [42, 43, 102, 103]. In this review, we failure [44, 58, 61, 66, 71]. Exercise training is
used various examples to illustrate the roles of recognized as an effective way to prevent cardio-
circular RNAs in cardiovascular system. For vascular disease [111–114]. A large number of
example, circular RNA HRCR, Cdr1as, and clinical studies have shown that exercise training
MFACR could act as endogenous miRNA can improve coronary blood flow by improving
sponges to inhibit miRNA activity [44, 58, 61]. vasodilation while reducing myocardial oxida-
Circular RNA FOXO3, Amotl1, and ANRIL have tive stress and preventing myocardial cell loss
been reported to interact with RBPs [66, 69, 71]. and cardiac fibrosis, thereby improving cardiac
RNA-binding protein Quaking and RBM20 function and reducing cardiovascular risk factors
could affect the biogenesis of circular RNAs in [115, 116]. Different from pathological myocar-
the heart [77, 78]. However, it should be noted dial hypertrophy, physiological cardiac hypertro-
that many questions underlying the regulatory phy usually occurs after regular exercise
role of circular RNAs remained; further investi- [117–123]. Some microRNAs have also been
gation of circular RNAs is urgently requested. reported to participate in the regulation of
Although circular RNAs were initially consid- exercise-­mediated physiological cardiac hyper-
ered to be noncoding RNAs, as the research pro- trophy and have protective functions for the heart
gressed, some circular RNAs were found to have [124–127]. However, the function of circular
the potential to encode proteins or polypeptides RNAs in exercise-mediated physiological cardiac
[104–106]. The presence of internal ribosome hypertrophy is still unknown. Therefore, explor-
entry sites (IRESs) and appropriate open reading ing the role of circular RNA in physiological car-
frame (ORF) makes the cap-independent transla- diac hypertrophy is also of great significance for
tion possible [107]. In addition, N6-­ the treatment of cardiovascular diseases and pre-
methyladenosine (m6A) modification sites of vention of heart failure.
200 L. Wang et al.

So far, only a few studies have been reported hundreds of human genes in diverse cell types. PLoS
One 7(2):e30733
in the study of circular RNAs in cardiovascular 5. Lu C, Shi X, Wang AY et al (2018) RNA-Seq profil-
system, and much of our understanding comes ing of circular RNAs in human laryngeal squamous
from a small number of circular RNAs. RNA cell carcinomas. Mol Cancer 17(1):86
sequencing data have been demonstrated that cir- 6. AbouHaidar MG, Venkataraman S, Golshani A
et al (2014) Novel coding, translation, and gene
cular RNAs are abundant in the heart [42, 43, expression of a replicating covalently closed cir-
103]. However, the underlying regulatory role of cular RNA of 220 nt. Proc Natl Acad Sci USA
most circular RNAs remains unclear. What can 111(40):14542–14547
be confirmed is that the expression level of circu- 7. Salzman J, Chen RE, Olsen MN et al (2013) Cell-­
type specific features of circular RNA expression.
lar RNAs has undergone a great change during PLoS Genet 9(9):e1003777
the occurrence of various cardiovascular dis- 8. Barrett SP, Wang PL, Salzman J (2015) Circular
eases. Many circular RNAs may play crucial RNA biogenesis can proceed through an exon-­
roles in the development of cardiovascular dis- containing lariat precursor. Elife 4:e07540
9. Starke S, Jost I, Rossbach O et al (2015) Exon circu-
ease. Therefore, to study the regulatory mecha- larization requires canonical splice signals. Cell Rep
nism of the circular RNAs in the cardiovascular 10(1):103–111
system is of great significance for the diagnosis, 10. Wilusz JE (2015) Repetitive elements regulate circu-
intervention, and treatment of cardiovascular lar RNA biogenesis. Mob Genet Elements 5(3):1–7
11. Liang D, Tatomer DC, Luo Z et al (2017) The out-
diseases. put of protein-coding genes shifts to circular RNAs
when the Pre-mRNA processing machinery is limit-
Acknowledgments The authors thank members of the ing. Mol Cell 68(5):940–954 e3
Cardiac Regeneration and Ageing Lab in Shanghai 12. Wilusz J (2015) Circular RNA and splicing: skip
University for the discussion. Due to space restrictions, happens. J Mol Biol 427(15):2411–2413
the authors cannot cite all the relevant literature in the 13. Liang D, Wilusz JE (2014) Short intronic repeat
field. The authors apologize to those colleagues whose sequences facilitate circular RNA production. Genes
work contributed significantly. This work was supported Dev 28(20):2233–2247
by the grants from the National Natural Science 14. Zhang XO, Wang HB, Zhang Y et al (2014)
Foundation of China (81722008, 91639101, and 81570362 Complementary sequence-mediated exon circular-
to JJ Xiao), the Innovation Program of Shanghai Municipal ization. Cell 159(1):134–147
Education Commission (2017-01-07-00-09-E00042 to JJ 15. Ashwal-Fluss R, Meyer M, Pamudurti NR et al
Xiao), the grant from the Science and Technology (2014) circRNA biogenesis competes with pre-­
Commission of Shanghai Municipality (17010500100 to mRNA splicing. Mol Cell 56(1):55–66
JJ Xiao), and the development fund for Shanghai talents 16. Conn SJ, Pillman KA, Toubia J et al (2015) The
(to JJ Xiao). RNA binding protein quaking regulates formation of
circRNAs. Cell 160(6):1125–1134
Competing Financial Interests The authors declare no 17. Ivanov A, Memczak S, Wyler E et al (2015) Analysis
competing financial interests. of intron sequences reveals hallmarks of circular
RNA biogenesis in animals. Cell Rep 10(2):170–177
18. Li Z, Huang C, Bao C et al (2015) Exon-intron cir-
cular RNAs regulate transcription in the nucleus. Nat
Struct Mol Biol 22(3):256–264
References 19. Zhang Y, Zhang XO, Chen T et al (2013) Circular
intronic long noncoding RNAs. Mol Cell
51(6):792–806
1. Capel B, Swain A, Nicolis S et al (1993) Circular
20. Yang L, Han B, Zhang Y et al (2018) Engagement
transcripts of the testis-determining gene Sry in
of circular RNA HECW2 in the nonautophagic role
adult mouse testis. Cell 73(5):1019–1030
of ATG5 implicated in the endothelial-mesenchymal
2. Cocquerelle C, Daubersies P, Majerus MA et al
transition. Autophagy 14(3):404–418
(1992) Splicing with inverted order of exons occurs
21. Zheng Q, Bao C, Guo W et al (2016) Circular RNA
proximal to large introns. EMBO J 11(3):1095–1098
profiling reveals an abundant circHIPK3 that regu-
3. Cocquerelle C, Mascrez B, Hetuin D et al (1993)
lates cell growth by sponging multiple miRNAs. Nat
Mis-splicing yields circular RNA molecules. FASEB
Commun 7:11215
J 7(1):155–160
22. Lin SP, Ye S, Long Y et al (2016) Circular
4. Salzman J, Gawad C, Wang PL et al (2012) Circular
RNA expression alterations are involved in
RNAs are the predominant transcript isoform from
­OGD/R-­induced neuron injury. Biochem Biophys
Res Commun 471(1):52–56
15 Circular RNAs in Cardiovascular Diseases 201

23. Veno MT, Hansen TB, Veno ST et al (2015) Spatio-­ 41. Burchfield JS, Xie M, Hill JA (2013) Pathological
temporal regulation of circular RNA expression dur- ventricular remodeling: mechanisms: part 1 of 2.
ing porcine embryonic brain development. Genome Circulation 128(4):388–400
Biol 16:245 42. Tan WL, Lim BT, Anene-Nzelu CG et al (2017) A
24. Szabo L, Morey R, Palpant NJ et al (2015) landscape of circular RNA expression in the human
Statistically based splicing detection reveals neural heart. Cardiovasc Res 113(3):298–309
enrichment and tissue-specific induction of circular 43. Werfel S, Nothjunge S, Schwarzmayr T et al (2016)
RNA during human fetal development. Genome Biol Characterization of circular RNAs in human, mouse
16:126 and rat hearts. J Mol Cell Cardiol 98:103–107
25. Lukiw WJ (2013) Circular RNA (circRNA) in 44. Wang K, Long B, Liu F et al (2016) A circular RNA
Alzheimer’s disease (AD). Front Genet 4:307 protects the heart from pathological hypertrophy
26. You X, Vlatkovic I, Babic A et al (2015) Neural cir- and heart failure by targeting miR-223. Eur Heart
cular RNAs are derived from synaptic genes and reg- J 37(33):2602–2611
ulated by development and plasticity. Nat Neurosci 45. Chen CZ, Li L, Lodish HF et al (2004) MicroRNAs
18(4):603–610 modulate hematopoietic lineage differentiation.
27. Qu S, Zhong Y, Shang R et al (2017) The emerging Science 303(5654):83–86
landscape of circular RNA in life processes. RNA 46. O’Connell RM, Zhao JL, Rao DS (2011)
Biol 14(8):992–999 MicroRNA function in myeloid biology. Blood
28. Salzman J (2016) Circular RNA expression: its 118(11):2960–2969
potential regulation and function. Trends Genet 47. Gao Y, Lin L, Li T et al (2017) The role of miRNA-
32(5):309–316 ­223 in cancer: function, diagnosis and therapy. Gene
29. Qu S, Yang X, Li X et al (2015) Circular RNA: 616:1–7
a new star of noncoding RNAs. Cancer Lett 48. Johnnidis JB, Harris MH, Wheeler RT et al (2008)
365(2):141–148 Regulation of progenitor cell proliferation and
30. Chen LL, Yang L (2015) Regulation of circRNA bio- granulocyte function by microRNA-223. Nature
genesis. RNA Biol 12(4):381–388 451(7182):1125–1129
31. Barrett SP, Salzman J (2016) Circular RNAs: analy- 49. Lu H, Buchan RJ, Cook SA (2010) MicroRNA-223
sis, expression and potential functions. Development regulates Glut4 expression and cardiomyocyte glu-
143(11):1838–1847 cose metabolism. Cardiovasc Res 86(3):410–420
32. Chen LL (2016) The biogenesis and emerging 50. Parikh NI, Gona P, Larson MG et al (2009) Long-­
roles of circular RNAs. Nat Rev Mol Cell Biol term trends in myocardial infarction incidence and
17(4):205–211 case fatality in the National Heart, Lung, and Blood
33. de Bruin RG, Rabelink TJ, van Zonneveld AJ et al Institute’s Framingham Heart study. Circulation
(2017) Emerging roles for RNA-binding proteins as 119(9):1203–1210
effectors and regulators of cardiovascular disease. 51. Saparov A, Ogay V, Nurgozhin T et al (2017) Role
Eur Heart J 38(18):1380–1388 of the immune system in cardiac tissue damage and
34. Fischer JW, Leung AK (2017) CircRNAs: a regula- repair following myocardial infarction. Inflamm Res
tor of cellular stress. Crit Rev Biochem Mol Biol 66(9):739–751
52(2):220–233 52. Oerlemans MI, Koudstaal S, Chamuleau SA et al
35. Viereck J, Thum T (2017) Circulating noncoding (2013) Targeting cell death in the reperfused heart:
RNAs as biomarkers of cardiovascular disease and pharmacological approaches for cardioprotection.
injury. Circ Res 120(2):381–399 Int J Cardiol 165(3):410–422
36. Du WW, Zhang C, Yang W et al (2017) Identifying 53. Nahrendorf M, Pittet MJ, Swirski FK (2010)
and characterizing circRNA-protein interaction. Monocytes: protagonists of infarct inflammation
Theranostics 7(17):4183–4191 and repair after myocardial infarction. Circulation
37. Cheng S, Vasan RS (2011) Advances in the epidemi- 121(22):2437–2445
ology of heart failure and left ventricular remodel- 54. Hansen TB, Kjems J, Damgaard CK (2013)
ing. Circulation 124(20):e516–e519 Circular RNA and miR-7 in cancer. Cancer Res
38. Dorn GW 2nd (2007) The fuzzy logic of physi- 73(18):5609–5612
ological cardiac hypertrophy. Hypertension 55. Peng L, Yuan XQ, Li GC (2015) The emerging land-
49(5):962–970 scape of circular RNA ciRS-7 in cancer (Review).
39. Han B, Chao J, Yao H (2018) Circular RNA and Oncol Rep 33(6):2669–2674
its mechanisms in disease: from the bench to the 56. Piwecka M, Glazar P, Hernandez-Miranda LR et al
clinic. Pharmacol Ther. https://doi.org/10.1016/j. (2017) Loss of a mammalian circular RNA locus
pharmthera.2018.01.010 causes miRNA deregulation and affects brain func-
40. Viereck J, Thum T (2017) Long noncoding RNAs tion. Science 357(6357)
in pathological cardiac remodeling. Circ Res 57. Hansen TB, Jensen TI, Clausen BH et al (2013)
120(2):262–264 Natural RNA circles function as efficient microRNA
sponges. Nature 495(7441):384–388
202 L. Wang et al.

58. Geng HH, Li R, Su YM et al (2016) The circular ration into the PeBoW-complex. Nucleic Acids Res
RNA Cdr1as promotes myocardial infarction by 34(10):3030–3043
mediating the regulation of miR-7a on its target 73. Rohrmoser M, Holzel M, Grimm T et al (2007)
genes expression. PLoS One 11(3):e0151753 Interdependence of Pes1, Bop1, and WDR12 con-
59. Tondera D, Czauderna F, Paulick K et al (2005) trols nucleolar localization and assembly of the
The mitochondrial protein MTP18 contributes to PeBoW complex required for maturation of the 60S
mitochondrial fission in mammalian cells. J Cell Sci ribosomal subunit. Mol Cell Biol 27(10):3682–3694
118(Pt 14):3049–3059 74. Holzel M, Grimm T, Rohrmoser M et al (2007) The
60. Tondera D, Santel A, Schwarzer R et al (2004) BRCT domain of mammalian Pes1 is crucial for
Knockdown of MTP18, a novel phosphatidylinositol nucleolar localization and rRNA processing. Nucleic
3-kinase-dependent protein, affects mitochondrial Acids Res 35(3):789–800
morphology and induces apoptosis. J Biol Chem 75. Errichelli L, Dini Modigliani S, Laneve P et al
279(30):31544–31555 (2017) FUS affects circular RNA expression in
61. Wang K, Gan TY, Li N et al (2017) Circular murine embryonic stem cell-derived motor neurons.
RNA mediates cardiomyocyte death via miRNA-­ Nat Commun 8:14741
dependent upregulation of MTP18 expression. Cell 76. Bayfield OW, Chen CS, Patterson AR et al (2012)
Death Differ 24(6):1111–1120 Trp RNA-binding attenuation protein: modifying
62. Segura AM, Frazier OH, Buja LM (2014) Fibrosis symmetry and stability of a circular oligomer. PLoS
and heart failure. Heart Fail Rev 19(2):173–185 One 7(9):e44309
63. Tang CM, Zhang M, Huang L et al (2017) 77. Gupta SK, Garg A, Bar C et al (2018) Quaking inhib-
CircRNA_000203 enhances the expression of its doxorubicin-mediated cardiotoxicity through reg-
fibrosis-­associated genes by derepressing targets of ulation of cardiac circular RNA expression. Circ Res
miR-26b-5p, Col1a2 and CTGF, in cardiac fibro- 122(2):246–254
blasts. Sci Rep 7:40342 78. Khan MA, Reckman YJ, Aufiero S et al (2016)
64. Zhou B, Yu JW (2017) A novel identified circular RBM20 regulates circular RNA production from the
RNA, circRNA_010567, promotes myocardial fibro- Titin gene. Circ Res 119(9):996–1003
sis via suppressing miR-141 by targeting TGF-beta1. 79. Guo W, Schafer S, Greaser ML et al (2012) RBM20,
Biochem Biophys Res Commun 487(4):769–775 a gene for hereditary cardiomyopathy, regulates titin
65. Du WW, Yang W, Liu E et al (2016) Foxo3 circular splicing. Nat Med 18(5):766–773
RNA retards cell cycle progression via forming ter- 80. Boeckel JN, Jae N, Heumuller AW et al (2015)
nary complexes with p21 and CDK2. Nucleic Acids Identification and characterization of hypoxia-­
Res 44(6):2846–2858 regulated endothelial circular RNA. Circ Res
66. Du WW, Yang W, Chen Y et al (2017) Foxo3 circular 117(10):884–890
RNA promotes cardiac senescence by modulating 81. Yan L, Feng J, Cheng F et al (2018) Circular RNA
multiple factors associated with stress and senes- expression profiles in placental villi from women
cence responses. Eur Heart J 38(18):1402–1412 with gestational diabetes mellitus. Biochem Biophys
67. Manning BD, Toker A (2017) AKT/PKB signaling: Res Commun 498(4):743–750
navigating the network. Cell 169(3):381–405 82. Xu H, Gong Z, Shen Y et al (2018) Circular
68. Seong HA, Jung H, Choi HS et al (2005) RNA expression in extracellular vesicles isolated
Regulation of transforming growth factor-beta from serum of patients with endometrial cancer.
signaling and PDK1 kinase activity by physical Epigenomics 10(2):187–197
interaction between PDK1 and serine-threonine 83. Li T, Shao Y, Fu L et al (2018) Plasma circular RNA
kinase receptor-­ associated protein. J Biol Chem profiling of patients with gastric cancer and their
280(52):42897–42908 droplet digital RT-PCR detection. J Mol Med (Berl)
69. Zeng Y, Du WW, Wu Y et al (2017) A circular RNA 96(1):85–96
binds to and activates AKT phosphorylation and 84. Zheng F, Yu X, Huang J et al (2017) Circular RNA
nuclear localization reducing apoptosis and enhanc- expression profiles of peripheral blood mononu-
ing cardiac repair. Theranostics 7(16):3842–3855 clear cells in rheumatoid arthritis patients, based
70. Burd CE, Jeck WR, Liu Y et al (2010) Expression on microarray chip technology. Mol Med Rep
of linear and novel circular forms of an INK4/ARF-­ 16(6):8029–8036
associated non-coding RNA correlates with athero- 85. Wu N, Jin L, Cai J (2017) Profiling and bioinformat-
sclerosis risk. PLoS Genet 6(12):e1001233 ics analyses reveal differential circular RNA expres-
71. Holdt LM, Stahringer A, Sass K et al (2016) Circular sion in hypertensive patients. Clin Exp Hypertens
non-coding RNA ANRIL modulates ribosomal RNA 39(5):454–459
maturation and atherosclerosis in humans. Nat 86. Alhasan AA, Izuogu OG, Al-Balool HH et al (2016)
Commun 7:12429 Circular RNA enrichment in platelets is a signature
72. Grimm T, Holzel M, Rohrmoser M et al (2006) of transcriptome degradation. Blood 127(9):e1–e11
Dominant-negative Pes1 mutants inhibit ribosomal 87. Li Y, Zheng Q, Bao C et al (2015) Circular RNA is
RNA processing and cell proliferation via incorpo- enriched and stable in exosomes: a promising bio-
marker for cancer diagnosis. Cell Res 25(8):981–984
15 Circular RNAs in Cardiovascular Diseases 203

88. Bahn JH, Zhang Q, Li F et al (2015) The landscape 103. Wu HJ, Zhang CY, Zhang S et al (2016) Microarray
of microRNA, Piwi-interacting RNA, and circular expression profile of circular RNAs in heart tissue of
RNA in human saliva. Clin Chem 61(1):221–230 mice with myocardial infarction-induced heart fail-
89. Hang D, Zhou J, Qin N et al (2018) A novel plasma ure. Cell Physiol Biochem 39(1):205–216
circular RNA circFARSA is a potential biomarker 104. Abe N, Matsumoto K, Nishihara M et al (2015)
for non-small cell lung cancer. Cancer Med. https:// Rolling circle translation of circular RNA in living
doi.org/10.1002/cam4.1514 human cells. Sci Rep 5:16435
90. Zhang S, Zeng X, Ding T et al (2018) Microarray 105. Abe N, Hiroshima M, Maruyama H et al (2013)
profile of circular RNAs identifies hsa_circ_0014130 Rolling circle amplification in a prokaryotic transla-
as a new circular RNA biomarker in non-small cell tion system using small circular RNA. Angew Chem
lung cancer. Sci Rep 8(1):2878 Int Ed Engl 52(27):7004–7008
91. Tan S, Gou Q, Pu W et al (2018) Circular RNA 106. Wang Y, Wang Z (2015) Efficient backsplic-
F-circEA produced from EML4-ALK fusion gene ing produces translatable circular mRNAs. RNA
as a novel liquid biopsy biomarker for non-small 21(2):172–179
cell lung cancer. Cell Res. https://doi.org/10.1038/ 107. Pamudurti NR, Bartok O, Jens M et al (2017)
s41422-018-0033-7 Translation of CircRNAs. Mol Cell 66(1):9–21 e27
92. Sun H, Tang W, Rong D et al (2018) Hsa_ 108. Yang Y, Fan X, Mao M et al (2017) Extensive
circ_0000520, a potential new circular RNA bio- translation of circular RNAs driven by N(6)-
marker, is involved in gastric carcinoma. Cancer methyladenosine. Cell Res 27(5):626–641
Biomark 21(2):299–306 109. Yang Y, Gao X, Zhang M et al (2018) Novel role of
93. Li H, Li K, Lai W et al (2018) Comprehensive cir- FBXW7 circular RNA in repressing Glioma tumori-
cular RNA profiles in plasma reveals that circular genesis. J Natl Cancer Inst 110(3)
RNAs can be used as novel biomarkers for systemic 110. Legnini I, Di Timoteo G, Rossi F et al (2017) Circ-­
lupus erythematosus. Clin Chim Acta 480:17–25 ZNF609 is a circular RNA that can be translated and
94. Chen B, Huang S (2018) Circular RNA: an emerg- functions in myogenesis. Mol Cell 66(1):22–37 e29
ing non-coding RNA as a regulator and biomarker in 111. Lenk K, Erbs S, Hollriegel R et al (2012) Exercise
cancer. Cancer Lett 418:41–50 training leads to a reduction of elevated myostatin
95. Iparraguirre L, Munoz-Culla M, Prada-Luengo I levels in patients with chronic heart failure. Eur
et al (2017) Circular RNA profiling reveals that J Prev Cardiol 19(3):404–411
circular RNAs from ANXA2 can be used as new 112. Wei X, Liu X, Rosenzweig A (2015) What do we
biomarkers for multiple sclerosis. Hum Mol Genet know about the cardiac benefits of exercise? Trends
26(18):3564–3572 Cardiovasc Med 25(6):529–536
96. Xuan L, Qu L, Zhou H et al (2016) Circular RNA: 113. Fleg JL, Cooper LS, Borlaug BA et al (2015)
a novel biomarker for progressive laryngeal cancer. Exercise training as therapy for heart failure: cur-
Am J Transl Res 8(2):932–939 rent status and future directions. Circ Heart Fail
97. Qin M, Liu G, Huo X et al (2016) Hsa_circ_0001649: 8(1):209–220
a circular RNA and potential novel biomarker 114. Lerchenmuller C, Rosenzweig A (2014) Mechanisms
for hepatocellular carcinoma. Cancer Biomark of exercise-induced cardiac growth. Drug Discov
16(1):161–169 Today 19(7):1003–1009
98. Li P, Chen S, Chen H et al (2015) Using circular 115. Fujimoto N, Prasad A, Hastings JL et al (2010)
RNA as a novel type of biomarker in the screening Cardiovascular effects of 1 year of progressive and
of gastric cancer. Clin Chim Acta 444:132–136 vigorous exercise training in previously sedentary
99. Vausort M, Salgado-Somoza A, Zhang L et al (2016) individuals older than 65 years of age. Circulation
Myocardial infarction-associated circular RNA 122(18):1797–1805
predicting left ventricular dysfunction. J Am Coll 116. Young DR, Reynolds K, Sidell M et al (2014) Effects
Cardiol 68(11):1247–1248 of physical activity and sedentary time on the risk of
100. Deng Y-YaZ, Weiping and She, Jianqing and Zhang, heart failure. Circ Heart Fail 7(1):21–27
Lisa and Chen, Tao and Zhou, Juan and Yuan, 117. Weeks KL, McMullen JR (2011) The athlete’s
Zu-Yi (2016) GW27-e1167 circular RNA related to heart vs. the failing heart: can signaling explain
PPARγ function as ceRNA of microRNA in human the two distinct outcomes? Physiology (Bethesda)
acute myocardial infarction. J Am Coll Cardiol 26(2):97–105
68:C51–C52 118. Calvert JW, Condit ME, Aragon JP et al (2011)
101. Zhao Z, Li X, Gao C et al (2017) Peripheral blood Exercise protects against myocardial ischemia-­
circular RNA hsa_circ_0124644 can be used as a reperfusion injury via stimulation of beta(3)-
diagnostic biomarker of coronary artery disease. Sci adrenergic receptors and increased nitric oxide
Rep 7:39918 signaling: role of nitrite and nitrosothiols. Circ Res
102. Jakobi T, Czaja-Hasse LF, Reinhardt R et al (2016) 108(12):1448–1458
Profiling and validation of the circular RNA reper- 119. McMullen JR, Amirahmadi F, Woodcock EA et al
toire in adult murine hearts. Genomics Proteomics (2007) Protective effects of exercise and phos-
Bioinformatics 14(4):216–223 phoinositide 3-kinase(p110alpha) signaling in
204 L. Wang et al.

dilated and hypertrophic cardiomyopathy. Proc Natl 124. Bernardo BC, Ooi JY, Lin RC et al (2015) miRNA
Acad Sci USA 104(2):612–617 therapeutics: a new class of drugs with potential
120. McMullen JR, Shioi T, Zhang L et al (2003) therapeutic applications in the heart. Future Med
Phosphoinositide 3-kinase(p110alpha) plays a criti- Chem 7(13):1771–1792
cal role for the induction of physiological, but not 125. Yang L, Li Y, Wang X et al (2016) Overexpression
pathological, cardiac hypertrophy. Proc Natl Acad of miR-223 tips the balance of pro- and anti-­
Sci USA 100(21):12355–12360 hypertrophic signaling cascades toward physi-
121. DeBosch B, Treskov I, Lupu TS et al (2006) Akt1 ologic cardiac hypertrophy. J Biol Chem
is required for physiological cardiac growth. 291(30):15700–15713
Circulation 113(17):2097–2104 126. Liu X, Xiao J, Zhu H et al (2015) miR-222 is nec-
122. Zou J, Li H, Chen X et al (2014) C/EBPbeta knock- essary for exercise-induced cardiac growth and pro-
down protects cardiomyocytes from hypertrophy via tects against pathological cardiac remodeling. Cell
inhibition of p65-NFkappaB. Mol Cell Endocrinol Metab 21(4):584–595
390(1–2):18–25 127. Shi J, Bei Y, Kong X et al (2017) miR-17-3p contrib-
123. Bostrom P, Mann N, Wu J et al (2010) C/EBPbeta utes to exercise-induced cardiac growth and protects
controls exercise-induced cardiac growth and pro- against myocardial ischemia-reperfusion injury.
tects against pathological cardiac remodeling. Cell Theranostics 7(3):664–676
143(7):1072–1083
 Circular RNAs and Neuronal
Development 16
Lena Constantin

Abstract 1 Introduction
Circular RNAs (circRNAs) are abundant in
the brain and are often expressed in complex High-throughput sequencing (RNA-seq) has
spatiotemporal patterns that coincide with dis- greatly expanded our understanding of circRNA
tinct developmental transitions. This suggests biology. The first half of this book chapter will
that circRNAs play a significant role in the review insights gained from the vast amounts of
central nervous system. This book chapter will RNA-seq data, including the general properties
review research progress into the function of for circRNA expression in the brain and their
circRNAs during neuronal development. The complex spatiotemporal expression patterns. The
major themes to be discussed are the enrich- second half will focus on two potential functions
ment of circRNAs in the synapse and their for circRNAs in the brain. The first in synaptic
possible contributions to synaptopathologies, learning, as evidenced by the enrichment (and
in addition to the findings that neural cir- activity-dependent transcription) of circRNAs in
cRNAs accumulate with age and appear ben- the synapse and their deregulation in some syn-
eficial for neuronal repair. Although more aptopathologies. Already, a convincing circRNA-
research is needed, some of the possible func- associated competing endogenous RNA network
tions of circRNAs with in the brain are already has been identified in Alzheimer’s disease. The
beginning to come to light. second in the neuroprotection of the ageing brain,
on the basis that neuronal circRNAs accumulate
Keywords with age, are spatiotemporally deregulated with
Circular RNA · Brain · Synapse · Ageing · age, and are differentially expressed during the
Alzheimer’s disease · Ischemic stroke recovery period after stroke or brain injury.
Although many unanswered questions
still remain, this book chapter will summarise
the current understanding of circRNA function
in neuronal development and will put for-
ward two potential biological roles for circRNAs
in the brain.
L. Constantin (*)
Faculty of Medicine, School of Biomedical Sciences,
The University of Queensland,
St Lucia 4072, QLD, Australia
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 205

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_16
206 L. Constantin

2  he Expression of Circular
T greatest number of distinct circRNAs: an average
RNAs in the Brain of 5925 circRNA candidates (from an average
total reads of 19,479,587) compared the second
2.1  ircular RNAs Are Lowly
C highest tissue, the testis, which expresses an
But Diversely Expressed average of 3018 circRNA candidates (out of an
in the Brain average total reads of 20,081,654) [1]. Similarly
in the human foetus, the brain produces the
Circular RNAs are expressed at very low levels in greatest variety of circRNAs out of any of the
all tissues. The same can be said for the brain. In other 14 tissues analysed [5]. Furthermore, one
mouse brain, less than 0.1% of RNA-seq reads study identified 65,731 distinct circRNAs in the
from circular junctions can be mapped back to human brain alone [2].
the reference genome [1]. In comparison, 44% of Two factors may help to explain why the brain
all protein-coding genes RNA-seq reads from the has a unique capacity to synthesise such a diverse
mouse brain are mappable [1]. The majority of set of circRNA species. One is that the population
the distinct circRNA species that make up this of circRNA-synthesising genes in the brain, on
small population of RNAs also appear to be average, synthesise multiple distinct circRNAs.
expressed at low levels. For example, 41,027 out For example, circRNA-host genes of the adult
of 65,731 circRNA candidates are supported by mouse brain are able to produce an average of 2.4
fewer than 10 RNA-seq reads in the human brain distinct circRNAs, compared to other tissues like
[2]. Similarly, in mouse brain, 10,081 out of the heart, liver, and lungs, which produce an
15,849 circRNA candidates are supported by average of 1.2–1.5 circRNAs per circRNA-host
fewer than 10 RNA-seq reads [2]. Further gene [1]. In the human brain, this appears to be
contributing to their low abundance is that only a even greater, where the average circRNA-­
small percentage of genes are able to synthesise producing host genes synthesise 6.4 distinct
circRNAs. This is also true for the brain, where circRNAs, or a median of three circRNAs per
21% of expressed genes produce circRNAs in the host gene [2]. The second factor is that many
adult mouse brain [1], 15.8% in the foetal pig circRNA-hosting genes are exclusively expressed
cortex [3], and 13% in the human brain [2]. in the brain. For example, 225 circRNA-­
However, the percentage of circRNA-hosting producing genes are exclusively expressed in the
genes in the brain is greater than in other tissues brain, relative to 140 in the testis, and fewer than
like the heart, liver, and lung, where less than 20 in the heart, liver, and lungs [1]. A similar
10% of expressed genes synthesise circRNAs trend has been reported in adult rat, where 60
[3]. Therefore, in the brain, a small number of circRNA-producing genes are exclusively
expressed genes are able to synthesise few but a expressed in the brain, ~35 in the testis, and fewer
diverse set of circRNAs. This is not unlike in than 10 in the heart, liver, and lung [1].
other tissues, although the relative abundance of Given that circRNA expression is dependent
circRNAs in the brain is much greater. on host gene transcription, it is not unexpected
The brain also stands apart from other tissues that different tissues should express different
in its ability to synthesise the greatest number of subsets of circRNAs. For example, the gene
distinct circRNAs. This is consistent across ontologies of linear transcripts derived from the
species. For example, the most comprehensive liver are enriched in liver-specific processes like
analysis of circRNA expression across tissues lipoprotein metabolism and extracellular
and developmental staging to date, which exosomes, while in the brain, the gene ontologies
involved mining 10 billion RNA-seq reads from are enriched in brain-specific processes such as
103 fly sample libraries, identified that 90–95% protein phosphorylation, postsynaptic density,
of all circRNAs are expressed in the head [4]. In and protein kinase activity [6]. Likewise, the host
adult mouse, the brain expresses by far the genes of tissue-specific circRNAs should also be
16 Circular RNAs and Neuronal Development 207

enriched in pathways specific to that tissue, and the median length of a fly intron is 96 bps, the
indeed this has been shown. For instance, the host median length of introns longer than 200 bps is
genes of brain-specific circRNAs are enriched in 1009 bp, while the median lengths of introns
pathways specific to neuron development, upstream and downstream of circRNAs are 4662
differentiation, and synaptic transmission, while and 2962 bps, respectively [4]. Moreover, splice
the host genes of liver-specific circRNAs are sites that are involved in the biogenesis of two or
enriched in ion transport, proton transport, and more circRNA isoforms tend to be flanked by
caton transport [5]. Therefore, it can be con- even longer introns than splice sites that drive a
cluded that the brain likely expresses the greatest single circRNA isoform [3]. Given that circRNA-­
number of circRNAs out of any tissue simply synthesising genes are flanked by much longer
because of the transcriptional properties of the introns, neuronal genes (particularly those related
linear circRNA-synthesising host genes. to the synapse) may have greater circRNA-­
synthesising capabilities simply because of their
increased length.
2.2  ircular RNA–Hosting Genes
C
May Be Enriched in Long
Neuronal Genes 2.3  ircular RNAs Are Actively
C
Regulated in the Brain
Another point of difference between the linear
transcripts of the brain and other tissues is that Although circRNAs are dependent on their
brain-specific genes tend to have much longer host gene for the initiating of transcription,
introns. This is particularly obvious when study- hundreds of neuronal circRNAs are expressed
ing topoisomerases, enzymes that catalyse the several times higher than their host genes [1–3,
winding and unwinding of supercoiled DNA 10]. In cases where circRNA-host genes are
strands during transcription or cell replication. equivalently expressed in the brain and other
Understandably, DNA topoisomerases are partic- tissues, the numbers of circRNAs produced
ularly important for long genes. This is exempli- from the brain host genes are significantly
fied by the strong negative correlation between higher [1]. This suggests that preferentially
the length of long genes (>67 kb) and their expres- expressed circRNAs may have biological roles
sion levels when DNA topoisomerases are inhib- independent from their linear host genes.
ited [7]. Topoisomerases have recently been CircRNAs are not only more relatively abun-
shown to regulate synaptic genes [7] and to dantly and diversely expressed in the brain,
maintain normal synaptic functions [8]. More they are also differentially expressed in the
specifically, the inhibition of topoisomerases at different regions and cell types of the brain.
excitatory synapses reduces the number of syn- For example, the adult mouse cortex and hip-
apses, while inhibition at inhibitory synapses pocampus share 4030 out of 6231 circRNA
interferes with the membrane trafficking of candidates; however, 2201 circRNAs are
GABAA receptor subunits. Therefore, neuronal differentially expressed [11]. Some specific
genes, particularly those involved in synaptic pro- examples of circRNAs with region- or cell-­
cesses, tend to be long (and therefore dependent specific expression profiles are circRims2 and
on DNA topoisomerases). circDym, which are expressed at greater than
Circular RNAs tend to be bracketed by longer 50% in the mouse adult cerebellum versus the
introns. The flanking introns of circRNAs in striatum, prefrontal cortex, olfactory bulb,
human forebrain neurons are on average five midbrain, and hippocampus [2]. Another
times longer than randomly selected introns [9]. example is circPlxnd1, which is predominantly
Furthermore, a comprehensive analysis of expressed in the prefrontal cortex (<60%) ver-
10 billion RNA-seq reads in a fly identified that sus the other aforementioned brain regions [2].
208 L. Constantin

Many circRNA isoforms, which are derived embryonic carcinoma cells upregulate 1116 and
from the same host gene splice acceptor or donor downregulate 238 (out of 2735) circRNAs [2].
sites, also display divergent expression profiles in Furthermore, metabolic tagging of differentiating
the brain [3]. An intriguing example in human human forebrain neuron progenitor cells revealed
cells are the circStau2a (containing exons 2–5) that 785 (out of 11,185) circRNA candidates are
and circStau2b (containing exons 2–3) isoforms, upregulated over 26 days of in vitro differentiation
which display inverse expression patterns. (27068474). The temporal expression patterns of
circStau2b is highly expressed in the adult brain some circRNAs in the brain appear highly
relative to almost all other tissues, while the coordinated and complex. For example, the 200
longer circStau2a isoform is highly expressed in most highly expressed circRNAs of the foetal pig
the adult lung and relatively lowly expressed cortex can be clustered into seven categories
elsewhere including the brain [2]. The divergent based on their temporal expression patterns. For
expression profiles of circRNAs in the brain instance, 29 circRNAs are expressed highly in
suggests that cis-regulatory elements and brain-­ the early half of gestation, 11 circRNAs in the
specific trans-acting factors may regulate these late half of gestation, and 130 at a specific
processes. developmental stage during mid-gestation [3].
There are many possibilities as to why some The tight temporal regulation of circRNAs in the
circRNAs are preferentially upregulated in the brain, which coincides with distinct
brain. First, the brain may be enriched for neuron-­ developmental transitions, strongly supports a
specific splicing factors and/or RNA-binding biological function.
proteins that regulate circRNA biogenesis. The
divergent expression profiles and levels of
circRNA isoforms from identical splice acceptor 3  ircular RNAs Appear
C
or donor sites in particular support this. Second, Important at the Synapse
as already discussed, circularised exons require
longer introns, and neuronal genes tend to fulfil 3.1  ircular RNAs Are Greatly
C
this requirement. Third, circRNAs have a half-­ Enriched in the Synapse
life almost five times longer than their host
transcripts [10], and in quiescent and postmitotic Given that the expression of a particular circRNA
tissues like neurons, this allows for circRNA to is dependent on the transcription of its host gene
accumulate. An accumulation of circRNAs (although, as already discussed, the relative
relative to their linear isoforms has already been abundance of circular and linear host transcripts
documented in the ageing fly brain [4]; conversely, can be regulated divergently), valuable insights
circRNAs tend to be reduced in cancers [12]. can be gained from analysing the molecular
Finally, highly polarised cells, such as neurons, function and biological process ontologies of
must regulate multiple cellular functions and host genes. Intriguing, the synapse is consistently
translate different combinations of proteins one of the most significantly enriched ontologies
within their different cellular compartments, and for circRNA-associated genes from the brain.
this is often mediated by RNA-dependent This is true for the set of circRNAs that are
mechanisms, not unlike circRNAs. One exciting consistently upregulated in mouse hippocampal
possibility is that circRNAs may form an cells during postnatal development [1] and for
additional layer of localised posttranscriptional the set of circRNAs most highly elevated with
regulation. age in the fly brain [4]. Further to this, circRNAs
Circular RNAs are also temporally regulated are highly expressed in synaptic subcellular
in the brain. Differentiating human primary fractions. The majority of circRNAs in the adult
cortical neurons upregulate 1926 and mouse brain (1117) are enriched in the
downregulate 797 (out of 5265) circRNA synaptoneurosomes, relative to the corresponding
candidates [2]. Similarly, differentiating mouse cytoplasmic (709) and whole brain (847) fractions
16 Circular RNAs and Neuronal Development 209

[2]. Similarly, in the mouse and rat hippocampus, the postsynaptic density, which activates group I
circRNAs are enriched in synaptoneurosomes mGluR signalling that in turn initiates homeostatic
and/or synaptic neuropils (which exhibit robust adaptation [13]. During homeostatic adaptation,
synaptic plasticity), compared to corresponding circHomer1_a is also synthesised. The synthesis
cell body or whole hippocampal homogenates of circHomer1_a requires the same splice sites as
[1]. The enrichment of circRNAs in synaptic those for Homer1b/c transcripts [1]. Therefore,
compartments has also been validated visually circHomer1_a and Homer1b/c transcripts cannot
using high-resolution in situ hybridisation, where be mutually expressed. Circular RNAs are
at least eight circRNAs derived from synapse-­ thought to regulate translation by competing with
related host genes were located throughout the the canonical splicing of the host gene [15]. This
dendritic arbours of the mouse hippocampus [1]. may be such an example. That is, the biogenesis
Evidence is also beginning to suggest that cir- of circHomer1_a may be actively and
cRNAs may play a direct role in homeostatic purposefully competing with the transcription of
scaling. Homeostatic adaptation, or scaling, is the Homer1b/c transcripts during homeostatic
ability of neurons to maintain excitability while adaptation, with the goal of reducing competition
the brain is adjusting to environmental change. It between Homer1a and Homer1b/c mRNA
involves a cell-wide increase or decrease in synthesis. Further experimental validation is
postsynaptic α-amino-3-hydroxy-5-methyl-4-­ required.
isoxazolepropionic acid (AMPA) receptor in
excitatory synapses; this scales all synapses by
the same multiplicative factor—to become 3.2 Circular RNAs May Be Linked
stronger or weaker—while maintaining their to Neurodegenerative
relative strengths. Bicuculline, a gamma-­ Synaptopathies
aminobutyric acid (GABA)A-receptor antagonist,
can also be used to induce homeostatic plasticity. Alzheimer’s disease (AD) is the most common
Recently, 37 circRNAs (versus 7 host genes) form of progressive dementia in the ageing brain.
were transcribed in response to bicuculline-­ The pathological features of AD are intracellular
induced homeostatic plasticity, while 5 circRNAs tau-containing neurofibrillary tangles and extra-
(versus 3 host genes) were downregulated [1]. cellular amyloid-β plaques, which accumulate in
These circRNAs are examples of synaptic vulnerable regions of the brain such as the cortex
activity-dependent transcription, which suggests and hippocampus. Alzheimer’s disease is thought
a role in neural plasticity. The most dramatically to be a synaptic pathology. Some evidence for this
upregulated transcripts were circHomer1_a and are that subtle alterations in the synaptic efficacy
its host gene, homer homolog 1 (Drosophila) of the hippocampus occur in AD patients prior to
(Homer1) [1]. the detection of neurofibrillary tangles and
Homer1 can be translated into three different amyloid-β plaques [16]. Also, patients within
protein products at the postsynaptic density: 2–4 years of the clinical onset of AD have reduced
Homer1a, Homer1b, and Homer1c. Homer1b/c numbers of spines per neuron in layers II–III
is constitutively expressed, while Homer1a is (38%) and V (30%) of the temporal cortex and in
transiently upregulated during increases to layer V (30%) of the frontal cortex [17].
network activity, such as those created by long-­ Furthermore, cognitive deficits associated with
term bicuculline treatment [13]. Under neutral AD are more strongly correlated to neocortical
network conditions, Homer1b/c interacts with synapse loss compared to the number of plaques
group I metabotropic glutamate receptor (mGluR) and tangles [18].
at the postsynaptic density; however, when Recently, circRNA dysfunction was identified
Homer1a is expressed, it interferes with native in a sporadic mouse model of AD [19]. More spe-
interactions between mGluR and Homer1b/c cifically, 94 and 141 circRNAs were, respectively,
[14]. This leads to a cell-wide reorganisation of up- and downregulated (out of 34,096) in the adult
210 L. Constantin

brain relative to controls. Based on this, in combi- in Alzheimer’s disease. Indeed, UBE2A is also
nation with RNA-seq reads on deregulated downregulated by 3.7-fold in the hippocampal
miRNAs and linear mRNAs, a circRNA-associ- CA1 region and by 2.8-fold in Brodmann area 22
ated-competing endogenous RNA network was [26]. Thus, a deficiency in ciRS-7 ‘miRNA spong-
built. This RNA regulatory network did not take ing’ would enable miR-7 to potentially and effi-
into account the number and density of miRNA- ciently downregulate target genes essential for the
binding seed sequences and therefore likely con- clearance of amyloid-β plaques. In a similar fash-
tains a large proportion of false-positive pairings. ion, ciRS-7 could be protective against Parkinson’s
Nonetheless, the network identified two interac- disease by preventing miR-7 from silencing
tions formally linked to AD involving the genes epidermal growth factor receptor (EGFR), alpha-
deiodinase, iodothyronine, type II (Dio2), and synuclein (SNCA), and insulin receptor substrate
high-mobility group box 2 (HMGB2). Dio2, 2 (IRS2) [27].
which activates myelination [20] and is reduced in The fervent sponging capacity of ciRS-7 is
AD [21], putatively associated with miR-122-5p not, however, a general feature of circRNAs. The
and five deregulated circRNAs. HMGB2, which majority of circRNAs expressed in the brain have
activates pathways involved in amyloid-β plaque comparable numbers of miRNA-binding sites as
clearance [22], was paired with let-7 g-3p and linear mRNAs and therefore would not make for
deregulated 3 circRNAs. These circRNAs may be strong ‘sponges’ [1, 23, 28]. Also, organisms
competitively modulating the activity of miR- completely devoid of RNA interference, such as
122-5p and let-7 g-3p, to affect the expression of the yeast Saccharomyces cerevisiae [29], clearly
Dio2 and HMGB2, in a process commonly produce circRNAs [30]. Hence, the function of
described as ‘miRNA sponging’ [23]. This study the majority of circRNAs must extend beyond
did not identify the most convincing circRNA- being competitive-binding moderators of the
mRNA pairing involved in human AD, as would RNA interference pathway.
be expected: ciRS-7 and ubiquitin-conjugating
enzyme E2A, RAD6 homolog (S. cerevisiae)
(UBE2A). 4  ircular RNAs Appear
C
The circRNA ciRS-7 is produced from the Important for the Ageing
antisense of the cerebellar degeneration-related Brain
protein 1 (CDR1 as) gene. ciRS-7 contains 74 tan-
dem seed matches to miR-7, and 63 of these are 4.1  ircular RNAs Accumulate
C
conserved from annelids to humans [24]. in the Ageing Brain
Unsurprisingly, ciRS-7 is a potent negative regula-
tor of miR-7. For example, based on RNA-­seq Circular RNAs accumulate in the ageing brain.
data, a single HEK293 cell is estimated to contain This is substantiated by the most comprehensive
~1400 ciRS-7 molecules which can sequester up survey of circRNA expression to date that mined
to 20,000 miR-7 molecules [24]. In patients with 103 fly tissues from various developmental stages
sporadic AD, ciRS-7 is downregulated by 5.4-fold [4]. More specifically, the total circRNA levels
in the hippocampal CA1 region [25, 26] and is were found to increase across fly embryo
significantly reduced by more than fivefold in the development and dramatically increase in the
superior temporal lobe neocortex (Brodmann area adult head relative to earlier time points from the
22). A decrease of ciRS-7 in hippocampal CA1 head or all other adult tissues [4]. Similarly in the
and Brodmann area 22 would enable the excess mouse, the global levels of circRNAs—measured
accumulation of miR-7 in these regions. Indeed, by transcript per kilobase million—significantly
miR-7 is upregulated in the brain of AD patients increase from 1 to 22 months in the cortex and
by an average of threefold [26]. Excesses of miR- hippocampus, but not in the heart [11]. CircRNAs
7 would in turn repress UBE2A, which is essential not only accumulate in the brain with time, but a
for the proteolytic clearance of amyloid-β peptides subset is differentially upregulated with age
16 Circular RNAs and Neuronal Development 211

independently from host genes. In a fly, 262 (out occlusions, called ischemic strokes. For some
of 2513) circRNA candidates are significantly ischemic stroke patients, the restoration of blood
upregulated by more than twofold in 20-day-old flow exacerbates the initial injury, producing a
heads compared to 1-day-old heads [4]. Similarly so-called cerebral reperfusion injury. The specific
in the ageing mouse hippocampus, 250 (out of role of circRNAs during ‘cerebral reperfusion
5528) circRNA candidates are significantly injury’ has been investigated by three different
upregulated at 22 months compared to 1 month, laboratories, using either the intraluminal middle
and in the ageing cortex, 258 (out of 4733) cerebral artery occlusion model in the mouse [32,
circRNA candidates are significantly upregulated 33] or oxygen-glucose deprivation and then
at 22 months compared to 1 month [11]. In the reoxygenation in cell culture [34]. The most
mouse cortex, the functional ontologies of host comprehensive analyses to date identified 283
genes synthesising age-upregulated circRNAs deregulated (out of 1064) circRNA candidates
are enriched for synapse assembly, synapse over a 6-, 12-, and/or 24-h time course, with 239
organisation, neurotransmitter secretion, and significantly altered at 6 h [32]. This great peak
neurotransmitter transport [11]. Alternatively, in of circRNA deregulation at 6 h, but not at 12 and
the mouse hippocampus, the host genes of age-­ 24 h, is suggestive of complex temporal regulatory
upregulated circRNAs are enriched for protein processes taking place. Another study profiled
and chromatin modifications [11]. circRNAs at 48 h after artery occlusion and
Given that linear RNA expression does not identified over a thousand deregulated circRNAs
change with age in the mouse brain [11] and that [33]. Only modest changes in the expression of
age-upregulated circRNAs were largely 15 circRNAs were found at 24 h in the cell culture
independent to the expression level of the host model [34]. These three studies lack consistencies
gene [4, 11], the mechanisms that drive circRNA in the particular circRNAs involved. For example,
age-accumulation are not host-dependent. One none of the circRNAs altered post-stroke were
mechanism may be that circRNAs are especially shared between the cell culture [34] and
stable in the quiescent and postmitotic cells of the intraluminal middle cerebral artery occlusion
ageing brain, which allow a greater proportion of model [32] at 24 h. Also, there was a far greater
circRNAs to accumulate over time. Another number of circRNAs deregulated at 48 h [33]
mechanism may be related to the phenomenon compared to before 24 h [32], perhaps
that more than one-third of genes expressed in highlighting differences in the models used and
the ageing human brain undergo changes to the brain regions assayed.
alternative splicing that encourage back-splicing The host genes of stroke-responsive circRNAs
[31]. The next steps would be to determine across these studies were, however, enriched for
whether the accumulation of circRNAs is repair processes. For example, at 48 h post-­
innocuous or serves a protective or detrimental stroke, host genes were most significantly
function to the brain. Recent findings that enriched for the cell survival and proliferation
circRNAs are deregulated following stroke pathways of Rap1 and Hippo signalling [33].
provide insights into the potential role of age-­ Similarly, the host genes of all stroke-responsive
upregulated circRNAs. circRNAs over the entire 24-h time course were
most significantly enriched for mitogen-activated
kinase signalling, cell cycle and actin cytoskeletal
4.2  ircular RNAs Are Linked
C regulators, and focal adhesions molecules that
to Neural Repair are related to cell growth, proliferation, and death
[32]. In a rat model of traumatic injury to the
Stroke is the third leading cause of death in the hippocampus, the host genes of injury-responsive
United States. It is an acute neurological event circRNAs were most significantly enriched for
that leads to the death of neural tissues. The neurogenesis, neuronal differentiation and
majority of strokes result from vascular development, and in cellular components related
212 L. Constantin

to the synapse (5 out of the top 10) [35]. 10. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular
RNAs are abundant, conserved, and associated with
Therefore, circRNAs that are deregulated ALU repeats. RNA 19(2):141–157
following brain injury are tightly associated with 11. Gruner H, Cortes-Lopez M, Cooper DA et al (2016)
neuronal repair processes. This tentatively CircRNA accumulation in the aging mouse brain. Sci
implies that age-upregulated circRNAs may Rep 6:38907
12. Bachmayr-Heyda A, Reiner AT, Auer K et al (2015)
serve as a biological function during recovery in Correlation of circular RNA abundance with prolif-
the injured brain, although much more research is eration—exemplified with colorectal and ovarian
required. cancer, idiopathic lung fibrosis, and normal human
tissues. Sci Rep 5:8057
13. Hu JH, Park JM, Park S et al (2010) Homeostatic scal-
Acknowledgement I would like to thank Professor
ing requires group I mGluR activation mediated by
Brandon J. Wainwright for encouraging me to pursue
Homer1a. Neuron 68(6):1128–1142
noncoding RNAs as a field of research.
14. Tu JC, Xiao B, Yuan JP et al (1998) Homer binds a
novel proline-rich motif and links group 1 metabo-
Competing Financial Interests The author declares no tropic glutamate receptors with IP3 receptors. Neuron
competing financial interests. 21 (4):717–726; Brakeman PR, Lanahan AA, O’Brien
R et al (1997) Homer: a protein that selectively
binds metabotropic glutamate receptors. Nature 386
(6622):284–288
15. Ashwal-Fluss R, Meyer M, Pamudurti NR et al (2014)
References circRNA biogenesis competes with pre-mRNA splic-
ing. Mol Cell 56(1):55–66
1. You X, Vlatkovic I, Babic A et al (2015) Neural 16. Selkoe DJ (2002) Alzheimer’s disease is a synaptic
circular RNAs are derived from synaptic genes and failure. Science 298(5594):789–791
regulated by development and plasticity. Nat Neurosci 17. Davies CA, Mann DM, Sumpter PQ et al (1987) A
18(4):603–610 quantitative morphometric analysis of the neuronal
2. Rybak-Wolf A, Stottmeister C, Glazar P et al (2015) and synaptic content of the frontal and temporal cor-
Circular RNAs in the mammalian brain are highly tex in patients with Alzheimer’s disease. J Neurol Sci
abundant, conserved, and dynamically expressed. gs. 78(2):151–164
Mol Cell 58(5):870–885 18. Terry RD, Masliah E, Salmon DP et al (1991)
3. Veno MT, Hansen TB, Veno ST et al (2015) Spatio-­ Physical basis of cognitive alterations in Alzheimer’s
temporal regulation of circular RNA expression dur- disease: synapse loss is the major correlate of cogni-
ing porcine embryonic brain development. Genome tive impairment. Ann Neurol 30(4):572–580
Biol 16:245 19. Zhang S, Zhu D, Li H et al (2017) Characterization
4. Westholm JO, Miura P, Olson S et al (2014) Genome-­ of circRNA-associated-ceRNA networks in a senes-
wide analysis of drosophila circular RNAs reveals their cence-accelerated mouse prone 8 brain. Mol Ther
structural and sequence properties and age-dependent 25(9):2053–2061
neural accumulation. Cell Rep 9(5):1966–1980 20. Calza L, Fernandez M, Giuliani A et al (2002)
5. Xia S, Feng J, Lei L et al (2017) Comprehensive Thyroid hormone activates oligodendrocyte precur-
characterization of tissue-specific circular RNAs sors and increases a myelin-forming protein and
in the human and mouse genomes. Brief Bioinform NGF content in the spinal cord during experimental
18(6):984–992 allergic encephalomyelitis. Proc Natl Acad Sci U S A
6. Li L, Zheng YC, Kayani MUR et al (2017) 99(5):3258–3263
Comprehensive analysis of circRNA expres- 21. Humphries CE, Kohli MA, Nathanson L et al (2015)
sion profiles in humans by RAISE. Int J Oncol Integrated whole transcriptome and DNA methylation
51(6):1625–1638 analysis identifies gene networks specific to late-onset
7. King IF, Yandava CN, Mabb AM et al (2013) Alzheimer’s disease. J Alzheimers Dis 44(3):977–987
Topoisomerases facilitate transcription of long genes 22. Yamanaka Y, Faghihi MA, Magistri M et al (2015)
linked to autism. Nature 501(7465):58–62 Antisense RNA controls LRP1 sense transcript
8. Mabb AM, Kullmann PH, Twomey MA et al (2014) expression through interaction with a chromatin-­
Topoisomerase 1 inhibition reversibly impairs associated protein, HMGB2. Cell Rep 11(6):967–976
synaptic function. Proc Natl Acad Sci U S A 23. Guo JU, Agarwal V, Guo H et al (2014) Expanded
111(48):17290–17295 identification and characterization of mammalian cir-
9. Zhang XO, Wang HB, Zhang Y et al (2014) cular RNAs. Genome Biol 15(7):409
Complementary sequence-mediated exon circulariza- 24. Memczak S, Jens M, Elefsinioti A et al (2013)
tion. Cell 159(1):134–147 Circular RNAs are a large class of animal RNAs with
regulatory potency. Nature 495(7441):333–338
16 Circular RNAs and Neuronal Development 213

25. Lukiw WJ (2013) Circular RNA (circRNA) in 31. Mazin P, Xiong J, Liu X et al (2013) Widespread
Alzheimer’s disease (AD). Front Genet 4:307 splicing changes in human brain development and
26. Zhao Y, Alexandrov PN, Jaber V et al (2016) aging. Mol Syst Biol 9:633; Harries LW, Hernandez
Deficiency in the ubiquitin conjugating enzyme D, Henley W et al (2011) Human aging is charac-
UBE2A in Alzheimer’s disease (AD) is linked to defi- terized by focused changes in gene expression and
cits in a natural circular miRNA-7 sponge (circRNA; deregulation of alternative splicing. Aging Cell 10
ciRS-7). Genes (Basel) 7(12):116 (5):868–878
27. Hansen TB, Jensen TI, Clausen BH et al (2013) 32. Mehta SL, Pandi G, Vemuganti R (2017) Circular
Natural RNA circles function as efficient RNA expression profiles alter significantly in
microRNA sponges. Nature 495 (7441):384–388; mouse brain after transient focal ischemia. Stroke
Junn E, Lee KW, Jeong BS et al (2009) Repression 48(9):2541–2548
of alpha-­ synuclein expression and toxicity by 33. Liu C, Zhang C, Yang J et al (2017) Screening circu-
microRNA-7. Proc Natl Acad Sci U S A 106 lar RNA expression patterns following focal cerebral
(31):13052–13057 ischemia in mice. Oncotarget 8(49):86535–86547
28. Guo J (2014) Transcription: the epicenter of gene 34. Lin SP, Ye S, Long Y et al (2016) Circular RNA
expression. J Zhejiang Univ Sci B 15(5):409–411 expression alterations are involved in OGD/R-induced
29. Drinnenberg IA, Fink GR, Bartel DP (2011) neuron injury. Biochem Biophys Res Commun
Compatibility with killer explains the rise of RNAi-­ 471(1):52–56
deficient fungi. Science 333(6049):1592 35. Xie B, Wang Y, Lin Y et al (2018) Circular RNA
30. Wang PL, Bao Y, Yee MC et al (2014) Circular RNA expression profiles alter significantly after trau-
is expressed across the eukaryotic tree of life. PLoS matic brain injury in rats. J Neurotrauma. https://doi.
One 9(6):e90859 org/10.1089/neu.2017.5468
 Circular RNAs in Cancer
17
Susanne Lux and Lars Bullinger

Abstract Keywords
Circular RNAs (circRNAs) constitute a class CircRNA · MicroRNA · Cancer · Solid tumor
of RNAs that only recently have come into the · Hematologic malignancies · Oncogene ·
focus of the scientific cancer community after Tumor suppressor
it was revealed that they are very abundant,
highly conserved across species and show tis-
sue- and developmental stage-specific expres- 1 Introduction
sion. This tightly regulated, dynamic circRNA
expression, in line with expression of messen- Despite the fact that in some cancers, e.g., in leu-
ger RNAs, microRNAs, and long noncoding kemias, usually only few gene mutations affect-
RNAs, is altered in both solid tumors and ing protein-coding sequences can be detected,
hematologic malignancies and most likely cancer is in general a disease accompanied by a
contributes to tumorigenesis. In this chapter, global deregulation of gene expression compared
we will review cancer-associated and cancer-­ to normal tissue. This leads to the assumption
specific circRNAs, some of which have onco- that in addition to genomic mutations, other
genic or tumor-suppressive potential. We will changes, such as deregulated epigenetic mecha-
specifically focus on circRNAs for which the nisms, contribute to deregulated gene expression,
role in cancer has been studied in more detail, which by itself might contribute to tumorigene-
and we will discuss the opportunity to use cir- sis. In addition to aberrant expression of genes in
cRNAs as biomarkers and potential therapeu- cancer, the expression of alternative splice vari-
tic targets in cancer. ants as well as noncoding RNAs (ncRNAs) is
involved in the global deregulation of gene
expression in cancer and might also play an
important role.
S. Lux
Department of Internal Medicine III, University Recent transcriptome studies have uncovered
Hospital of Ulm, Ulm, Germany that, while only 2% of the genome encodes for
L. Bullinger (*) proteins, more than 60% of the genome is tran-
Department of Internal Medicine III, University scribed and can be detected as RNA transcripts
Hospital of Ulm, Ulm, Germany [1, 2]. This implies an important biological func-
Department of Hematology, Oncology and tion of the noncoding transcriptome, and in
Tumorimmunology, Charité – Universitätsmedizin, accordance, several studies have pointed to an
Berlin, Germany important role of noncoding RNAs such as
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 215

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_17
216 S. Lux and L. Bullinger

microRNAs (miRNA) and long noncoding RNAs This chapter aims to give an overview of cir-
(lncRNA) in cancer [3–6]. cRNAs that have been implicated in cancer, both
MiRNAs represent a well-studied subgroup of in solid tumors and in hematologic malignancies,
small ncRNAs that is implicated in several with focus on cancer-associated and cancer-­
pathologies. Aberrant miRNA expression pro- specific circRNAs and variants for which the
files are found in various types of cancer and can functional impact on carcinogenesis has been
be correlated with clinical parameters like overall more comprehensively studied.
survival [7]. In addition, selected miRNAs have
been shown to influence apoptosis, proliferation,
and differentiation of cancer cells [8, 9]. 2 CircRNAs in Cancer
Furthermore, lncRNAs do also show cell type-­
specific expression [10] and are regulated during So far, studies have focused both on the charac-
development [11], and selected lncRNAs have terization of the circular RNAome in hematopoi-
been proven as regulators of self-renewal and etic malignancies and in solid tumors. While
pluripotency in stem cells [12]. Not surprisingly, many studies are mainly descriptive, listing cir-
some lncRNAs are also implicated in cancer, e.g., cRNAs differentially expressed in tumor versus
lncRNA URHC regulates cell proliferation and healthy tissue, others investigated candidate cir-
apoptosis in hepatocellular carcinoma [13], cRNAs in detail, further elucidating the function
lncRNA HIS-1 is overexpressed in murine of the respective “mysterious” circles in the cell.
myeloid leukemia [14], and the oncogenic In general, studies followed mainly two
lncRNA HOTAIR is upregulated in breast cancer approaches to screen the circular RNAome of
and promotes tumor metastasis [15]. Moreover, cancer cells: (1) several groups performed ribo-
lncRNAs can regulate DNA methylation [16] and somal RNA-depleted RNA sequencing
coactivate proteins involved in transcriptional (RNAseq), comparing cancer samples with
regulation [17], thereby possibly contributing to healthy cells or different subsets of patients. This
the global transcriptional dysregulation in approach allows the simultaneous quantification
cancer. of linear parental gene expression and detection
Like the other ncRNAs, circular RNAs (cir- of novel, so far unknown circRNAs; (2) the sec-
cRNAs) are a class of RNAs that has long been ond approach makes use of the Arraystar Human
dismissed as rare splicing errors or even experi- Circular RNA Microarray which can detect
mental artifacts. Only recently it was revealed 13,617 circRNAs that have been previously
that they are very abundant, are highly conserved found in other studies.
across species, and show tissue- and develop- Then, expression of differentially expressed
mental stage-specific expression [18–22]. With circRNAs was often correlated with clinical
regard to circRNA function, some circRNAs parameters like patient survival, cancer grading,
have been shown to serve as miRNA sponges and metastasis. Moreover, the impact of candi-
[23], and recent studies show that at least some of date circRNAs on the hallmarks of cancer [29],
the circRNAs can be translated into proteins [24, including proliferation, apoptosis, self-renewal,
25] which adds another level of complexity on metastasis, and angiogenesis, was investigated.
this matter and challenges the classification of This chapter provides an overview of current
circRNAs as “noncoding” RNAs. Due to their important findings on circRNAs in hematologic
structure, circRNAs are very stable and protected malignancies (2.1) and solid tumors (2.2).
from exonuclease degradation, and they can be Furthermore, this chapter about circRNAs in can-
detected in serum and saliva [26–28] which com- cer aims to give a summary of putative oncogenic
prises an important criterion for a potential circRNAs (Table 17.1) as well as putative tumor-­
biomarker. suppressive circRNAs (Table 17.2).
Table 17.1 Conceivably oncogenic circRNAs and their potential functions in cancer
Parental gene CircBase [30] ID Cancer Expr. Function miRNA References
ABCB10 hsa_circ_0008717 BRCA High miRNA sponge, proliferation ↑, cancer progression ↑ miR-1271 [31]
ACP6 hsa_circ_0013958 LUAD High miRNA sponge, apoptosis ↓, proliferation ↑, invasion ↑ miR-134 [32]
AMOTL1 hsa_circ_0004214 BRCA High Tumorigenesis ↑, c-myc nuclear translocation ↑ [33]
CCDC66 CRC High miRNA sponge, proliferation ↑, cancer growth ↑, metastasis ↑ miR-33b, [34]
miR-93, miR-185
CDR1AS / hsa_circ_0001946 HCC High miRNA sponge, EGFR and RAF1 levels↑ miR-7 [35]
17 Circular RNAs in Cancer

ciRS-7 CRC High [36]

GBM Low miR-671-5p-mediated degradation [37]
FLT3 hsa_circ_0100163 hsa_ AML High [38]
circ_0100164
HIAT1a hsa_circ_0000096 GC Low Proliferation ↑, migration ↑, cyclin D1↑, CDK6 ↑, MMP-2 ↑, MMP-9 ↑ [39]
HIPK3a hsa_circ_0000284 CRC High miRNA sponge, apoptosis ↓, proliferation ↑, cancer growth ↑, miR-7 [40]
metastasis ↑
HCC High miRNA sponge, cell growth ↑ miR-124 [41]
IKBKB hsa_circ_0001793 CRC High [42]
KCNH1 hsa_circ_0016347 OSA High miRNA sponge, proliferation ↑, invasion ↑, metastasis ↑ miR-214 [43]
LARP1B hsa_circ_0070933 CSCC High [44]
hsa_circ_0070934
LINC00340 hsa_circ_0075825 BCC High [45]
hsa_circ_0075828
MLL/AF9 AML with High Proliferation ↑, progression ↑, therapy resistance ↑, MAPK and AKT1 [46]
t(9;11) signaling ↑
MYLK hsa_circ_0002768 Bladder High miRNA sponge, proliferation ↑, migration ↑, progression ↑ through miR-29a [47]
VEGFA/VEGFR2, epithelial-mesenchymal transition ↑
NPM1 hsa_circ_0075001 AML High TLR signaling ↓ [38]
PML/RARA APL with High Proliferation ↑ [46]
t(15;17)
PRKCI hsa_circ_0067934 ESCC High Proliferation ↑ [48]
PVT1 GC High miRNA sponge, proliferation ↑ miR-125 [49]
MM High therapy resistance ↑, proliferation ↑, apoptosis ↓ [50]
SLC30A7 hsa_circ_0013339 OSCC High miRNA sponge, CDK6 ↑, proliferation ↑ miR-29 [51]
(continued)
217
Table 17.1 (continued)
218

Parental gene CircBase [30] ID Cancer Expr. Function miRNA References

TCF25 hsa_circ_0041103 Bladder High miRNA sponge, CDK6 ↑, proliferation ↑, migration ↑ miR-103a-3p [52]
miR-107
TTBK2 hsa_circ_0000594 Glioma High miRNA sponge, HNF1β ↑, proliferation ↑, migration ↑, apoptosis ↓, miR-217 [53]
UBAP2 OSA High miRNA sponge, cancer progression ↑ miR-143 [54]
VCAN Glioma High [55]
Abbreviations: AML Acute Myeloid Leukemia, APL Acute Promyelocytic Leukemia, BCC Basal Cell Carcinoma, BRCA Breast Invasive Carcinoma, CRC Colorectal Carcinoma,
CSCC Cutaneous Squamous Cell Carcinoma, ESCC Esophageal Squamous Cell Carcinoma, Expr. Expression, GBM Glioblastoma Multiforme, GC Gastric Cancer, HCC
Hepatocellular Carcinoma, LUAD Lung Adenocarcinoma, miRNA microRNA, MM Multiple Myeloma, OSA Osteosarcoma, OSCC Oral Squamous Cell Carcinoma, Ref.
References, TLR Toll-like Receptor
a
Oncogenic/tumor-suppressive role is context-specific/inconclusive; ↑ increase; ↓ decrease
S. Lux and L. Bullinger
Table 17.2 Conceivably tumor-suppressive circRNAs and their potential functions in cancer
Parental
gene CircBase [30] ID Cancer Expr. Function miRNA References
FADS2 hsa_ BCC Low [45]
circ_00223883 CSCC Low [44]
FBXW7 GBM Low Proliferation ↓, cell cycle ↓, Encoding [56]
FBXW7-185aa, USP28-induced c-Myc
stabilization ↓
FOXO3 BRCA Low miR-22, miR-136, miR-138, miR-149, miR-433, [57]
17 Circular RNAs in Cancer

miRNA sponge, tumor growth ↓, proliferation

↓, cancer cell survival ↓, FOXO3 translation ↑ miR-762, miR-3614, miR-3622
Cell cycle progression ↓ via forming a [58]
complex with p21 and CDK2
BRCA Low Tumor growth ↓, MDM2-induced p53 [59]
degradation ↑, PUMA-mediated apoptosis ↑
HIPK3a hsa_circ_0000284 Bladder Low miRNA sponge, heparanase expression ↓, miR-588 [60]
migration ↓, invasion ↓, angiogenesis ↓
ITCH ESCC Low miRNA sponge, ITCH levels ↑, WNT/ miR-7, miR-17, miR-214 [61]
CRC Low beta-catenin pathway ↓ [62]
HCC Low [63]
METTL3 hsa_circ_0000523 CRC Low [42]
hsa_circ_006229
MTO1 hsa_circ_0007874 HCC Low miRNA sponge, cancer progression ↓ miR-9 [64]
RNF13 hsa_circ_0001346 CRC Low [42]
SMARCA5 hsa_circ_0001445 HCC Low miRNA sponge, growth ↓, metastasis ↓, miR-17-3p [65]
expression of tumor suppressor TIMP3 ↑ miR-181b-5p
SMYD4 hsa_circ_0004018 HCC Low [66]
WDR37 hsa_circ_0004277 AML Low [67]
ZKSCAN1 hsa_circ_0001727 HCC Low Growth ↓, migration ↓, invasion ↓ [68]
Abbreviations: AML Acute Myeloid Leukemia, BCC Basal Cell Carcinoma, BRCA Breast Invasive Carcinoma, CRC Colorectal Carcinoma, CSCC Cutaneous Squamous Cell
Carcinoma, ESCC Esophageal Squamous Cell Carcinoma, Expr. Expression, GBM Glioblastoma Multiforme, GC Gastric Cancer, HCC Hepatocellular Carcinoma, miRNA
microRNA, MM Multiple Myeloma, OSA Osteosarcoma, Ref. References
a
Oncogenic/tumor suppressive role is context-specific/inconclusive; ↑ increase; ↓ decrease
219
220 S. Lux and L. Bullinger

2.1  he Role of CircRNAs

T healthy cells. One of them, hsa_circ_0075001,
in the Pathology showed differential expression between different
of Hematologic Malignancies AML cell lines and AML patients, independent
of the NPM1 mutational status [38]. Its expres-
Expression of circRNAs is known to be differen- sion correlated with a distinct gene expression
tiation stage-specific [19] in many tissues, and signature characterized by a downregulation of
changes in the circRNA repertoire have been Toll-like receptor (TLR) signaling.
extensively studied during neuronal development Similar to NPM1 mutations, chromosomal
[69] and epithelial-mesenchymal transition [70]. translocations represent another pathomecha-
Blood cells express circRNAs at levels compara- nism of leukemic transformation. These translo-
ble to that in the cerebellum, a tissue that is cations result in oncogenic fusion proteins, i.e.,
known to be rich in circRNAs [26]. In the hema- the KMT2A-MLLT3 (alias MLL-AF9) fusion in
topoietic system, we examined the circular t(9;11) AML or the PML-RARA fusion in acute
RNAome during myeloid differentiation and promyelocytic leukemia (APL) with t(15;17).
detected changes upon leukemic transformation Guarnerio and colleagues found that also fusion
[38] using ribosomal RNA-depleted RNA-Seq. circRNAs (f-circRNAs) are transcribed from
The circRNA repertoire of mature myeloid cells these fusion genes which can promote prolifera-
including metamyelocytes and neutrophils dif- tion, viability, and transformation of the cells and
fered from circRNAs expressed in more imma- contribute to therapy resistance in vivo [46]. The
ture cells including myeloblasts and circular KMT2A-MLLT3 transcript exerted its
promyelocytes, which is in line with findings in oncogenic activity via mitogen-activated protein
other tissue types (Fig. 17.1). kinase (MAPK) and RAC-alpha serine/threonine-­
protein kinase (AKT1) signaling.
2.1.1 Acute Leukemias Salzman and colleagues intended to investi-
CircRNA expression was found altered in leuke- gate intragenic rearrangements in childhood
mic blasts of acute myeloid leukemia (AML) acute lymphoblastic leukemia (ALL). To their
patients, which constitute a very immature cell surprise, they found thousands of genes that pro-
type. Nevertheless, the leukemic circular duced transcripts with a scrambled exon order
RNAome does not merely resemble that of and estimated that for many genes, around 10%
healthy immature cells but is distinct and differ- of the transcripts were circular-derived [18].
ent AML subtypes, e.g., patients with mutation of However, the most abundant circRNAs they
the nucleophosmin (NPM1) gene could be distin- detected and validated were also present in
guished from healthy hematopoietic cells healthy peripheral blood cells and H9 embryonic
(Fig. 17.1). In healthy blood cells, for 44% of the stem cells and thus not leukemia-specific.
highly expressed genes (~5000 out of 11,000),
circRNAs were detectable, whereas in AML the 2.1.2 Other Hematological
percentage was slightly higher (47%). Malignancies
In an analysis of a relatively small sample In multiple myeloma (MM), a hematological
cohort comparing healthy and leukemic samples, malignancy that develops in plasma cells, cir-
27 genes were associated with differentially cPVT1 transcribed from the nonprotein-coding
expressed circRNAs between the healthy and PVT1 oncogene locus was shown to promote
leukemic samples, with 14 circRNAs being therapy resistance to glucocorticoid treatment,
higher in AML than in healthy cells, among them and knockdown of circPVT1 resulted in enhanced
is circFLT3 transcribed from the fms-related apoptosis and reduced proliferation of resistant
tyrosine kinase 3 (FLT3) gene that is commonly myeloma cells in vitro and in vivo [50].
mutated in AML. Similarly, NPM1, which is Considering the current evidence and the gen-
commonly mutated in AML, was linked to a vari- eral splicing deregulation in leukemia and
ety of different circRNAs, both in leukemic and myeloma [71–75], it is reasonable to speculate
17 Circular RNAs in Cancer 221

Fig. 17.1 Distinct circRNA expression changes during expression data derived from RNA-Seq analysis of n = 9
hematopoietic differentiation and leukemic transforma- FACS-sorted healthy control samples (including different
tion. (a) Principal component analysis (PCA) and (b) myeloid differentiation stages as indicated) and n = 7
unsupervised hierarchical clustering based on circRNA NPM1mut AML patients [38].

that aberrant circRNA expression, in line with focus on the characterization of circRNAs in
deregulation of mRNAs, miRNAs, and lncRNAs, hematological malignancies including chronic
contributes to the pathogenesis of hematologic leukemias, myeloproliferative diseases, and non-­
malignancies. Currently, many additional studies Hodgkin lymphoma (NHL).
222 S. Lux and L. Bullinger

2.2  he Role of CircRNAs

T Similarly, binding miR-138 the circular sex-­
in the Pathology of Solid determining gene (Sry) transcript can also func-
Tumors tion as such a “miRNA sponge” [23]. In
undifferentiated human embryonic stem cells,
In solid tumors, first studies have shown that circBIRC6, a circRNA of the BIR repeat-­
based on their function, circRNAs can be catego- containing ubiquitin-conjugating enzyme
rized into (i) circRNAs that were experimentally (BIRC6) gene, was found to bind miR-34a and
proven to serve as miRNA sponges (Sect. 2.2.1) miR-145, two miRNAs that target genes main-
and (ii) circRNAs that have other miRNA-­ taining the pluripotent state. By sponging these
independent functions or were not associated to miRNAs, circBIRC6 was able to suppress differ-
interact with miRNAs yet (Sect. 2.2.1.1). entiation [80].
Moreover, in this chapter we further grouped all However, sequence analysis of over 7000
circRNAs according to their putative oncogenic newly annotated circRNAs showed that a large
or tumor-suppressive properties. part of known circRNAs contains only few,
between zero to eight, miRNA-binding sites [81,
2.2.1 CircRNAs Functioning as MiRNA 82]. Nevertheless, miRNA-binding sites in cir-
Sponges in Solid Tumors cRNAs are often depleted of polymorphisms,
MiRNAs represent a well-studied subgroup of hinting at an underlying selective pressure con-
small ncRNAs that is implicated in several serving functional binding sites [83]. Today,
pathologies, including cancer. They target mRNA already many studies have investigated the inter-
transcripts in a sequence-specific manner and are action of circRNAs with miRNAs, often in the
able to block mRNA translation or trigger its deg- context of cancer as outlined below.
radation. Aberrant miRNA expression profiles
were found in various types of cancer [3, 4, 76]. CircRNAs with Oncogenic Properties
In addition to an impact on clinical parameters, CircRNAs Targeting miR-7 In hepatocellular
miRNAs were also shown to have the potential to carcinoma (HCC), CDR1as/ciRS-7 (hsa_
influence proliferation and differentiation of the circ_0001946), which is probably the most well-­
cancer cells. For example, the oncogenic miR-­ studied circRNA that serves as a miRNA sponge,
155 is overexpressed in many solid tumors, in was highly expressed [35, 84] like in colorectal
lymphoma, and in AML patients with FLT3-ITD cancer (CRC) [36]. In this study, ciRS-7
mutations, where it can induce myeloid prolifera- increased EGFR and RAF1 expression levels in
tion [7, 77]. In AML patients carrying NPM1 part via binding of the tumor-suppressive miR-7.
mutations, elevated levels of miR-10a/b and In glioblastoma (GBM), however, ciRS-7 levels
miR-196a/b were found [8, 77], which led to were low as the circRNA might be targeted by
enhanced proliferation and a block in differentia- miR-­671-­5p, which can result in ciRS-7 degra-
tion of hematopoietic progenitor cells. dation [37].
Conversely, the tumor-suppressive miR-29 fam-
ily is downregulated in high-risk CLL, lung can- In CRC, another circRNA, circHIPK3 (hsa_
cer, and invasive breast cancer [6, 78]. circ_000284) from the homeodomain-interacting
Today, some cytoplasmic circRNAs have been protein kinase 3 (HIPK3) gene, was also reported
shown to modulate miRNA activity by function- to be highly expressed and associated with metas-
ing as a miRNA sponge. A prime example is cir- tasis and poor patient survival [40]. Knockdown
cRNA CDR1as/ciRS-7 that contains 74 of circHIPK3 decreased proliferation and migra-
evolutionary conserved miR-7-binding sites and tion and increased apoptosis in CRC cells in vitro
that can be cleaved by AGO proteins [79]. and could suppress cancer growth and metastasis
Interestingly, CDR1as overexpression resulted in in vivo. This could also be mimicked in vitro by
a phenotype similar to that of miR-7 knockdown overexpression of miR-7, a miRNA that is
and led to impaired midbrain development [19]. sequestered by circHIPK3. Vice versa, overex-
17 Circular RNAs in Cancer 223

pression of circHIPK3 led to increased levels of oncogenes that were otherwise repressed by these
proto-oncogene targets of miR-7, including FAK, miRNAs; among them are DNMT3B, EZH2, and
IGF1R, EGFR, and YY1. Moreover, expression MYC.
of circHIPK3 was also high in HCC where it pro-
moted cell growth, and the circRNA could bind In bladder cancer, circMYLK (hsa_
also to other miRNAs such as e.g. miR-124 [41]. circ_0002768), transcribed from the myosin light
Thus, while miR-7 seems to be a prominent chain kinase (MYLK) gene, served as a sponge
target for circRNAs in cancer, these first studies for miR-29a, thereby increasing proliferation,
clearly demonstrate that the function of circRNAs migration, and tumor progression through
is complex and that several different miRNAs can VEGFA/VEGFR2 deregulation [47]. Another
be sponged by individual circRNAs suggesting circRNA (hsa_circ_0013339) sponging members
that many other miRNAs are deregulated by this of the tumor-suppressive miR-29 family is highly
mechanism in cancer. A summary of circRNAs expressed in oral squamous cell carcinoma
with conceivably oncogenic function is given in (OSCC). Transcribed from the solute carrier
Table 17.1. family 30 member 7 (SLC30A7) gene expression
of this circRNA increased levels of CDK6 and
CircRNAs Targeting Other miRNAs Similar to promoted proliferation in OSCC [51]. Similarly,
its elevated expression in multiple myeloma (see circTCF25 (hsa_circ_00411103), a circRNA of
above), circPVT1 was highly expressed in gastric the transcription factor 25 (TCF25) gene, was
cancer where it could sponge miR-125 and highly expressed in bladder cancer and increased
increase proliferation [49]. In a different context, migration, proliferation, and CDK6 levels in can-
circPVT1 expression was found to be low in cer cells by sponging miR-103a-3p and miR-107
senescent fibroblasts, and senescence could be [52].
triggered in proliferating fibroblasts by knocking In addition to carcinoma, circRNA-mediated
down circPVT1 [85]. In proliferating fibroblasts, deregulation of miRNAs does also play a role in
circPVT1 could sponge miRNA let-7, thereby mesenchymal tumors such as glioma and sar-
targeting several proliferative proteins including coma. For example, a circRNA of the Tau tubulin
IGF2BP1, KRAS, and HMGA2. In breast cancer, kinase 2 (TTBK2) gene was found highly
circABCB10 (hsa_circ_0008717) transcribed expressed in glioma, in which circTTBK2 (hsa_
from the ATP-binding cassette subfamily B mem- circ_0000594) increased proliferation and
ber 10 (ABCB10) gene was highly expressed, and decreased apoptosis [53]. As a sponge for miR-­
it induced proliferation and cancer progression 217, it increased levels of HNF1β. In osteosar-
by sponging miR-1271 [31]. Similarly, in lung coma (OSA), a circRNA of the potassium
adenocarcinoma (LUAC), circACP6 (hsa_ voltage-gated channel subfamily H member 1
circ_0013958) transcribed from the lysophos- (KCNH1) gene, hsa_circ_0016347, was found
phatidic acid phosphatase 6 (ACP6) gene was highly expressed and increased proliferation,
found highly expressed, which in turn induced invasion, and metastasis of OSA cells by spong-
proliferation and invasion and reduced apoptosis ing miR-214 [43]. Similarly, circUBAP2 tran-
of the cancer cells. As a sponge for miR-134, cir- scribed from the ubiquitin-associated protein 2
cACP6 could increase levels of the oncogenic (UBAP2) gene was highly expressed and pro-
cyclin D1 [32]. In CRC, circCCDC66 of the moted OSA progression as a sponge of miR-143
coiled-coil domain containing 66 (CCDC66) [54].
locus, but not the linear mRNA, was highly
expressed and associated with poor patient sur- CircRNAs with Tumor-Suppressive
vival. CircCCDC66 promoted cancer growth and Properties
metastasis by sequestering tumor-suppressive While for the abovementioned circRNAs, high
miRNAs, including miR-33b, miR-93, and miR-­ expression was associated with tumor growth;
185 [34]. This led to an upregulation of several there have been also several circRNAs reported
224 S. Lux and L. Bullinger

which due to their tumor suppressive properties sponge for miR-588 [60], a miRNA associated
are often downregulated in cancer (Table 17.2). with invasion and angiogenesis [87]. Low circH-
A prime example is a circRNA of the forkhead IPK3 was associated with high-grade bladder
box O3 (FOXO3) gene that can act as a sponge cancer characterized by vascular invasion, and
for multiple miRNAs that regulate FOXO3 overexpression of circHIPK3 could inhibit blad-
mRNA expression, including miR-22, miR-136, der cancer growth and metastasis. Thus, by
miR-138, miR-149, miR-433, miR-762, miR-­ sponging miR-588 circHIPK3 exhibited tumor-­
3614, and miR-3622 [57]. High expression of suppressive properties in bladder cancer.
circFOXO3 leads to increased translation of the However, circHIPK3 can also serve as a sponge
FOXO3 protein by taking away the miRNA bur- for miR-7 in CRC and miR-124 in HCC in which
den of the FOXO3 mRNA, resulting in decreased it then promotes tumor growth. This clearly dem-
cellular growth, proliferation, and survival. onstrates that the function of circRNAs is cell
Moreover, ectopic expression of circFOXO3 in context-specific and that circRNAs, like many
cancer or non-cancer cell lines blocked cell cycle genes, can have both tumor-suppressive and
progression by forming a complex with the oncogenic function.
cyclin-dependent kinase 2 (CDK2) and the Another potential double-edged sword is cir-
cyclin-dependent kinase inhibitor 1 (p21), thus cRNA from the itchy E3 ubiquitin protein ligase
preventing CDK2 from interaction with cyclin A (ITCH) gene that was not only expressed at low
and cyclin E resulting in a block of G1 to S tran- levels in HCC but also in esophageal squamous
sition [58]. In accordance, breast cancer cir- cell carcinoma (ESCC) and CRC [61–63].
cFOXO3 expression was reported to be very low CircITCH could sequester miR-7, which is usu-
[59] and through protein-binding circFOXO3 ally involved in tumor progression (see above)
was shown to increase MDM2-induced p53 ubiq- but also miR-17 and miR-214 and increase ITCH
uitination and subsequent degradation [59]. levels itself. This results in increased ubiquitina-
However, it also increased stress-induced apopto- tion and degradation of phosphorylated dishev-
sis since higher levels of FOXO3 protein resulted eled segment polarity protein 2 (Dvl2) which
in upregulation of PUMA, a pro-apoptotic pro- resulted in inhibition of the WNT/beta-catenin
tein that can induce apoptosis independent of p53 pathway. This pathway plays a role in tumorigen-
[86]. esis and progression, as well as metastasis [88].
In line, circMTO1 (hsa_circ_0007874), a cir-
cRNA of the mitochondrial tRNA translation 2.2.2 Other CircRNAs Impacting Solid
optimization 1 (MTO1) gene, was shown to bind Tumor Pathology
the oncogenic miR-9 which targets p21 [64].
Silencing of circMTO1 promoted HCC cell pro- CircRNAs with Oncogenic Properties
liferation and invasion via a miR-9-mediated In addition to the potential miRNA sponge func-
downregulation of p21, and low circMTO1 levels tion, there have been several other mechanisms
were associated with inferior survival of HCC of action assigned to circRNAs (Table 17.1). For
patients. Similarly, circRNA (hsa_circ_0001445) example, highly expressed in breast cancer cir-
derived from the SWI/SNF-related matrix-­ cAMOTL1 (hsa_circ_0004214) transcribed from
associated actin-dependent regulator of chroma- the angiomotin like 1 (AMOTL1) gene can pro-
tin subfamily A member 5 (SMARCA5) gene is mote tumorigenesis by increasing the nuclear
lowly expressed in HCC [65]. As a sponge for translocation of the MYC proto-oncogene [33].
miR-17-3p and miR-181b-5p, circSMARCA5 This led to increased stability of c-MYC and
increased levels of the tumor suppressor TIMP3 higher affinity to its target promoters. However,
and inhibited tumor growth and metastasis. for other oncogenic circRNAs, the exact mecha-
Additional “tumor suppressive” circRNAs nism of action was not entirely resolved, and an
have been found in bladder cancer, in which involvement of miRNAs or other ncRNAs cannot
circHIPK3 (hsa_circ_0000284) serves as a be excluded. In gastric cancer, knockdown of
17 Circular RNAs in Cancer 225

circHIAT1 (hsa_circ_0000096), a circRNA circ_0001727, was shown to inhibit growth,

derived from the hippocampus abundant tran- migration, and invasion and was expressed at low
script 1 (HIAT1) gene inhibited migration and levels in HCC [68]. Overexpression of circZK-
proliferation of cancer cells in vitro and in vivo SCAN1 could inhibit HCC progression in vivo,
by decreasing protein levels of matrix metallo- but the exact mechanism of action was not inves-
proteinase-­2 (MMP-2), MMP-9, cyclin D1, and tigated. Moreover, hsa_circ_0004018 levels, a
cyclin-dependent kinase 6 (CDK6) [39]. circRNA of the SET and MYND domain contain-
A circRNA of the inhibitor of nuclear factor ing 4 (SMYD4) gene, were also low in HCC tis-
kappa B kinase subunit beta (IKBKB) gene (hsa_ sues and cell lines [66].
circ_0001793) was shown to be highly expressed Other circRNAs also showed low expression
in colorectal cancer; however, the circRNA’s in different cancer types when compared to
effect on NFκB signaling was not investigated healthy tissue (Table 17.2). However, since the
[42]. Two circRNAs of La ribonucleoprotein function of these circRNAs was not investigated
domain family member 1B (LARP1B), i.e., hsa_ in detail, it can only be assumed that they have
circ_0070933 and hsa_circ_0070934), were tumor-suppressive properties. Expression of hsa_
highly expressed in cutaneous squamous cell car- circ_00223883, a circRNA of the fatty acid
cinoma (CSCC) [44], and the expression of two desaturase 2 (FADS2) gene, was low in BCC and
circRNAs (hsa_circ_0075825, hsa_ CSCC [44, 45], and expression of two circRNAs
circ_0075828) from the LINC00340 locus, also of the methyltransferase-like 3 (METTL3) gene
known as the nonprotein-coding cancer suscepti- (hsa_circ_0000523, hsa_circ_0006229), as well
bility 15 locus, was high in basal cell carcinoma as hsa_circ_0001346 of the ring finger protein 13
(BCC) [45]. The gene also produces a long non-­ (RNF13) gene, was low in CRC [42]. In AML,
coding RNA that is thought to promote cell pro- low expression of hsa_circ_0004277 of the WD
liferation in HCC. Expression of circPRKCI repeat domain 37 (WDR37) gene was detected
(hsa_circ_0067934) from the protein kinase C [67].
Iota (PRKCI) gene was high in ESCC and could
increase proliferation of the cancer cells [48]. In
glioma, a highly expressed circRNA derived 2.3 Methylation and Translation
from the versican (VCAN) gene, a gene that is of CircRNAs
known to be involved in processes like prolifera-
tion, migration, and angiogenesis [55]. CircRNAs have long been classified as non-­
coding, but it was recently found that at least
CircRNAs with Tumor–Suppressive some of the circRNAs can be translated into pro-
Properties teins [24, 25], such as circFBXW7 mentioned
A protein-coding circRNA of the F-box and WD above, which adds another level of complexity
repeat domain containing 7 (FBXW7) gene and challenges the classification of circRNAs as
encoded a novel 185 amino acid (aa) 21-kDa pro- “non-coding” RNAs. The prerequisite for cir-
tein that inhibited proliferation and cell cycle cRNA translation is the presence of an internal
progression in cancer cells [56]. By preventing ribosomal entry site (IRES) [89]. To date, the
stabilization of c-MYC through USP28, FBXW7-­ only circRNA-encoded protein that was more
185aa was able to reduce the half-life of extensively studied in the context of cancer is
c-MYC. In GBM, circFBXW7 and FBXW7-­ indeed FBXW7-185aa which seems to have
185aa levels were decreased compared with tumor-suppressive properties, and thus, it was
healthy tissue. found downregulated in GBM [56].
For other circRNAs, the tumor-suppressive Yang and colleagues have shown that the
mode of action is less well-studied (Table 17.2). translation of circular RNAs can be promoted by
A circRNA of the zinc finger with KRAB and methylation of adenosine at position 6, N6-­
SCAN domains 1 (ZKSCAN1) gene, hsa_ methyladenosine (m6A), which is the most com-
226 S. Lux and L. Bullinger

mon posttranscriptional internal modification cRNAs have been detected in several types of
found in eukaryotic mRNAs [90]. The methyla- cancer; however, only few studies comprehen-
tion pattern of circRNAs was found to be cell-­ sively characterized the possible functions of cir-
type-­specific and also distinct from the respective cRNA candidates in tumorigenesis.
mRNA [91], and m6A-circRNAs interacted with One potential pitfall in studies detecting dif-
YTHDF1/YTHDF2 m6A-reader proteins. ferentially expressed circRNAs between cancer-
However, for m A-circRNAs, unlike mRNAs,
6
ous and healthy tissue is the fact that cancer cells
interaction with YTHDF2 did not promote RNA often bear properties of immature, undifferenti-
degradation. In comparison to circRNAs without ated cells. However, only few groups have explic-
methylation, regions flanking m6A-circRNAs itly investigated differentiation-independent
were enriched for transposable elements, and it is changes in the tumor cells when compared to
known that the methylation writer complex healthy tissue. Thus, it is possible that the expres-
METTL3/14 binds to transposable elements (TE) sion of some of these “tumor-associated” cir-
[92]. TEs in flanking introns represent one mech- cRNAs simply reflects the differentiation status
anism known to promote exon circularization of the cell, and in accordance the expression of
[93], and TEs are enriched in flanking regions of the respective circRNAs do not necessarily have
m6A-circRNAs. to be functionally relevant to tumorigenesis in
The METTL3 gene itself also produces a cir- these cases.
cular RNA that was downregulated in colorectal Nevertheless, the functional characterization
cancer while linear METTL3 levels remained of single circRNA candidates has shown that cir-
unchanged [42]. Interestingly, m6A-methylation cRNAs can affect more or less all cancer-related
is known to play a role in the pathogenesis of processes including proliferation, apoptosis, self-­
cancer. To give some examples, components of renewal, metastasis, and angiogenesis. In some
the m6A methyltransferase complex are differen- instances, only few miRNA-binding sites within
tially regulated throughout healthy hematopoie- a circRNA were sufficient to impact miRNA
sis, and their expression is altered in AML. In function and to affect its downstream targets. As
particular, METTL14 is highly expressed in a consequence, deregulation of a single circRNA
AML, where it is required for the development can impact the expression of a broad range of
and maintenance of the disease by conferring genes within a cell, thereby having the potential
self-renewal properties to the leukemia stem cells to play an important role in tumorigenesis.
[94]. METTL3 was shown to increase translation
of oncogenes in lung cancer [95]. Changes of
m6A-methylation in the pathogenesis of cancer 3.1 Diagnostic and Prognostic
have not been linked to altered translation of cir- Potential of CircRNAs
cRNAs yet, but this connection should be investi- in Cancer
gated in the future given the importance of altered
epigenetic and epitranscriptomic mechanisms in Many groups are examining the potential of cir-
tumorigenesis. cRNAs as biomarkers for diagnosis and progno-
sis of cancer. Due to their circular structure,
circRNAs are protected from exonuclease
3 Conclusion and Outlook ­degradation, and their half-life is much longer
than that of the parental mRNA, i.e., around 48 h
In line with the transcriptome, the circular compared to 10 h for mRNAs [82]. Moreover,
RNAome of cancer cells is distinct from that of they are stably detected in body fluids like saliva
healthy cells. Moreover, the circRNA repertoire and blood [26, 27], fluids that could be accessed
can vary in different tumor subtypes and might in a noninvasive way for diagnostic purposes, so-­
change with cancer stage and the development of called liquid biopsies. CircRNAs could further be
therapy resistance. Differentially expressed cir- detected in serum exosomes, and analysis of the
17 Circular RNAs in Cancer 227

circRNA content in these vesicles could distin- In summary, first studies on circRNAs in can-
guish colon cancer patients from healthy indi- cer have shown promising results and ongoing
viduals [28]. Some circRNAs have already been comprehensive analyses will improve our under-
shown to be of prognostic value, either alone or standing of the impact of deregulated circRNA
synergistically with known cancer markers [36, expression on tumorigenesis, which in turn will
49]. Especially for cases lacking traditional bio- also increase our general understanding of the
markers, circRNAs could be of value, but the role of circRNAs. In the long term, circRNAs
reproducibility and specificity of circRNA detec- might also become part of the routine clinical
tion have yet to be confirmed in larger studies work-up and contribute to improved patient man-
before circRNAs might make their way into can- agement by offering new perspectives to improve
cer diagnostics and might serve as variables that diagnosis, to individualize outcome prediction,
could enter prognostic scores. and to provide targets for innovative therapeutic
approaches, both in solid tumors and hemato-
logic malignancies.
3.2 CircRNAs as Therapeutic
Targets in Cancer Acknowledgments This work was supported by the
Deutsche Forschungsgemeinschaft (SFB 1074 project B3
to LB) and the Studienstiftung des Deutschen Volkes
Both oncogenic and tumor-suppressive circRNAs (doctoral grant to SL).
have the potential to serve as therapeutic targets
in cancer. The backsplice junction sequence is Competing Financial Interests The authors declare no
unique for a certain circRNA and makes it pos- competing financial interests.
sible to specifically target a particular circRNA
without affecting the parental mRNA. Oncogenic
circRNAs could therefore be targeted by siRNAs
and subjected to Ago2-mediated degradation References
[79]. Vice versa, ectopic expression of tumor-­
suppressive circRNAs could be achieved by 1. Djebali S, Davis CA, Merkel A et al (2012) Landscape
of transcription in human cells. Nature 489(7414):101
using expression vectors with long reverse com- 2. ENCODE Project Consortium (2012) An integrated
plementary sequences which facilitate circular- encyclopedia of DNA elements in the human genome.
ization of the enclosed sequence [96]. Since Nature 489(7414):57
mammalian cells are able to discriminate “self- 3. Rücker FG, Russ AC, Cocciardi S et al (2013) Altered
miRNA and gene expression in acute myeloid leu-
“circRNAs from foreign ones [97], the use of kemia with complex karyotype identify networks of
in vitro-generated circRNAs should be avoided prognostic relevance. Leukemia 27(2):353–361
to circumvent activation of the innate immune 4. Russ AC, Sander S, Luck SC et al (2011) Integrative
system. The fact that not all circRNAs are non-­ nucleophosmin mutation-associated microRNA and
gene expression pattern analysis identifies novel
coding adds another level of complexity to this microRNA-target gene interactions in acute myeloid
field, and the first circRNA-derived protein with leukemia. Haematologica 96(12):1783–1791
tumor-suppressive properties has already been 5. Garzon R, Volinia S, Papaioannou D et al (2014)
identified [56]. Thus, another therapeutic strategy Expression and prognostic impact of lncRNAs in
acute myeloid leukemia. Proc Natl Acad Sci U S A
could be to make use of translatable circRNAs. 111(52):18679–18684
CircRNAs containing an IRES can be translated 6. Yanaihara N, Caplen N, Bowman E et al (2006)
using the normal ribosome machinery of the cell. Unique microRNA molecular profiles in lung cancer
CircRNAs with an infinite open-reading frame diagnosis and prognosis. Cancer Cell 9(3):189–198
7. Garzon R, Volinia S, Liu C-G et al (2008) MicroRNA
without stop codon were able to produce a repeat- signatures associated with cytogenetics and prognosis
ing sequence of peptides in a “rolling circle” in acute myeloid leukemia. Blood 111(6):3183–3189
manner [98]. This could be exploited to produce 8. Bryant A, Palma CA, Jayaswal V et al (2012) miR-­
peptide sequences that bind and disable onco- 10a is aberrantly overexpressed in Nucleophosmin1
gene proteins.
228 S. Lux and L. Bullinger

mutated acute myeloid leukaemia and its suppression 25. Pamudurti NR, Bartok O, Jens M et al (2017)
induces cell death. Mol Cancer 11(1):8 Translation of circRNAs. Mol Cell 66(1):9–21. e7
9. Mott JL, Kobayashi S, Bronk SF et al (2007) mir-29 26. Memczak S, Papavasileiou P, Peters O et al (2015)
regulates Mcl-1 protein expression and apoptosis. Identification and characterization of circular RNAs
Oncogene 26(42):6133 as a new class of putative biomarkers in human blood.
10. Mercer TR, Dinger ME, Sunkin SM et al (2008) PLoS One 10(10):e0141214
Specific expression of long noncoding RNAs 27. Bahn JH, Zhang Q, Li F et al (2015) The landscape of
in the mouse brain. Proc Natl Acad Sci U S A microRNA, Piwi-interacting RNA, and circular RNA
105(2):716–721 in human saliva. Clin Chem 61(1):221–230
11. Dinger ME, Amaral PP, Mercer TR et al (2008) 28. Li Y, Zheng Q, Bao C et al (2015) Circular RNA is
Long noncoding RNAs in mouse embryonic stem enriched and stable in exosomes: a promising bio-
cell pluripotency and differentiation. Genome Res marker for cancer diagnosis. Cell Res 25(8):981
18(9):1433–1445 29. Hanahan D, Weinberg RA (2011) Hallmarks of can-
12. Guttman M, Amit I, Garber M et al (2009) Chromatin cer: the next generation. Cell 144(5):646–674
signature reveals over a thousand highly con- 30. Glažar P, Papavasileiou P, Rajewsky N (2014)
served large non-coding RNAs in mammals. Nature circBase: a database for circular RNAs. RNA
458(7235):223 20(11):1666–1670
13. Xu W-H, Zhang J-B, Dang Z et al (2014) Long non-­ 31. Liang H-F, Zhang X-Z, Liu B-G et al (2017) Circular
coding RNA URHC regulates cell proliferation and RNA circ-ABCB10 promotes breast cancer prolifera-
apoptosis via ZAK through the ERK/MAPK signal- tion and progression through sponging miR-1271. Am
ing pathway in hepatocellular carcinoma. Int J Biol J Cancer Res 7(7):1566
Sci 10(7):664 32. Zhu X, Wang X, Wei S et al (2017) hsa_circ_0013958:
14. Askew DS, Li J, Ihle JN (1994) Retroviral insertions a circular RNA and potential novel biomarker for lung
in the murine His-1 locus activate the expression of adenocarcinoma. FEBS J 284(14):2170–2182
a novel RNA that lacks an extensive open reading 33. Yang Q, Du WW, Wu N et al (2017) A circular RNA
frame. Mol Cell Biol 14(3):1743–1751 promotes tumorigenesis by inducing c-myc nuclear
15. Gupta RA, Shah N, Wang KC et al (2010) Long non-­ translocation. Cell Death Differ 24(9):1609
coding RNA HOTAIR reprograms chromatin state to 34. Hsiao K-Y, Lin Y-C, Gupta SK et al (2017)
promote cancer metastasis. Nature 464(7291):1071 Noncoding effects of circular RNA CCDC66 pro-
16. Imamura T, Yamamoto S, Ohgane J et al (2004) mote colon cancer growth and metastasis. Cancer Res
Non-coding RNA directed DNA demethylation of 77(9):2339–2350
Sphk1 CpG island. Biochem Biophys Res Commun 35. Yu L, Gong X, Sun L et al (2016) The circular RNA
322(2):593–600 Cdr1as act as an oncogene in hepatocellular carci-
17. Lanz RB, McKenna NJ, Onate SA et al (1999) A ste- noma through targeting miR-7 expression. PLoS One
roid receptor coactivator, SRA, functions as an RNA 11(7):e0158347
and is present in an SRC-1 complex. Cell 97(1):17–27 36. Weng W, Wei Q, Toden S et al (2017) Circular RNA
18. Salzman J, Gawad C, Wang PL et al (2012) Circular ciRS-7—a promising prognostic biomarker and a
RNAs are the predominant transcript isoform from potential therapeutic target in colorectal cancer. Clin
hundreds of human genes in diverse cell types. PLoS Cancer Res 23(14):3918–3928
One 7(2):e30733 37. Barbagallo D, Condorelli A, Ragusa M et al (2016)
19. Memczak S, Jens M, Elefsinioti A et al (2013) Dysregulated miR-671-5p/CDR1-AS/CDR1/
Circular RNAs are a large class of animal RNAs with VSNL1 axis is involved in glioblastoma multiforme.
regulatory potency. Nature 495(7441):333 Oncotarget 7(4):4746
20. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular 38. Hirsch S, Blätte TJ, Grasedieck S et al (2017) Circular
RNAs are abundant, conserved, and associated with RNAs of the nucleophosmin (NPM1) gene in acute
ALU repeats. RNA 19(2):141–157 myeloid leukemia. Haematologica 102:2039–2047
21. You X, Vlatkovic I, Babic A et al (2015) Neural 39. Li P, Chen H, Chen S et al (2017) Circular RNA
circular RNAs are derived from synaptic genes and 0000096 affects cell growth and migration in gastric
regulated by development and plasticity. Nat Neurosci cancer. Br J Cancer 116(5):626
18(4):603 40. Zeng K, Chen X, Xu M et al (2018) CircHIPK3 pro-
22. Gruner H, Cortés-López M, Cooper DA et al (2016) motes colorectal cancer growth and metastasis by
CircRNA accumulation in the aging mouse brain. Sci sponging miR-7. Cell Death Dis 9(4):417
Rep 6:38907 41. Zheng Q, Bao C, Guo W et al (2016) Circular RNA
23. Hansen TB, Jensen TI, Clausen BH et al (2013) profiling reveals an abundant circHIPK3 that regu-
Natural RNA circles function as efficient microRNA lates cell growth by sponging multiple miRNAs. Nat
sponges. Nature 495(7441):384 Commun 7:11215
24. Legnini I, Di Timoteo G, Rossi F et al (2017) Circ-­ 42. Bachmayr-Heyda A, Reiner AT, Auer K et al (2015)
ZNF609 is a circular RNA that can be translated and Correlation of circular RNA abundance with prolifer-
functions in myogenesis. Mol Cell 66(1):22–37. e9 ation–exemplified with colorectal and ovarian cancer,
17 Circular RNAs in Cancer 229

idiopathic lung fibrosis, and normal human tissues. nary complexes with p21 and CDK2. Nucleic Acids
Sci Rep 5:8057 Res 44(6):2846–2858
43. Jin H, Jin X, Zhang H et al (2017) Circular RNA 59. Du WW, Fang L, Yang W et al (2017) Induction of
hsa-circ-0016347 promotes proliferation, invasion tumor apoptosis through a circular RNA enhancing
and metastasis of osteosarcoma cells. Oncotarget Foxo3 activity. Cell Death Differ 24(2):357
8(15):25571 60. Li Y, Zheng F, Xiao X et al (2017) CircHIPK3
44. Sand M, Bechara FG, Gambichler T et al (2016) sponges miR-558 to suppress heparanase expression
Circular RNA expression in cutaneous squamous cell in bladder cancer cells. EMBO Rep 18:e201643581
carcinoma. J Dermatol Sci 83(3):210–218 61. Li F, Zhang L, Li W et al (2015) Circular RNA ITCH
45. Sand M, Bechara FG, Sand D et al (2016) Circular has inhibitory effect on ESCC by suppressing the
RNA expression in basal cell carcinoma. Epigenomics Wnt/β-catenin pathway. Oncotarget 6(8):6001
8(5):619–632 62. Huang G, Zhu H, Shi Y et al (2015) cir-ITCH plays an
46. Guarnerio J, Bezzi M, Jeong JC et al (2016) inhibitory role in colorectal cancer by regulating the
Oncogenic role of fusion-circRNAs derived from Wnt/β-catenin pathway. PloS One 10(6):e0131225
cancer-associated chromosomal translocations. Cell 63. Guo W, Zhang J, Zhang D et al (2017) Polymorphisms
165(2):289–302 and expression pattern of circular RNA circ-ITCH
47. Zhong Z, Huang M, Lv M et al (2017) Circular RNA contributes to the carcinogenesis of hepatocellular
MYLK as a competing endogenous RNA promotes carcinoma. Oncotarget 8(29):48169
bladder cancer progression through modulating 64. Han D, Li J, Wang H et al (2017) Circular RNA MTO1
VEGFA/VEGFR2 signaling pathway. Cancer Lett acts as the sponge of miR-9 to suppress hepatocellular
403:305–317 carcinoma progression. Hepatology 66(4):1151
48. Xia W, Qiu M, Chen R et al (2016) Circular RNA 65. Yu J, Xu Q-g, Wang Z-g et al (2018) Circular RNA
has_circ_0067934 is upregulated in esophageal squa- cSMARCA5 inhibits growth and metastasis in hepa-
mous cell carcinoma and promoted proliferation. Sci tocellular carcinoma. J Hepatol 68:1214
Rep 6:35576 66. Fu L, Yao T, Chen Q et al (2017) Screening dif-
49. Chen J, Li Y, Zheng Q et al (2017) Circular RNA ferential circular RNA expression profiles reveals
profile identifies circPVT1 as a proliferative factor hsa_circ_0004018 is associated with hepatocellular
and prognostic marker in gastric cancer. Cancer Lett carcinoma. Oncotarget 8(35):58405
388:208–219 67. Li W, Zhong C, Jiao J et al (2017) Characterization
50. Wan X-Y, Chu Z-B, Hu Y et al (2017) CircPVT1 of hsa_circ_0004277 as a new biomarker for acute
inhibit apoptosis and enhance drug resistance in mul- myeloid leukemia via circular RNA profile and bioin-
tiple myeloma. Blood 130:3085 formatics analysis. Int J Mol Sci 18(3):597
51. Chen L, Zhang S, Wu J et al (2017) circRNA_100290 68. Yao Z, Luo J, Hu K et al (2017) ZKSCAN1 gene and
plays a role in oral cancer by functioning as a sponge its related circular RNA (circZKSCAN1) both inhibit
of the miR-29 family. Oncogene 36(32):4551 hepatocellular carcinoma cell growth, migration, and
52. Zhong Z, Lv M, Chen J (2016) Screening differential invasion but through different signaling pathways.
circular RNA expression profiles reveals the regula- Mol Oncol 11(4):422–437
tory role of circTCF25-miR-103a-3p/miR-107-CDK6 69. Rybak-Wolf A, Stottmeister C, Glažar P et al (2015)
pathway in bladder carcinoma. Sci Rep 6:30919 Circular RNAs in the mammalian brain are highly
53. Zheng J, Liu X, Xue Y et al (2017) TTBK2 circular abundant, conserved, and dynamically expressed.
RNA promotes glioma malignancy by regulating Mol Cell 58(5):870–885
miR-217/HNF1β/Derlin-1 pathway. J Hematol Oncol 70. Conn SJ, Pillman KA, Toubia J et al (2015) The RNA
10(1):52 binding protein quaking regulates formation of cir-
54. Zhang H, Wang G, Ding C et al (2017) Increased cRNAs. Cell 160(6):1125–1134
circular RNA UBAP2 acts as a sponge of miR-143 71. Le Gouill S, Podar K, Harousseau J-L et al (2004)
to promote osteosarcoma progression. Oncotarget Mcl-1 regulation and its role in multiple myeloma.
8(37):61687 Cell Cycle 3(10):1259–1262
55. Song X, Zhang N, Han P et al (2016) Circular RNA 72. Dolnik A, Engelmann JC, Scharfenberger-Schmeer
profile in gliomas revealed by identification tool M et al (2012) Commonly altered genomic regions
UROBORUS. Nucleic Acids Res 44(9):e87–e87 in acute myeloid leukemia are enriched for somatic
56. Yang Y, Gao X, Zhang M et al (2018) Novel role of mutations involved in chromatin remodeling and
FBXW7 circular RNA in repressing glioma tumori- splicing. Blood 120(18):e83–e92
genesis. JNCI: J Natl Cancer Inst 110(3) 73. Moalli PA, Pillay S, Krett NL et al (1993) Alternatively
57. Yang W, Du WW, Li X et al (2016) Foxo3 activity spliced glucocorticoid receptor messenger RNAs in
promoted by non-coding effects of circular RNA and glucocorticoid-resistant human multiple myeloma
Foxo3 pseudogene in the inhibition of tumor growth cells. Cancer Res 53(17):3877–3879
and angiogenesis. Oncogene 35(30):3919 74. Sun L, Goodman PA, Wood CM et al (1999)
58. Du WW, Yang W, Liu E et al (2016) Foxo3 circular Expression of aberrantly spliced oncogenic ikaros
RNA retards cell cycle progression via forming ter- isoforms in childhood acute lymphoblastic leukemia.
J Clin Oncol 17(12):3753–3766
230 S. Lux and L. Bullinger

75. Janssen JW, Ludwig W-D, Borkhardt A et al (1994) 86. Yu J, Zhang L (2009) PUMA, a potent killer with or
Pre-pre-B acute lymphoblastic leukemia: high fre- without p53. Oncogene 27(S1):S71
quency of alternatively spliced ALL1-AF4 transcripts 87. Qu H, Zheng L, Pu J et al (2015) miRNA-558 pro-
and absence of minimal residual disease during com- motes tumorigenesis and aggressiveness of neuro-
plete remission. Blood 84(11):3835–3842 blastoma cells through activating the transcription of
76. Trang P, Weidhaas JB, Slack FJ (2017) MicroRNAs heparanase. Hum Mol Genet 24(9):2539–2551
and cancer. In: The molecular basis of human cancer. 88. Polakis P (2000) Wnt signaling and cancer. Genes
Springer, New York, pp 277–286 Dev 14(15):1837–1851
77. Jongen-Lavrencic M, Sun SM, Dijkstra MK et al 89. Chen C-Y, Sarnow P (1995) Initiation of protein syn-
(2008) MicroRNA expression profiling in relation to thesis by the eukaryotic translational apparatus on cir-
the genetic heterogeneity of acute myeloid leukemia. cular RNAs. Science 268(5209):415–417
Blood 111(10):5078–5085 90. Yang Y, Fan X, Mao M et al (2017) Extensive transla-
78. Calin GA, Ferracin M, Cimmino A et al (2005) A tion of circular RNAs driven by N 6-methyladenosine.
MicroRNA signature associated with prognosis and Cell Res 27(5):626
progression in chronic lymphocytic leukemia. New 91. Zhou C, Molinie B, Daneshvar K et al (2017)
Engl J Med 353(17):1793–1801 Genome-wide maps of m6A circRNAs identify wide-
79. Hansen TB, Wiklund ED, Bramsen JB et al (2011) spread and cell-type-specific methylation patterns that
miRNA-dependent gene silencing involving are distinct from mRNAs. Cell Rep 20(9):2262–2276
Ago2-mediated cleavage of a circular antisense 92. Kelley DR, Hendrickson DG, Tenen D et al (2014)
RNA. EMBO J 30(21):4414–4422 Transposable elements modulate human RNA abun-
80. Yu C-Y, Li T-C, Wu Y-Y et al (2017) The circular dance and splicing via specific RNA-protein interac-
RNA circBIRC6 participates in the molecular cir- tions. Genome Biol 15(12):537
cuitry controlling human pluripotency. Nat Commun 93. Ivanov A, Memczak S, Wyler E et al (2015) Analysis
8(1):1149 of intron sequences reveals hallmarks of circular RNA
81. Guo JU, Agarwal V, Guo H et al (2014) Expanded biogenesis in animals. Cell Rep 10(2):170–177
identification and characterization of mammalian cir- 94. Shi H, Skibbe J, Shen C et al (2018) METTL14 inhib-
cular RNAs. Genome Biol 15(7):409 its hematopoietic stem/progenitor differentiation and
82. Jeck WR, Sharpless NE (2014) Detecting and charac- promotes Leukemogenesis via mRNA m 6 A modifi-
terizing circular RNAs. Nat Biotechnol 32(5):453 cation. Cell Stem Cell 22:1–15
83. Thomas LF, Sætrom P (2014) Circular RNAs are 95. Lin S, Choe J, Du P et al (2016) The m 6 A methyl-
depleted of polymorphisms at microRNA binding transferase METTL3 promotes translation in human
sites. Bioinformatics 30(16):2243–2246 cancer cells. Mol Cell 62(3):335–345
84. Xu L, Zhang M, Zheng X et al (2017) The circular 96. Zhang X-O, Wang H-B, Zhang Y et al (2014)
RNA ciRS-7 (Cdr1as) acts as a risk factor of hepatic Complementary sequence-mediated exon circulariza-
microvascular invasion in hepatocellular carcinoma. tion. Cell 159(1):134–147
J Cancer Res Clin Oncol 143(1):17–27 97. Chen YG, Kim MV, Chen X et al (2017) Sensing self
85. Panda AC, Grammatikakis I, Kim KM et al (2016) and foreign circular RNAs by intron identity. Mol
Identification of senescence-associated circular Cell 67(2):228–238. e5
RNAs (SAC-RNAs) reveals senescence suppressor 98. Abe N, Matsumoto K, Nishihara M et al (2015)
CircPVT1. Nucleic Acids Res 45(7):4021–4035 Rolling circle translation of circular RNA in living
human cells. Sci Rep 5:16435
 Circular RNAs in Brain Physiology
and Disease 18
S. Gokul and G. K. Rajanikant

Abstract Abbreviations
Circular RNAs (circRNAs) are endogenously
expressed non-coding RNAs discovered in the circRNAs Circular RNAs
early 1990s as a transcriptional by-product of GBM Glioblastoma multiforme
little importance. It was only recently that they MDD Major depressive disorder
were identified as a key player in regulating microRNAs MicroRNAs
the gene expression by targeting and modulat- mRNA Messenger RNA
ing the functions of microRNA, a process MSA Multiple system atrophy
known as microRNA sponging. They are dis- ncRNA Non-coding RNA
tributed throughout the system in a tissue-­ pre-mRNA Precursor mRNA
specific manner showing abundant enrichment SRSF1 Serine- and arginine-rich splicing
in neuronal tissue. Their physiological func- factor 1
tions in the brain such as neuronal maturation, TBI Traumatic brain injury
differentiation, etc. as well as their implica-
tions in numerous brain-related disorders have
made its entry into the spotlight. Yet the wider
scope and molecular mechanism of circRNAs 1 Introduction
still remain elusive. In this chapter, we
describe in detail the functional aspects and The widespread notion that the circular RNAs
importance of circRNAs in the human brain (circRNAs) are simply cellular artefacts or tran-
and how it is associated with various neuro- scriptional noises has changed recently, in the
logical diseases. light of several studies of late portraying their
diverse physiological functions. Circular RNAs
Keywords are one of the emerging classes of non-coding
Circular RNA · Brain · Ischemic stroke · RNAs, with a widespread expression across the
Neurodegenerative diseases · Brain tumour evolutionary tree of life [1]. The biogenesis of
circRNA follows a back-splicing mechanism,
wherein a covalent bond is formed between 5′
and 3′ splice sites of a pre-mRNA [2, 3]. This
S. Gokul · G. K. Rajanikant (*) unique structural conformation and higher stabil-
School of Biotechnology, National Institute of ity compared to other RNA species have raised
Technology Calicut, Calicut, India an immense interest to characterize and
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 231

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_18
232 S. Gokul and G. K. Rajanikant

u­ nderstand its mechanism within the scientific of circZfp609 was expressed at a constant level.
community. The expression of these RNAs in Intriguingly, the abundant expression of cir-
humans is said to be in a tissue type and develop- cRNAs in the brain has been attributed to a num-
mental stage-specific manner, with a relatively ber of reasons. First, there are a number of genes
high abundance in the brain, especially in neuro- that are hosts to circRNAs such as cevircRims2,
nal cells [4, 5]. Here we outline how circRNAs circTulp4, circElf2, circPhf21a and circMyst4,
modulate and regulate the normal physiology and and most of these host genes are expressed exclu-
development of brain and how their dysregula- sively in the brain [8]. Secondly, there is a high
tion is associated with various neurological dis- resemblance in the intronic regions of both cir-
orders. We conclude this chapter with a brief look cRNAs and neuronal genes, wherein, the latter
into the future of circRNA research in the context often have long intronic regions, while they are
of neurological diseases. highly conserved during evolution in the former.
Finally, the regulatory elements of circRNA bio-
genesis: cis- and trans-factors [9].
2 circRNAs in Brain Physiology It is previously known that evolutionarily con-
served genes display a striking functional simi-
2.1 Enriched circRNAs in the Brain larity and physiological significance. In this
regard, circRNAs present a well-conserved
The complex morphological nature of brain expression pattern in both mouse and human neu-
demands tight orchestrated regulation and con- ronal cells [7, 10, 11]. Besides, there is an
trol of gene expression throughout the entire extremely conserved exonic sequence around the
lifespan of an organism. This tight regulation can head-to-tail junctions of both mouse and rat [7].
be seen at various developmental stages such as All these evidences point to the fact that cir-
synaptogenesis, neurite outgrowth, neuronal mat- cRNAs might have a potential role in normal
uration, synaptic plasticity etc. Non-coding physiological functions of the brain.
RNAs such as microRNAs, long non-coding
RNAs, small interfering RNAs have immense
roles in regulating these developmental pro- 2.2 Neuronal Distribution
cesses. Recently, circRNAs have emerged as a of circRNAs
novel regulatory ncRNA with a higher level of
expression in the brain as well-documented by The structure of neurons is unique in its own way
several in vitro and in vivo studies [5–7]. Even with a long axon having protective covering
though 20% of the protein-coding genes in the (myelin sheath), large cell body and cellular pro-
brain produce circRNAs, their expression pattern jections (dendrites). The synapses or neuronal
is not uniformly distributed throughout the brain, junction, where neurons communicate with each
as some regions are more enriched than the oth- other, is considered as a circRNA-enriched
ers [8]. For example, circRims2, circEl2 and region, and it tends to accumulate with age [12].
circDym are highly expressed in the cerebellum, Since the synaptic density in the human brain is
whereas circPlxnd1 is enriched in the cortex. higher than in mouse, this is speculated to be one
Other regions of the brain showing high expres- of the reasons for the higher expression of cir-
sion patterns are the striatum, olfactory bulb and cRNAs in humans compared to the mouse brain
hippocampus [5]. However, the expression level [13]. This suggests that these highly enriched cir-
of circRNAs does not always correlate with the cRNAs might have some functional significance
expression of their linear isoforms. circRNAs in the neurons. Some of these circRNAs such as
such as circMyst4, circKlhl2 and circAagab CDR1as/CiRS-7, circRTN4, circTULP4 and cir-
whose expressions are significantly upregulated cRIMS2 were found to be involved in neuronal
during synaptogenesis had their mRNA level differentiation and developmental processes
downregulated [7], whilst the mRNA transcript
18 Circular RNAs in Brain Physiology and Disease 233

Fig. 18.1 Circular RNAs in the brain and their functional roles

whilst some in synaptogenesis, and they tend to microRNA sponging. circRNAs have multiple
show an increase in their expression level [5, 7]. miRNA-binding sites; for instance, one of the
Neuroplasticity is the most salient feature of highly abundant circRNA in neurons CDR1as/
the human brain, and the neural networks are ciRS-7 harbour 74 binding sites for miR-7 by
kept tightly under control and regulated through- which they competitively bind and supress their
out the lifespan of humans. circRNAs are one of function. CDR1as/ciRS-7 is one of the best stud-
the regulatory RNAs involved in this process as ied circRNA where they bind to the miRNA
evidenced by the latest report stating a dynamic effector protein Argonaute in an miR-7-­dependent
change in their expression level after neural plas- manner [14]. Besides, there are some reports
ticity. The circRNA which is linked to this pro- regarding circRNAs’ ability to regulate the tran-
cess was found to be circHomer1_a [7]. The scriptional activity in neurons and also in protein
summarized details of all the circRNAs involved and/or RNA transport within the cell [15]. Even
in brain physiology is given in (Fig. 18.1). though thousands of circRNAs were identified,
the functional role of only a handful of them is
known.
2.3  hysiological Role of circRNAs
P
in the Brain
3 CircRNAs and CNS Disease
The biological processes in which circRNAs are
involved have been mentioned in the previous 3.1 Cerebrovascular Diseases
section. Besides, most of the genes that host cir-
cRNAs upon gene ontology analysis were found Cerebrovascular diseases are group of events that
to be enriched in the process related to nervous affect the blood flow to the brain resulting in per-
system development, neurogenesis and differen- manent brain damage or even cell death. Ischemic
tiation. The most important role of circRNAs so stroke is one of the most common cerebrovascu-
far documented is to modulate the function of lar diseases and a leading cause of mortality and
microRNAs (miRNAs), a process known as disability, affecting millions of people ­worldwide.
234 S. Gokul and G. K. Rajanikant

Table 18.1 Circular RNAs in major brain-related disorders

circRNAs Disease Regulation References
mmu_circRNA_40001 Ischemic stroke Upregulated [16]
mmu_circRNA_013120
mmu_circRNA_40806 Downregulated
mmu_circRNA-015947 Ischemic stroke Upregulated [17]
circ_008018 circ_015350 circ_016128 Ischemic stroke Upregulated [18]
circ_011137 circ_001729 circ_006696 Ischemic stroke Downregulated [18]
circDLGAP4 Ischemic stroke Upregulated [19]
circ-TTBK2 Glioma Upregulated [20]
circSMARCA5 GBM Downregulated [24]
circBRAF Glioma Downregulated [23]
cZNF292 Glioma Upregulated [21]
circ-FBXW7 Glioma Downregulated [22]
IQCK MSA Upregulated [28]
MAP4K3 EFCAB11
DTNA
MCTP1
circzip-2 Parkinson’s disease Downregulated [27]
ciRS-7 Alzheimer’s disease Downregulated [25, 26]
circ_0005402 circ_0035560 MS Downregulated [30, 31]
hsa_circRNA_103636 MDD Downregulated [29]

The landscape of stroke research has expanded in They act as a microRNA-143 sponge (Table 18.2)
the last decade, allowing researchers to under- thereby inhibiting endothelial-mesenchymal
stand the molecular mechanism underlying its transition. This suggests a potential therapeutic
pathology. An extensive research relating non- role of circDLGAP4 for acute ischemic injury
coding RNAs role in stroke pathology can be [19].
seen lately, with circRNAs in particular gaining
wider attention amongst all of them. Association
of circRNAs with ischemic stroke pathology has 3.2 Cancer
recently been revealed and was found to be sig-
nificantly altered in the ischemic brain Cancer has been in the forefront of research for
(Table 18.1), indicating its significance as poten- more than two decades, due to its complex
tial biomarker (mmu_circRNA_40001, mmu_ molecular interactions and unique ability to
circRNA_013120 and mmu_circRNA_40806) escape and survive the normal homeostatic con-
for stroke. Furthermore, their gene targets were trol. circRNAs being one of the regulatory RNAs
involved in various signalling pathways such as have been linked to various cancer-related events
cell survival, death and neuroinflammation [16]. like cellular proliferation, migration and inva-
Mouse circRNA-015947 and circ_016423 were sion. For example, circ-TTBK2 in glioma tissue
significantly altered during ischaemia-­reperfusion acts as miR-217 sponge, thereby inhibiting its
injury indicating its impact in post-stroke pathol- function in a sequence specific manner, resulting
ogy. circRNA-015947 could interact with five in increased level of HNF1β (a direct target of
microRNA targets, thereby enhancing their target miR-217) that has an oncogenic role [20]. In con-
gene expression, whilst circ_016423 showed 625 trast, circRNAs like cZNF292 and circ-FBXW7
miRNA-binding sites that can bind to 521 differ- exert an inhibitory role preventing cell cycle pro-
ent miRNAs [17, 18]. circRNA DLGAP4 gression and cell proliferation in glioma cells
(circDLGAP4) upregulation significantly [21, 22]. circBRAF identified as a negatively
reduced infarction size and blood-brain barrier downregulated circRNA in glioblastoma multi-
damage in a mouse model for ischemic stroke. forme (GBM) has been associated with
18 Circular RNAs in Brain Physiology and Disease 235

Table 18.2 circRNAs and their known miRNA targets

S.No. Circular RNA miRNA Reference
1 ciRS-7 miR-7a [25]
2 circ-TTBK2 miR-217 [20]
3 mmu-circRNA-015947a mmu-miR-188-3p [17]
mmu-miR-329-5p
mmu-miR-3057-3p
mmu-miR-5098
mmu-miR-683
4 circDLGAP4 miR-143 [19]
5 mmu_circRNA_40806a miR-149-5p [16]
mmu-miR-346-3p
mmu-miR-20a-3p
6 circSMARCA5 RBP SRSF1 [24]
7 circzip-2a miR-60-3p [27]
8 hsa_circRNA_103636a hsa-miR-890 [29]
hsa-miR-617
hsamiR-520a
hsa-miR-15b-3p
hsa-miR-103a-2–5p
Predicted targets
a

t­umorigenesis and development of GBM, thus 3.4 Other Maladies

making it a potential biomarker to assess the
prognosis in glioma patients [23]. circSMARCA5 Circular RNAs have been linked to numerous
overexpression has been linked to decreased other nervous system-related disease such as
migration of cells in glioma cell lines. They have traumatic brain injury (TBI), major depressive
a higher binding affinity to serine- and arginine- disorder (MDD) and multiple sclerosis (MS).
rich splicing factor 1 (SRSF1), a positive control- MDD or major depressive disorder is one of the
ler of cell migration often overexpressed in most common problem affecting millions of peo-
GBM. Thus, targeting circSMARCA5 may be a ple worldwide. It is characterized by low mood
viable therapeutic choice in GBM [24]. swing and causes a serious threat to the health of
a person. Microarray analysis reports a consider-
able number of circRNAs to be differentially
3.3 Neurodegenerative Diseases regulated, in which hsa_circRNA_103636 has
been considered as a potential biomarker for
Neurodegenerative diseases are a group of condi- MDD [29]. Similarly circ_0005402 and
tions characterized by progressive degeneration circ_0035560 are considered as potential bio-
of structure and functions of neurons in the brain. marker for multiple sclerosis, a debilitating auto-
The most common of them are Alzheimer’s dis- immune disorder that destroys the myelin sheath
ease, Parkinson's disease and Huntington’s dis- in neurons [30, 31]. Astrocyte activation can be
ease. Some of the important circRNAs linked to seen in numerous neurological diseases and
neurodegenerative diseases are CDR1as/ciRS-7 in potentially worsen the inflammatory reactions
Alzheimer’s [25, 26], circzip-2 in Parkinson’s and neuronal tissue damage. Therapeutic strate-
[27] and IQCK, MAP4K3, EFCAB11, DTNA gies to alleviate astrocyte activation have been
and MCTP1 circRNAs in multiple system atrophy shown to have a significant impact on
(MSA) [28]. These studies suggest a strong con- ­experimental neuroinflammatory models. circH-
nection between circRNAs and neurodegenera- IPK2 has recently been proven to inhibit astro-
tive diseases, making them a potential therapeutic cyte activation by targeting MIR124-2HG via the
target and a disease biomarker. regulation of autophagy and endoplasmic reticu-
236 S. Gokul and G. K. Rajanikant

lum (ER) stress, thus acting as a potential thera- 10. Chen BJ, Yang B, Janitz M (2018) Region-specific
expression of circular RNAs in the mouse brain.
peutic target in the context of neuroinflammatory Neurosci Lett 666:44–47
disorders and drug abuse [32]. 11. Reddy AS, O'Brien D, Pisat N et al (2017) A com-
prehensive analysis of cell type-specific nuclear RNA
from neurons and glia of the brain. Biol Psychiatry
81(3):252–264
4 Conclusion 12. Gruner H, Cortes-Lopez M, Cooper DA et al (2016)
CircRNA accumulation in the aging mouse brain. Sci
Circular RNAs are one of the regulatory ncRNAs Rep 6:38907
that have been in the spotlight of research as of 13. Daniel C, Silberberg G, Behm M et al (2014) Alu
elements shape the primate transcriptome by cis-­
late yet still in its preliminary stages. To the best regulation of RNA editing. Genome Biol 15(2):R28
extent, what we know so far is the functional 14. Hansen TB, Jensen TI, Clausen BH et al (2013)
importance of circRNAs in developmental stages Natural RNA circles function as efficient microRNA
of the human brain and in disease pathology. The sponges. Nature 495(7441):384–388
15. van Rossum D, Verheijen BM, Pasterkamp RJ (2016)
wider picture lurks clarity, as more experimental Circular RNAs: novel regulators of neuronal develop-
studies are needed to identify their target and its ment. Front Mol Neurosci 9:74
mechanism of regulation. By understanding the 16. Liu C, Zhang C, Yang J et al (2017) Screening
complex interlink and regulatory actions of these circular RNA expression patterns following
focal cerebral ischemia in mice. Oncotarget
RNAs, better therapeutic approaches can be 8(49):86535–86547
taken towards nervous system-related diseases. 17. Lin SP, Ye S, Long Y et al (2016) Circular RNA
expression alterations are involved in OGD/R-induced
Competing Financial Interests The authors declare no neuron injury. Biochem Biophys Res Commun
competing financial interests. 471(1):52–56
18. Mehta SL, Pandi G, Vemuganti R (2017) Circular
RNA Expression Profiles Alter Significantly in
Mouse Brain After Transient Focal Ischemia. Stroke
References 48(9):2541–2548
19. Bai Y, Zhang Y, Han B et al (2018) Circular RNA
1. Wang PL, Bao Y, Yee MC et al (2014) Circular RNA DLGAP4 ameliorates ischemic stroke outcomes
is expressed across the eukaryotic tree of life. PLoS by targeting miR-143 to regulate endothelial-­
One 9(6):e90859 mesenchymal transition associated with blood-brain
2. Chen LL (2016) The biogenesis and emerging roles of barrier integrity. J Neurosci 38(1):32–50
circular RNAs. Nat Rev Mol Cell Biol 17(4):205–211 20. Zheng J, Liu X, Xue Y et al (2017) TTBK2 circular
3. Greene J, Baird AM, Brady L et al (2017) Circular RNA promotes glioma malignancy by regulating
RNAs: biogenesis, function and role in human dis- miR-217/HNF1beta/Derlin-1 pathway. J Hematol
eases. Front Mol Biosci 4:38 Oncol 10(1):52
4. Salzman J, Chen RE, Olsen MN et al (2013) Cell-type 21. Yang P, Qiu Z, Jiang Y et al (2016) Silencing of
specific features of circular RNA expression. PLoS cZNF292 circular RNA suppresses human glioma
Genet 9(9):e1003777 tube formation via the Wnt/beta-catenin signaling
5. Rybak-Wolf A, Stottmeister C, Glazar P et al (2015) pathway. Oncotarget 7(39):63449–63455
Circular RNAs in the mammalian brain are highly 22. Yang Y, Gao X, Zhang M et al (2018) Novel role of
abundant, conserved, and dynamically expressed. FBXW7 circular RNA in repressing glioma tumori-
Mol Cell 58(5):870–885 genesis. J Natl Cancer Inst 110(3)
6. Westholm JO, Miura P, Olson S et al (2014) Genome-­ 23. Zhu J, Ye J, Zhang L et al (2017) Differential expres-
wide analysis of drosophila circular RNAs reveals their sion of circular RNAs in glioblastoma multiforme
structural and sequence properties and age-dependent and its correlation with prognosis. Transl Oncol
neural accumulation. Cell Rep 9(5):1966–1980 10(2):271–279
7. You X, Vlatkovic I, Babic A et al (2015) Neural 24. Barbagallo D, Caponnetto A, Cirnigliaro M et al
circular RNAs are derived from synaptic genes and (2018) CircSMARCA5 inhibits migration of glioblas-
regulated by development and plasticity. Nat Neurosci toma multiforme cells by regulating a molecular axis
18(4):603–610 involving splicing factors SRSF1/SRSF3/PTB. Int
8. Hanan M, Soreq H, Kadener S (2017) CircRNAs in J Mol Sci 19(2)
the brain. RNA Biol 14(8):1028–1034 25. Shi Z, Chen T, Yao Q et al (2017) The circular
9. Chen W, Schuman E (2016) Circular RNAs in RNA ciRS-7 promotes APP and BACE1 degrada-
brain and other tissues: a functional enigma. Trends tion in an NF-kappaB-dependent manner. FEBS
Neurosci 39(9):597–604 J 284(7):1096–1109
18 Circular RNAs in Brain Physiology and Disease 237

26. Zhao Y, Alexandrov PN, Jaber V et al (2016) therapeutic biomarker in Major depressive disorder.
Deficiency in the ubiquitin conjugating enzyme Biomark Med 10(9):943–952
UBE2A in Alzheimer’s Disease (AD) is linked to def- 30. Cardamone G, Paraboschi EM, Rimoldi V et al (2017)
icits in a natural circular miRNA-7 sponge (circRNA; The characterization of GSDMB splicing and backs-
ciRS-7). Genes (Basel) 7(12):116 plicing profiles identifies novel isoforms and a circu-
27. Kumar L, Shamsuzzama, Jadiya P et al (2018) lar RNA that are dysregulated in multiple sclerosis.
Functional characterization of novel circular Int J Mol Sci 18(3)
RNA molecule, circzip-2 and its synthesizing 31. Iparraguirre L, Munoz-Culla M, Prada-Luengo I
gene zip-2 in C. elegans model of Parkinson’s et al (2017) Circular RNA profiling reveals that cir-
disease. Mol Neurobiol. https://doi.org/10.1007/ cular RNAs from ANXA2 can be used as new bio-
s12035-018-0903-5 markers for multiple sclerosis. Hum Mol Genet
28. Chen BJ, Mills JD, Takenaka K et al (2016) 26(18):3564–3572
Characterization of circular RNAs landscape 32. Huang R, Zhang Y, Han B et al (2017) Circular RNA
in multiple system atrophy brain. J Neurochem HIPK2 regulates astrocyte activation via cooperation
139(3):485–496 of autophagy and ER stress by targeting MIR124-­
29. Cui X, Niu W, Kong L et al (2016) hsa_cir- 2HG. Autophagy 13(10):1722–1741
cRNA_103636: potential novel diagnostic and
 Circular RNA and Alzheimer’s
Disease 19
Rumana Akhter

Abstract relationship of circRNA with AD and inter-

Circular RNAs (circRNAs) represent a special play of miRNA-mRNA-mediated genetic reg-
group of noncoding single-stranded highly ulatory networks. Our conceptual
stable ribonucleic acid entities abundant in the understanding also suggests that circRNA can
eukaryotic transcriptome. These circular be considered as a potential biomarker and
forms of RNAs are significantly enriched in therapeutic target in AD diagnosis and
human brain and retinal tissues. However, the treatment.
biological evolution and function of these cir-
cRNAs are poorly understood. Recent reports Keywords
showed circRNA to be an important player in circRNA · Alzheimer’s disease · Amyloid ·
the development of neurodegenerative dis- CDR1as miR-7 · UBE2A
eases like Alzheimer’s disease. With the pro-
gression of age, circRNA level increases in the
brain and also in age-associated neurological
disorder like Alzheimer’s disease (AD), 1 Introduction
Parkinson’s disease, inflammatory neuropa-
thy, nervous system neoplasms, and prion dis- Alzheimer’s disease (AD) is a progressive neuro-
eases. One highly represented circRNA in the degenerative disorder and the most common
human brain and retina is a ciRS-7 (CDR1as) cause of dementia in aging population. AD is
which acts as an endogenous, anticomplemen- clinically apparent as the insidious impairment of
tary miRNA inhibitor or “sponge” to quench higher intellectual functions that ultimately leads
the normal functioning of miRNA-7. Low to death from complete brain failure. There are
CDR1as level can lead to increase in miR-7 regions of the brain that are specifically affected
expression which downregulates the activity sites of neuropathology in Alzheimer’s disease,
of ubiquitin protein ligase A (UBE2A), an and they include the hippocampus, the amygdala,
important AD target, functionally involved in the temporal cortex, and the frontal cortex.
clearing toxic amyloid peptides from AD Complex multifactorial interactions among
brain. This chapter focuses on the functional genetic, epigenetic, and environmental compo-
nents are responsible for causation of AD. Recent
R. Akhter (*) research has come up with an interesting entity of
Cleveland Clinic Lerner Research Institute, RNA, an endogenous noncoding circular RNA
Cleveland, OH, USA (circRNA) abundantly expressed in eukaryotes,
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 239

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_19
240 R. Akhter

which has important functions in many disease-­ prominence of stable circRNAs in the synapse
associated gene regulations including AD. These provides both the stability and flexibility of neu-
structures have the 3′ and 5′ ends joined together ronal networks which are vital to all behavior,
by covalent bonds giving a circular appearance including learning and memory. Future func-
which is unlike linear RNA. These molecules tional research should be directed in understand-
which are evolutionarily conserved had been dis- ing the effect of genetic perturbation of specific
missed as a rare, exotic RNA species for decades. circRNAs followed by phenotypic examination
which will address circRNA function in the ner-
vous system in respect to molecular memory.
2 circRNAs in Neural
Development
3 Role of circRNA
circRNAs are enriched in the brain. These neural in Alzheimer’s Disease
circRNAs are derived from synaptic genes named
Dscam and Homer1 or from genes with impor- circRNAs tend to accumulate during normal
tant functions in early neural development, such process of brain aging and thus make suscepti-
as genes involved in axon guidance, Wnt and ble to age-related neurodegenerative diseases
TGF-β signaling [1, 2]. These are localized like AD. This disease is the most common cause
mostly in neuronal cell bodies and neuropil [2] of dementia in elderly population characterized
and found to be highly enriched in hippocampal by the presence of neurotoxic senile amyloid
synaptosomes. The developmental role of expres- plaques, hyper-phosphorylated tau tangles, mas-
sion of circRNAs in the brain is determined by sive neuron death, and neuro-inflammation.
profiling the circRNA population in the hippo- According to the amyloid hypothesis, accumu-
campus over several stages: embryonic (E18), lation of Aβ in the brain is central to AD patho-
early postnatal (P1), postnatal at the beginning genesis [6]. Amyloid is a general term for
of synapse formation (P10), and late postnatal protein fragments of albuminoid proteinaceous
hippocampus following the establishment of material that the body produces normally.
mature neural circuits (P30) [2]. The existence β-amyloid is a 36–43 amino acid peptide frag-
of circRNA in vivo as well as in primary neuron ment clipped from amyloid precursor protein
cultures and cell lines suggests a diverse distri- (APP). Most cases of Alzheimer’s belong to the
bution of circRNAs within neurons. This also late-onset category, which occurs after age 60.
cites an important role for circRNAs in neuronal The reasons of late-­onset Alzheimer’s are not
development and plasticity [2, 3]. yet completely elucidated, but they include a
circRNAs are flanked by introns. Neuronal combination of genetic, environmental, and life-
genes usually have long (> 10 kb) introns which style factors that have an important influence on
are highly conserved in evolution. So the cir- disease susceptibility of a person. The single-
cRNAs flanked by long introns are mostly evolu- gene mutations are mostly directly responsible
tionarily conserved [4]. Alternative explanation for early-onset Alzheimer’s disease but do not
for conserved circRNA could be that introns in seem to be involved in late-­onset Alzheimer’s,
these genes are long for other reasons and that and thus a specific gene mutation does not cause
circRNA is produced due to recursive splicing, the late onset of the disease complex multifacto-
i.e., circRNAs are produced as by-product of the rial interactions among genetic, epigenetic, and
complex splicing [5]. environmental components which is responsible
The function of circRNA in mammalian brain for causation of AD. Familial AD is character-
remains to be defined. Since circRNAs are highly ized by the genetic mutations involved in Aβ
stable, they could serve as topologically complex peptide biogenesis which consists of four well-
platforms for protein or RNA transportation. The studied AD genes, the APP, PS1, PS2, and
19 Circular RNA and Alzheimer’s Disease 241

APOE. These genes exhibit mutations that 4 circRNA in Other

enhance the relative rate of generation of Aβ42, Neuropathies
the longer form of the p­ eptide that is much more
prone to oligomerization and fibrillation than circRNA and Parkinson’s disease Parkinson’s
Aβ40 [7].The spatial and temporal patterns of disease (PD) is a neurodegenerative disorder that
senile plaques consisting of fibrillar Aβ do not affects mainly dopamine-producing neurons in a
equate very well with the degree of dementia in specific area of the brain, substantia nigra pars
AD, and thus the traditional amyloid hypothesis compacta. CDR1as downregulates miR-7 [14]. It
remains debatable. In contrast, cognitive mal- has already been reported that miR-7 can inhibit
functioning displays a profound relationship the expression of α-synuclein, a crucial constitu-
with most common type of sporadic AD which ent of Lewy bodies in the PD brain. α-Synuclein
remains largely unknown. circRNA expressed protein is expressed highly in the diseased brain
in the human brain might play a causative role and considered as hallmark feature in PD patho-
in AD and other neurodegenerative conditions. genesis. Moreover, downregulation of α-synuclein
Interestingly, circRNA could emerge as a poten- by miR-7 protects cells against oxidative stress
tial therapeutic target in AD diagnosis and [15]. miRNA-7 can provide protection against
treatment. neuron death caused by 1-methyl-4-phenylpyri-
Although circRNA has been reported in many dinium (MPP+) by targeting the nuclear factor
diseases, their role in Alzheimer’s disease (NF)-κB signaling pathway [14, 16].
remains unclear. Interestingly, evolutionarily
conserved microRNA-7 which is highly abun-
dant in human brain is associated with a circRNA circRNA and Neoplasms High expression of
for miRNA-7 (ciRS-7, also known as CDR1as). CDR1as is evident in the brain cerebrum.
ciRS-7 contains multiple, tandem anti-miRNA-7 CDR1as is highly expressed in neuroblastomas
sequences that thereby act as an endogenous, and astrocytoma [17]. miR-7 was found to be
anticomplementary miRNA “sponge” to adsorb downregulated in astrocytoma and neuroblas-
and hence quench normal miRNA-7 functions [8, toma compared to other brain tissue. Another
9]. In the hippocampal CA1 region of sporadic study indicated that miR-7 could suppress EGFR
AD patients, miR-7 circRNA system is dysregu- expression in a glioblastoma cell line and down-
lated which is confirmed by Northern blot hybrid- regulate IRS-1 and IRS-2 expression by repress-
ization techniques and the circularity-sensitive ing protein kinase B [18]. CDR1as acts as
circRNA probe RNase R [8]. Downregulation of negative regulator of miR-7 [9, 14]. These evi-
ciRS-7 and ciRS-7 “sponging activities” might dences indicate possible role of circRNAs in the
increase endogenous miRNA-7 levels in AD pathophysiology of nervous system neoplasms.
[10]. The elevated miRNA-7, due to inhibition in
ciRS-7 “sponging” effects, can downregulate
AD-associated targets like ubiquitin protein circRNA and Neuro-inflammation Virus-­
ligase, UBE2A, and an autophagic, phagocytic associated miRNA binding sites are present in
protein essential in the clearance of amyloid pep- some circRNA which plays vital role in immuno-
tides in AD brain [11, 12]. Such miRNA-mRNA regulation as, for instance, hsa-circRNA 2149
regulatory systems may represent another crucial contains 13 unique, head-to-tail spanning reads.
aspect of epigenetic control over gene expression Hsa-circRNA 2149 is present in CD19+ leuko-
in health and disease. Inhibition of “miRNA cytes but not in CD341 leukocytes or neutrophils.
sponging systems” and increase of specific induc- Another circRNA, circRNA100783, has implica-
ible miRNAs might be a reason for downregula- tion in chronic CD28-associated CD8(+)T cell
tion of important genes related to sporadic AD aging which could be utilized as an important
brain [8, 13]. biomarker for this disease [14, 19]. circRNA
242 R. Akhter

from SRY can repress miR-138 activity. miR-138 deficient brains, indicating a possible molecular
downregulates runt-related transcription factor 3 link to the behavioral phenotype [22].
(RUNX3) which plays an essential role in the The most popular technology microarray was
regulation of T helper cells [14, 20]. These stud- a preferred way for global RNA expression anal-
ies provide indication of association of circRNA ysis before the arrival of next-generation sequenc-
with neuro-inflammation. ing (NGS), but it was not convenient to screen
circRNAs or its expression from linear counter-
parts. The high-throughput NGS technology has
circRNA and Prion Diseases Progressive neuro- provided a competent way to detect circRNAs.
degeneration is evident in Prion diseases with Many software packages came to the rescue to
neuronal loss and a failure to induce inflamma- decipher circRNAs from RNA-Seq data.
tory response. These diseases include Common RNA-seq protocols have limitations as
Creutzfeldt–Jakob disease (CJD), Gerstmann-­ it may introduce technical artifacts that can result
Straussler syndrome (GSS), and also fatal famil- in wrong identification of circRNA isoforms as
ial insomnia. Prion diseases are mainly caused by exonucleases might act upon some circRNA and
alteration of a normal cell-surface glycoprotein inhibit their expression [1, 23]. In mouse, deep
(PrPC) into a modified isoform (PrPSc) that ren- sequencing of multiple organs reveals signifi-
ders infectious nature to PrP in the absence of cantly greater fraction of circRNA junctional
nucleic acid. Reports show PrPC overexpression reads. Similar reports were found in human tis-
induces CDR1as expression [21]. Thus, CDR1as sues [3, 24]. circRNAs show different patterns of
might have some implication in the etiology of expression respective to brain areas which
prion diseases. include the striatum, prefrontal cortex, olfactory
cortex, cerebellum, and hippocampus. A gene
ontology analysis of the transcripts generating
5 circRNA Detection circRNAs reveals synaptic genes encoding pre-
in the Brain and postsynaptic functional groups are fortified
as circRNA host genes and thus provide an
Numerous strategies are employed to detect important reasoning for the abundance of cir-
genome-wide circRNA expression over the past cRNA in the brain. Along with that, many cir-
few years, but little is known to find out the accu- cRNAs are highly distributed in synaptic fractions
racy of these approaches. Experimental and bio- and synaptosomes [24]. Rare circRNA localiza-
informatic tools along with accurate statistical tion in cell body and dendrites of cultured hip-
approaches can help address these objectives of pocampal neurons and hippocampal slices can be
circRNA detection. Recently, scientists have targeted by high-resolution in situ hybridization
shown that circular RNA is associated with brain technique [24].
functions. When CDR1as, an RNA molecule
highly expressed in human and mouse brain, was
deleted from the genome of mice, the animal 6 Conclusion
brain failed to retain important information and
disregard the unnecessary ones like in other men- circRNA function and their relationships with
tal disorders [22]. Copious circRNAs are highly Alzheimer’s disease and other neuropathies
abundant in mammalian brain expression with remain to be fully elucidated. circRNAs are usu-
conserved expression. The well-known CDR1as ally abundant and found to be stable in vivo,
is strongly bound by miR-7 and miR-671 in the which might attribute to their importance in
human and mouse brain. Expression of these two molecular diagnostics. Until then, the role of cir-
microRNAs was posttranscriptionally dysregu- cRNA in gene regulation may be utilized as
lated in all brain regions. Early genes such as Fos, imperative treatment option. Importantly, the
a direct miR-7 target, were enhanced in CDR1as- potential role of circRNAs as miRNA sponges
19 Circular RNA and Alzheimer’s Disease 243

can be utilized as an innovative approach to regu- with intraneuronal amyloid-beta are associated
with autophagic defects in Alzheimer’s disease.
late gene expression. Further research on cir- J Alzheimers Dis 33(1):231–247
cRNA will enhance our understanding in relation 13. Ginsberg SD, Alldred MJ, Che S (2012) Gene expres-
to neuropathies like AD and lead to new diagnos- sion levels assessed by CA1 pyramidal neuron and
tic biomarkers and promising therapeutic options. regional hippocampal dissections in Alzheimer’s dis-
ease. Neurobiol Dis 45(1):99–107
14. Shao Y, Chen Y (2016) Roles of circular RNAs in
neurologic disease. Front Mol Neurosci 9:25
References 15. Junn E, Lee KW, Jeong BS et al (2009) Repression
of alpha-synuclein expression and toxic-
1. Veno MT, Hansen TB, Veno ST et al (2015) Spatio-­ ity by microRNA-7. Proc Natl Acad Sci USA
temporal regulation of circular RNA expression dur- 106(31):13052–13057
ing porcine embryonic brain development. Genome 16. Choi DC, Chae YJ, Kabaria S et al (2014) MicroRNA-7
Biol 16:245 protects against 1-methyl-4-­ phenylpyridinium-
2. You X, Vlatkovic I, Babic A et al (2015) Neural induced cell death by targeting RelA. J Neurosci
circular RNAs are derived from synaptic genes and 34(38):12725–12737
regulated by development and plasticity. Nat Neurosci 17. Chen YT, Rettig WJ, Yenamandra AK et al (1990)
18(4):603–610 Cerebellar degeneration-related antigen: a highly
3. Rybak-Wolf A, Stottmeister C, Glazar P et al (2015) conserved neuroectodermal marker mapped to chro-
Circular RNAs in the mammalian brain are highly mosomes X in human and mouse. Proc Natl Acad
abundant, conserved, and dynamically expressed. Sci USA 87(8):3077–3081; Dropcho EJ, Chen YT,
Mol Cell 58(5):870–885 Posner JB et al (1987) Cloning of a brain protein iden-
4. Ivanov A, Memczak S, Wyler E et al (2015) Analysis tified by autoantibodies from a patient with paraneo-
of intron sequences reveals hallmarks of circular RNA plastic cerebellar degeneration. Proc Natl Acad Sci
biogenesis in animals. Cell Rep 10(2):170–177; Jeck USA 84(13):4552–4556
WR, Sorrentino JA, Wang K et al (2013) Circular 18. Liu Z, Jiang Z, Huang J et al (2014) miR-7 inhibits
RNAs are abundant, conserved, and associated with glioblastoma growth by simultaneously interfering
ALU repeats. RNA 19(2):141–157 with the PI3K/ATK and Raf/MEK/ERK pathways. Int
5. Burnette JM, Miyamoto-Sato E, Schaub MA et al J Oncol 44(5):1571–1580
(2005) Subdivision of large introns in Drosophila by 19. Wang YH, Yu XH, Luo SS et al (2015)
recursive splicing at nonexonic elements. Genetics Comprehensive circular RNA profiling reveals that
170(2):661–674 circular RNA100783 is involved in chronic CD28-
6. Hardy J, Selkoe DJ (2002) The amyloid hypothesis associated CD8(+)T cell ageing. Immun Ageing
of Alzheimer’s disease: progress and problems on the 12:17
road to therapeutics. Science 297(5580):353–356 20. Fu D, Yu W, Li M et al (2015) MicroRNA-138 regu-
7. Bertram L, Tanzi RE (2008) Thirty years of lates the balance of Th1/Th2 via targeting RUNX3 in
Alzheimer’s disease genetics: the implications psoriasis. Immunol Lett 166(1):55–62
of systematic meta-analyses. Nat Rev Neurosci 21. Satoh J, Obayashi S, Misawa T et al (2009) Protein
9(10):768–778 microarray analysis identifies human cellular prion
8. Lukiw WJ (2013) Circular RNA (circRNA) in protein interactors. Neuropathol Appl Neurobiol
Alzheimer’s disease (AD). Front Genet 4:307 35(1):16–35; Satoh J, Yamamura T (2004) Gene
9. Hansen TB, Jensen TI, Clausen BH et al (2013) expression profile following stable expression
Natural RNA circles function as efficient microRNA of the cellular prion protein. Cell Mol Neurobiol
sponges. Nature 495(7441):384–388 24(6):793–814
10. Cogswell JP, Ward J, Taylor IA et al (2008) 22. Piwecka M, Glazar P, Hernandez-Miranda LR et al
Identification of miRNA changes in Alzheimer’s dis- (2017) Loss of a mammalian circular RNA locus
ease brain and CSF yields putative biomarkers and causes miRNA deregulation and affects brain func-
insights into disease pathways. J Alzheimers Dis tion. Science 357(6357)
14(1):27–41 23. Guo JU, Agarwal V, Guo H et al (2014) Expanded
11. Bingol B, Sheng M (2011) Deconstruction for recon- identification and characterization of mammalian cir-
struction: the role of proteolysis in neural plasticity cular RNAs. Genome Biol 15(7):409
and disease. Neuron 69(1):22–32 24. Chen W, Schuman E (2016) Circular RNAs in
12. Lonskaya I, Shekoyan AR, Hebron ML et al (2013) brain and other tissues: a functional enigma. Trends
Diminished parkin solubility and co-localization Neurosci 39(9):597–604
 Circular RNA in Liver: Health
and Diseases 20
Meiyi Song, Lu Xia, Mengxue Sun,
Changqing Yang, and Fei Wang

Abstract 1 Introduction
Circular RNA (circRNA) is an important class
of noncoding RNA characterized by covalently Circular RNAs (CircRNAs) are a class of non-
closed continuous loop structures. In recent coding RNAs (ncRNAs) that forms covalently
years, the various functions of circRNAs have closed continuous loop structures, lacking the
been continuously documented, including terminal 5′ and 3′ ends [1, 2]. For quite a long
effects on cell proliferation and apoptosis and time, circRNA was excluded from the scope of
nutrient metabolism. The liver is the largest research for being regarded as low abundant
solid organ in mammals, and it also performs RNA and a splicing error without any function.
many functions in the body, which is considered In a variety of organisms, including plants, nema-
to be the busiest organ in the body. At the same todes, drosophila, mice, and human, there are
time, the liver is vulnerable to multiple patho- abundant circRNA expression [3–7]. Most of
genic factors, causing various acute and chronic human circRNAs are transcribed from single or
liver diseases. The pathogenesis of liver disease multiple exons, known as exonic circRNAs.
is still not fully understood. As a rising star for However, many studies have shown that the intri-
the past few years, circRNAs have been proven cate mechanisms of circRNA splicing, including
involved in the regulation of liver homeostasis inverted repeated ALU pairs [8–10], reverse
and disease. This chapter will explain the role of complementary sequences and exon skipping
circRNAs in liver health and diseases and sort [11–14]. CircRNAs can be produced by multiple
out the confusion in the present study. genetic structures, while the same location of a
gene can produce different types of circRNAs.
Keywords Recent researches have shown that circular RNAs
Circular RNA · Liver · MicroRNA · mainly function in four ways: (1) circRNAs may
Noncoding RNA function as miRNA sponges; (2) circular RNAs
are involved in transcription and translation regu-
lation; (3) circRNAs may inhibit the splicing of
Authors Meiyi Song and Lu Xia have been contributed
linear RNA; and (4) circRNAs regulate gene
equally to this chapter.
expression as sponges of molecules (RBPs), such
M. Song · L. Xia · M. Sun · C. Yang (*) · F. Wang (*)
as RNA-binding protein components [15–17].
Division of Gastroenterology and Hepatology,
Digestive Disease Institute, Shanghai Tongji Hospital, The latest research also suggests that circRNAs
Tongji University School of Medicine, can also encode proteins, which raises questions
Shanghai, China about whether they are still classified as noncod-
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 245

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_20
246 M. Song et al.

ing RNAs [18]. CircRNAs function in the growth the authors still believe that the expression of cir-
or development of tissues or organs and regula- cRNAs needed to be confirmed in a larger num-
tion of cell cycle [19, 20], cell apoptosis [21, 22], ber of human samples. Although there have not
cell stress [23, 24], cellular senescence [25, 26], been many studies describing the expression pat-
and inflammatory responses [22, 27]. tern of the liver under a physiological state no
The liver is the largest solid organ of the matter in mice or human, the rich expression of
human body, with the functions of immunity, circRNAs in the liver has been basically revealed,
detoxification, biosynthesis, and metabolism. It which indicates that circRNAs may be closely
is responsible for the metabolism of lipids, sug- related to liver function.
ars, proteins, and vitamins in the body, which is a
veritable metabolism factory [28–30]. In addi-
tion, the liver is the only organ with regenerative 3  ircular RNAs in Liver
C
function in the body, less than 25% of the normal Damage and Repair
liver can form a complete liver. But even then, the
liver is susceptible to various factors that cause Many pathogenic factors including drugs, alco-
acute or chronic liver injury, and chronic liver hol, and virus can lead to liver injuries [33–35].
disease can lead to liver failure or liver cancer, Since the liver itself has a great capacity to repair
which is a serious threat to human health. There after damage, the process of regeneration can be
is a lot of evidence to explain the relationship initiated immediately since the attack of the
between circRNAs and the liver homeostasis and pathogenic factors. Damage in the liver results in
diseases; here we review the role of circRNAs in injury-related diseases. Poor or insufficient
the liver. regeneration in the liver will eventually lead to
liver fibrosis, irreversible liver cirrhosis, and even
liver cancer, so-called chronic liver diseases.
2 Identification of Circular CircRNAs are a group of important members of
RNAs in the Liver the noncoding RNAs involved in both the dam-
age and the repair process in the liver. The role of
The existence of circRNA was reported as early circRNAs is gradually revealed (Fig. 20.1).
as the 1990s [31, 32]. The latest research on cir-
cRNA expression of six adult and fetal normal
tissues by RNA-seq found that 36.97–50.04% of 3.1  onalcoholic Fatty Liver
N
circRNAs detected were organ-specific; mean- Disease (NAFLD)
while, 33 circRNAs were universally expressed
in almost all tested tissues. Six hundred sixty-­ Nonalcoholic fatty liver disease (NAFLD) is the
eight circRNAs were specifically expressed in most common liver disease worldwide. Between
liver tissues. Further analysis of the regulatory 75 million and 100 million individuals are esti-
network found that 17 circRNAs, 22 miRNAs, mated to suffer from NAFLD in America [36–
and 90 mRNAs constitute a tight network, form- 38]. Similar to alcoholic liver disease and viral
ing a circRNA-miRNA-mRNA interaction net- liver disease, liver inflammation usually occurs in
work. In contrast with fatal tissue, the particular NAFLD, which can lead to liver cirrhosis, liver
circRNA is higher expressed in the same tissue of cancer, and even liver failure [39]. NAFLD is
adults. Lin Li et al. used RAISE to detect cir- divided into nonalcoholic fatty liver and nonalco-
cRNAs in RNA-seq data. They identified RAISE holic steatohepatitis (NASH), which is associated
circRNA candidates in 61 human liver rRNA-­ with histological feature of liver fibrosis/cirrhosis
depleted samples. Finally, 8270 circRNA candi- from mild to severe [40, 41]. The typical patho-
dates for advanced analysis were authenticated, logical change of nonalcoholic fatty liver includes
and 59,128 circRNA candidates were identified mild nonspecific inflammation in the liver and
in HCC and adjacent non-tumor tissues. However, isolated hepatic steatosis [42–44]. When it comes
20 Circular RNA in Liver: Health and Diseases 247

Fig. 20.1 CircRNAs in liver diseases. HCC hepatocellular carcinoma, NAFLD nonalcoholic fatty liver disease, PPARα
peroxisome proliferators-activated receptors α

to the stage of nonalcoholic steatohepatitis, liver undergo acetylation and form fatty acyl-CoA and
inflammation will progress with or without fibro- then participate in the esterification reaction or
sis [45]. Hepatic steatosis with inflammation has β-oxidation [57]. Genetic and epigenetic regula-
a distinct and more progressive natural history tions are equally important in process of NAFLD
than isolated hepatic steatosis as the evidence [58–60].
from clinical and experimental research [46, 47]. Fatty acid synthesis and degradation [61, 62],
NAFLD/NASH is a multifactor comprehen- insulin resistance (IR) [63], oxidative stress [64,
sive disease, including metabolic, genetic, envi- 65], and hepatofibrogenesis [66–68] are regu-
ronmental, gut microbial [48], and epigenetic lated by the epigenetics regulation in hepatocyte.
factors. At present, scholars have established the Noncoding RNAs play pivotal roles through epi-
theory of “multiple-hit,” instead of the previous genetics regulation in NAFLD. Recent researches
“two-hit” pathogenesis [49], which is more have drawn attention to a novel class of noncod-
appropriate to describe the molecular and meta- ing RNA in NAFLD. The expression profile of
bolic changes in the occurrence of NAFLD [50, circRNA in NAFLD was first determined in 2016
51]. The hypothesis proposes that multiple [69]. The authors used MCD diet-induced NASH
attacks mentioned above are genetically predis- mice model and preformed circRNA and mRNA
posed to NAFLD and provide a more accurate microarray at the same time. Sixty-nine up- and
explanation for the pathogenesis of NAFLD [52]. 63 down-regulated circRNAs were found; mean-
Hepatic triglyceride accumulation is one of while, 2760 mRNAs up and 2465 mRNAs
the most significant features in NASH. Epigenetics down regulated in the same sample. They further
regulation plays a key role in the development of searched for the relationship between circRNAs
NAFLD. The esterification of glycerol and free and mRNAs changes based on circRNA-miRNA-­
fatty acids (FFAs) forms triglycerides [53]. The mRNA network. Four pathways were built,
synthetic triglycerides enter the storage or secre- including circ_002581-miR-122-Slc1a5,
tion pool, with varying degrees of lost [54–56]. circ_002581-miR-122-Plp2,
FFAs derived from dietary or adipose tissue ­circ_002581-miR-­122-Cpeb1, and circ_007585-
through fat decomposition and/or hepatic DNL miR-326-UCP2 based on the downstream miR-
248 M. Song et al.

NAs analysis, pathway analysis, and PCR involvement of circ_0046366 in the progress of
verification. Since circRNAs may function as NAFLD.
competitors of pre-­mRNA, they also described Besides, the functions of other circRNAs are
some possible circRNA-­ mRNA pairs, such as worthwhile to mention here. The circular RNA
circ_011775-­ Rn45s and circ_004300-Malat1. ciRS-7 can increase both insulin content in islet
Another research used metformin to treat high- cells and promote its secretion [72]. These pro-
fat diet (HFD)-induced NAFLD mice and then cesses may affect by miR-7 and its targets,
identified related noncoding RNAs. Focusing on including Myrip. It suggests that ciRS-7 may
circRNA response to HFD and metformin, they play a considerable role in regulating blood glu-
revealed 396 (231 up- and 165 downregulated) cose and further orchestrate the occurrence of
circRNAs and 222 (126 up- and 96 downregu- NAFLD.
lated) circRNAs respond to HFD and metformin,
respectively [70]. Unfortunately, they did not
observe a direct association between circRNAs 3.2 Liver Regeneration
and metabolism in liver diseases, whereas they
suggested the extensive interaction between The potential of the liver to repair damage or
some of the miRNAs and circRNAs was involved recover weight loss is marvelous. This process is
in the development of disease and required future always defective under fat stimulation or some
investigation. The function of circRNAs has been serious lesions, leading to fibrosis/cirrhosis of the
uncovered in hepatocytes steatosis. The regula- liver or other undesirable phenotypes. It is often
tion effect of circ_0046367 on NAFLD was thought that growth and differentiation are mutu-
revealed by functional experiment in human ally exclusive processes; as an exception, the
hepatocytes. Results showed decreased expres- liver maintains vital, highly differentiated func-
sion of circ_0046367 in FFA-induced hepato- tions throughout its growth. The majority of pres-
cytes steatosis. To explore the underlying ent studies are based on the in vivo model of liver
mechanism, they searched for the potential miR- cell proliferation post-partial hepatectomy.
NAs that regulated hepatocytes steatosis and However, liver cell damage and death and the
found that miR-34a could bind to circ_0046367 characteristics of nonsurgical resection of liver
and hence mediated the effect of the cir- cell regeneration are also involved in the process
cRNA. The inhibition of miR-34a on its target of liver regeneration (LR). The partial hepatec-
gene PPARα was also reduced by circ_0046367. tomy is the most commonly used model to study
In addition, lipid metabolism-related genes which the liver regeneration process. The process of
were associated with PPARα were significantly liver damage and repair is often interwoven, and
activated by circ_0046367 resulting in alleviated many factors can act double-edged simultane-
intracellular lipid accumulation, lipid peroxida- ously, such as tumor necrosis factor-α (TNF-α).
tion, and even mitochondrial dysfunction. In When liver resection is performed, the whole
patient with hepatic steatosis, changes in expres- organ enters a transient regeneration process,
sion levels of circ_0046367 and miR-34a are in without a stage of chronic injury. The process of
the opposite directions which confirm the inter- liver regeneration can be divided into three main
action of each other. Thereafter the same research phases: priming phase, proliferation phase, and
group identified another circRNA, circ_0046366 termination phase [73]. Various cytokines or hor-
acting as miR-34a sponge [71]. Circ_0046366 mones appear to serve as modulators of liver
decreasing is a significant feature in FFA-induced regeneration. Insulin and glucagon are involved
steatosis of hepatocytes. The increasing of miR-­ in energy metabolism. TNF-α, interleukin-6 (IL-­
34a and corresponding inhibition of its target 6), and other cytokines begin to be secreted in
genes as a result of circ_0046366 deficiency may priming phase [74, 75]. Most quiescent
lead to the transcriptional inhibition of lipometa- ­hepatocytes enter back into the cell cycle rapidly
bolic genes. This is the possible mechanism for from G0 phase and initiate the liver regeneration
20 Circular RNA in Liver: Health and Diseases 249

process. Within approximate 4 h after the injury, 10 years. Extracellular matrix (ECM) accumula-
the liver began to initiate its regeneration process tion leads to structural modification and dysfunc-
in mice [76]. During the proliferation phase, tion of the liver. Activated hepatic stellate cells
remaining hepatocytes start to mitose, under the (HSCs) make great contributions to the synthesis
stimulation of growth factors and metabolic sig- of ECM. In addition, the role of activation of
naling. The primary stage of liver weight gain hepatic Kupffer cells and lack of liver sinusoidal
occurs within 3 days after the partial hepatec- endothelial cells(LSECs) should be mentioned
tomy, and the liver can basically restore to its [78]. Inflammation and activation of Kupffer
original weight between 5 and 7 days after sur- cells prevent ECM secretion and deposition in
gery. Finally, various regulatory factors are HSCs [79]. Different from Kupffer cells, LSECs
applied to the liver to modulate the weight of the are regarded as a gatekeeper in the fibrotic pro-
liver and liver regeneration in termination phase. cess [80]. LSEC differentiation occurs during
The priming phase drew closer attention because capillarization before the fibrosis development,
in this phase the normally quiescent hepatocytes which has the function of allowing and promot-
reenter the cell cycle to proliferate in response to ing the activation of HSC [78]. Therefore, the
injury or infection, determining the fate of the anfractuous interaction between non-­
liver. Lifei Li et al. described the expression pat- parenchymal cells participates in the process of
tern of circRNAs during priming phase of rat LR liver fibrosis via affecting the activation of HSC.
(0, 2, and 6 h after surgery). Using high-­ In 1876, HSCs were observed and described
throughput RNA sequencing technology, 2412 for the first time by von Kupffer, between liver
circRNAs were identified. Three hundred of the sinusoidal endothelial cells (LSECs) and hepato-
circRNAs were detected at all three time points, cytes in the subendothelial space of Disse, which
among which 15 circRNAs were downregulated represent about 10% of all resident liver cells [81,
and 28 circRNAs were upregulated significantly 82]. HSC can transdifferentiate from quiescent,
in both 2 and 6 h after PH [77]. The following vitamin-A-storing cells to the myofibroblast
functional analysis of circRNAs and their host acquiring the function of proliferation, contractil-
genes was performed using gene ontology (GO) ity, fibrogenesis, altered matrix degradation, and
and Kyoto Encyclopedia of Genes and Genomes chemotaxis. TGF-β1 is essential for the activa-
(KEGG) biological pathway analysis to predict tion of HSCs, which result in expression of
possible pathway involved. The results indicated α-smooth muscle actin (α-SMA) and secretion of
that steroid hormone biosynthesis, inflammatory ECM proteins, mostly collagen Type I (ColI) [83,
mediator regulation of TRP channels were 84]. Targeting HSC has been the research focus
involved in. According to expression and struc- to regress liver fibrosis during the last half-­
ture verification, they finally verified two down- century [85]. Noncoding RNA is an especially
regulated circRNAs (circ_137 and circ_2270) important regulator of HSC activation, since reg-
and four upregulated circRNAs (circ_432, ulatory activities must occur quickly response to
circ_2077, circ_1366, and circ_15) played a cru- wounding stimulates, and directly activate or
cial role during the process of liver regeneration. inhibit target genes or function through post-­
They also highlighted the co-expression of miR- transcription regulation. MicroRNA and lncRNA
NAs with selected circRNAs. are reported to be involved in the activation of
HSC and thereby participate in the process of
liver fibrosis/cirrhosis [86]. Knockdown of Lnc-­
3.3 Liver Fibrosis/Cirrhosis LFAR1, a liver-enriched lncRNA, can impair the
activation of HSC, protect TGF-β-induced apop-
Liver fibrosis is the end-stage hepatopathy result- tosis of hepatocytes, and attenuate CCl4- and bile
ing from chronic liver injury due to any common duct ligation-induced liver fibrosis in vivo by
etiology. It can progress and develop irreversible activating TGF-β and Notch pathways. As
cirrhosis with a survival less than 50% in another member of the noncoding RNA, cir-
250 M. Song et al.

cRNAs are predicted to be associated with the types of cancer, including digestive system
HSC activation and may be a promising target for malignancy [95]. circ_100269 shows an inhibi-
treatment of liver fibrosis [87]. Compared to qui- tion effect on gastric cancer cell proliferation by
escent HSCs, 179 circRNAs were found to be targeting miR-630 [96]. Investigation of hepato-
upregulated, and 630 circRNAs were downregu- cellular carcinoma (HCC)-associated circRNAs
lated during irradiation-induced activation [88]. is undergoing. CiRS-7 (also named as CDR1) is
A pathway analysis was performed on the down- one of the famous intensive-studied circRNAs,
regulated circRNAs, and eight significantly asso- which function as a sponge of miR-7 [72, 97–99].
ciated pathways were revealed, including Yu et al. enrolled 35 HCC patients in the cohort
glycosaminoglycan degradation, pentose phos- and found an increased expression of ciRS-7 in
phate pathway, and phosphatidylinositol signal- human HCC tissue corresponding with downreg-
ing system. They detected that three circRNAs ulated miR-7 expression levels [100]. In addition,
(hsa_circ_0072765, hsa_circ_0071410, hsa_ they found that ciRS-7 accelerated the prolifera-
circ_0054345) were significantly upregulated tion and invasion of liver cancer cell by decreas-
and three circRNAs (hsa_circ_0070963, hsa_ ing the expression of miR-7 and relieving its
circ_0061893, hsa_circ_0013255) were signifi- inhibitory effect on its target gene CCNE1 and
cantly downregulated which was confirmed by PIK3CD. In another cohort consisting of 108
PCR results. The research further explored the HCC patients, the upregulation of ciRS-7 was
function of hsa_circ_0071410 in HSC activation; detected compared to the adjacent non-tumor tis-
three putative binding sites of miR-9-5p were sues [101]. Meanwhile, increased expression of
predicted. Suppression of miR-9-5p reversed the ciRS-7 in HCC tissues was significantly associ-
effect of hsa_circ_0071410 knockdown on pro- ated with serum AFP level (P < 0.01), hepatic
moting HSC activation. These preliminarily MVI (P = 0.03), and those with early onset
prompted that circRNAs may be involved in HSC (P = 0.02) and the deterioration of HCC
activation and thereby affect the process of liver (P = 0.08). The consistent expression change of
fibrosis development. Other circRNAs such as ciRS-7 and three target genes of miR-7, PIK3CD,
circ_000203 and circ_010567 have also been p70S6K, and mTOR were detected in the tissue
reported to be regulators implicated in the heart as well. However, the median time of disease-­
fibrogenesis [89, 90], but whether they are also free survival (DFS) had no significant difference
involved in liver fibrosis is worth further in ciRS-7 higher- or lower-expressed patients.
investigation. Decreased circ-MTO1 (mitochondrial translation
optimization 1 homologue) was detected in the
human tumor tissues, which was also named as
4  ircular RNAs in Liver
C hsa_circ_0007874/hsa_circ_104135 and related
Neoplasm to poor prognosis in patients [102]. Furthermore,
inhibition of circMTO1 accelerated hepatocellu-
4.1  he Roles of CircRNAs in Liver
T lar cancer cell proliferation by targeting miR-9
Carcinogenesis and abolished its inhibition on downstream target
genes P21 in vivo and in vitro. These all indicate
Liver cancer is the fifth most common cancer and potential therapeutic effect of circRNA on
the third leading cause of cancer death in the HCC. Both linear and circular (cir-cZKSCAN1)
world [91]. Proteins encoded by different genes forms of ZKSCAN1 RNA were detected down-
have been shown to play a role in the develop- regulated during HCC development [103]. The
ment and progression of liver cancer. These genes downregulation of cir-cZKSCAN1 was corre-
directly or indirectly regulate the cell cycle, cell lated with histological grade of tumor (National
apoptosis, DNA damage response, and cell sur- Comprehensive Cancer Network Clinical
vival signaling [92–94]. The function of circular Practice Oncology Guidelines), tumor numbers,
RNAs in cancers has been confirmed in various and underlying cirrhosis, vascular, invasion, or
20 Circular RNA in Liver: Health and Diseases 251

microscopic vascular invasion. In vivo and Intrahepatic cholangiocarcinoma (ICC) is the pri-
in vitro experiment showed that inhibition of mary malignant tumor in the liver after HCC with
both ZKSCAN1 mRNA and cir-cZKSCAN1 an increasing incidence in recent years [107,
accelerated tumor progression by facilitating cell 108]. It accounts for about 10–15% of the pri-
proliferation, migration, and invasion. mary malignant tumor of the liver. A TGF-β-­
Interestingly, linear and circular RNAs function induced long noncoding RNA (TLINC) was
through entirely different mechanisms without identified in ICC, also known as cancer suscepti-
interference on expression of HCC. Cir-­ bility candidate 15 (CASC15) [109]. Long and
cZKSCAN1 affected cancer-related signaling short TLINC isoforms existed and were related to
pathways including TGF-β1, ITGB4, and the epithelial to mesenchymal transition pheno-
CXCR4, while linear ZKSCAN1 mRNA was type. Circular isoforms of TLINC was verified in
found to be associated with cell metabolism. On TGF-β-induced Huh28 cells and human ICC tis-
the other hand, circRNA also reported to be a tar- sue, which was more stable to be a marker for the
get regulated by other protein of noncoding diagnosis.
RNAs during HCC genesis. The function of clus- Hepatoblastoma is the most common malig-
ter of differentiation (CD)90 (Thy-1) on liver nant liver cancer in children [110, 111].
cancer cell viability, migration, and invasive abil- Circ_0015756 was screened from 869 differen-
ities was revealed. After that, two specific cir- tially expressed circRNAs in human hepatoblas-
cRNAs, hsa_circ_0067531 and hsa_circ_0057096, toma, according to functional analysis of the host
were identified to be differently expressed in genes of above circRNAs [112]. Downregulation
CD90 expression or deletion cells [104]. of circ_0015756 reduced the cell viability of hep-
Additionally, the expression level of hsa_ atoblastoma and suppressed their proliferation
circ_0067531 was decreased in human HCC tis- and invasion in vitro. Functioning as a molecular
sue, while hsa_circ_0057096 had no significant sponge of miR-1250-3p, circ_0015756 shows a
difference in expression. But the regulatory effect bright prospect of curing hepatoblastoma.
and diagnosis value of the circRNAs should be
further studied. The interactions between in
RhoA, circ_000839, and miR-200b have been 4.2 Circular RNA as Biomarker
expounded. miR-200b occupies the core position for Liver Cancer
in the regulation network, which inhibits the
migration and invasion of HCC cell in vitro Searching for efficient biomarkers is the focus of
[105]. The expression of circ_000839 has been current translational medicine and is the most
found negatively related to that of miR-200b and concerned topic in clinical practice, especially
positively related to that of its target gene RhoA, for cancer [113]. On the one hand, the biomark-
encouraging us to study the function of ers in peripheral blood and body fluids are of
circ_000839 in detail in the future. Another cir- great significance for the early noninvasive diag-
cRNA has been reported to be involved in nosis of diseases; on the other hand, the discov-
HCC. Increased expression of circ_0067934 has ery of histological biomarkers is important to
been detected in HCC tissue [106]. Inhibition of pathological classification and the prognosis pre-
circ_0067934 can repress the proliferation, diction of diseases. Abundant, conserved, and
migration, and invasion of liver cancer cell. dynamic expression of circRNAs has been
Besides, the abovementioned effect of reported for many times, which indicate circRNA
circ_0067934 was exerted as a sponge of miR-­ is an emerging biomarker for cancer and related
1324 and constitutes a circ_0067934/miR-1324/ diseases, including HCC. Liyun Fu et al. estab-
FZD5/Wnt/β-catenin axis for regulating liver lished the expression pattern in HCC and explored
cancer. the association of particular circRNAs and the
Except HCC, the function of circRNAs in characteristics of HCC patients [114]. Hsa_
other types of liver cancer has also been reported. circ_0004018 was selected to estimate its value
252 M. Song et al.

for HCC diagnosis and evaluation. Low expres- one hand, we need to clarify which cell type cir-
sion of hsa_circ_0004018 was correlated with cRNAs mostly affect. Sometimes several cells
tumor size (P = 0.045), degree of tumor differen- play a synergistic role in mediating the same
tiation (P = 0.006), serum alpha-fetoprotein pathological process. For example, HSC activa-
(AFP) level (P = 0.027), Barcelona Clinic Liver tion is the key process in the pathogenesis of liver
Cancer (BCLC) stage (P = 0.040), and tumor fibrosis [119, 120]. Inflammation and activation
node metastasis (TNM) stage (P = 0.029) in of Kupffer cells and differentiation of LSECs can
patient with HCC. Further research found that indirectly activate HSC [121]. At the same time,
hsa_circ_0004018 was differently expressed in hepatocyte also transfers to myofibroblast
diverse chronic liver disease including chronic through epithelial-mesenchymal transition
hepatitis, cirrhosis, and HCC tissue (P < 0.001). (EMT) [122]. We believe that the identified cir-
It refers as a liver cancer specificity circRNA cRNAs differentially expressed in cirrhosis or
with an area under receiver operating characteris- fibrosis should be further accessed in specific cell
tic curve (ROC) of 0.848 (95% CI = 0.803–0.894, types and clarify its function at the level of indi-
P < 0.001). Compared to its adjacent non-tumor vidual cell type. On the other hand, it can be pru-
tissues, hsa_circ_0001649 is significantly down- dent to choose the strategy when the molecular
regulated in the liver cancer with an AUC of 0.63 mechanism is comprehensively understood.
[115]. Furthermore, hsa_circ_0001649 expres- There are four major ways in which circRNAs
sion in HCC tissue was related to tumor diameter regulate the biological process, including tran-
(P = 0.045), the cancer embolus (P = 0.017), and scription, translation, and post-transcriptional
metastasis. regulation. Moreover, recent research has
revealed the ability of circRNAs in coding pro-
tein. Based on up-to-date researches, circRNAs
5 Challenge of Research functioning as miRNA sponges have been
in CircRNAs reported the most [12, 123–126]. It is not only
because of the ubiquity of the regulation but also
In terms of what is mentioned above, there have because of the support of bioinformatics [127–
been many researches on the expression and 129]. Hundreds of binding sites of miRNA have
function of circRNAs in liver physiology and dif- been predicted on circRNAs, providing theoreti-
ferent liver pathological conditions. However, cal basis for further research. High-throughput
current studies still do not have a complete pic- test, bioinformatics analysis and the expression
ture of the source of circRNAs and whereabouts of linear RNA from the host gene, is still the key
in the liver and the underlying mechanisms. means of cracking the puzzle of circRNAs in
There is still a long way towards the practice of liver pathology and physiological processes.
applicable researches. First challenge we come The last question is the general problem of cir-
across is the complexity of the liver itself, includ- cRNA researches. It is well known that circRNAs
ing the complexity of its composition and dis- are characterized as tissue-specific and species-­
ease. There are many types of cells in the liver, specific [130–134]. The translational value of cir-
including parenchymal cells, at least three types, cRNAs may be greatly limited. The experimental
mesenchymal cells, and immune cells [116–118]. results obtained in mice, rats, and other model
They are all involved in multiple physiological organisms have encountered difficulties in the
and pathological processes of the liver, which application on human beings. The benefit is that
plays important roles in maintenance of liver organ specificity makes the treatment of targeting
homeostasis and development of liver diseases. circRNAs less interfered by off-target effect. In
These aspects increase the difficulty of the study terms of the homology of circRNAs, one is
on circRNAs, because a lot of works are required through the homology of host genes, and the
even after a circRNA is confirmed involved in a other is through the homology of circRNAs. In
certain pathological process in the liver. On the future studies on circRNAs, it is necessary to fur-
20 Circular RNA in Liver: Health and Diseases 253

ther address the homologous problems of cir- 10. Jeck WR, Sorrentino JA, Wang K et al (2013)
Circular RNAs are abundant, conserved, and associ-
cRNAs and to establish the bridge of circRNAs ated with ALU repeats. RNA 19(2):141–157
between different species. 11. Ivanov A, Memczak S, Wyler E et al (2015) Analysis
of intron sequences reveals hallmarks of circular
Acknowledgments This work was supported by the RNA biogenesis in animals. Cell Rep 10(2):170–177
grants from National Natural Science Foundation of 12. Andres-Leon E, Nunez-Torres R, Rojas AM (2016)
China (81670571 and 81370559 C. Yang; 81400635 to miARma-Seq: a comprehensive tool for miRNA,
F. Wang), Joint Projects in Major Diseases funding from mRNA and circRNA analysis. Sci Rep 6:25749
Shanghai Municipal Commission of Health and Family 13. Starke S, Jost I, Rossbach O et al (2015) Exon circu-
Planning (2014ZYJB0201 to C. Yang), Joint Projects for larization requires canonical splice signals. Cell Rep
Novel Frontier Technology in Shanghai Municipal 10(1):103–111
Hospital from Shanghai Municipal Commission of Health 14. Vicens Q, Westhof E (2014) Biogenesis of circular
and Family Planning (SHDC12014122 to C. Yang), RNAs. Cell 159(1):13–14
Shanghai Medical Guide Project from Shanghai Science 15. Legnini I, Di Timoteo G, Rossi F et al (2017) Circ-­
and Technology Committee (14411971500 to F. Wang), ZNF609 is a circular RNA that can be translated
grants from Chinese Foundation for Hepatitis Prevention and functions in Myogenesis. Mol Cell 66(1):22–
and Control (TQGB20140141 to F. Wang), and funds 37 e29
from Shanghai Innovation Program (12431901002 to 16. Pamudurti NR, Bartok O, Jens M et al (2017)
C. Yang). Translation of CircRNAs. Mol Cell 66(1):9–21
e27
17. Peng L, Yuan XQ, Li GC (2015) The emerging land-
Competing Financial Interests The authors declare no
scape of circular RNA ciRS-7 in cancer (review).
competing financial interests.
Oncol Rep 33(6):2669–2674
18. Han B, Chao J, Yao H (2018) Circular RNA and its
mechanisms in disease: from the bench to the clinic.
Pharmacol Ther 187:31. https://doi.org/10.1016/j.
pharmthera.2018.01.010
References 19. Du WW, Yang W, Liu E et al (2016) Foxo3 circular
RNA retards cell cycle progression via forming ter-
1. Errichelli L, Dini Modigliani S, Laneve P et al nary complexes with p21 and CDK2. Nucleic Acids
(2017) FUS affects circular RNA expression in Res 44(6):2846–2858
murine embryonic stem cell-derived motor neurons. 20. Xue J, Liu Y, Luo F et al (2017) Circ100284, via
Nat Commun 8:14741 miR-217 regulation of EZH2, is involved in the
2. Tan WL, Lim BT, Anene-Nzelu CG et al (2017) A arsenite-accelerated cell cycle of human kerati-
landscape of circular RNA expression in the human nocytes in carcinogenesis. Biochim Biophys Acta
heart. Cardiovasc Res 113(3):298–309 1863(3):753–763
3. Pan T, Sun X, Liu Y et al (2018) Heat stress 21. Zhang R, Xu J, Zhao J et al (2018) Silencing of hsa_
alters genome-wide profiles of circular RNAs in circ_0007534 suppresses proliferation and induces
Arabidopsis. Plant Mol Biol 96(3):217–229 apoptosis in colorectal cancer cells. Eur Rev Med
4. Hansen TB, Jensen TI, Clausen BH et al (2013) Pharmacol Sci 22(1):118–126
Natural RNA circles function as efficient microRNA 22. Cheng X, Zhang L, Zhang K et al (2018) Circular
sponges. Nature 495(7441):384–388 RNA VMA21 protects against intervertebral
5. Cortes-Lopez M, Gruner MR, Cooper DA et al disc degeneration through targeting miR-200c
(2018) Global accumulation of circRNAs during and X linked inhibitor-of-apoptosis protein. Ann
aging in Caenorhabditis elegans. BMC Genomics Rheum Dis 77:770. https://doi.org/10.1136/
19(1):8 annrheumdis-2017-212056
6. Kramer MC, Liang D, Tatomer DC et al (2015) 23. Holdt LM, Stahringer A, Sass K et al (2016) Circular
Combinatorial control of Drosophila circular RNA non-coding RNA ANRIL modulates ribosomal RNA
expression by intronic repeats, hnRNPs, and SR pro- maturation and atherosclerosis in humans. Nat
teins. Genes Dev 29(20):2168–2182 Commun 7:12429
7. Memczak S, Jens M, Elefsinioti A et al (2013) 24. Place RF, Noonan EJ (2014) Non-coding RNAs turn
Circular RNAs are a large class of animal RNAs up the heat: an emerging layer of novel regulators
with regulatory potency. Nature 495(7441):333–338 in the mammalian heat shock response. Cell Stress
8. Liang D, Wilusz JE (2014) Short intronic repeat Chaperones 19(2):159–172
sequences facilitate circular RNA production. Genes 25. Huang JL, Qin MC, Zhou Y et al (2018)
Dev 28(20):2233–2247 Comprehensive analysis of differentially expressed
9. Zhang XO, Wang HB, Zhang Y et al (2014) profiles of Alzheimer’s disease associated c­ircular
Complementary sequence-mediated exon circular- RNAs in an Alzheimer’s disease mouse model.
ization. Cell 159(1):134–147 Aging (Albany NY) 10(2):253–265
254 M. Song et al.

26. Qu S, Zhong Y, Shang R et al (2017) The emerging 40. Del Campo JA, Gallego-Duran R, Gallego P et al
landscape of circular RNA in life processes. RNA (2018) Genetic and epigenetic regulation in nonal-
Biol 14(8):992–999 coholic fatty liver disease (NAFLD). Int J Mol Sci
27. Yang X, Wang J, Zhou Z et al (2018) Silica-induced 19(3)
initiation of circular ZC3H4 RNA/ZC3H4 pathway 41. Krenkel O, Puengel T, Govaere O et al (2017)
promotes the pulmonary macrophage activation. Therapeutic inhibition of inflammatory monocyte
FASEB J 32:3264. https://doi.org/10.1096/fj.20170 recruitment reduces steatohepatitis and liver fibro-
1118R:fj201701118R sis. Hepatology 67:1270. https://doi.org/10.1002/
28. Abd Majid FA, Ab Rahman NS, Abd Wahid E hep.29544
et al (2018) Evaluation of herb-drug interaction of 42. Tarantino G, Caputi A (2011) JNKs, insulin resis-
SynacinnTM and individual biomarker through tance and inflammation: a possible link between
cytochrome 450 inhibition assay. Drug Metab Lett NAFLD and coronary artery disease. World
12. https://doi.org/10.2174/18723128126661803141 J Gastroenterol 17(33):3785–3794
12457 43. Tilg H, Moschen AR (2011) IL-1 cytokine family
29. Ahmad T, Valentovic MA, Rankin GO (2018) members and NAFLD: neglected in metabolic liver
Effects of cytochrome P450 single nucleotide poly- inflammation. J Hepatol 55(5):960–962
morphisms on methadone metabolism and pharma- 44. Vespasiani-Gentilucci U, Carotti S, Perrone G et al
codynamics. Biochem Pharmacol 153:196. https:// (2015) Hepatic toll-like receptor 4 expression is
doi.org/10.1016/j.bcp.2018.02.020 associated with portal inflammation and fibrosis in
30. Ahn SB, Jun DW, Jang K et al (2018) Duodenal patients with NAFLD. Liver Int 35(2):569–581
Niemann-Pick C1-like 1 expression was negatively 45. Neuschwander-Tetri BA (2014) NASH: the trib-
correlated with liver X receptor expression in non- ulations of conducting NASH trials. Nat Rev
alcoholic fatty liver disease. Korean J Intern Med. Gastroenterol Hepatol 11(5):274–276
https://doi.org/10.3904/kjim.2017.100 46. Diehl AM, Day C (2017) Cause, pathogenesis, and
31. Capel B, Swain A, Nicolis S et al (1993) Circular treatment of nonalcoholic steatohepatitis. N Engl
transcripts of the testis-determining gene Sry in J Med 377(21):2063–2072
adult mouse testis. Cell 73(5):1019–1030 47. Caussy C, Hsu C, Lo MT et al (2018) Novel link
32. Cocquerelle C, Daubersies P, Majerus MA et al between gut-microbiome derived metabolite and
(1992) Splicing with inverted order of exons occurs shared gene-effects with hepatic steatosis and fibro-
proximal to large introns. EMBO J 11(3):1095–1098 sis in NAFLD. Hepatology. https://doi.org/10.1002/
33. Chiu YJ, Chou SC, Chiu CS et al (2018) hep.29892
Hepatoprotective effect of the ethanol extract of 48. Arab JP, Martin-Mateos RM, Shah VH (2018) Gut-­
Polygonum orientale on carbon tetrachloride-­ liver axis, cirrhosis and portal hypertension: the
induced acute liver injury in mice. J Food Drug Anal chicken and the egg. Hepatol Int 12(Suppl 1):24–33
26(1):369–379 49. Buzzetti E, Pinzani M, Tsochatzis EA (2016)
34. George N, Chen M, Yuen N et al (2018) Interplay The multiple-hit pathogenesis of non-alco-
of gender, age and drug properties on reporting fre- holic fatty liver disease (NAFLD). Metabolism
quency of drug-induced liver injury. Regul Toxicol 65(8):1038–1048
Pharmacol 94:101–107 50. Younossi ZM, Loomba R, Anstee QM et al (2017)
35. Su Q, Kumar V, Sud N et al (2018) MicroRNAs Diagnostic modalities for non-alcoholic fatty liver
in the pathogenesis and treatment of progres- disease (NAFLD), non-alcoholic steatohepatitis
sive liver injury in NAFLD and liver fibrosis. Adv (NASH) and associated fibrosis. Hepatology. https://
Drug Deliv Rev 129:54. https://doi.org/10.1016/j. doi.org/10.1002/hep.29721
addr.2018.01.009 51. Younossi ZM, Loomba R, Rinella ME et al (2017)
36. Rahmoune H, Boutrid N, Bioud B (2018) Current and future therapeutic regimens for non-­
Nonalcoholic steatohepatitis. N Engl J Med alcoholic fatty liver disease (NAFLD) and non-­
378(8):780–781 alcoholic steatohepatitis (NASH). Hepatology.
37. Aspichueta P (2018) Lipid-rich environ- https://doi.org/10.1002/hep.29724
ment: a key role promoting carcinogenesis in 52. Beier JI, Banales JM (2018) Pyroptosis: an inflam-
obesity-­related non-alcoholic fatty liver disease. matory link between NAFLD and NASH with
Gut:gutjnl-2018-316047. https://doi.org/10.1136/ potential therapeutic implications. J Hepatol 68:643.
gutjnl-2018-316047 https://doi.org/10.1016/j.jhep.2018.01.017
38. Fukushima H, Yamashina S, Arakawa A et al (2018) 53. Ruiz-Ramirez A, Chavez-Salgado M, Peneda-Flores
The formation of p62-positive inclusion body is JA et al (2011) High-sucrose diet increases ROS gen-
associated with macrophage polarization in non-­ eration, FFA accumulation, UCP2 level, and proton
alcoholic fatty liver disease. Hepatol Res. https://doi. leak in liver mitochondria. Am J Physiol Endocrinol
org/10.1111/hepr.13070 Metab 301(6):E1198–E1207
39. Wesolowski SR, Kasmi KC, Jonscher KR et al (2017) 54. Kawano Y, Cohen DE (2013) Mechanisms of
Developmental origins of NAFLD: a womb with a hepatic triglyceride accumulation in non-alcoholic
clue. Nat Rev Gastroenterol Hepatol 14(2):81–96 fatty liver disease. J Gastroenterol 48(4):434–441
20 Circular RNA in Liver: Health and Diseases 255

55. Bril F, Barb D, Portillo-Sanchez P et al (2017) motes severe NAFLD and early fibrosis. Am J Physiol
Metabolic and histological implications of intrahe- Gastrointest Liver Physiol 301(5):G865–G876
patic triglyceride content in nonalcoholic fatty liver 68. Al-Serri A, Anstee QM, Valenti L et al (2012) The
disease. Hepatology 65(4):1132–1144 SOD2 C47T polymorphism influences NAFLD
56. Alves-Bezerra M, Cohen DE (2017) Triglyceride fibrosis severity: evidence from case-control and
metabolism in the liver. Compr Physiol 8(1):1–8 intra-familial allele association studies. J Hepatol
57. Ezquerro S, Mendez-Gimenez L, Becerril S et al 56(2):448–454
(2016) Acylated and desacyl ghrelin are associated 69. Guo XY, Chen JN (2017) circRNA_0046367 pre-
with hepatic lipogenesis, beta-oxidation and autoph- vents hepatoxicity of lipid peroxidation: an inhibi-
agy: role in NAFLD amelioration after sleeve gas- tory role against hepatic steatosis. Oxid Med Cell
trectomy in obese rats. Sci Rep 6:39942 Longev 2017:3960197
58. Zhang Y, Cheng X, Lu Z et al (2013) Upregulation 70. Guo J, Zhou Y, Cheng Y et al (2018) Metformin-­
of miR-15b in NAFLD models and in the serum of induced changes of the coding transcriptome and
patients with fatty liver disease. Diabetes Res Clin non-coding RNAs in the livers of non-alcoholic
Pract 99(3):327–334 fatty liver disease mice. Cell Physiol Biochem
59. An X, Yang Z, An Z (2017) MiR-149 compromises 45(4):1487–1505
the reactions of liver cells to fatty acid via its poly- 71. Guo XY, Sun F, Chen JN et al (2018) cir-
morphism and increases non-alcoholic fatty liver cRNA_0046366 inhibits hepatocellular steato-
disease (NAFLD) risk by targeting methylene tet- sis by normalization of PPAR signaling. World
rahydrofolate reductase (MTHFR). Med Sci Monit J Gastroenterol 24(3):323–337
23:2299–2307 72. Xu H, Guo S, Li W et al (2015) The circular RNA
60. Ghareghani P, Shanaki M, Ahmadi S et al (2018) Cdr1as, via miR-7 and its targets, regulates insulin
Aerobic endurance training improves nonalcoholic transcription and secretion in islet cells. Sci Rep
fatty liver disease (NAFLD) features via miR-33 5:12453
dependent autophagy induction in high fat diet fed 73. Argemi J, Kress TR, Chang HCY et al (2017)
mice. Obes Res Clin Pract 12(1S1):80–89 X-box binding protein 1 regulates unfolded pro-
61. Drygalski K, Berk K, Charytoniuk T et al (2017) tein, acute-phase, and DNA damage responses dur-
Does the enterolactone (ENL) affect fatty acid trans- ing regeneration of mouse liver. Gastroenterology
porters and lipid metabolism in liver? Nutr Metab 152(5):1203–1216 e1215
(Lond) 14:69 74. Schmidt-Arras D, Rose-John S (2016) IL-6 path-
62. Cheng J, Liu C, Hu K et al (2017) Ablation of sys- way in the liver: from physiopathology to therapy.
temic SIRT1 activity promotes nonalcoholic fatty J Hepatol 64(6):1403–1415
liver disease by affecting liver-mesenteric adipose 75. Paschos KA, Bird NC (2010) Liver regeneration and
tissue fatty acid mobilization. Biochim Biophys Acta its impact on post-hepatectomy metastatic tumour
1863(11):2783–2790 recurrence. Anticancer Res 30(6):2161–2170
63. Salgado AL, Carvalho L, Oliveira AC et al (2010) 76. Tao Y, Wang M, Chen E et al (2017) Liver regen-
Insulin resistance index (HOMA-IR) in the differ- eration: analysis of the main relevant signaling mol-
entiation of patients with non-alcoholic fatty liver ecules. Mediat Inflamm 2017:4256352
disease and healthy individuals. Arq Gastroenterol 77. Li L, Guo J, Chen Y et al (2017) Comprehensive
47(2):165–169 CircRNA expression profile and selection of key
64. Klisic A, Isakovic A, Kocic G et al (2017) CircRNAs during priming phase of rat liver regen-
Relationship between oxidative stress, inflam- eration. BMC Genomics 18(1):80
mation and dyslipidemia with fatty liver index 78. DeLeve LD (2015) Liver sinusoidal endothelial cells
in patients with type 2 diabetes mellitus. Exp in hepatic fibrosis. Hepatology 61(5):1740–1746
Clin Endocrinol Diabetes 126:371. https://doi. 79. Pradere JP, Kluwe J, De Minicis S et al (2013)
org/10.1055/s-0043-118667 Hepatic macrophages but not dendritic cells con-
65. Firidin G (2018) Oxidative stress parameters, induc- tribute to liver fibrosis by promoting the survival of
tion of lipid peroxidation, and ATPase activity in the activated hepatic stellate cells in mice. Hepatology
liver and kidney of Oreochromis niloticus exposed 58(4):1461–1473
to lead and mixtures of lead and zinc. Bull Environ 80. Cai C, Zhu X, Li P et al (2017) NLRP3 deletion
Contam Toxicol 100(4):477–484 inhibits the non-alcoholic steatohepatitis develop-
66. Francque S, De Maeght S, Adler M et al (2011) ment and inflammation in Kupffer cells induced by
High prevalence of advanced fibrosis in associa- palmitic acid. Inflammation 40(6):1875–1883
tion with the metabolic syndrome in a Belgian pro- 81. Li Z, Ji L, Su S et al (2018) Leptin up-regulates
spective cohort of NAFLD patients with elevated microRNA-27a/b-3p level in hepatic stellate cells.
ALT. Results of the Belgian NAFLD registry. Acta Exp Cell Res 366:63. https://doi.org/10.1016/j.
Gastroenterol Belg 74(1):9–16 yexcr.2018.03.015
67. Tan TC, Crawford DH, Jaskowski LA et al (2011) 82. Qu C, Zheng D, Li S et al (2018) Tyrosine kinase
Altered lipid metabolism in Hfe-knockout mice pro- SYK is a potential therapeutic target for liver
256 M. Song et al.

fibrosis. Hepatology. https://doi.org/10.1002/ tumor cell growth by targeting miR-630. Aging

hep.29881 (Albany NY) 9(6):1585–1594
83. Bu FT, Chen Y, Yu HX et al (2018) SENP2 alleviates 97. Geng HH, Li R, Su YM et al (2016) The circular
CCl4-induced liver fibrosis by promoting activated RNA Cdr1as promotes myocardial infarction by
hepatic stellate cell apoptosis and reversion. Toxicol mediating the regulation of miR-7a on its target
Lett 289:86–98 genes expression. PLoS One 11(3):e0151753
84. Coppola N, Perna A, Lucariello A et al (2018) 98. Barrett SP, Parker KR, Horn C et al (2017) ciRS-7
Effects of treatment with Maraviroc a CCR5 inhibi- exonic sequence is embedded in a long non-coding
tor on a human hepatic stellate cell line. J Cell RNA locus. PLoS Genet 13(12):e1007114
Physiol 233:6224. https://doi.org/10.1002/jcp.26485 99. Zhao ZJ, Shen J (2017) Circular RNA participates
85. Kiagiadaki F, Kampa M, Voumvouraki A et al (2018) in the carcinogenesis and the malignant behavior of
Activin-a causes hepatic stellate cell activation via cancer. RNA Biol 14(5):514–521
the induction of TNFalpha and TGFbeta in Kupffer 100. Yu L, Gong X, Sun L et al (2016) The circular RNA
cells. Biochim Biophys Acta 1864(3):891–899 Cdr1as act as an oncogene in hepatocellular carci-
86. Lan T, Li C, Yang G et al (2018) Sphingosine kinase noma through targeting miR-7 expression. Hepatol
1 promotes liver fibrosis by preventing miR-19b-3p-­ Res Treat 11(7):e0158347
mediated inhibition of CCR2. Hepatology. https:// 101. Xu L, Zhang M, Zheng X et al (2017) The circular
doi.org/10.1002/hep.29885 RNA ciRS-7 (Cdr1as) acts as a risk factor of hepatic
87. Zhang K, Han X, Zhang Z et al (2017) The liver-­ microvascular invasion in hepatocellular carcinoma.
enriched lnc-LFAR1 promotes liver fibrosis by acti- J Cancer Res Clin Oncol 143(1):17–27
vating TGFbeta and Notch pathways. Nat Commun 102. Han D, Li J, Wang H et al (2017) Circular RNA
8(1):144 circMTO1 acts as the sponge of microRNA-9 to
88. Chen Y, Yuan B, Wu Z et al (2017) Microarray pro- suppress hepatocellular carcinoma progression.
filing of circular RNAs and the potential regulatory Hepatology 66(4):1151–1164
role of hsa_circ_0071410 in the activated human 103. Yao Z, Luo J, Hu K et al (2017) ZKSCAN1 gene
hepatic stellate cell induced by irradiation. Gene and its related circular RNA (circZKSCAN1) both
629:35–42 inhibit hepatocellular carcinoma cell growth, migra-
89. Tang CM, Zhang M, Huang L et al (2017) tion, and invasion but through different signaling
CircRNA_000203 enhances the expression of pathways. Mol Oncol 11(4):422–437
fibrosis-­associated genes by derepressing targets of 104. Zhang K, Che S, Su Z et al (2018) CD90 promotes
miR-26b-5p, Col1a2 and CTGF, in cardiac fibro- cell migration, viability and sphere forming ability
blasts. Sci Rep 7:40342 of hepatocellular carcinoma cells. Int J Mol Med
90. Zhou B, Yu JW (2017) A novel identified circular 41(2):946–954
RNA, circRNA_010567, promotes myocardial fibro- 105. Wang BG, Li JS, Liu YF et al (2017)
sis via suppressing miR-141 by targeting TGF-beta1. MicroRNA-200b suppresses the invasion and
Biochem Biophys Res Commun 487(4):769–775 migration of hepatocellular carcinoma by downreg-
91. Forner A, Reig M, Bruix J (2018) Hepatocellular ulating RhoA and circRNA_000839. Tumour Biol
carcinoma. Lancet (18):30010–30012. https://doi. 39(7):1010428317719577
org/10.1016/S0140-6736 106. Zhu Q, Lu G, Luo Z et al (2018) CircRNA
92. Abitbol S, Dahmani R, Coulouarn C et al (2018) circ_0067934 promotes tumor growth and metasta-
AXIN deficiency in human and mouse hepatocytes sis in hepatocellular carcinoma through regulation
induces hepatocellular carcinoma in the absence of of miR-1324/FZD5/Wnt/beta-catenin axis. Biochem
beta-catenin activation. J Hepatol 68:1203. https:// Biophys Res Commun 497(2):626–632
doi.org/10.1016/j.jhep.2017.12.018 107. Dong LQ, Shi Y, Ma LJ et al (2018) Spatial and tem-
93. Cai Z, Qian ZY, Jiang H et al (2018) hPCL3s pro- poral clonal evolution of intrahepatic cholangiocar-
motes hepatocellular carcinoma metastasis by acti- cinoma. J Hepatol 69:89. https://doi.org/10.1016/j.
vating beta-catenin signaling. Cancer Res 78:2536. jhep.2018.02.029
https://doi.org/10.1158/0008-5472.CAN-17-0028 108. Lee DD, Croome KP, Musto KR et al (2018) Caveat
94. Buschauer S, Koch A, Wiggermann P et al (2018) emptor, liver transplantation for intrahepatic chol-
Hepatocellular carcinoma cells surviving doxorubi- angiocarcinoma – a diagnostic conundrum for 44
cin treatment exhibit increased migratory potential patients at a single center. Liver Transpl 24:634.
and resistance to doxorubicin re-treatment in vitro. https://doi.org/10.1002/lt.25052
Oncol Lett 15(4):4635–4640 109. Merdrignac A, Angenard G, Allain C et al (2018) A
95. Yang F, Liu DY, Guo JT et al (2017) Circular novel transforming growth factor beta-induced long
RNA circ-LDLRAD3 as a biomarker in diagno- noncoding RNA promotes an inflammatory micro-
sis of pancreatic cancer. World J Gastroenterol environment in human intrahepatic cholangiocarci-
23(47):8345–8354 noma. Hepatol Commun 2(3):254–269
96. Zhang Y, Liu H, Li W et al (2017) CircRNA_100269 110. Xie L, Onysko J, Morrison H (2018) Childhood
is downregulated in gastric cancer and suppresses cancer incidence in Canada: demographic and
20 Circular RNA in Liver: Health and Diseases 257

g­ eographic variation of temporal trends (1992-2010). 122. Li TZ, Kim SM, Hur W et al (2017) Elk-3 contrib-
Health Promot Chronic Dis Prev Can 38(3):79–115 utes to the progression of liver fibrosis by regulat-
111. Zhong S, Zhao Y, Fan C (2018) Hepatoblastoma ing the epithelial-mesenchymal transition. Gut Liver
with pure fetal epithelial differentiation in a 10-year-­ 11(1):102–111
old boy: a rare case report and review of the litera- 123. Caiment F, Gaj S, Claessen S et al (2015) High-­
ture. Medicine (Baltimore) 97(2):e9647 throughput data integration of RNA-miRNA-­
112. Liu BH, Zhang BB, Liu XQ et al (2018) circRNA reveals novel insights into mechanisms of
Expression profiling identifies circular RNA sig- benzo[a]pyrene-induced carcinogenicity. Nucleic
nature in hepatoblastoma. Cell Physiol Biochem Acids Res 43(5):2525–2534
45(2):706–719 124. Jin X, Feng CY, Xiang Z et al (2016) CircRNA
113. Li P, Chen S, Chen H et al (2015) Using circular expression pattern and circRNA-miRNA-mRNA
RNA as a novel type of biomarker in the screening network in the pathogenesis of nonalcoholic steato-
of gastric cancer. Clin Chim Acta 444:132–136 hepatitis. Oncotarget 7(41):66455–66467
114. Fu L, Yao T, Chen Q et al (2017) Screening dif- 125. Zhao Y, Alexandrov PN, Jaber V et al (2016)
ferential circular RNA expression profiles reveals Deficiency in the ubiquitin conjugating enzyme
hsa_circ_0004018 is associated with hepatocellular UBE2A in Alzheimer’s disease (AD) is linked to
carcinoma. Oncotarget 8(35):58405–58416 deficits in a natural circular miRNA-7 sponge (cir-
115. Xu Y, Yao Y, Zhong X et al (2018) Downregulated cRNA; ciRS-7). Genes (Basel) 7(12)
circular RNA hsa_circ_0001649 regulates pro- 126. Andres-Leon E, Nunez-Torres R, Rojas AM (2018)
liferation, migration and invasion in cholangio- Corrigendum: miARma-Seq: a comprehensive tool
carcinoma cells. Biochem Biophys Res Commun for miRNA, mRNA and circRNA analysis. Sci Rep
496(2):455–461 8:46928
116. Blaya D, Aguilar-Bravo B, Hao F et al (2018) 127. Hansen TB (2018) Improved circRNA identification
Expression of microRNA-155 in inflammatory by combining prediction algorithms. Front Cell Dev
cells modulates liver injury. Hepatology. https://doi. Biol 6:20
org/10.1002/hep.29833 128. Lei W, Feng T, Fang X et al (2018) Signature of cir-
117. Bourgine M, Crabe S, Lobaina Y et al (2018) Nasal cular RNAs in human induced pluripotent stem cells
route favors the induction of CD4(+) T cell responses and derived cardiomyocytes. Stem Cell Res Ther
in the liver of HBV-carrier mice immunized with a 9(1):56
recombinant hepatitis B surface- and core-based 129. Liang J, Wu X, Sun S et al (2018) Circular RNA
therapeutic vaccine. Antivir Res 153:23–32 expression profile analysis of severe acne by RNA-­
118. Neumann K, Rudolph C, Neumann C et al (2015) seq and bioinformatics. J Eur Acad Dermatol
Liver sinusoidal endothelial cells induce immuno- Venereol. https://doi.org/10.1111/jdv.14948
suppressive IL-10-producing Th1 cells via the Notch 130. Jakobi T, Dieterich C (2018) Deep computational
pathway. Eur J Immunol 45(7):2008–2016 circular RNA analytics from RNA-seq data. Methods
119. Zhuang S, Hua X, He K et al (2018) Inhibition of Mol Biol 1724:9–25
GSK-3beta induces AP-1-mediated osteopontin 131. Quan G, Li J (2018) Circular RNAs: biogenesis,
expression to promote cholestatic liver fibrosis. expression and their potential roles in reproduction.
FASEB J:fj.201701137R. https://doi.org/10.1096/fj. J Ovarian Res 11(1):9
201701137R:fj201701137R 132. Gao Y, Wang J, Zheng Y et al (2016) Comprehensive
120. Kostallari E, Hirsova P, Prasnicka A et al (2018) identification of internal structure and alternative
Hepatic stellate cell-derived PDGFRalpha-enriched splicing events in circular RNAs. Nat Commun
extracellular vesicles promote liver fibrosis in mice 7:12060
through SHP2. Hepatology. https://doi.org/10.1002/ 133. Zhao J, Li L, Wang Q et al (2017) CircRNA expres-
hep.29803 sion profile in early-stage lung adenocarcinoma
121. O’Leary JG, Demetris AJ, Philippe A et al (2017) patients. Cell Physiol Biochem 44(6):2138–2146
Non-HLA antibodies impact on C4d staining, stel- 134. Xia S, Feng J, Chen K et al (2018) CSCD: a database
late cell activation and fibrosis in liver allografts. for cancer-specific circular RNAs. Nucleic Acids
Transplantation 101(10):2399–2409 Res 46(D1):D925–D929
 Circular RNAs in Organ Fibrosis
21
Jianhua Yao, Qiying Dai, Zhuyuan Liu, Lei Zhou,
and Jiahong Xu

Abstract studies about how circRNA dysregulation

Fibrosis refers to a process involving the accu- contributes to organ fibrosis. Finally, the
mulation of extracellular matrix components. advantages and potential challenges of
It could happen in chronic organ injury or dur- circRNA-­based therapeutics for the treatment
ing the recovery of acute organ injury. The of fibroproliferative diseases will be
severity of fibrosis interferes with the function discussed.
of the organ involved. Numerous studies have
been carried out to explore the mechanism of Keywords
fibrosis, including parenchyma injury, fibrillar circRNA · Fibrosis
ECM accumulation, fibroblast activation,
microvasculature rarefaction, and a mononu-
clear infiltrate. Unfortunately, its underlying
mechanism is at largely unknown. The study- 1 Introduction
ing of noncoding RNAs has provided novel
insight for circRNA-miRNA-mRNA in learn- Cell damage and tissue repair is a complex bio-
ing disease progress. Emerging evidence has logical process [1]. A successful repairment
shown that circRNA is related to fibrosis involves at least three distinct phases: (1) an elimi-
activity and could potentially be a monitoring nation phase, in which damaged or dead cells are
factor for fibrosis or, more excitingly, could be eliminated; (2) a regenerative phase, in which the
a target for treatment. In this chapter, we will eliminated cells are replaced by cells of the same
first present the basic mechanism of organ or different origin; and (3) a fibrogenic phase, in
fibrosis. Then we will focus on the recent which fibrosis tissue is generated to partially
restore the function of the organ [2–7]. The fibro-
Author contributed equally with all other contributors. genic phase will continue until d­ amaged tissue or
Jianhua Yao and Qiying Dai
J. Yao
Z. Liu · L. Zhou (*)
Department of Cardiology, Shanghai Tenth People’s
Department of Cardiology, First Affiliated Hospital of
Hospital, Tongji University School of Medicine,
Nanjing Medical University, Nanjing, China
Shanghai, China
J. Xu (*)
Q. Dai
Department of Cardiology, Shanghai Tongji Hospital,
MetroWest Medical Center, Framingham, MA, USA
Tongji University School of Medicine,
Department of Cardiology, First Affiliated Hospital of Shanghai, China
Nanjing Medical University, Nanjing, China e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 259

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_21
260 J. Yao et al.

lost cells are regenerated. Ideally, temporary fibro- (circular RNA sponge for miR-7) in neuronal tis-
sis tissue will be replaced by fully functional cells. sues [34, 36] is significantly reduced in sporadic
In some cases, like chronic tissue injury or exten- Alzheimer disease [37]. Moreover, a lot of other
sive injury, the generation of fibrotic tissue over- circRNAs were found to be significantly altered
grows that of the original tissue. And then, the during cancer-associated cell proliferation [38].
function of the injured site would be impaired The organ specific expression makes cir-
because of the replacement by the nonfunctional cRNAs a candidate for disease biomarkers [22,
fibrosis tissue [8, 9]. 39, 40]. Unlike other noncoding RNAs, cir-
It is a common phenomenon in both chronic cRNAs are highly expressed and concentrated
and acute diseases, and it can occur in nearly all due to their resistance to RNA exonucleases or
organs including the heart, liver, kidney, lung, RNase R [41, 42]. What’s more, circRNA appears
and so on [10–14]. Chronic stimuli such as per- to be highly conserved among mammals [43]. In
sistent infections and hormonal changes and mice, 4% of orthologous genes can generate cir-
acute stimuli such as the sudden change of blood cRNAs [31], and approximately 5–30% of these
dynamics can induce organ fibrosis [15–17]. The circRNAs are completely preserved in human
progressive organ fibrosis can eventually cause and mice [36, 41, 44]. About 5–10% of human
organ dysfunction or even failure [18–20]. A brain circRNAs are also expressed in the porcine
great amount of efforts have been put into explor- brain [43, 45]. Fibrosis, an irreversible damage to
ing the nature of fibrosis with minimal achieve- organs, has been the focus of studies for years.
ment, and no effective management has been The discovery of circRNA has created a new path
available either [21]. in understanding this process in a different level.
The prevalence of non-polyadenylated tran- Here we will start with the basic mechanism of
scriptome data has prospered the study in the organ fibrosis and then take a deep look into how
field of circRNAs, which are found to modulate circRNAs affect this process.
disease progress, including fibrosis. CircRNAs
are single-stranded RNAs which are detectable in
humans, mice, rats, fungi, and other organisms 2 Mechanism of Organ Fibrosis
[22–35]. There are over ten thousands of differ-
ent circRNAs in metazoans, from insects [22] to Fibrosis is the formation of excess fibrous tissue
mammals [23, 24]. Moreover, over 10% of during tissue repair. It can lead to malfunction of
expressed genes in examined cells and tissues the organ after replacing the original cells. Organs
can produce circRNAs [25]. However, only a few like the liver, kidneys, heart, and lungs which
number of circRNAs have been found. contain a lot of parenchymal cells are most vul-
Considering their low expression, it was hypoth- nerable to fibrosis [55–57]. Other organs might
esized that they were produced by aberrant splic- be affected as well. For example, fibrosis involv-
ing [26–30].While approximately 50 genes have ing the skin and joints will impair movement and
circRNAs that are expressed more abundantly decrease life quality. Bone marrow fibrosis could
than their linear isoforms [31], few numbers of lead to cytopenia [46–50]. In addition to normal
circRNAs have been convinced to be highly organs, tumors, especially solid ones, could also
expressed in a cell-type-specific or organ-specific be affected by fibrosis [51–53].
manner [32, 33]. The high ubiquitousness of the
co-expression of circRNA and its linear isoforms
suggests that circRNAs are not simply by-­ 2.1  ommon Features of Tissue
C
products. Several circRNAs have been found to Fibrosis (Fig. 21.1)
participate in disease progress. For example, cir-
cRNA from sex-determining region Y (circSRY) The fibrosis in all organs shares similar histomor-
has a key role in testes development in mouse phology, both macroscopically and microscopi-
[34]. CircRNAs from ANRIL positively affect cally [54]. The fibrosis organs are stiff, pale, and
the development of atherosclerosis [35]. CiRS-7 uneven. The accumulation of excessive extracel-
21 Circular RNAs in Organ Fibrosis 261

Fig. 21.1 Common features of tissue fibrosis consists of functional parenchyma, connective tissue, ves-
The fibrosis in all organs shared similar histomorphology, sels, and fibroblasts. In the heart, the functional paren-
and macroscopically and microscopically fibrosis organs chyma is cardiomyocytes. In the liver, the functional
share commonalities. The fibrosis organs are stiff, pale, parenchyma is hepatocytes, and in the lung, the functional
and uneven. The histopathological analysis reveals that parenchymas are endothelial cells and smooth muscle
organ fibrosis is associated with the parenchyma injury, cells. The fibrotic organs have more connective tissue than
fibrillar ECM accumulation, fibroblast activation, micro- normal organs. The connective tissue mainly contains
vasculature rarefaction, and a mononuclear infiltrate. The fibroblasts, myofibroblasts, collagens, and mononuclear
schematic diagram showed the common mechanisms of infiltrate
organ fibrosis in the heart, liver, and lung. Normal organ

lular matrix (ECM) causes the stiffness [55]. tively high regenerative capacity such as hepato-
Destruction of the vasculature makes them look cytes [60–62]. The balance of regeneration breaks
pale. Contraction of the fibroblasts contributes to when the injury stimulation exceeds the rate of
the uneven surface. The histopathological analy- repair, causing a suppression on the regeneration
sis reveals that organ fibrosis is associated with capacity [53, 63–65]. This in turn facilitates the
the parenchyma injury, fibrillar ECM accumula- formation of the fibrous tissue. Meanwhile, the
tion, fibroblast activation, microvasculature rar- injured parenchymal cells would secrete chemo-
efaction, and a mononuclear infiltrate [56–59]. kines like growth factors and other profibrotic
metabolites to promote fibrogenesis. The typical
pathway is TGF-β signaling pathway. TGF-β was
2.2 Parenchymal Injury initially identified as a hallmark of malignant
transformation in embryonic kidney fibroblasts
Each parenchymal organ has its specific paren- [66]. It turns out that TGF-β is overexpressed in
chymal cells. Different parenchymal cells have fibrosis tissues as well. It induces collagen pro-
different potentials to regenerate. Some paren- duction both in vitro and in vivo. Fibrotic activity
chymal cells have low regeneration capacity, was noticed to be suppressed after TGF-β neu-
such as cardiomyocytes, while some have rela- tralization [67–70].
262 J. Yao et al.

2.3 Fibrillar ECM Accumulation of the biggest challenges that make it difficult to
study fibroblasts is their heterogeneity.
Fibroblasts are the major source of extracellular Myofibroblasts are the principal mediators of
matrix in fibrosis [71, 72]. The main goal of anti-­ fibrogenesis with unique α-SMA-expressing fea-
fibrosis therapy is to inhibit fibroblasts, thus sup- tures. Using FSP1 as a marker, the group of
pressing the accumulation of extracellular matrix. FSP1+ fibroblasts are found to contribute for kid-
Little is known in this field because of the com- ney fibrosis [90]. Due to the large number of dif-
plexity and multiple factors it involves [73]. ferent types of myofibroblasts, it is a big challenge
to carry on the study on myoblasts. Moreover, the
dynamic changes of these cells during aging
2.4 Fibroblast Activation make it more difficult to study [91–94]. Searching
simple markers remains the most significant task.
Fibroblast activation is the first step in the pro- On the other hand, the origination of activated
cess of fibrosis. Fortunately, fibroblasts are easy fibroblasts remains controversial. It was previ-
to culture, making it convenient for us to do ously acknowledged that these cells were gener-
in vitro experiment [74, 75]. Fibroblasts are usu- ated from the in situ fibroblasts until bone
ally found in connective tissue. They have promi- marrow-derived fibroblasts were discovered [95].
nent cytoskeleton and endoplasmic reticulum and Later on, it was found that epithelial cells, peri-
specific leaflet-shaped cytoplasm [76, 77]. Based cytes, and even vascular smooth muscle cells par-
on their location and basic structure, we can iden- ticipate in fibrosis [96–99]. Moreover, by using
tify them through the light microscope. Activated fate-mapping strategies and marker analysis, it
fibroblasts and myofibroblasts, which have great was found that mechanisms of fibrosis might be
potential for biosynthesis and proliferation, con- different in different organs and diseases [53, 65,
sist of the major source of fibrosis. They are char- 100–111].
acterized by pronounced rough endoplasmic
reticulum, stress fibers, and large nucleolus.
Cell biomarkers are frequently used in 2.5 Microvasculature Rarefaction
fibroblast-­related researches. Both fibroblasts
and activated fibroblasts can be labelled with Microvasculature rarefaction causes the pale look
α-smooth muscle actin (α-SMA) [78–80]. Even in the fibrotic tissues by decreasing perfusion
so, specific markers are still needed because of chronically. The microvascular system plays an
the existence of subpopulations. The types of important role in transporting oxygen. The
fibroblasts vary among different organs. In addi- ­ declination of these small vessels creates a
tion, there are heterogeneous phenotypes within chronic hypoxic environment, which promotes
the same organ. For example, in the kidney, the fibrosis by activating HIF-α signaling pathway
lipid-laden fibroblasts in medulla are different [112–114]. In turn, fibrosis can attenuate the
from the ones in the cortex [81]. Functions of capacity of vessel regeneration through decreas-
fibroblasts vary according to their origins. ing the expression of proangiogenic molecules
Fibroblasts in the kidneys could transform into and increasing the expression of antiangiogenic
erythropoietin-producing cells [82, 83]. A part of molecules. Pro-angiogenic therapy, for example,
the cardiac conduction system is atrial fibroblasts VEGF administration, has been reported to have
[84–87]. Tracking the expression of Hox gene, cardiac protection through decreasing fibrosis in
the original position of dermal fibroblasts can be experimental models [115–118].
found [88]. In order to study the functions of Interestingly, cell loss by either apoptosis or
these subtypes, additional markers have been necrosis is related to microvascular rarefaction
developed, like α-SMA, vimentin, fibroblast-­ [119]. One of the well-known explanations is
specific protein 1 (FSP1), desmin, etc. [89]. One endothelial-mesenchymal transition (EndMT)
21 Circular RNAs in Organ Fibrosis 263

[109, 120–122]. It was discovered in cardiac 128]. Of note, anti-inflammatory regimens are
development in which mesenchymal cells are responsible for anti-fibrogenesis in clinical prac-
found to be derived from the endocardium tice, suggesting that the anti-fibrogenesis could
through a process called EndMT [123–125]. The be an independent process.
mesenchymal cells are group of cells which form
the atrioventricular cushion, the primordia of
valve, and the septa of the heart. During EndMT, 3 CircRNA in Organ Fibrosis
endothelial cells lose their own characteristic
markers and acquire a mesenchymal phenotype. Fibrosis has different characteristics and clini-
After the transformation, they will initiate the cal effects in different organ systems
fibrosis process by producing mesenchymal cell (Table 21.1). Kidney fibrosis decreases erythro-
products. Transforming growth factor-β (TGF-β) poietin production and causes anemia. Liver
is one of the well-studied EndMT factors [122]. fibrosis impairs the metabolism of lots of medi-
In the heart and kidney, it was found that inhibi- cation. Electromechanical coupling is debili-
tion of EndMT could attenuate fibrosis [105]. tated when cardiac fibrosis occurred. It will lead
This provides a potential therapeutic target for to all kinds of arrhythmias and also affect the
fibrosis. heart pump function. Causes of fibrosis vary in
Fibrosis can influence inflammatory process these organs too. Infections have been reported
by affecting mononuclear infiltrate. It is common the most common cause of fibrosis in the liver.
to see the coexistence of inflammation and fibro- Also, fibrosis in organs like the liver and lung
sis, especially in chronic diseases such as viral can potentially lead to cancer.
hepatitis, schistosomiasis, or bacterial Emerging evidence has suggested that cir-
pyelonephritis-­induced fibrosis. Whether inflam- cRNA participates in this process. Here we will
mation is a constituent of fibrogenesis or they are discuss the role of circRNA in fibrosis in differ-
independent of each other is controversial [126– ent organ systems.

Table 21.1 Specific aspects of organ fibrosis

Kidney Liver Heart Lung Skin
Etiology Diabetes mellitus viral hepatitis Hypertension Idiopathic Physical injury
Hypertension alcohol-induced Coronary artery pulmonary Idiopathic
Glomerulonephritis steatohepatitis disease fibrosis scleroderma
Nonalcoholic Aortic stenosis Occupational
steatohepatitis disease
Sarcoidosis
Diagnosis Kidney Liver function Cardiac function X-ray Manifestations
function(BUN, GFR, MRI (stiffness,
serum creatinine) Liver biopsy elasticity)
Kidney biopsy
Specific Kidney fibrosis Stellate cells Cardiac fibrosis can Survival mean Skin fibrosis is
features causes anemia participate in result in atrial time of IPF associated with
because of cessation liver fibrosis fibrillation because patients is <6 skin color and
of erythropoietin Liver fibrosis has fibroblasts are years keloid formation
production detrimental required for
microvascular electromechanical
shunts which coupling
cause porto-­
venous
hypertension
264 J. Yao et al.

3.1 Liver Fibrosis 3.2 Cardiac Fibrosis

Liver fibrosis has devastating effects on metab- Unlike hepatocytes, cardiomyocytes have less
olism and has been a concern worldwide. The capacity to regenerate; fibrosis can occur after a
major causes of liver fibrosis are alcohol, non- sudden loss of cardiomyocytes. Ischemia is the
alcoholic steatohepatitis (NASH), and viral major trigger of the process. After an episode of
hepatitis. The major cells involved in liver acute myocardial infarction, the dead portion will
fibrosis are stellate cells, which are derived be ultimately replaced by scar tissue.
from neurocrest. They can be transdifferenti- Apart from the acute causes, chronic causes
ated into myofibroblasts, which, in turn, can like aging, hypertension, valvular disease, and
activate common fibroblasts and peripheral cardiotoxic medication also contribute to cardiac
fibroblasts. Picrosirius red staining and trans- fibrosis. They can induce the perivascular and
mission electron microscopy are often used to interstitial collagen production and accumula-
identify cross-linked and stable collagen fibers tion. Aging is an unpreventable cause for fibrosis.
to assess liver fibrosis. During aging, the cardiac functional cell loss and
In cancer patients, radiation-induced liver fibrotic tissue replacement increase. These
injury accounts for most of liver fibrosis [143– changes could impair cardiac function. Stimuli
146]. It is recognized as radiation-induced liver which change the hemodynamic balance in the
fibrosis (RILF). RILF is a complex process. heart could cause more extensive damage. For
Radiation can facilitate the transformation of qui- instance, uncontrolled hypertension and valvular
escent hepatic stellate cells (HSC) into diseases like aortic stenosis generate stimuli on
myofibroblast-­like cells (activated HSC) which cardiomyocytes by increasing intraventricular
are mediated by TGF-β signaling pathway [129, pressure. Due to the limited regenerative ability
130]. In the hepatic microenvironment, quiescent of cardiomyocytes, fibrous tissue grows to com-
hepatic stellate cells can also be activated via pro- pensate for the elevated workload. The major
longed stimulation from other inflammatory fac- consequence is increased stiffness of the ventric-
tors, which result in excessive extracellular ular wall. Heart failure would eventually occur.
matrix accumulation. On the whole, the activated Chemotherapy with cardiotoxicity could cause
HSCs contribute to the progression of liver fibro- heart injury as well.
sis [131]. It is known that metabolic disease like diabe-
CircRNA plays an important role in various tes could induce heart failure independent of any
biological processes through acting as microRNA coronary artery disease or hypertension. Fibrosis
(miRNA) sponges. Recently, it has been reported was observed in both metabolic disease patients
that the expression of circRNA was correlated and experimental models [133]. In diabetes-­
with RILF. By using circRNA microarray, cir- related heart fibrosis, hyperglycemia, increased
cRNA expression was profiled in HSCs treated fatty acid metabolism, insulin resistance, and
with or without irradiation. It was found that the microcirculatory changes all contribute to the
expressions of 179 circRNAs were significantly cardiac remodeling by fibrosis [134–138].
increased, while the expressions of 630 circRNAs Myocardial fibrosis increases myocardial stiff-
were decreased in irradiated HSCs compared ness, disturbs relaxation, and causes cardiac dys-
with normal HSCs. Bioinformatic analyses indi- function. It also damages conduction system and
cated these expression changes might be the results in atrial and ventricular arrhythmias [139].
response to irradiation and the product of fibrotic There are two main types of fibrotic lesions in
process. In another similar study, researcher the heart—perivascular fibrosis and interstitial
reported that inhibiting the expression of hsa_ fibrosis. The latter is also known as “subendocar-
circ_0071410 increased the expression of miR-­ dial fibrosis” or “endomyocardial fibrosis.” It is
9-­5p and attenuated irradiation-induced HSC characterized by subendocardial deposition of
activation [132]. elastic fibers and is usually found in newborns
21 Circular RNAs in Organ Fibrosis 265

with congenital heart defects [152–155]. The emphysema, pneumonia, and so on [108, 143–
unique feature of cardiac fibroblasts is that they 147]. These are usually considered as secondary
are able to function as mechanoelectrical trans- pulmonary fibrosis. Primary pulmonary fibrosis,
ducers [84–87, 140]. Thus, as the most abundant also known as idiopathic pulmonary fibrosis,
cell type in the cardiac, the cardiac fibroblasts could be the sole presentation or a part of other
play important physiological roles. systemic diseases [147, 148]. Inflammation is the
CircRNAs keep the balance of fibrosis by main cause of pulmonary fibrosis, and anti-­
interacting with miRNAs. A good number of inflammatory treatment is the standard treatment
miRNAs have been recognized to enhance or in most circumstances. However, in primary pul-
suppress the activation of fibroblasts. CircRNAs monary fibrosis, anti-inflammatory therapy does
act as “miRNA sponges” to help fibrosis not lead to favorable outcome [127].
modulation. Silicosis is a chronic fibrotic pulmonary dis-
In diabetes mouse model, circRNA_000203 ease caused by the inhalation of silica. It is one of
was found to be increased in the myocardium as the most serious occupational diseases in the
well as in Ang II-induced cardiac fibroblasts world. Silicosis is characterized by chronic pro-
in vivo. Upregulated circRNA_000203 enhanced gressive pulmonary fibrosis. Diagnosis is usually
the expression of fibrosis markers (Col1a2, made when patients enter the late stage. Early
Col3a1, and α-SMA) in mouse cardiac fibro- diagnosis remains challenging [127, 149, 150].
blasts. In addition, circRNA_000203 further Unfortunately, there is no specific therapy for this
facilitated fibrosis by blocking the effect of miR-­ disease. Avoidance of the substance, smoking
26b-­5p which is known to downregulate Col1a2, cessation, and symptomatic treatment with bron-
CTGF, Col3a1, and α-SMA expression in cardiac chodilators are the major management for mild
fibroblasts. Consistently, the anti-fibrosis effect disease. For more advanced silicosis, lung trans-
of miR-26b-5p could be eliminated by the over- plantation is the only way to go [151, 152]. Lung
expression of circRNA_000203 in cardiac fibro- fibrosis complications include but not limited to
blasts. In this case, circRNA and its downstream lung infections, rheumatoid disorders, airway
miRNA formed a circRNA_000203/miR-26b-5p obstruction, and even cancer [153]. Fibrosis in
axis to modulate cardiac fibrosis [141]. silicosis is initiated with the ingestion of silica
CircRNA_010567 is another circRNA that is dust by macrophages. Stimulated by silica dust,
found to regulate cardiac fibrosis. MiR-26b-5p these macrophages will produce cytokines to
and miR-141 are identified to be the downstream attract fibroblasts. Collagens recruited by macro-
miRNAs of circRNA_010567. miR-141 is also phages participate in “eating” silica dust and pro-
an anti-fibrotic factor which directly targets TGF-­ duce inflammatory reaction. The inflammatory
b1. Subsequent study showed that the block of reaction further attracts more collagen and mac-
circRNA_010567 upregulated miR-141 and thus rophages. Fibrotic nodules are produced in this
decreased TGF-b1 and other fibrosis-associated vicious cycle. On the other hand, EMT also pro-
protein in CFs, including Col1, Col3, and motes the accumulation of fibroblasts.
a-SMA. Taken together, circRNA_010567/miR-­ Researches on macrophages found that circ-
141/TGF-b1 axis prohibited cardiac fibrosis and ZC3H4 and its downstream product ZC3H4 were
provides a novel insight for circRNA-miRNA-­ positively related to SiO2 stimulation in macro-
mRNA in cardiac fibrosis [142]. phages. Fibroblast proliferation and migration
can be activated through the circZC3H4 RNA/
ZC3H4 pathway. This study suggests that ZC3H4
3.3 Lung Fibrosis could be the candidate target for fibrosis therapy.
Study on the EMT in silicosis confirmed the
Lung fibrosis is a devastating condition which occurrence of the EMT in response to SiO2 expo-
could be resulted from any chronic lung diseases sure both in vivo and in vitro. In addition, cir-
including obstructive pulmonary diseases, cHECTD1/HECTD1 pathway was reported to
266 J. Yao et al.

regulate silica-induced fibrosis. SiO2 modulated thus interfering disease progress. Ideally, miRNA
the recruitment of cell proliferation and migra- manipulation could produce the similar effects.
tion by targeting circHECTD1 or upregulating However, previous researchers have found that
HECTD1 [154]. miRNAs and small interfering RNAs have inevi-
table off-target effects due to their short lengths
[165–167]. Unlike miRNAs or small interfering
4 Clinical Applications RNAs, circRNAs have low off-target effects and
of circRNAs in Fibrosis more stable structure.
Circular RNAs provide potential drug targets
4.1 Implications of circRNA for fibrosis therapy. However, this approach will
Therapies for Anti-Fibrosis base on innovations in drug synthetics to allow
clinical applications. Given the rapid progress in
While the anti-fibrosis studies have been prosper- these areas, circRNA-based therapeutics will
ous, investigations for according therapy has undoubtedly enrich the development of anti-­
been lagged behind. This is due to the challenges fibrotic drugs in the coming years.
of finding the common targets in fibrosis. Organ
fibrosis is not a disease; it is a common and pro-
gressive process in chronic disease. It involves 4.2 Implications of Diagnostic
various complicated interactions. Currently, a and Prognostic circRNA
great number of strategies have been reported to Markers in Fibrosis
be useful in experimental fibrosis models in vivo
and in vitro. The anti-fibrosis strategies proposed CircRNAs could also be candidate markers for
nowadays are majorly based on the following disease. As mentioned above, circRNAs have
mechanisms: (1) modulation of the accompany- more stable structures and are resistant to nucle-
ing inflammation, (2) pharmacological preven- ase [168–170]. The half-lives of circRNAs can be
tion of ECM deposition, (3) correction of the as long as 50 h. On average, the half-lives of cir-
altered epigenome, and (4) inhibition of profi- cRNAs are about 2.5-fold longer than their linear
brotic growth factors, especially TGF-β [155– counterparts [42, 171]. Also, the expressions of
158]. Among these, direct inhibition of TGF-β is circRNAs in the blood are generally more stable
arguable since studies have shown that it might than linear mRNAs. All the above advantages
not be safe [21, 159]. Apart from inducing fibro- make circRNAs attractive diagnostic and prog-
sis, TGF-β performs anti-inflammatory and anti- nostic tools [172, 173].
cancer activities [160, 161]. TGF-β1 knockout CircRNAs have been reported to be potential
mice were found to have more generalized biomarkers for several types of cancer [174, 175].
inflammation. What’s more, TGF-β1 mutations In liver cancer, it was found that the expression of
are associated with certain gastrointestinal can- hsa_circ_0001649 (circSHPRH) is significantly
cers [162–164]. downregulated. What’s more, the level of circSH-
With the emergence of researches on cir- PRH is associated with the size of the tumor as
cRNAs, more attention has shifted toward cir- well as the occurrence of tumor embolus.
cRNAs. Compared with synthetic molecules The expression levels of circRNAs correlate
(modified chemical drugs, RNA interference with fibrosis too. CircRNA expression profiles
constructs), circRNAs might have reduced side revealed that expressions of over 800 circRNAs
effects. The main function of circRNAs is spong- were changed during liver fibrosis. Among these
ing miRNA, which is found to play a significant circRNAs, hsa_circ_0071410 performed anti-­
role in all kinds of diseases, including fibrosis. fibrosis effects [199].
CircRNAs could modulate protein (especially Determining fibrosis status appears to be more
inflammatory cytokines and other protein sig- challenging in certain organs, such as the heart,
nals) expression by interacting with miRNAs, in which case functional test seems more preva-
21 Circular RNAs in Organ Fibrosis 267

lent. However, functional test is not sensitive Competing Financial Interests The authors declare no
competing financial interests.
enough to detect early fibrosis. Progress has been
made to identify serum markers in early cardiac
fibrosis. The expression of circRNA_010567 and
circRNA_000203 changed during cardiac fibro-
sis. CircRNA_010567 is markedly increased in
References
Ang II-induced cardiac fibrosis through cir- 1. Murakami M (1991) Study of the mechanism of sup-
cRNA_010567/miR-141/TGF-b1 axis [200]. porting cells repairing the organ of Corti in terms of
Similarly, circRNA_000203/miR-26b-5p axis cell kinetics–nuclear DNA synthesis of supporting
was found to effect fibroblasts in diabetic myo- cell of the organ of Corti in the cochlea damaged by
nitromin administration. Nihon Jibiinkoka Gakkai
cardium [201]. These findings provide a novel Kaiho 94(3):386–395
insight for circRNA-miRNA-mRNA in cardiac 2. Muller I, Vogl T, Pappritz K et al (2017) Pathogenic
fibrosis and their potential role in diagnosis and role of the damage-associated molecular patterns
prognosis. However, cardiac fibrosis is still less S100A8 and S100A9 in coxsackievirus B3-induced
myocarditis. Circ Heart Fail 10(11)
assessed in the lab setting, compared with other 3. Guerrero-Juarez CF, Plikus MV (2017) Gli-fully
organs. Until now no blood test has been approved halting the progression of fibrosis. Cell Stem Cell
for clinical use yet. 20(6):735–736
4. Nightingale S, Stormon MO, O’Loughlin EV et al
(2017) Early posthepatoportoenterostomy predictors
of native liver survival in biliary atresia. J Pediatr
5 Conclusions Gastroenterol Nutr 64(2):203–209
5. Schwab ME (2002) Increasing plasticity and func-
CircRNAs are a class of noncoding RNAs that tional recovery of the lesioned spinal cord. Prog
Brain Res 137:351–359
regulate the expression of genes by a variety of 6. Tampe B, Steinle U, Tampe D et al (2017) Low-dose
mechanisms. They might also have the potential hydralazine prevents fibrosis in a murine model of
roles in regulating protein expression. The func- acute kidney injury-to-chronic kidney disease pro-
tions of circRNAs are so complicated, and the gression. Kidney Int 91(1):157–176
7. Zoubek ME, Trautwein C, Strnad P (2017) Reversal
regulation mechanisms have not yet been com- of liver fibrosis: from fiction to reality. Best Pract
pletely understood. Current studies mainly focus Res Clin Gastroenterol 31(2):129–141
on their expression changes during physiological 8. Friedman SL, Sheppard D, Duffield JS et al (2013)
and pathological process. Generally speaking, Therapy for fibrotic diseases: nearing the starting
line. Sci Transl Med 5(167):167sr161
circRNA has emerged as a new tool to under- 9. Wynn TA (2007) Common and unique mechanisms
stand different diseases and their progressions. regulate fibrosis in various fibroproliferative dis-
Efforts have been made to explore their potential eases. J Clin Invest 117(3):524–529
use in detecting disease status and gene therapy. 10. Fani F, Regolisti G, Delsante M et al (2017) Recent
advances in the pathogenetic mechanisms of sepsis-­
The function study of circRNAs in organ fibrosis associated acute kidney injury. J Nephrol. https://
is a rapid developing field. CircRNAs involved in doi.org/10.1007/s40620-017-0452-4
fibrosis have been discovered, but these studies 11. Fischereder M, Schroppel B (2009) The role of
are still in their initial stage. Difficulties and chal- chemokines in acute renal allograft rejection and
chronic allograft injury. Front Biosci (Landmark Ed)
lenges are on the way. CircRNAs will be impor- 14:1807–1814
tant players in genetic regulation in the future. 12. Gomez H, Kellum JA, Ronco C (2017) Metabolic
reprogramming and tolerance during sepsis-induced
Acknowledgments This work was supported by the AKI. Nat Rev Nephrol 13(3):143–151
grants from National Natural Science Foundation of 13. Kusano KF, Pola R, Murayama T et al (2005) Sonic
China (81370280 and 81570332 to LZ, 81470515 and hedgehog myocardial gene therapy: tissue repair
81670362 to JH Xu), Shanghai Medical Guide Project through transient reconstitution of embryonic signal-
from Shanghai Science and Technology Committee ing. Nat Med 11(11):1197–1204
(134119a3000 to Jiahong Xu), and the grant from Jiangsu 14. Li X, Shu R, Filippatos G et al (2004) Apoptosis in
Province’s Key Provincial Talents Program lung injury and remodeling. J Appl Physiol (1985)
(ZDRCA2016019 to LZ). 97(4):1535–1542
268 J. Yao et al.

15. Mehrotra P, Collett JA, Gunst SJ et al (2018) Th17 31. Salzman J, Chen RE, Olsen MN et al (2013) Cell-­
cells contribute to pulmonary fibrosis and inflamma- type specific features of circular RNA expression.
tion during chronic kidney disease progression after PLoS Genet 9(9):e1003777
acute ischemia. Am J Physiol Regul Integr Comp 32. Liang D, Wilusz JE (2014) Short intronic repeat
Physiol 314(2):R265–R273 sequences facilitate circular RNA production. Genes
16. Niroomand F, Kubler W (1994) Hibernating, stun- Dev 28(20):2233–2247
ning and ischemic preconditioning of the myo- 33. Starke S, Jost I, Rossbach O et al (2015) Exon circu-
cardium: therapeutic implications. Clin Investig larization requires canonical splice signals. Cell Rep
72(10):731–736 10(1):103–111
17. Popovic B, Sutic I, Skocibusic N et al (2015) 34. Hansen TB, Jensen TI, Clausen BH et al (2013)
Cholestasis and Inflammation of the Pancreas in Natural RNA circles function as efficient microRNA
family medicine. Acta Med Croatica 69(4):319–326 sponges. Nature 495(7441):384–388
18. Ranganathan P, Jayakumar C, Ramesh G (2013) 35. Burd CE, Jeck WR, Liu Y et al (2010) Expression
Proximal tubule-specific overexpression of netrin-1 of linear and novel circular forms of an INK4/ARF-­
suppresses acute kidney injury-induced interstitial associated non-coding RNA correlates with athero-
fibrosis and glomerulosclerosis through suppres- sclerosis risk. PLoS Genet 6(12):e1001233
sion of IL-6/STAT3 signaling. Am J Physiol Renal 36. Memczak S, Jens M, Elefsinioti A et al (2013)
Physiol 304(8):F1054–F1065 Circular RNAs are a large class of animal RNAs
19. Tewes S, Gueler F, Chen R et al (2017) Functional with regulatory potency. Nature 495(7441):333–338
MRI for characterization of renal perfusion impair- 37. Lukiw WJ (2013) Circular RNA (circRNA) in
ment and edema formation due to acute kid- Alzheimer’s disease (AD). Front Genet 4:307
ney injury in different mouse strains. PLoS One 38. Bachmayr-Heyda A, Reiner AT, Auer K et al (2015)
12(3):e0173248 Correlation of circular RNA abundance with pro-
20. Xiao Y, Yang N, Zhang Q et al (2014) Pentraxin 3 liferation--exemplified with colorectal and ovarian
inhibits acute renal injury-induced interstitial fibro- cancer, idiopathic lung fibrosis, and normal human
sis through suppression of IL-6/Stat3 pathway. tissues. Sci Rep 5:8057
Inflammation 37(5):1895–1901 39. Rybak-Wolf A, Stottmeister C, Glazar P et al (2015)
21. Wynn TA, Ramalingam TR (2012) Mechanisms of Circular RNAs in the mammalian brain are highly
fibrosis: therapeutic translation for fibrotic disease. abundant, conserved, and dynamically expressed.
Nat Med 18(7):1028–1040 Mol Cell 58(5):870–885
22. Westholm JO, Miura P, Olson S et al (2014) 40. You X, Vlatkovic I, Babic A et al (2015) Neural cir-
Genome-wide analysis of drosophila circular RNAs cular RNAs are derived from synaptic genes and reg-
reveals their structural and sequence properties ulated by development and plasticity. Nat Neurosci
and age-dependent neural accumulation. Cell Rep 18(4):603–610
9(5):1966–1980 41. Jeck WR, Sorrentino JA, Wang K et al (2013)
23. Salzman J, Gawad C, Wang PL et al (2012) Circular Circular RNAs are abundant, conserved, and associ-
RNAs are the predominant transcript isoform from ated with ALU repeats. RNA 19(2):141–157
hundreds of human genes in diverse cell types. PLoS 42. Suzuki H, Zuo Y, Wang J et al (2006) Characterization
One 7(2):e30733 of RNase R-digested cellular RNA source that con-
24. Yang L, Zou M, Fu B et al (2013) Genome-wide sists of lariat and circular RNAs from pre-mRNA
identification, characterization, and expression anal- splicing. Nucleic Acids Res 34(8):e63
ysis of lineage-specific genes within zebrafish. BMC 43. Barrett SP, Salzman J (2016) Circular RNAs: analy-
Genomics 14:65 sis, expression and potential functions. Development
25. Kelly S, Greenman C, Cook PR et al (2015) Exon 143(11):1838–1847
Skipping Is Correlated with Exon Circularization. 44. Guo JU, Agarwal V, Guo H et al (2014) Expanded
J Mol Biol 427(15):2414–2417 identification and characterization of mammalian
26. Nigro JM, Cho KR, Fearon ER et al (1991) circular RNAs. Genome Biol 15(7):409
Scrambled exons. Cell 64(3):607–613 45. Veno MT, Hansen TB, Veno ST et al (2015) Spatio-­
27. Cocquerelle C, Daubersies P, Majerus MA et al temporal regulation of circular RNA expression dur-
(1992) Splicing with inverted order of exons occurs ing porcine embryonic brain development. Genome
proximal to large introns. EMBO J 11(3):1095–1098 Biol 16:245
28. Capel B, Swain A, Nicolis S et al (1993) Circular 46. Bellemare J, Roberge CJ, Bergeron D et al (2005)
transcripts of the testis-determining gene Sry in Epidermis promotes dermal fibrosis: role in the
adult mouse testis. Cell 73(5):1019–1030 pathogenesis of hypertrophic scars. J Pathol
29. Cocquerelle C, Mascrez B, Hetuin D et al (1993) 206(1):1–8
Mis-splicing yields circular RNA molecules. FASEB 47. Holvoet T, Devriese S, Castermans K et al (2017)
J 7(1):155–160 Treatment of intestinal fibrosis in experimen-
30. Pasman Z, Been MD, Garcia-Blanco MA (1996) tal inflammatory bowel disease by the pleio-
Exon circularization in mammalian nuclear extracts. tropic actions of a local rho kinase inhibitor.
RNA 2(6):603–610 Gastroenterology 153(4):1054–1067
21 Circular RNAs in Organ Fibrosis 269

48. Alcalde O, Cabrera Gomez S, Valles Gras E et al teral obstruction: a new look at an old model. Am
(2017) Rheumatoid arthritis with severe atrial fibro- J Physiol Renal Physiol 301(1):F110–F117
sis and multiple atrial arrhythmias: chronic atrial 63. Akram KM, Lomas NJ, Spiteri MA et al (2013)
myocarditis? Rev Esp Cardiol (Engl Ed). https://doi. Club cells inhibit alveolar epithelial wound repair
org/10.1016/j.rec.2017.04.003 via TRAIL-dependent apoptosis. Eur Respir
49. Abdul N, Dixon D, Walker A et al (2015) Fibrosis J 41(3):683–694
is a common outcome following total knee arthro- 64. Aravinthan A, Scarpini C, Tachtatzis P et al (2013)
plasty. Sci Rep 5:16469 Hepatocyte senescence predicts progression in
50. Ramos F, Robledo C, Izquierdo-Garcia FM et al non-alcohol-related fatty liver disease. J Hepatol
(2016) Bone marrow fibrosis in myelodysplastic 58(3):549–556
syndromes: a prospective evaluation including muta- 65. Rowe RG, Lin Y, Shimizu-Hirota R et al (2011)
tional analysis. Oncotarget 7(21):30492–30503 Hepatocyte-derived Snail1 propagates liver fibrosis
51. Dvorak HF, Harvey VS, McDonagh J (1984) progression. Mol Cell Biol 31(12):2392–2403
Quantitation of fibrinogen influx and fibrin deposi- 66. Roberts AB, Anzano MA, Lamb LC et al (1982)
tion and turnover in line 1 and line 10 guinea pig Isolation from murine sarcoma cells of novel trans-
carcinomas. Cancer Res 44(8):3348–3354 forming growth factors potentiated by EGF. Nature
52. Dvorak HF (1986) Tumors: wounds that do not heal. 295(5848):417–419
Similarities between tumor stroma generation and 67. Border WA, Okuda S, Languino LR et al (1990)
wound healing. N Engl J Med 315(26):1650–1659 Suppression of experimental glomerulonephritis by
53. Iwano M, Plieth D, Danoff TM et al (2002) Evidence antiserum against transforming growth factor beta 1.
that fibroblasts derive from epithelium during tissue Nature 346(6282):371–374
fibrosis. J Clin Invest 110(3):341–350 68. Denis M (1994) Neutralization of transforming
54. Jian XC, Liu SF, Shen ZH et al (1988) growth factor-beta 1 in a mouse model of immune-­
Histomorphology of oral submucous fibrosis. Report induced lung fibrosis. Immunology 82(4):584–590
of 24 cases. Chin Med J (Engl) 101(7):505–509 69. Kuwahara F, Kai H, Tokuda K et al (2002)
55. Zhang SS, Gong ZJ, Xiong W et al (2016) A rat Transforming growth factor-beta function block-
model of oral submucous fibrosis induced by bleo- ing prevents myocardial fibrosis and diastolic dys-
mycin. Oral Surg Oral Med Oral Pathol Oral Radiol function in pressure-overloaded rats. Circulation
122(2):216–223 106(1):130–135
56. Harwood IR, Greene LM, Kozakowski-Koch JA et al 70. Roberts AB, Sporn MB, Assoian RK et al (1986)
(1992) New peripherally inserted midline catheter: a Transforming growth factor type beta: rapid induc-
better alternative for intravenous antibiotic therapy tion of fibrosis and angiogenesis in vivo and stimula-
in patients with cystic fibrosis. Pediatr Pulmonol tion of collagen formation in vitro. Proc Natl Acad
12(4):233–239 Sci USA 83(12):4167–4171
57. Jiao LR, Seifalian AM, Habib N et al (1999) The 71. Crosas-Molist E, Fabregat I (2015) Role of NADPH
effect of mechanically enhancing portal venous oxidases in the redox biology of liver fibrosis. Redox
inflow on hepatic oxygenation, microcirculation, Biol 6:106–111
and function in a rabbit model with extensive hepatic 72. Kendall RT, Feghali-Bostwick CA (2014)
fibrosis. Hepatology 30(1):46–52 Fibroblasts in fibrosis: novel roles and mediators.
58. Perry KA, Banarjee A, Liu J et al (2013) Gastric Front Pharmacol 5:123
ischemic conditioning increases neovascularization 73. Kalluri R, Zeisberg M (2006) Fibroblasts in cancer.
and reduces inflammation and fibrosis during gas- Nat Rev Cancer 6(5):392–401
troesophageal anastomotic healing. Surg Endosc 74. Kalluri R, Zeisberg E (2006) Controlling angiogen-
27(3):753–760 esis in heart valves. Nat Med 12(10):1118–1119
59. Tucker RD, Gibbs GE, Christensen MB (1979) 75. Kalluri R, Zeisberg M (2003) Exploring the connec-
Cystic fibrosis serum effect on the short circuit cur- tion between chronic renal fibrosis and bone morpho-
rent of rat jejunum. Pediatr Res 13(12):1371–1374 genic protein-7. Histol Histopathol 18(1):217–224
60. Poling J, Gajawada P, Lorchner H et al (2012) The 76. Moore-Morris T, Cattaneo P, Puceat M et al (2016)
Janus face of OSM-mediated cardiomyocyte dedif- Origins of cardiac fibroblasts. J Mol Cell Cardiol
ferentiation during cardiac repair and disease. Cell 91:1–5
Cycle 11(3):439–445 77. Zeisberg EM, Kalluri R (2010) Origins of cardiac
61. Hano H, Takasaki S (2003) Three-dimensional fibroblasts. Circ Res 107(11):1304–1312
observations on the alterations of lobular architec- 78. Dufourcq P, Louis H, Dandre F et al (1997)
ture in chronic hepatitis with special reference to Phenotypic modification of arterial smooth muscle
its angioarchitecture for a better understanding of cells in response to medial dissection. Coron Artery
the formal pathogenesis of liver cirrhosis. Virchows Dis 8(3–4):163–170
Arch 443(5):655–663 79. Shujia J, Haider HK, Idris NM et al (2008) Stable
62. Forbes MS, Thornhill BA, Chevalier RL (2011) therapeutic effects of mesenchymal stem cell-based
Proximal tubular injury and rapid formation of multiple gene delivery for cardiac repair. Cardiovasc
atubular glomeruli in mice with unilateral ure- Res 77(3):525–533
270 J. Yao et al.

80. Tanaka Y, Nouchi T, Yamane M et al (1991) population that mediates tissue repair. Mol Med
Phenotypic modulation in lipocytes in experimental 1(1):71–81
liver fibrosis. J Pathol 164(3):273–278 96. Lin SL, Kisseleva T, Brenner DA et al (2008)
81. Grupp C, Lottermoser J, Cohen DI et al (1997) Pericytes and perivascular fibroblasts are the primary
Transformation of rat inner medullary fibroblasts to source of collagen-producing cells in obstructive
myofibroblasts in vitro. Kidney Int 52(5):1279–1290 fibrosis of the kidney. Am J Pathol 173(6):1617–1627
82. Maxwell PH, Osmond MK, Pugh CW et al 97. Ronnov-Jessen L, Petersen OW (1996) A func-
(1993) Identification of the renal erythropoietin-­ tion for filamentous alpha-smooth muscle actin:
producing cells using transgenic mice. Kidney Int retardation of motility in fibroblasts. J Cell Biol
44(5):1149–1162 134(1):67–80
83. Paliege A, Rosenberger C, Bondke A et al (2010) 98. Ronnov-Jessen L, Petersen OW, Koteliansky VE
Hypoxia-inducible factor-2alpha-expressing inter- et al (1995) The origin of the myofibroblasts in
stitial fibroblasts are the only renal cells that express breast cancer. Recapitulation of tumor environment
erythropoietin under hypoxia-inducible factor stabi- in culture unravels diversity and implicates con-
lization. Kidney Int 77(4):312–318 verted fibroblasts and recruited smooth muscle cells.
84. Camelliti P, Devlin GP, Matthews KG et al (2004) J Clin Invest 95(2):859–873
Spatially and temporally distinct expression of fibro- 99. Ronnov-Jessen L, Petersen OW (1993) Induction of
blast connexins after sheep ventricular infarction. alpha-smooth muscle actin by transforming growth
Cardiovasc Res 62(2):415–425 factor-beta 1 in quiescent human breast gland fibro-
85. Camelliti P, Green CR, Kohl P (2006) Structural and blasts. Implications for myofibroblast generation in
functional coupling of cardiac myocytes and fibro- breast neoplasia. Lab Invest 68(6):696–707
blasts. Adv Cardiol 42:132–149 100. Boutet A, De Frutos CA, Maxwell PH et al (2006)
86. Camelliti P, Green CR, LeGrice I et al (2004) Snail activation disrupts tissue homeostasis
Fibroblast network in rabbit sinoatrial node: struc- and induces fibrosis in the adult kidney. EMBO
tural and functional identification of homoge- J 25(23):5603–5613
neous and heterogeneous cell coupling. Circ Res 101. Broekema M, Harmsen MC, van Luyn MJ et al
94(6):828–835 (2007) Bone marrow-derived myofibroblasts con-
87. Tveito A, Lines G, Artebrant R et al (2011) Existence tribute to the renal interstitial myofibroblast popu-
of excitation waves for a collection of cardiomyo- lation and produce procollagen I after ischemia/
cytes electrically coupled to fibroblasts. Math Biosci reperfusion in rats. J Am Soc Nephrol 18(1):165–175
230(2):79–86 102. Humphreys BD, Lin SL, Kobayashi A et al (2010)
88. Chang HY, Chi JT, Dudoit S et al (2002) Diversity, Fate tracing reveals the pericyte and not epithelial
topographic differentiation, and positional memory origin of myofibroblasts in kidney fibrosis. Am
in human fibroblasts. Proc Natl Acad Sci USA J Pathol 176(1):85–97
99(20):12877–12882 103. Kim KK, Wei Y, Szekeres C et al (2009) Epithelial
89. Sugimoto H, Mundel TM, Kieran MW et al cell alpha3beta1 integrin links beta-catenin and Smad
(2006) Identification of fibroblast heterogeneity signaling to promote myofibroblast formation and
in the tumor microenvironment. Cancer Biol Ther pulmonary fibrosis. J Clin Invest 119(1):213–224
5(12):1640–1646 104. Kisseleva T, Uchinami H, Feirt N et al (2006) Bone
90. Iwano M, Fischer A, Okada H et al (2001) marrow-derived fibrocytes participate in pathogen-
Conditional abatement of tissue fibrosis using nucle- esis of liver fibrosis. J Hepatol 45(3):429–438
oside analogs to selectively corrupt DNA replication 105. Li J, Qu X, Yao J et al (2010) Blockade of
in transgenic fibroblasts. Mol Ther 3(2):149–159 endothelial-­mesenchymal transition by a Smad3
91. Bayreuther K, Francz PI, Rodemann HP (1992) inhibitor delays the early development of strep-
Fibroblasts in normal and pathological terminal dif- tozotocin-induced diabetic nephropathy. Diabetes
ferentiation, aging, apoptosis and transformation. 59(10):2612–2624
Arch Gerontol Geriatr 15(Suppl 1):47–74 106. Osterreicher CH, Penz-Osterreicher M, Grivennikov
92. Bayreuther K, Francz PI, Rodemann HP (1995) SI et al (2011) Fibroblast-specific protein 1 identifies
Fibroblasts in normal and pathological terminal dif- an inflammatory subpopulation of macrophages in
ferentiation, aging, apoptosis and transformation. the liver. Proc Natl Acad Sci USA 108(1):308–313
Ontogenez 26(1):22–37 107. Quaggin SE, Kapus A (2011) Scar wars: mapping
93. Bayreuther K, Rodemann HP, Francz PI et al (1988) the fate of epithelial-mesenchymal-myofibroblast
Differentiation of fibroblast stem cells. J Cell Sci transition. Kidney Int 80(1):41–50
Suppl 10:115–130 108. Rock JR, Barkauskas CE, Cronce MJ et al (2011)
94. Bayreuther K, Rodemann HP, Hommel R et al Multiple stromal populations contribute to pul-
(1988) Human skin fibroblasts in vitro differenti- monary fibrosis without evidence for epithelial to
ate along a terminal cell lineage. Proc Natl Acad Sci mesenchymal transition. Proc Natl Acad Sci USA
USA 85(14):5112–5116 108(52):E1475–E1483
95. Bucala R, Spiegel LA, Chesney J et al (1994) 109. Zeisberg EM, Potenta SE, Sugimoto H et al
Circulating fibrocytes define a new leukocyte sub- (2008) Fibroblasts in kidney fibrosis emerge via
21 Circular RNAs in Organ Fibrosis 271

endothelial-­to-mesenchymal transition. J Am Soc 123. Eisenberg LM, Markwald RR (1995) Molecular

Nephrol 19(12):2282–2287 regulation of atrioventricular valvuloseptal morpho-
110. Zeisberg M, Yang C, Martino M et al (2007) genesis. Circ Res 77(1):1–6
Fibroblasts derive from hepatocytes in liver fibro- 124. Liebner S, Cattelino A, Gallini R et al (2004) Beta-­
sis via epithelial to mesenchymal transition. J Biol catenin is required for endothelial-mesenchymal
Chem 282(32):23337–23347 transformation during heart cushion development in
111. Zhou B, von Gise A, Ma Q et al (2010) Genetic fate the mouse. J Cell Biol 166(3):359–367
mapping demonstrates contribution of epicardium-­ 125. Nakajima Y, Yamagishi T, Hokari S et al (2000)
derived cells to the annulus fibrosis of the mamma- Mechanisms involved in valvuloseptal endocar-
lian heart. Dev Biol 338(2):251–261 dial cushion formation in early cardiogenesis:
112. Distler O, Distler JH, Scheid A et al (2004) roles of transforming growth factor (TGF)-beta
Uncontrolled expression of vascular endothelial and bone morphogenetic protein (BMP). Anat Rec
growth factor and its receptors leads to insufficient 258(2):119–127
skin angiogenesis in patients with systemic sclero- 126. Murray PJ, Wynn TA (2011) Protective and patho-
sis. Circ Res 95(1):109–116 genic functions of macrophage subsets. Nat Rev
113. Higgins DF, Kimura K, Bernhardt WM et al (2007) Immunol 11(11):723–737
Hypoxia promotes fibrogenesis in vivo via HIF-1 127. Wynn TA (2011) Integrating mechanisms of pulmo-
stimulation of epithelial-to-mesenchymal transition. nary fibrosis. J Exp Med 208(7):1339–1350
J Clin Invest 117(12):3810–3820 128. Wynn TA, Barron L, Thompson RW et al (2011)
114. Moon JO, Welch TP, Gonzalez FJ et al (2009) Quantitative assessment of macrophage functions
Reduced liver fibrosis in hypoxia-inducible factor-­ in repair and fibrosis. Curr Protoc Immunol Chapter
1alpha-­deficient mice. Am J Physiol Gastrointest 14:Unit14.22
Liver Physiol 296(3):G582–G592 129. Bataller R, Brenner DA (2005) Liver fibrosis. J Clin
115. Corpechot C, Barbu V, Wendum D et al (2002) Invest 115(2):209–218
Hypoxia-induced VEGF and collagen I expres- 130. Bataller R, Sancho-Bru P, Gines P et al (2005)
sions are associated with angiogenesis and fibro- Liver fibrogenesis: a new role for the renin-
genesis in experimental cirrhosis. Hepatology angiotensin system. Antioxid Redox Signal
35(5):1010–1021 7(9-10):1346–1355
116. Ioannou M, Pyrpasopoulou A, Simos G et al 131. Seki E, Schwabe RF (2015) Hepatic inflammation
(2013) Upregulation of VEGF expression is associ- and fibrosis: functional links and key pathways.
ated with accumulation of HIF-1alpha in the skin Hepatology 61(3):1066–1079
of naive scleroderma patients. Mod Rheumatol 132. Chen Y, Yuan B, Wu Z et al (2017) Microarray pro-
23(6):1245–1248 filing of circular RNAs and the potential regulatory
117. Maeshima Y, Makino H (2010) Angiogenesis and role of hsa_circ_0071410 in the activated human
chronic kidney disease. Fibrogenesis Tissue Repair hepatic stellate cell induced by irradiation. Gene
3:13 629:35–42
118. Yoon YS, Uchida S, Masuo O et al (2005) 133. Kong P, Christia P, Frangogiannis NG (2014) The
Progressive attenuation of myocardial vascular pathogenesis of cardiac fibrosis. Cell Mol Life Sci
endothelial growth factor expression is a seminal 71(4):549–574
event in diabetic cardiomyopathy: restoration of 134. Goyal BR, Mehta AA (2013) Diabetic cardiomy-
microvascular homeostasis and recovery of cardiac opathy: pathophysiological mechanisms and cardiac
function in diabetic cardiomyopathy after replen- dysfuntion. Hum Exp Toxicol 32(6):571–590
ishment of local vascular endothelial growth factor. 135. Poornima IG, Parikh P, Shannon RP (2006) Diabetic
Circulation 111(16):2073–2085 cardiomyopathy: the search for a unifying hypoth-
119. Worda M, Sgonc R, Dietrich H et al (2003) In vivo esis. Circ Res 98(5):596–605
analysis of the apoptosis-inducing effect of anti-­ 136. Suys BE, Katier N, Rooman RP et al (2004) Female
endothelial cell antibodies in systemic sclerosis children and adolescents with type 1 diabetes
by the chorionallantoic membrane assay. Arthritis have more pronounced early echocardiographic
Rheum 48(9):2605–2614 signs of diabetic cardiomyopathy. Diabetes Care
120. Rieder F, Kessler SP, West GA et al (2011) 27(8):1947–1953
Inflammation-induced endothelial-to-mesenchymal 137. Aneja A, Tang WH, Bansilal S et al (2008) Diabetic
transition: a novel mechanism of intestinal fibrosis. cardiomyopathy: insights into pathogenesis, diag-
Am J Pathol 179(5):2660–2673 nostic challenges, and therapeutic options. Am
121. Zeisberg EM, Potenta S, Xie L et al (2007) Discovery J Med 121(9):748–757
of endothelial to mesenchymal transition as a source 138. Tarquini R, Lazzeri C, Pala L et al (2011) The dia-
for carcinoma-associated fibroblasts. Cancer Res betic cardiomyopathy. Acta Diabetol 48(3):173–181
67(21):10123–10128 139. Dzeshka MS, Lip GY, Snezhitskiy V et al (2015)
122. Zeisberg EM, Tarnavski O, Zeisberg M et al (2007) Cardiac fibrosis in patients With atrial fibrillation:
Endothelial-to-mesenchymal transition contributes mechanisms and clinical implications. J Am Coll
to cardiac fibrosis. Nat Med 13(8):952–961 Cardiol 66(8):943–959
272 J. Yao et al.

140. Tveito A, Lines GT, Edwards AG et al (2012) Slow promotes the pulmonary macrophage activation.
Calcium-Depolarization-Calcium waves may initi- FASEB J. https://doi.org/10.1096/fj.201701118R:fj
ate fast local depolarization waves in ventricular tis- 201701118R
sue. Prog Biophys Mol Biol 110(2–3):295–304 155. Kawaratani H, Moriya K, Namisaki T et al (2017)
141. Tang CM, Zhang M, Huang L et al (2017) Therapeutic strategies for alcoholic liver disease:
CircRNA_000203 enhances the expression of focusing on inflammation and fibrosis (Review). Int
fibrosis-­associated genes by derepressing targets of J Mol Med 40(2):263–270
miR-26b-5p, Col1a2 and CTGF, in cardiac fibro- 156. Nogueira A, Pires MJ, Oliveira PA (2017)
blasts. Sci Rep 7:40342 Pathophysiological mechanisms of renal fibrosis: a
142. Zhou B, Yu JW (2017) A novel identified cir- review of animal models and therapeutic strategies.
cular RNA, circRNA_010567, promotes myo- In Vivo 31(1):1–22
cardial fibrosis via suppressing miR-141 by 157. Owens GM (2017) Strategies to manage costs in
targeting TGF-beta1. Biochem Biophys Res idiopathic pulmonary fibrosis. Am J Manag Care
Commun 487(4):769–775 23(11 Suppl):S191–S196
143. Barkauskas CE, Noble PW (2014) Cellular mecha- 158. Stasi C, Milani S (2017) Evolving strategies for liver
nisms of tissue fibrosis. 7. New insights into the fibrosis staging: non-invasive assessment. World
cellular mechanisms of pulmonary fibrosis. Am J Gastroenterol 23(2):191–196
J Physiol Cell Physiol 306(11):C987–C996 159. Friedman SL (2004) Mechanisms of disease:
144. Noble PW, Barkauskas CE, Jiang D (2012) mechanisms of hepatic fibrosis and therapeutic
Pulmonary fibrosis: patterns and perpetrators. J Clin implications. Nat Clin Pract Gastroenterol Hepatol
Invest 122(8):2756–2762 1(2):98–105
145. Churg A, Wright JL, Tazelaar HD (2011) Acute 160. Bayomi HS, Elsherbiny NM, El-Gayar AM et al
exacerbations of fibrotic interstitial lung disease. (2013) Evaluation of renal protective effects of
Histopathology 58(4):525–530 inhibiting TGF-beta type I receptor in a cisplatin-­
146. Tazelaar HD, Wright JL, Churg A (2011) induced nephrotoxicity model. Eur Cytokine Netw
Desquamative interstitial pneumonia. 24(4):139–147
Histopathology 58(4):509–516 161. Xu Y, Lou Z, Lee SH (2017) Arctigenin represses
147. Wright JL, Tazelaar HD, Churg A (2011) Fibrosis TGF-beta-induced epithelial mesenchymal transi-
with emphysema. Histopathology 58(4):517–524 tion in human lung cancer cells. Biochem Biophys
148. Mathai SK, Gulati M, Peng X et al (2010) Res Commun 493(2):934–939
Circulating monocytes from systemic sclero- 162. Munjal C, Opoka AM, Osinska H et al (2014) TGF-­
sis patients with interstitial lung disease show beta mediates early angiogenesis and latent fibrosis
an enhanced profibrotic phenotype. Lab Invest in an Emilin1-deficient mouse model of aortic valve
90(6):812–823 disease. Dis Model Mech 7(8):987–996
149. Carneiro PJ, Clevelario AL, Padilha GA et al (2017) 163. Katz LH, Likhter M, Jogunoori W et al (2016) TGF-­
Bosutinib therapy ameliorates lung inflammation beta signaling in liver and gastrointestinal cancers.
and fibrosis in experimental silicosis. Front Physiol Cancer Lett 379(2):166–172
8:159 164. Rao S, Zaidi S, Banerjee J et al (2017) Transforming
150. Miao R, Ding B, Zhang Y et al (2016) Proteomic growth factor-beta in liver cancer stem cells and
profiling differences in serum from silicosis and regeneration. Hepatol Commun 1(6):477–493
chronic bronchitis patients: a comparative analysis. 165. Kiryu H, Terai G, Imamura O et al (2011) A
J Thorac Dis 8(3):439–450 detailed investigation of accessibilities around tar-
151. Rosengarten D, Fox BD, Fireman E et al (2017) get sites of siRNAs and miRNAs. Bioinformatics
Survival following lung transplantation for artificial 27(13):1788–1797
stone silicosis relative to idiopathic pulmonary fibro- 166. Samuel P, Pink RC, Brooks SA et al (2016) miR-
sis. Am J Ind Med 60(3):248–254 NAs and ovarian cancer: a miRiad of mechanisms
152. Zhao MM, Cui JZ, Cui Y et al (2013) Therapeutic to induce cisplatin drug resistance. Expert Rev
effect of exogenous bone marrowderived mes- Anticancer Ther 16(1):57–70
enchymal stem cell transplantation on silicosis 167. Scimeca JC, Verron E (2017) The multiple therapeu-
via paracrine mechanisms in rats. Mol Med Rep tic applications of miRNAs for bone regenerative
8(3):741–746 medicine. Drug Discov Today 22(7):1084–1091
153. Gungen AC, Aydemir Y, Coban H et al (2016) 168. Chen Y, Li C, Tan C et al (2017) Circular RNA
Lung cancer in patients diagnosed with silicosis in human disease and their potential clinic sig-
should be investigated. Respir Med Case Rep nificance. Zhonghua Yi Xue Yi Chuan Xue Za Zhi
18:93–95 34(1):133–137
154. Yang X, Wang J, Zhou Z et al (2018) Silica-induced 169. Park WS, Miyano-Kurosaki N, Abe T et al (1999)
initiation of circular ZC3H4 RNA/ZC3H4 pathway Properties of circular dumbbell RNA/DNA chimeric
21 Circular RNAs in Organ Fibrosis 273

oligonucleotides containing antisense phosphodies- 172. Chen LL (2016) The biogenesis and emerging
ter oligonucleotides. Nucleic Acids Symp Ser doi roles of circular RNAs. Nat Rev Mol Cell Biol
42:225–226 17(4):205–211
170. Yamakawa H, Abe T, Saito T et al (1998) Properties 173. Greene J, Baird AM, Brady L et al (2017) Circular
of nicked and circular dumbbell RNA/DNA chime- RNAs: biogenesis, function and role in human dis-
ric oligonucleotides containing antisense phospho- eases. Front Mol Biosci 4:38
diester oligodeoxynucleotides. Bioorg Med Chem 174. Qin M, Liu G, Huo X et al (2016) Hsa_circ_0001649:
6(7):1025–1032 a circular RNA and potential novel biomarker
171. Zuo H, Suzuki S, Sotoda M et al (2006) New tech- for hepatocellular carcinoma. Cancer Biomark
nique for visualizing cerebral vessels in MR angio- 16(1):161–169
graphic images using three-dimensional discrete 175. Bao C, Lyu D, Huang S (2016) Circular RNA
wavelet transform. Igaku Butsuri 26(2):65–74 expands its territory. Mol Cell Oncol 3(2):e1084443
 Circular RNAs in Metabolic
Diseases 22
Tianhui Wang, Wen Pan, Jun Hu,
Zhongrong Zhang, Guoping Li, and Yajun Liang

Abstract molecular diagnostics, nucleic acid therapy,

Metabolic diseases include diabetes mellitus and biomarkers.
(DM), obesity, metabolic syndrome, and non-­
alcoholic fatty liver disease (NAFLD). Keywords
Circular RNA is a new type of RNA that is Circular RNA · Metabolic diseases
different from traditional linear RNA and has
a closed loop structure. However, the function
of circular RNA is not yet well elucidated in 1 Introduction
metabolic diseases. Only a few studies have
reported about the relationship between circu- Since the first discovery of circular RNAs in
lar RNA and metabolic diseases such as DM plants in 1976 [1], thousands of circular RNAs
and NAFLD. This chapter presents a brief have been demonstrated to express in eukaryotic
review of epidemiology, pathophysiology, or cells [2–5]. Generally, the biological functions of
treatment of DM and NAFLD and then dis- circular RNA mainly include the following six
cusses the relationship between circular RNA aspects: (1) microRNA sponge, (2) RBP sponge,
and DM or NAFLD. Besides, this chapter fur- (3) regulator of transcription, (4) interaction with
ther provides an updated discussion of the long non-coding RNAs, (5) interaction with
most relevant discoveries regarding circular mRNAs, and (6) secreted into exosomes. To date,
RNA and their potential applications in the function of circular RNAs serving as

Author contributed equally with all other contributors.

Tianhui Wang, Wen Pan and Jun Hu

T. Wang · Y. Liang (*)

Cardiac Regeneration and Ageing Lab, Institute of
Cardiovascular Sciences, School of Life Science,
Shanghai University, Shanghai, China
J. Hu
Shanghai Key Laboratory of Bio-Energy Crops, Department of Pediatric Surgery, The First Affiliated
School of Life Sciences, Shanghai University, Hospital of Wenzhou Medical University,
Shanghai, China Wenzhou, China
W. Pan · Z. Zhang G. Li
Cardiac Regeneration and Ageing Lab, Institute of Cardiovascular Division of the Massachusetts
Cardiovascular Sciences, School of Life Science, General Hospital, Harvard Medical School,
Shanghai University, Shanghai, China Boston, MA, USA

© Springer Nature Singapore Pte Ltd. 2018 275

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_22
276 T. Wang et al.

Table 22.1 Circular RNAs in metabolic diseases

Circular RNA Biological process Reference
Diabetes mellitus
ciRS-7/CDR1as Sponge for miR-7 and involved in the regulation of insulin secretion [39]
circRNA HIPK3 Enhance retinal vascular disorders via sponging miR-30a-3p [6, 39]
Sponge for miR-7 and involved in the regulation of insulin secretion
Cdr1as Increases the insulin secretion via targeting miR-7 in mouse β cells [40]
hsa_circ_0054633 Capable of predicting prediabetes in peripheral blood [57]
hsa_ Differentially expressed in the peripheral blood of diabetes mellitus patients or [58]
circRNA11783-­2 control individuals
circRNA_010567 circRNA_010567/miR-141/TGF-β1 axis plays an important regulatory role in the [37]
diabetic mice myocardial fibrosis model
circRNA_000203 Eliminate the anti-fibrosis effect of miR-26b-5 in the diabetic mice myocardial [59]
fibrosis model
circ_0005015 Facilitate retinal endothelial angiogenic function via sponging miR-519d-3p in [61]
retinal endothelial cell
circRNA WDR77 Regulate proliferation and migration via circRNA WDR77-miR-124-FGF2 axis [62]
in high-glucose-induced VSMCs
Non-alcoholic fatty liver disease
circRNA_021412 Lead to hepatic steatosis via circRNA_021412/miR-1972/LPIN1 axis in HepG2 [85]
cells induced by high-fat mixture
circRNA_0046366 Lead to hepatocyte steatosis via circRNA_0046366 / miR-34a/PPARα axis in [89]
HepG2 cells induced by high-fat mixture
circRNA_0046367 Facilitate steatosis resolution via circRNA_0046367/miR-34a/PPARα axis in [90]
HepG2 cells induced by high-fat mixture

microRNA sponge is well elucidated in accumu- 2 Diabetes Mellitus (DM)

lating studies [6–9]. Metabolic diseases, includ-
ing diabetes mellitus, obesity, metabolic 2.1 The Epidemiology
syndrome, and non-alcoholic fatty liver disease and Pathophysiology
(NAFLD), are commonly caused by metabolic of Diabetes Mellitus
homeostasis imbalance such as glucose
­metabolism, lipid metabolism, insulin signaling, 2.1.1 Diabetes Mellitus Epidemiology
or metabolism-related genes dysregulation [10– Recently, diabetes mellitus has emerged as a
12]. Although numerous studies have suggested major threat to human health worldwide. 425
that microRNAs played vital roles in metabolic million adults were diagnosed with diabetes all
diseases [13], the potential roles of circular RNAs over the world, with an estimated increase to 629
in metabolic diseases are less mentioned. Given million by 2045. And there are 352 million peo-
the close relationships between circular RNAs ple who have the risk of developing type 2 diabe-
and microRNAs, researchers need to pay more tes mellitus (T2DM) [14]. The International
attention on the functions of circular RNAs in Diabetes Federation (IDF) has predicted the
metabolism diseases. In this chapter, we will prevalence of diabetes mellitus in many coun-
introduce the functions of microRNAs and circu- tries. It is reported that the morbidity and preva-
lar RNAs in metabolic diseases. And then we will lence of diabetes mellitus vary in different
discuss the potential applications of circular geographical regions [14, 15]. The burden of dia-
RNAs in molecular diagnostics, nucleic acid betes has grown faster in low-income and middle-­
therapy, and biomarkers (Table 22.1). income countries than in high-income countries.
22 Circular RNAs in Metabolic Diseases 277

About 79% of the people who were diagnosed gested that new therapeutic strategies should
with diabetes live in low- or middle-income focus on ameliorating the function of β cells [31].
countries [16].

2.1.2  iabetes Mellitus Etiology,

D 2.2 MicroRNAs and Diabetes
Pathophysiology, Mellitus
and Diagnosis
It is reported that type 1 diabetes mellitus (T1DM) Numerous studies have demonstrated the func-
is caused by the absolute lack of insulin which is tions of non-coding RNAs in diabetes mellitus.
due to autoimmune-mediated destruction of pan- Non-coding RNAs, especially microRNAs, have
creatic β cells. Ninety percent of diabetes are been reported to play important roles in diabetes
T2DM which is characterized by relative insulin mellitus pathophysiology. For example, miR-375
deficiencies and abnormal hyperglycemia [17, is specifically expressed in pancreas cells and
18]. Long-term dysregulation of metabolism-­ affects insulin secretion [32]. MiR-124a modu-
related substances will lead to damages to organs, lates insulin secretion by regulating the expres-
including the heart, brain, kidney, liver, etc. [19– sion of Rab27a in pancreatic β cells [33]. Let-7
21]. The World Health Organization defines the family members regulate glucose homeostasis
clinical diagnosis of diabetes mellitus based on and insulin sensitivity by directly targeting sev-
the clinical symptoms and values of plasma glu- eral components of insulin-signaling pathway,
cose [22]. such as Insr, Igf1r, Pik3ip1, and Irs2 [34]. In
However, the diagnosis does not provide any addition to the above microRNAs, other microR-
clues to the causes or progression of the disease. NAs, including miR-29 family [35], miR-107
In addition, almost half of the cases have not [36], miR-33 family [37], Let-7 family [38],
been diagnosed, leading to increased risk of miR-221/222 family [39], and miR-223 family
organ damages through epigenetic mechanisms. [40], are also reported to affect β-cell prolifera-
Therefore, identifying novel biomarkers that can tion, insulin secretion, and exocytosis.
indicate the early stages of disease is highly
needed.
People who are chronically suffering diabetes 2.3  ircular RNAs and Diabetes
C
mellitus are at high risk of developing severe Mellitus
microvascular or macro-vascular complications
[23–25]. If patients are not diagnosed early or 2.3.1 Circular RNAs and Diseases
treated appropriately, these complications will It has been investigated that circular RNAs play
eventually develop into cardiovascular disease an important role of being “super sponge” in
(diabetic cardiomyopathy), kidney disease (dia- other diseases. Circular RNAs-microRNAs-­
betic nephropathy), diabetic eye disease (diabetic mRNAs interaction network widely exists in
retinopathy), or diabetic foot [26–28]. multiple pathophysiology processes. For
Interestingly, some studies proposed a β cell-­ instance, defects in ciRS-7-mediated “sponging
centric model to clarify that all persistent dysgly- events” lead to the dysregulation of ciRS-7-­
cemia shares a common feature: the damaged or miRNA-7-UBE2A circuit in neocortex and hip-
abnormal β cells [29, 30]. There were several pocampal CA1, which eventually results in
mechanisms that mediated the abnormal glucose sporadic Alzheimer’s disease [6]. In colorectal
metabolism in β cells, such as insulin resistance, cancer, hsa_circ_001569, acting as a positive
insulin signal dysregulation, inflammation dys- regulator of cell proliferation and invasion, is
regulation, and immunity dysregulation [29, 31]. identified as a sponge of miR-145 [41]. Circular
Due to the heterogeneity and complexity of dia- RNA CER affects cartilage-related extracellular
betic complications, it is needed to diagnose the matrix degradation by sponging miR-136 in
disease elaborately and accurately. Studies sug- chondrocyte [7]. In vivo and in vitro experiments
278 T. Wang et al.

demonstrate that circMTO1 sponges microRNA-9 [50]. In a cohort of 1112 patients with CAD, Cox
to modulate p21 expression and inhibit hepato- regression analyses suggest that miR-132, miR-­
cellular carcinoma progression [42]. It is reported 140-­3p, and miR-210 are correlated with cardio-
that circRNA_010567/miR-141/TGF-β1 axis vascular death events [51]. Kaplan-Meier analysis
regulates myocardial fibrosis in diabetic mice and Cox proportional hazards model reveal that
[43]. On the whole, circular RNAs, as newly dis- miR-425-5p is an independent prognostic factor
covered non-coding RNAs, are involved in mul- for cervical cancer [52].
tiple pathophysiological processes of diseases. Circular RNAs have the potential to be dis-
ease biomarkers due to the following character-
2.3.2  ircular RNAs and Pancreatic β
C istics: (1) circular RNAs are covalently closed
Cells RNAs which are resistant to nucleases; (2) circu-
Pancreatic β cells are the only source of insulin, lar RNAs are relatively abundant, stable, and
and the secreted insulin is essential for maintain- conserved in the blood [2]; (3) circular RNAs are
ing blood glucose homeostasis. Insulin secretion enriched and stable in exosomes [53]; (4) circu-
defection will lead to chronic hyperglycemia and lar RNAs can be detected not only in tumor tis-
then the development of diabetes [44]. The func- sues but also in the blood,cerebral spinal fluid,
tions of circular RNAs in β cells are studied by and saliva [54, 55]; and (s5) compared with lin-
several researchers. CiRS-7 and circHIPK3 are ear RNAs, the half-lives of circular RNAs are
reported to be involved in β-cell function regula- much longer [56]. Recent studies have demon-
tion and the development of diabetes [45]. strated that circular RNAs could be utilized as
Another study reveals that circular RNA Cdr1as potential biomarkers for the diagnosis or prog-
is able to improve insulin secretion by targeting nosis of diseases, such as cancers [57–60], coro-
miR-7, Pax6, and Myrip in mouse β cells. The nary artery disease [61], and central nerve system
results indicate that Cdr1as might be a potential disease [62]. In addition to the above studies, it
therapeutic target in diabetes [46]. is reported that 489 circular RNAs were found to
be differentially expressed in the peripheral
2.3.3  ircular RNAs as Potential
C blood of patients with T2DM. Moreover, hsa_
Biomarkers in Diabetes Mellitus circ_0054633 is reported to be a promising diag-
Biomarkers are defined as biological molecules nostic biomarker for prediabetes and T2DM
that are the hallmarks of normal or abnormal bio- [63]. In another study, real-time polymerase
logical process, or of healthy condition or disease chain reaction is used to explore the differen-
[47]. Circulating circular RNAs derived from tially expressed circular RNAs in the peripheral
blood can indicate the physiological changes of blood from diabetes mellitus patients and control
the whole body. Biomarkers function in the fol- individuals. They verify that the expression of
lowing four aspects: (1) identify people at risk of hsa-­circRNA11783-­2 correlates with T2DM
developing diseases, (2) diagnose diabetes or [64]. However, most patients are asymptomatic
other metabolic diseases, (3) predict the develop- in the early stages of T2DM, which makes it dif-
ment of complications, and (4) monitor the ficult to diagnose the disease. A convenient, spe-
response to treatments [48]. As we all know, the cific, and sensitive diagnostic method is urgently
dynamic interplay of circular RNAs, microR- needed to facilitate the early diagnosis of T2DM.
NAs, and long non-coding RNAs is needed to
regulate cellular homeostasis [49]. Numerous 2.3.4  ircular RNAs and Diabetes
C
studies focus on developing microRNAs and Mellitus-Related Complications
long non-coding RNAs as diagnosis or prognosis The potentials of circular RNAs as biomarkers in
biomarkers of diseases. For instance, circulating predicting diabetes mellitus complications
miR-203 associates with poor survival and remain in further investigations. Several studies
metastasis in patients with colorectal carcinoma are devoted to identify the relationships between
22 Circular RNAs in Metabolic Diseases 279

diabetes-related complications and circular 3  on-alcoholic Fatty Liver

N
RNAs. Disease (NAFLD)
Zhou et al. report that circRNA_010567 is
upregulated in diabetic mice myocardium. The 3.1 NAFLD Epidemiology
knockdown of circRNA_010567 suppresses
myocardium fibrosis. Further studies d­ emonstrate Non-alcoholic fatty liver disease (NAFLD) has
that circRNA_010567/miR-141/TGF-β1 axis become the most common cause for chronic liver
plays an important role in diabetic mice myocar- disease, cirrhosis, and hepatocellular carcinoma
dium [43]. Tang et al. discover that cir- worldwide [69]. Up to 20% of NAFLD have non-­
cRNA_000203 is upregulated in the diabetic alcoholic steatohepatitis (NASH) with active
mouse myocardium. Further studies demonstrate hepatic necrotizing inflammation and injury,
that circRNA_000203 can specifically sponge which lead to progressive hepatic fibrosis and
miR-26b-5p, and the overexpression of cir- complications. It is now believed that NAFLD is
cRNA_000203 eliminates the anti-fibrosis effects the representative liver metabolic syndrome (MS)
of miR-26b-5p [65]. and affects liver metabolic homeostasis [70].
Diabetic vascular complications are the main NAFLD increases liver-related morbidity, liver-­
cause of disability and mortality in diabetic related mortality, and the risk of other metabolic
patients [66]. These complications include dia- complications such as type 2 diabetes [71], obe-
betic cardiomyopathy, diabetic nephropathy, sity [72], and metabolic syndrome [73]. The
diabetic retinopathy, diabetic cardiomyopathy, global prevalence rate of NAFLD in different
and diabetic foot. Diabetic retinopathy (DR) is a regions is as follows: Middle East, 32%; the
frequent diabetic vascular complication. In a United States, 30%; South America, 30%; Asia,
recent study, Zhang et al.. use circular RNA 27%; Europe, 24%; and Africa, 13% [74–76].
microarrays to identify the differential expres- 18–33% of NAFLD cases are found to have type
sion profiles of diabetic retinas and nondiabetic 2 diabetes (T2DM), and 66–83% of NAFLD
retinas. They find that circ_0005015 is signifi- cases are identified with insulin resistance (IR).
cantly upregulated in the diabetic retina.
Functional analysis reveals that circ_0005015
regulates retinal endothelial cell proliferation, 3.2 NAFLD Pathogenesis
migration, and tube formation by sponging
miR-519d-3p [67]. Another study suggests that NAFLD affects many liver-related symptoms
circHIPK3 is a potential therapeutic target for such as asymptomatic steatosis, steatohepatitis,
alleviating diabetes-related retinopathy. In fibrosis, cirrhosis, and hepatocellular carcinoma
mechanism, gain-of-function and loss-of-func- [77]. The pathogenesis of NAFLD is still contro-
tion assays reveal that circHIPK3 affect diabetic versial: according to traditional double-hit the-
retinopathy by sponging miR-­30a-­3p and regu- ory, the first hit of NAFLD is the abnormal
lating downstream genes VEGFC, FZD4, and accumulation of triglycerides in liver cells. The
WNT2 [8]. In addition, researchers identify that second hit is that oxidative stress and inflamma-
circWDR77-miR-­ 124-FGF2 regulatory path- tory mediators promote liver steatosis to steato-
way plays a role in VSMC proliferation and hepatitis, fibrosis, and cirrhosis [75, 78].
migration. It provides the theoretical basis for However, recent findings suggest that “multiple
diabetes mellitus-related vasculopathy treat- parallel hits” hypothesis is more reasonable to
ment [68]. In conclusion, more efforts are NAFLD pathogenesis. Multiple pathogenic fac-
needed to study the functions of circular RNAs tors, such as insulin resistance, oxidative stress,
in diabetes-related complications. mitochondrial dysfunction, lipotoxicity, endo-
280 T. Wang et al.

plasmic reticulum stress, adipose tissue dysfunc- ous injection of liraglutide which is a
tion, altered innate immune regulation, and glucagon-like peptide-1 (GLP-1) analogue leads
cytokine secretion, contribute to liver injury [79]. to histological improvement of non-alcoholic ste-
Therefore, the pathogenesis of NAFLD is com- atohepatitis [89]. Given the complexity and het-
plex and heterogeneous. The underlying mecha- erogeneity of NAFLD, more efforts should be
nisms of NAFLD are still needed to be elucidated. expended on the drug development.
In addition, the early diagnosis of non-alcoholic
steatohepatitis may prevent disease progression
and improve patient outcomes [80]. 3.5 Circular RNAs and NAFLD

Insulin sensitizers, such as metformin and thia-

3.3 MicroRNAs and NAFLD zolidinedione are commonly used for those
NAFLD patients who are resistant to insulin.
Recently, researchers study the potential roles of Researchers detect the circular RNA expression
microRNAs in NAFLD pathogenesis. It has changes in metformin-treated or high-fat diet
recently been reported that microRNAs serve as mice livers. They find that the altered circular
key regulators in NAFLD. For instance, liver-­ RNAs by metformin treatment can modulate
specific miR-122 knockout mice display abnor- NAFLD-related pathways through interacting
mal triglyceride accumulation in the liver due to with downstream microRNAs [90]. NAFLD
the increased triglyceride synthesis and reduced starts with hepatic steatosis symptom and then
triglyceride secretion [81]. Compared with progresses to steatohepatitis. 5–20% steatosis
healthy people, NASH patients have elevated patients will develop non-alcoholic steatohepati-
expression of miR-122 [82]. MiR-34a regulates tis (NASH). 10–20% of NASH patients will
VLDL metabolism, fatty acid β oxidation, cho- develop higher-grade fibrosis, and less than 5%
lesterol synthesis, and liver injury by targeting of them will progress to full-blown cirrhosis.
several key transcription factors such as HNF4α, Therefore, numerous studies focus on uncovering
PPARα, Sirt1, and p53 [83, 84]. MiR-155 plays a the underlying mechanisms of NAFLD
protective role for NAFLD by targeting LXRα-­ progression.
SREBP-­1c pathway [85]. A recent study reports that 357 circular RNAs
are associated with hepatic steatosis. Further
study demonstrates that the circRNA_021412-­
3.4 NAFLD Treatment miR-­1972-LPIN1 signaling plays an important
role in liver metabolism [91]. Peroxisome
Non-alcoholic steatohepatitis is a more aggres- proliferator-­ activated receptors (PPARs) are
sive form of NAFLD and is also considered as a reported as fatty acid regulators which control
first hit to liver. Currently, no Food and Drug lipid metabolism and inflammation [92]. PPARα
Administration (FDA)-approved medicines are signaling pathway is widely involved in various
applied in the treatment of non-alcoholic steato- physiological processes, including glucose
hepatitis. NAFLD-related medicine development metabolism; lipid metabolism; inflammation;
focuses on the following fields: medicines that cell proliferation, differentiation, and apoptosis;
regulate nuclear transcription factors, medicines and aging [93]. It is reported that PPARα signal-
that target lipotoxicity and oxidative stress, and ing pathway is inhibited by PPAR1 in NAFLD
medicines that regulate cellular energy homeo- patients. They suggest that the restoration of
stasis [86, 87]. PPARα signaling is important for NAFLD treat-
In a double-blind clinical study, NGM282, an ment [94].
antidiabetic drug, is reported to reduce hepatic fat In HepG2 cells, circRNA_0046366 is identi-
contents in patients with non-alcoholic steato- fied as antagonist of miR-34a. The inhibition of
hepatitis [88]. In another clinical trial, subcutane- circRNA_0046366 to miR-34a restores the
22 Circular RNAs in Metabolic Diseases 281

expression of PPARα, which suppresses hepato- 4.2 Perspective

cytes steatosis. These findings suggest that
circRNA_0046366-miR-34a-PPARα pathway Circular RNAs, acting as microRNA sponges,
plays an important role in hepatocyte steatosis RBP sponges, or regulators of transcription, are
regulation [95]. Another study reveals that cir- involved in the progression of various diseases.
cRNA_0046367 abolishes the inhibitory effects The fact that circular RNAs are detected in saliva
of miR-34a on PPARα, which improves the [27], exosomes [83], and the blood [86] encour-
­transcriptional activation lipid metabolism asso- ages researchers to develop circular RNAs as
ciated genes and steatosis [96]. diagnostic biomarkers for metabolic diseases.
Further studies demonstrate that circular RNAs
biomarkers can also be detected in urine and
4 Conclusion and Perspective cerebrospinal fluid. Although a growing number
of studies focus on the functions of circular
4.1 Conclusion RNAs, several issues should be elucidated such
as the biosynthesis of circular RNAs, the metabo-
In brief, circular RNAs play important roles in lisms of circular RNAs, and the regulation of cir-
the regulation of metabolic diseases. In particu- cular RNAs to downstream genes.
lar, circular RNAs are involved in the develop- It is reported that circular RNAs are enriched in
ment of diabetes and are closely correlated with exosomes. Exosomes are filled with nucleic acids
the secretion of insulin by β cells. Moreover, cir- such as messenger RNAs, microRNAs, circular
cular RNAs take part in the development of RNAs, and long non-coding RNAs [53, 97, 98].
diabetes-­related complications such as diabetic They can be detected in serum and serve as media-
retinopathy. Circulating circular RNAs can be tors for cell or organ communication. The inter-
used as biomarkers for early diabetes plays between circular RNAs and exosomes are
assessment. manifested in the following aspects: (1) exosomes
On the other hand, circular RNAs are involved transport circular RNAs from one metabolic organ
in NAFLD pathogenesis. The current research to another, (2) circular RNAs in exosomes can be
mainly focuses on the regulation of circular a biomarker for some metabolic diseases, and (3)
RNAs to NASH. Circular RNAs regulate NASH exosomes may help to eliminate circular RNAs.
pathogenesis mainly through circular RNA/ Until now, the researches on circular RNAs still
microRNA/mRNA axis. At present, NAFLD is remain in laboratory studies. There is still a long
still lack of effective drug therapies. In-depth way to go before we apply circular RNAs as diag-
study of the molecular mechanisms in NAFLD is nostic and prognostic biomarkers.
highly needed. Nucleic acid treatments were first proposed 30
Circular RNAs are evolutionarily conserved years ago. RNA interference technology repre-
and resistant to nuclease digestion. The half-life sents a revolutionary breakthrough in the phar-
of circular RNAs is more than 48 h which makes maceutical field and has become a frontier area
it more stable than other non-coding RNAs. for drug development [99]. Nucleic acid drugs
Based on the above characteristics, circular have many advantages compared with traditional
RNAs are proposed to be the potential diagnos- medicines, such as nucleic acid drugs do not pro-
tic markers and therapeutic targets in various duce exogenous proteins and they are safer than
diseases. Nowadays, only several groups focus conventional drugs. To date, microRNAs are
on the studies of circular RNAs in metabolic used to target downstream genes in two forms,
diseases. The functions of circular RNAs in namely, microRNA antagonists and microRNA
metabolism disease pathogenesis remain to be mimics [100]. For hepatitis C treatment, miR-­
elucidated [6]. 122 antagonists have entered phase 1B clinical
282 T. Wang et al.

trials [5, 101]. And for treatment of patients with 9. Wang K, Long B, Liu F et al (2016) A circular RNA
protects the heart from pathological hypertrophy
advanced solid tumors, MRX34, a liposomal and heart failure by targeting miR-223. Eur Heart
miR-34a mimic, have entered phase I clinical tri- J 37(33):2602–2611
als [102]. Until now, no therapeutic circular 10. Rottiers V, Naar AM (2012) MicroRNAs in metabo-
RNAs have been produced or approved. However, lism and metabolic disorders. Nat Rev Mol Cell Biol
13(4):239–250
with the development of high-throughput 11. Perry RJ, Peng L, Barry NA et al (2016) Acetate medi-
sequencing, increasing circular RNAs will be ates a microbiome-brain-beta-cell axis to promote
discovered and the circular RNA mysteries will metabolic syndrome. Nature 534(7606):213–217
finally be uncovered. In the future, circular RNAs 12. Afdhal NH (2012) Management of nonalcoholic
fatty liver disease: a 60-year-old man with probable
will be used in clinical treatment. Based on their nonalcoholic fatty liver disease: weight reduction,
important biological functions, circular RNAs liver biopsy, or both? JAMA 308(6):608–616
are proposed to be the most promising field in 13. Fernandez-Hernando C, Ramirez CM, Goedeke
nucleic acid therapy. L et al (2013) MicroRNAs in metabolic disease.
Arterioscler Thromb Vasc Biol 33(2):178–185
14. Cho NH, Shaw JE, Karuranga S et al (2018) IDF dia-
Competing Financial Interests The authors declare no
betes Atlas: global estimates of diabetes prevalence
competing financial interests.
for 2017 and projections for 2045. Diabetes Res Clin
Pract 138:271–281
15. Worldwide trends in diabetes since 1980: a
pooled analysis of 751 population-based studies
with 4·4 million participants (2016). The Lancet
References 387(10027):1513–1530.
16. Benjamin EJ, Virani SS, Callaway CW et al (2018)
1. Sanger HL, Klotz G, Riesner D et al (1976) Heart disease and stroke statistics-2018 update:
Viroids are single-stranded covalently closed cir- a report from the American Heart Association.
cular RNA molecules existing as highly base-­ Circulation 137:e67–e492
paired rod-like structures. Proc Natl Acad Sci USA 17. Chatterjee S, Khunti K, Davies MJ (2017) Type 2
73(11):3852–3856 diabetes. The Lancet 389(10085):2239–2251
2. Jeck WR, Sorrentino JA, Wang K et al (2013) 18. Koekkoek PS, Kappelle LJ, van den Berg E et al
Circular RNAs are abundant, conserved, and associ- (2015) Cognitive function in patients with diabetes
ated with ALU repeats. RNA 19(2):141–157 mellitus: guidance for daily care. Lancet Neurol
3. Memczak S, Jens M, Elefsinioti A et al (2013) 14(3):329–340
Circular RNAs are a large class of animal RNAs 19. Collier A, Ghosh S, Hair M et al (2015) Impact of
with regulatory potency. Nature 495(7441):333–338 socioeconomic status and gender on glycaemic con-
4. Afonina ZA, Myasnikov AG, Shirokov VA et al trol, cardiovascular risk factors and diabetes compli-
(2014) Formation of circular polyribosomes on cations in type 1 and 2 diabetes: a population based
eukaryotic mRNA without cap-structure and analysis from a Scottish region. Diabetes Metab
poly(A)-tail: a cryo electron tomography study. 41(2):145–151
Nucleic Acids Res 42(14):9461–9469 20. Augustin K, Khabbush A, Williams S et al (2018)
5. Jopling CL, Schutz S, Sarnow P (2008) Position-­ Mechanisms of action for the medium-chain triglyc-
dependent function for a tandem microRNA miR-­ eride ketogenic diet in neurological and metabolic
122-­binding site located in the hepatitis C virus disorders. Lancet Neurol 17(1):84–93
RNA genome. Cell Host Microbe 4(1):77–85 21. Kaczorowska M, Ryterska K, Ossowski P et al
6. Zhao Y, Alexandrov PN, Jaber V et al (2016) (2016) Metabolic risk factors of coronary heart dis-
Deficiency in the ubiquitin conjugating enzyme ease in relation to anthropometric measures in nonal-
UBE2A in Alzheimer’s Disease (AD) is linked to coholic fatty liver disease patients following dietary
deficits in a natural circular miRNA-7 sponge (cir- intervention. Pomeranian J Life Sci 62(2):8–14
cRNA; ciRS-7). Genes (Basel) 7(12) 22. Organization WH Definition, Diagnosis and classifi-
7. Liu Q, Zhang X, Hu X et al (2016) Circular RNA cation of diabetes mellitus and its complications part
related to the chondrocyte ECM regulates MMP13 1: diagnosis and classification of diabetes mellitus
expression by functioning as a MiR-136 ‘Sponge’ in provisional report of a WHO consultation
human cartilage degradation. Sci Rep 6:22572 23. Ghaemi N, Hasanabadi H, Ashrafzadeh F et al
8. Shan K, Liu C, Liu BH et al (2017) Circular (2018) Peripheral neuropathy in children and ado-
noncoding RNA HIPK3 mediates retinal vascu- lescents with insulin-dependent diabetes Mellitus.
lar dysfunction in diabetes mellitus. Circulation Iran J Child Neurol 12(2):83–90
136(17):1629–1642
22 Circular RNAs in Metabolic Diseases 283

24. Packer M (2018) Heart failure: the most important, activity exacerbates neointimal hyperplasia in diabe-
preventable, and treatable cardiovascular complica- tes mellitus. Atherosclerosis 269:71–78
tion of type 2 diabetes. Diabetes Care 41(1):11–13 40. Long Y, Zhan Q, Yuan M et al (2017) The expression
25. Shanbhogue VV, Hansen S, Frost M et al (2017) of microRNA-223 and FAM5C in cerebral infarction
Bone disease in diabetes: another manifestation of patients with diabetes mellitus. Cardiovasc Toxicol
microvascular disease? Lancet Diabetes Endocrinol 17(1):42–48
5(10):827–838 41. Xie H, Ren X, Xin S et al (2016) Emerging roles
26. Kahn SE, Cooper ME, Del Prato S (2014) of circRNA_001569 targeting miR-145 in the prolif-
Pathophysiology and treatment of type 2 diabetes: eration and invasion of colorectal cancer. Oncotarget
perspectives on the past, present, and future. Lancet 7(18):26680–26691
383(9922):1068–1083 42. Han D, Li J, Wang H et al (2017) Circular RNA
27. Kobayashi M, Zochodne DW (2018) Diabetic circMTO1 acts as the sponge of microRNA-9 to
neuropathy and the sensory neuron: new aspects suppress hepatocellular carcinoma progression.
of pathogenesis and their treatment implications. Hepatology 66(4):1151–1164
J Diabetes Investig. 43. Zhou B, Yu JW (2017) A novel identified circular
28. Kanwar YS, Sun L, Xie P et al (2011) A glimpse RNA, circRNA_010567, promotes myocardial fibro-
of various pathogenetic mechanisms of diabetic sis via suppressing miR-141 by targeting TGF-beta1.
nephropathy. Annu Rev Pathol 6:395–423 Biochem Biophys Res Commun 487(4):769–775
29. Schwartz SS, Epstein S, Corkey BE et al (2016) 44. Heit JJ, Karnik SK, Kim SK (2006) Intrinsic regula-
The time is right for a new classification system tors of pancreatic beta-cell proliferation. Annu Rev
for diabetes: rationale and Implications of the beta-­ Cell Dev Biol 22:311–338
cell-­
centric classification schema. Diabetes Care 45. Stoll L, Sobel J, Rodriguez-Trejo A et al (2018)
39(2):179–186 Circular RNAs as novel regulators of beta-cell func-
30. Ashcroft FM, Rorsman P (2012) Diabetes mel- tions in normal and disease conditions. Mol Metab
litus and the beta cell: the last ten years. Cell 9:69–83
148(6):1160–1171 46. Xu H, Guo S, Li W et al (2015) The circular RNA
31. Schwartz SS, Epstein S, Corkey BE et al (2017) A Cdr1as, via miR-7 and its targets, regulates insulin
unified pathophysiological construct of diabetes transcription and secretion in islet cells. Sci Rep
and its complications. Trends Endocrinol Metab 5:12453
28(9):645–655 47. Henry NL, Hayes DF (2012) Cancer biomarkers.
32. Poy MN, Eliasson L, Krutzfeldt J et al (2004) A Mol Oncol 6(2):140–146
pancreatic islet-specific microRNA regulates insulin 48. Parrizas M, Novials A (2016) Circulating microR-
secretion. Nature 432(7014):226–230 NAs as biomarkers for metabolic disease. Best Pract
33. Lovis P, Gattesco S, Regazzi R (2008) Regulation Res Clin Endocrinol Metab 30(5):591–601
of the expression of components of the exocytotic 49. Guil S, Esteller M (2015) RNA-RNA interactions in
machinery of insulin-secreting cells by microRNAs. gene regulation: the coding and noncoding players.
Biol Chem 389(3):305–312 Trends Biochem Sci 40(5):248–256
34. Frost RJA, Olson EN (2011) Control of glu- 50. Hur K, Toiyama Y, Okugawa Y et al (2017)
cose homeostasis and insulin sensitivity by the Circulating microRNA-203 predicts prognosis
Let-7 family of microRNAs. Proc. Natl Acad Sci and metastasis in human colorectal cancer. Gut
108(52):21075–21080 66(4):654–665
35. Pullen TJ, da Silva Xavier G, Kelsey G et al (2011) 51. Karakas M, Schulte C, Appelbaum S et al (2017)
miR-29a and miR-29b contribute to pancreatic beta-­ Circulating microRNAs strongly predict cardiovas-
cell-­
specific silencing of monocarboxylate trans- cular death in patients with coronary artery disease-­
porter 1 (Mct1). Mol Cell Biol 31(15):3182–3194 results from the large AtheroGene study. Eur Heart
36. Trajkovski M, Hausser J, Soutschek J et al (2011) J 38(7):516–523
MicroRNAs 103 and 107 regulate insulin sensitivity. 52. Sun L, Jiang R, Li J et al (2017) MicoRNA-425-5p is
Nature 474(7353):649–653 a potential prognostic biomarker for cervical cancer.
37. Davalos A, Goedeke L, Smibert P et al (2011) Ann Clin Biochem 54(1):127–133
miR-33a/b contribute to the regulation of fatty acid 53. Li Y, Zheng Q, Bao C et al (2015) Circular RNA is
metabolism and insulin signaling. Proc Natl Acad enriched and stable in exosomes: a promising bio-
Sci USA 108(22):9232–9237 marker for cancer diagnosis. Cell Res 25(8):981–984
38. Zhang J, Zhang L, Fan R et al (2013) The polymor- 54. Bonizzato A, Gaffo E, Te Kronnie G et al (2016)
phism in the let-7 targeted region of the Lin28 gene CircRNAs in hematopoiesis and hematological
is associated with increased risk of type 2 diabetes malignancies. Blood Cancer J 6(10):e483
mellitus. Mol Cell Endocrinol 375(1–2):53–57 55. Bahn JH, Zhang Q, Li F et al (2015) The landscape
39. Lightell DJ Jr, Moss SC, Woods TC (2018) of microRNA, Piwi-interacting RNA, and circular
Upregulation of miR-221 and -222 in response to RNA in human saliva. Clin Chem 61(1):221–230
increased extracellular signal-regulated kinases 1/2 56. Enuka Y, Lauriola M, Feldman ME et al (2016)
Circular RNAs are long-lived and display only mini-
284 T. Wang et al.

mal early alterations in response to a growth factor. in atients with type 2 diabetes Mellitus and normal
Nucleic Acids Res 44(3):1370–1383 plasma aminotransferase levels. J Clin Endocrinol
57. Yang F, Liu D-Y, Guo J-T et al (2017) Circular Metab 100(6):2231–2238
RNA circ-LDLRAD3 as a biomarker in diagno- 72. Henao-Mejia J, Elinav E, Jin C et al (2012)
sis of pancreatic cancer. World J Gastroenterol Inflammasome-mediated dysbiosis regulates
23(47):8345–8354 progression of NAFLD and obesity. Nature
58. Chen S, Li T, Zhao Q et al (2017) Using circular 482(7384):179–185
RNA hsa_circ_0000190 as a new biomarker in 73. Adams LA, Waters OR, Knuiman MW et al (2009)
the diagnosis of gastric cancer. Clin Chim Acta NAFLD as a risk factor for the development of dia-
466:167–171 betes and the metabolic syndrome: an eleven-year
59. Xu L, Zhang M, Zheng X et al (2017) The circular follow-up study. Am J Gastroenterol 104(4):861–867
RNA ciRS-7 (Cdr1as) acts as a risk factor of hepatic 74. Pappachan JM, Babu S, Krishnan B et al (2017)
microvascular invasion in hepatocellular carcinoma. Non-alcoholic fatty liver disease: a clinical update.
J Cancer Res Clin Oncol 143(1):17–27 J Clin Transl Hepatol 5(4):384–393
60. Li P, Chen S, Chen H et al (2015) Using circular 75. Carr RM, Oranu A, Khungar V (2016) Nonalcoholic
RNA as a novel type of biomarker in the screening fatty liver disease: pathophysiology and manage-
of gastric cancer. Clin Chim Acta 444:132–136 ment. Gastroenterol Clin North Am 45(4):639–652
61. Zhao Z, Li X, Gao C et al (2017) Peripheral blood 76. Vernon G, Baranova A, Younossi ZM (2011)
circular RNA hsa_circ_0124644 can be used as a Systematic review: the epidemiology and natu-
diagnostic biomarker of coronary artery disease. Sci ral history of non-alcoholic fatty liver disease and
Rep 7:39918 non-alcoholic steatohepatitis in adults. Aliment
62. Westholm JO, Miura P, Olson S et al (2014) Pharmacol Ther 34(3):274–285
Genome-wide analysis of drosophila circular RNAs 77. Angulo P (2002) Nonalcoholic fatty liver disease. N
reveals their structural and sequence properties Engl J Med 346(16):1221–1231
and age-dependent neural accumulation. Cell Rep 78. Farrell GC, Larter CZ (2006) Nonalcoholic fatty
9(5):1966–1980 liver disease: from steatosis to cirrhosis. Hepatology
63. Zhao Z, Li X, Jian D et al (2017) Hsa_circ_0054633 in 43(2 Suppl 1):S99–s112
peripheral blood can be used as a diagnostic bio- 79. Birkenfeld AL, Shulman GI (2014) Nonalcoholic
marker of pre-diabetes and type 2 diabetes mellitus. fatty liver disease, hepatic insulin resistance, and
Acta Diabetol 54(3):237–245 type 2 diabetes. Hepatology 59(2):713–723
64. Li X, Zhao Z, Jian D et al (2017) Hsa-­ 80. Rinella ME (2015) Nonalcoholic fatty liver disease:
circRNA11783-­2 in peripheral blood is correlated a systematic review. JAMA 313(22):2263–2273
with coronary artery disease and type 2 diabetes 81. Hsu SH, Wang B, Kota J et al (2012) Essential
mellitus. Diab Vasc Dis Res 14(6):510–515 metabolic, anti-inflammatory, and anti-­tumorigenic
65. Tang CM, Zhang M, Huang L et al (2017) functions of miR-122 in liver. J Clin Invest
CircRNA_000203 enhances the expression of 122(8):2871–2883
fibrosis-­associated genes by derepressing targets of 82. Auguet T, Aragones G, Berlanga A et al (2016)
miR-26b-5p, Col1a2 and CTGF, in cardiac fibro- miR33a/miR33b* and miR122 as Possible
blasts. Sci Rep 7:40342 Contributors to Hepatic Lipid Metabolism in Obese
66. Sena CM, Pereira AM, Seica R (2013) Endothelial Women with Nonalcoholic Fatty Liver Disease. Int
dysfunction – a major mediator of diabetic vascular J Mol Sci 17(10)
disease. Biochim Biophys Acta 1832(12):2216–2231 83. Yamada H, Suzuki K, Ichino N et al (2013)
67. Zhang SJ, Chen X, Li CP et al (2017) Identification Associations between circulating microRNAs (miR-­
and characterization of circular RNAs as a new class 21, miR-34a, miR-122 and miR-451) and non-­
of putative biomarkers in diabetes retinopathy. Invest alcoholic fatty liver. Clin Chim Acta 424:99–103
Ophthalmol Vis Sci 58(14):6500–6509 84. Xu Y, Zalzala M, Xu J et al (2015) A metabolic
68. Chen J, Cui L, Yuan J et al (2017) Circular RNA stress-inducible miR-34a-HNF4alpha pathway regu-
WDR77 target FGF-2 to regulate vascular smooth lates lipid and lipoprotein metabolism. Nat Commun
muscle cells proliferation and migration by spong- 6:7466
ing miR-124. Biochem Biophys Res Commun 85. Miller AM, Gilchrist DS, Nijjar J et al (2013) MiR-­
494(1-2):126–132 155 has a protective role in the development of
69. Bhala N, Angulo P, van der Poorten D et al (2011) non-alcoholic hepatosteatosis in mice. PLoS One
The natural history of nonalcoholic fatty liver disease 8(8):e72324
with advanced fibrosis or cirrhosis: an international 86. Gastaldelli A (2017) Insulin resistance and reduced
collaborative study. Hepatology 54(4):1208–1216 metabolic flexibility: cause or consequence of
70. Vanni E, Bugianesi E, Kotronen A et al (2010) From NAFLD? Clin Sci (Lond) 131(22):2701–2704
the metabolic syndrome to NAFLD or vice versa? 87. Musso G, Cassader M, Gambino R (2016) Non-­
Dig Liver Dis 42(5):320–330 alcoholic steatohepatitis: emerging molecular tar-
71. Portillo-Sanchez P, Bril F, Maximos M et al (2015) gets and therapeutic strategies. Nat Rev Drug Discov
High prevalence of nonalcoholic fatty liver disease 15(4):249–274
22 Circular RNAs in Metabolic Diseases 285

88. Harrison SA, Rinella ME, Abdelmalek MF et al 95. Guo XY, Sun F, Chen JN et al (2018) cir-
(2018) NGM282 for treatment of non-alcoholic cRNA_0046366 inhibits hepatocellular steato-
steatohepatitis: a multicentre, randomised, double-­ sis by normalization of PPAR signaling. World
blind, placebo-controlled, phase 2 trial. Lancet J Gastroenterol 24(3):323–337
391(10126):1174–1185 96. Guo XY, Chen JN, Sun F et al (2017) cir-
89. Armstrong MJ, Gaunt P, Aithal GP et al (2016) cRNA_0046367 Prevents hepatoxicity of lipid
Liraglutide safety and efficacy in patients with non-­ peroxidation: an inhibitory role against hepatic ste-
alcoholic steatohepatitis (LEAN): a multicentre, atosis. Oxid Med Cell Longev 2017:3960197
double-blind, randomised, placebo-controlled phase 97. Zhao R, Zhou J, Dong X et al (2018) Circular RNA
2 study. Lancet 387(10019):679–690 expression alteration in exosomes from the brain
90. Guo J, Zhou Y, Cheng Y et al (2018) Metformin-­ extracellular space after traumatic brain injury in
induced changes of the coding transcriptome and mice. J Neurotrauma
non-coding RNAs in the livers of non-alcoholic 98. Lasda E, Parker R (2016) Circular RNAs co-­
fatty liver disease mice. Cell Physiol Biochem precipitate with extracellular vesicles: a possible
45(4):1487–1505 mechanism for circRNA clearance. PLoS One
91. Guo XY, He CX, Wang YQ et al (2017) Circular 11(2):e0148407
RNA profiling and bioinformatic modeling identify 99. Sridharan K, Gogtay NJ (2016) Therapeutic nucleic
its regulatory role in hepatic steatosis. Biomed Res acids: current clinical status. Br J Clin Pharmacol
Int 2017:5936171 82(3):659–672
92. Tanaka N, Aoyama T, Kimura S et al (2017) 100. Bouchie A (2013) First microRNA mimic enters
Targeting nuclear receptors for the treatment of fatty clinic. Nat Biotechnol 31(7):577
liver disease. Pharmacol Ther 179:142–157 101. van der Ree MH, de Vree JM, Stelma F et al (2017)
93. Samuel VT, Shulman GI (2018) Nonalcoholic fatty Safety, tolerability, and antiviral effect of RG-101 in
liver disease as a Nexus of metabolic and hepatic patients with chronic hepatitis C: a phase 1B,
diseases. Cell Metab 27(1):22–41 double-blind, randomised controlled trial. Lancet
94. Huang K, Du M, Tan X et al (2017) PARP1-mediated 389(10070):709–717
PPARalpha poly(ADP-ribosyl)ation suppresses fatty 102. Beg MS, Brenner AJ, Sachdev J et al (2017) Phase
acid oxidation in non-alcoholic fatty liver disease. I study of MRX34, a liposomal miR-34a mimic,
J Hepatol 66(5):962–977 administered twice weekly in patients with advanced
solid tumors. Invest New Drugs 35(2):180–188
 Circular RNAs in Vascular
Functions and Diseases 23
Shengguang Ding, Yujiao Zhu, Yajun Liang,
Haitao Huang, Yiming Xu, and Chongjun Zhong

Abstract tension, carotid atherosclerotic disease,

Vascular disease is one of the top five causes hepatic vascular invasion in hepatocellular
of death and affects a variety of other diseases, carcinoma, aortic aneurysm, coronary artery
such as heart, nervous system, and metabolic disease, and type 2 diabetes mellitus.
disorders. Vascular dysfunction is a hallmark
of ischemia, cancer, and inflammatory dis- Keywords
eases and can accelerate the progression of Circular RNAs · Vascular function · Vascular
diseases. Circular RNAs (circRNAs) are a diseases
new type of noncoding RNAs with covalent
bond ring structure, which have been reported
to be abnormally expressed in many human
diseases. circRNAs regulate gene expression 1 Introduction
through the sponging of microRNAs (miR-
NAs) and can also be used as disease biomark- Noncommunicable diseases (NCDs) cause a
ers. Here we will summarize the functions of great deal of disabilities and deaths in the world
circRNAs in vascular diseases, including vas- [1]. Among NCDs, vascular diseases rank on the
cular dysfunction, atherosclerosis, diabetes top of the list and are among the top five causes
mellitus-related retinal vascular dysfunction, of patient deaths, which affect many other dis-
chronic thromboembolic pulmonary hyper- eases such as the cardiac, nervous, and metabolic
diseases [2, 3]. Of the 57 million global deaths,
60% are caused by NCDs. About 17.5 million
Authors Shengguang Ding, Yujiao Zhu and Yajun Liang NCDs deaths are caused by cardiovascular dis-
have been equally contributed to this chapter.
eases, including 6.7 million cases of stroke [4].
S. Ding · H. Huang · Y. Xu · C. Zhong (*) Vascular disorders affect not only cardiac dis-
Department of Thoracic and Cardiovascular Surgery,
The Second Affiliated Hospital of Nantong eases but also nervous system disorders.
University, Nantong, China Hypertension and elevated serum lipids are clas-
Y. Zhu · Y. Liang sical risk factors for stroke, a disorder that affects
Cardiac Regeneration and Ageing Lab, Institute of approximately 15 million individuals every year
Cardiovascular Sciences, School of Life Science, and leads to an annual cost of 75 billion dollars in
Shanghai University, Shanghai, China the United States [5]. Other NCDs such as diabe-
Shanghai Key Laboratory of Bio-Energy Crops, tes mellitus (DM) can also lead to neurodegen-
School of Life Sciences, Shanghai University, erative diseases. Moreover, DM significantly
Shanghai, China

© Springer Nature Singapore Pte Ltd. 2018 287

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_23
288 S. Ding et al.

impacts vascular diseases, atherosclerosis, endo- to express in viruses, archaea, and animals [23].
thelial cell senescence, endothelial cell injury, circRNAs, formed by non-sequential back splic-
endothelial progenitor cell dysfunction, impaired ing of pre-messenger RNA (pre-mRNA) tran-
angiogenesis, and cardiovascular disease [6]. In scripts, have been widely studied in recent years
particular, 82% of NCDs deaths occur in low-­ [24, 25]. Unlike linear RNAs, circRNAs have 5′
income or middle-income countries [7]. caps and 3′ tails. Its 5′ caps and 3′ tails bond
Blood vessels are conduits for the transport of together covalently and form a ring. Recent
blood, which can be divided into arteries, veins, reports reveal that circRNAs can function as
and microvessels. Arteries carry blood from the competing endogenous RNAs or microRNA
heart to organs [8], while veins carry blood back sponges to regulate gene expression [26]. With
to the heart. Arteries and veins are connected by the development of high-throughput RNA
self-organization capillaries, which are the main sequencing technology, a large number of cir-
places for material exchanges [9, 10]. The aorta cRNAs are discovered. circRNAs are found to
walls are thick and contain rich elastic fibers, express endogenously in mammalian cells and be
which are expansive and elastic [11, 12]. When involved in multiple diseases, such as atheroscle-
the left ventricle shoots blood, the pressure in the rosis, neurological disorder, prion disease, and
artery rises, which pushes the blood forward in cancer [27].
the artery. After the aortic valve is closed, the
expansion of the aorta and artery is still able to
maintain the potential energy from the left ven- 2  ircular RNAs and Vascular
C
tricular systole, so they are called elastic reser- Diseases
voir vessels [13]. As the artery branches, the
walls of arteries thin and the elastic fibers are 2.1  he Roles of circRNAs
T
gradually reduced. Smooth muscles become the in Vascular Dysfunction
main components of the arteriole walls. The con-
traction and relaxation of smooth muscle endow Endothelium and smooth muscles are essential
resistance to blood flow [14]. The resistance to components for vascular function [28, 29].
blood flow mostly occurs in small arteries, espe- Vascular dysfunction is the classical symptom of
cially the microarteries, which are called resis- ischemia, cancer, and inflammatory diseases
tance vessels [15]. Meanwhile, the resistance of [30]. Vascular dysfunction is often associated
peripheral vascular to blood flow is called periph- with abnormal gene regulation and endothelial
eral resistance or capillary vascular resistance cell dysfunction [31]. Aberrant circRNA expres-
[16, 17]. sions have been detected in cardiovascular dis-
In all kinds of blood vessels, blood capillaries eases and cancers, which are usually accompanied
have the smallest caliber and the largest number. with vascular dysfunction [32].
The total cross-sectional area of blood capillaries The host gene, ZNF609, is an important mem-
is the largest, and the blood flow velocity is the ber of zinc finger protein family [33]. Zinc finger
slowest. The blood walls of capillaries consist of protein plays an important role in DNA identifi-
monolayer endothelial cells and basement mem- cation, RNA packaging, cell apoptosis regula-
brane, which create conditions for material tion, protein folding and assembly, and other
exchanges [18]. When capillaries converge into a biological processes [34]. In endothelium and
vein, the vessel wall has a complete smooth mus- smooth muscle, circular RNAs cZNF609 and
cle layer [19]. Compared with arteries, the veins ZNF609 derive from the same transcripts.
have larger diameter and thinner tube wall [20]. Researchers report that cZNF609 is one of the
Usually, veins are in quiescent state, and they top ten abundantly expressed circular RNAs in
contain 60–70% of the circulating blood [21, 22]. endothelial cells. CircBase data shows that the
In the early 1990s, circRNAs were first found sequence of cZNF609 is homologous between
in plants. Since then circRNAs have been found mouse and human genome. In vivo experiments
23 Circular RNAs in Vascular Functions and Diseases 289

demonstrate that cZNF609 silence decreases the controlling ribosomal RNA processing or inhibit-
loss of retinal vessels and pathologic angiogene- ing cell proliferation.
sis [35]. Mechanistically, cZNF609 acts as an
endogenous miR-615 sponge to sequester and
inhibit miR-615 activity, which leads to the 2.3  he Roles of circRNAs
T
increased expression of MEF2A [36]. In sum- in Diabetes Mellitus-Induced
mary, cZNF609/miR-615-5p/MEF2A regulatory Retinal Vascular Dysfunction
network plays an important role in vascular regu-
lation [37]. Diabetic retinal dysfunction is a common and
severe microvascular complication [47–49].
Hyperglycemia can cause retinal vascular dam-
2.2  he Roles of circRNAs
T age and blood–retinal barrier damage, which is
in Atherosclerosis the main cause of morbidity and mortality in dia-
betic patients [50]. The vascular endothelium is
Atherosclerotic lesions start with endothelial composed of a layer of endothelial cells. The fis-
dysfunction in arterial vascular system, and then sures between adjacent endothelial cells are
circulating monocytes and lipoprotein particles closely connected by the transmembrane protein
accumulate in the subendothelial space to block complex, such as occludins, claudins, and zonula
the blood flow [51]. Atherosclerosis is a chronic occludens [51, 52]. These complexes contribute
inflammatory disease and is the leading cause of to the paracellular barriers, such as the blood–
global vascular deaths [38–40]. The risk factors brain barrier and blood–retinal barrier [53].
of atherosclerosis include tobacco use, obesity, Based on the above information, endothelial cell
and hyperlipidemia. In addition, genetic factors barrier can be used as therapeutic targets for vas-
are proved to be associated with atherosclerosis cular permeability-associated diabetic retinopa-
[41]. The main symptoms of atherosclerosis thy [54].
include coronary heart disease (CHD), ischemic Homeodomain-interacting protein kinase 3
stroke, and peripheral artery disease [42, 43]. In (HIPK3) is highly expressed in heart and mus-
the United States, the prevalence of CHD in cle tissue and localized in nuclear speckles.
adults is estimated to be 6.2%, and the annual HIPK3, as a corepressor for homeodomain tran-
cost of CHD and strokes is approximately $317 scription factor, involves in cell cycle regulation
billion [44]. [55, 56]. circHIPK3 derives from the second
The INK4/ARF locus encodes three tumor exon of HIPK3 gene and has long introns on
suppressor genes, namely, p16INK4a, p15INK4b, and both sides, which include many complementary
ARF. In addition, a long noncoding RNA called Alu repeats [24]. circHIPK3 affects the activity,
antisense noncoding RNA in the INK4 locus proliferation, migration, and tube formation of
(ANRIL) is also encoded from INK4/ARF locus retinal endothelial cells [57]. circHIPK3 is
[45, 46]. INK4/ARF protein is indispensable in found to be upregulated in diabetic retinas [55].
the development of normal mammals, and it In vivo silence of circHIPK3 alleviates retinal
plays an important role in inhibiting abnormal vascular dysfunction symptoms, such as retinal
proliferation. circANRIL is reported to bind to acellular capillary decrease, vascular leakage,
pescadillo homologue 1 (PES1) which is an and inflammation [58]. Furthermore, circHIPK3
essential 60S-preribosomal assembly factor. The acts as an endogenous miR-30a-3p sponge to
binding of circANRIL to PES1 impairs sequester and inhibit miR-30a-3p activity,
exonuclease-­mediated pre-rRNA processing and which leads to the increased expression of vas-
ribosome biogenesis in vascular smooth muscle cular endothelial growth factor-­C, FZD4, and
cells [44]. In addition, circANRIL is reported to WNT2. Overall, circHIPK3 plays a role in dia-
induce p53 activation and promote cell apoptosis. betic retinopathy and is a potential target for
In sum, circANRIL protects atherosclerosis by diabetic retinopathy treatment.
290 S. Ding et al.

2.4  he Roles of Circular RNAs

T activates the downstream ErbB signaling path-
in Chronic Thromboembolic way, which leads to increased cell apoptosis. The
Pulmonary Hypertension hsa_circ_0022342/hsa-miR-940/CRKL/ErbB
signaling pathway is proposed to be an important
Chronic thromboembolic pulmonary hyperten- pathway in CTEPH development [79].
sion (CTEPH) is a rare but debilitating and life-­
threatening complication of acute pulmonary
embolism, which is caused by pulmonary arterial 2.5  he Roles of Circular RNAs
T
obstruction and progressive vascular remodeling in Carotid Atherosclerotic
[59–62]. The incidence of CTEPH is about 0.1– Disease
9.1% in the 2 years after acute pulmonary embo-
lism [63]. Risk factors for CTEPH include Early detection of acute ischemic stroke may
inflammatory bowel disease, splenectomy, and reduce morbidity and mortality in patients with
myeloproliferative disease. According to the advanced carotid atherosclerosis [80]. Currently,
recent ESC/ERS guidelines for pulmonary hyper- there are still no biomarkers for atherosclerotic
tension, CTEPH is divided into four types [64, plaque rupture and stroke [81]. In stable carotid
65]. CTEPH is a life-threatening disease, and if it atherosclerotic plaques, the fibrous cap is com-
is not treated appropriately, CTEPH will develop posed of vascular smooth muscle cells (VSMCs)
into refractory right ventricular failure [66]. and collagen-enriched matrix. Here, the fibrous
Patients with CTEPH exhibit a poor prognosis cap is a kind of artery inflammation, which is
unless they receive treatment at an early stage caused by intimal thickening. In vulnerable
[67]. Thus, the early diagnosis and treatment of plaques, fibrous cap is thin and is composed with
CTEPH are important. CTEPH is a pan-vascular decreased VSMCs and increased inflammatory
disease of the pulmonary arteries. The feature of cells [82]. As a result, identifying circular bio-
CTEPH is the nonlinear arterial pulmonary pres- markers for fibrous cap change provides a new
sure increase caused by major vessels obstruction strategy for carotid artery prediction.
[67–70]. CTEPH is also the complication of pul- Recently, it is reported that the expressions of
monary hypertension (PH), which is caused by miR-221 and miR-222 are reduced in acute isch-
pulmonary endarterectomy (PEA) [71–73]. emic cerebrovascular-related carotid plaques
Twenty years ago, circRNAs were considered [83]. miR-221 and miR-222 promote intimal
as scrambled exons, most of which were misread thickening through depressing the expression of
as splicing errors [74]. Until recently, circRNAs p27Kip1 and inhibiting VSMC cell cycle progres-
are gradually recognized as regulatory transcripts sion [83]. Memczak et al. report that there are
that derive from protein-coding exons [75]. cir- several circRNAs with miR-221 and miR-222
cRNAs are reported to function in CTEPH mainly binding sites. Among them, circRNA-284 is
by affecting ribonucleotide biosynthesis, the cel- demonstrated to be able to regulate the activities
lular response to stress, the response to DNA of miR-221 and miR-222, which indicate that
damage, and gene expression. Microarray analy- circRNA-­284 may function in cerebrovascular-­
sis is performed to identify the differential related carotid plaques [42].
expression of circRNAs between CTEPH patients
and control individuals [76]. Statistically, 351
circRNAs are abnormally expressed in CTEPH 2.6  he Roles of Circular RNAs
T
patients. Among these circRNAs, hsa_ in Hepatic Vascular Invasion
circ_0022342 is found to play an important role
in CTEPH by targeting hsa-miR-940 [77, 78]. Hepatocellular carcinoma (HCC) is the most
CDKN1A is reported as the downstream gene of common malignant tumor and is a major health
hsa-miR-940 in CTEPH. And then, CDKN1A problem worldwide. Surgical treatment, such as
23 Circular RNAs in Vascular Functions and Diseases 291

hepatectomy and transplantation, remains the 2.8  he Roles of Circular RNAs

T
most effective treatment for patients with early in Coronary Artery Disease
liver cancer [84]. However, due to the high inci- and Type 2 Diabetes Mellitus
dence of postoperative recurrence, the prognosis
of HCC treatment is still not satisfactory [85]. In 2015, the mortality rate of coronary artery dis-
Multiple studies have demonstrated that micro- ease (CAD) was 2.6% higher than that in 2013.
vascular infiltration (MVI) is the most important The death rate of cardiovascular disease (CVD)
prognostic factor for HCC recurrence and sur- rapidly increases with population aging [106,
vival [86, 87]. 107]. According to the International Diabetes
It is reported that cirs-7 can promote HCC cell Federation (IDF) data, there are nearly 410 mil-
proliferation [88–90]. In a further study, the lion diabetic patients (DM) worldwide, and about
expression of cirs-7 is investigated in 108 paired 46.5% of them have not been diagnosed [108,
HCC tissues. They find that the expression of 109]. A Chinese survey reports the clinical data
cirs-7 is correlated with the clinical pathological of 3513 CVD patients from 7 cities across the
parameters of HCC and the deterioration of the country. They find that 77% of CVD patients
disease [91]. Another study analyzes the relation- have hyperglycemia [110]. The high morbidity
ships between cirs-7 and MVI. Univariate and and mortality of CAD and T2DM have brought
multivariate analyses suggest that the elevated enormous social and economic burdens world-
cirs-7 expression is an independent risk factor for wide [109].
MVI in HCC [92–95]. Hsa-circRNA11783–2 is located in chromo-
some 18, and its parental gene is
ENST00000251081 [111, 112]. Mutations of the
2.7  he Roles of Circular RNAs
T parental gene are linked to cardiovascular dis-
in Aortic Aneurysm ease. A recent study reports that hsa-­
circRNA11783–2 is also correlated with CAD
Aortic aneurysm is one of the leading causes of and T2DM. Most circRNAs regulate gene expres-
cardiovascular deaths [96]. Aortic aneurysm is sion by acting as “miRNA sponges.” In this con-
divided into true aortic aneurysm and aortic pseu- text, hsa-circRNA11783–2 might take part in
doaneurysm [97]. In true aortic aneurysm, the disease regulation by targeting miR-608 or miR-­
pathological aorta dilation is greater than 50% 3907 [113–115]. Overall, the specific role of hsa-­
compared with normal blood vessel diameter. In circRNA11783–2 in CAD and T2DM requires
contrast, the pathological aorta dilation in aortic further exploration [116].
pseudoaneurysm is less than 50% [98–100].
Considerable progress has been made to improve
cardiovascular disease survival over the past few 2.9  he Roles of Circular RNAs
T
decades. However, the 5 year survival rate of car- in Essential Hypertension
diovascular disease remains below 35% [101–
103]. Circular RNAs have been reported to Essential hypertension (EH) is one of the most
function in cardiovascular diseases. However, common diseases in human cardiovascular sys-
their expression and functions in aortic aneurysm tem [116–118]. EH is a kind of global disease. In
remain elusive. Hsa-circ-000595 is reported to be China, it is estimated that a third of Chinese have
located on chromosome 14 and regulate the func- EH. Since the symptoms of EH are not obvious,
tion of miR-19a [104]. In addition, it is indicated it is also known as the silent killer [119]. Risk
that hsa-circ-000595 may take part in the regula- factors of EH mainly include age, gender, and
tion of aortic aneurysm by preventing cell apop- regional and socioeconomic conditions [120].
tosis [105]. Here, we will focus on the function of circular
RNAs in EH [121–123].
292 S. Ding et al.

Table 23.1 Circular RNA associated with vascular function and disease
Vascular diseases Vascular function Circular RNA Target genes References
Retinopathy of prematurity diabetic Apoptosis cZNF609 miR-615-5p, MEF2A [30]
retinopathy
Ischemia-reperfusion injury Apoptosis circRNAs Cdr1 miR-7a [137]
Atherosclerosis Apoptosis circANRIL PES1 [44]
Heart failure Proliferation circ-Foxo3 CDK2,p21 [138]
Retinal vascular dysfunction induced Proliferation circHIPK3 miR-30a-3p, FZD4, [55]
by diabetes mellitus WNT2
Chronic thromboembolic pulmonary Proliferation hsa_circ_0022342 hsa-miR-940, CRKL, [67]
hypertension ErbB
Carotid atherosclerotic disease Necrosis circR-284 miR-221,miR-222 [83]
Hepatic vascular invasion in Apoptosis ciRS-7 miR-7,PIK3CD, [93]
hepatocellular carcinoma p70S6K, mTOR
Aortic aneurysm Proliferation hsa-circ-000595 miR-19a [100]
Diabetes mellitus Migration circWDR77 miR-124, FGF-2 [139]
Hypertension Migration and hsa_circ_0037911 miR-637 [116]
invasion
Coronary artery disease and type 2 Proliferation hsa-­ miR-608 [140]
diabetes mellitus circRNA11783–2

Hsa_circ_0037911 is located in chromosome might help the early diagnosis of diseases [35,
16 and contains four exons [124]. According to 135, 136]. With the rapid development of RNA
the bioinformatic analysis, hsa_circ_0037911 is sequencing technology, more and more circular
proposed to be associated with EH. Evidence RNAs are identified. In future, the circular RNAs
demonstrates that hsa_circ_0037911 acts as miR-­ study will gain more and more attention. In the
637 sponge to regulate the pathogenesis of EH. In past decade, a number of circRNAs are reported
previous studies, miR-637 has been shown to to be associated with vascular diseases. However,
function as a tumor suppressor in hepatocellular there is still a long way to go before we uncover
carcinoma, breast cancer, and follicular thyroid the underlying mechanisms of circular RNAs in
cancer [125–128]. The upregulation of miR-637 vascular diseases (Table 23.1).
decreases the activity of Akt1 and inhibits the
migration and invasion of cancer cells [129, 130]. Competing Financial Interests The authors declare no
competing financial interests.

3 Prospectives
References
In this chapter, we have introduced the functions
1. Maiese K (2017) Harnessing the power of SIRT1
of circRNAs in multiple vascular-related dis- and non-coding RNAs in vascular disease. Curr
eases. Intervention of certain circRNAs might Neurovasc Res 14(1):82–88
improve certain vascular diseases [131, 132]. 2. Arokiasamy P, Uttamacharya KP et al (2017)
Vascular disease-related deaths have been Chronic noncommunicable diseases in 6 low- and
middle-income countries: findings from wave 1 of
increasing. Clarifying the relationships between the World Health Organization’s study on global
circular RNAs and vascular diseases will provide ageing and adult health (SAGE). Am J Epidemiol
new methods for vascular disease treatment 185(6):414–428
[133]. At present, the early diagnosis of vascular 3. Strazhesko I, Tkacheva O, Boytsov S et al (2015)
Association of insulin resistance, arterial stiffness
diseases is urgently needed [134]. It has been and telomere length in adults free of cardiovascular
reported that some circRNAs are specifically diseases. PLoS One 10(8):e0136676
expressed in certain vascular diseases. Identifying 4. Minino AM, Murphy SL (2012) Death in the United
circular RNAs as vascular disease biomarkers States, 2010. NCHS Data Brief 99:1–8
23 Circular RNAs in Vascular Functions and Diseases 293

5. Turtle JR (2000) The economic burden of insulin associated with end-stage cardiomyopathy. J Heart
resistance. Int J Clin Pract Suppl(113):23–28 Lung Transplant 12(5):804–807
6. Mels CM, Schutte AE, Schutte R et al (2013) The 20. Scherbakov N, Sandek A, Martens-Lobenhoffer
link between vascular deterioration and branched J et al (2012) Endothelial dysfunction of the periph-
chain amino acids in a population with high gly- eral vascular bed in the acute phase after ischemic
cated haemoglobin: the SABPA study. Amino Acids stroke. Cerebrovasc Dis 33(1):37–46
45(6):1405–1413 21. Song MG, Yang HS, Lee DH et al (2014) Mid-term
7. Yao Z, Luo J, Hu K et al (2017) ZKSCAN1 gene results in patients having tricuspidization of the
and its related circular RNA (circZKSCAN1) both quadricuspid aortic valve. J Cardiothorac Surg 9:29
inhibit hepatocellular carcinoma cell growth, migra- 22. Zeng ZY, Gui C, Li L et al (2016) Effects of
tion, and invasion but through different signaling percutaneous coronary intervention on serum
pathways. Mol Oncol 11(4):422–437 Angiopoietin-2 in patients with coronary heart dis-
8. Siddiqui R, Saleem S, Khan NA (2016) The effect ease. Chin Med J 129(6):631–635
of peptidic and non-peptidic proteasome inhibitors 23. Chen Y, Li C, Tan C et al (2017) Circular RNA
on the biological properties of Acanthamoeba cas- in human disease and their potential clinic sig-
tellanii belonging to the T4 genotype. Exp Parasitol nificance. Zhonghua Yi Xue Yi Chuan Xue Za Zhi
168:16–24 34(1):133–137
9. Christensen HM, Schou M, Goetze JP et al (2013) 24. Zheng Q, Bao C, Guo W et al (2016) Circular RNA
Body mass index in chronic heart failure: associa- profiling reveals an abundant circHIPK3 that regu-
tion with biomarkers of neurohormonal activation, lates cell growth by sponging multiple miRNAs. Nat
inflammation and endothelial dysfunction. BMC Commun 7:11215
Cardiovasc Disord 13:80 25. Militello G, Weirick T, John D et al (2017) Screening
10. Boeckel JN, Jae N, Heumuller AW et al (2015) and validation of lncRNAs and circRNAs as miRNA
Identification and characterization of hypoxia-­ sponges. Brief Bioinform 18(5):780–788
regulated endothelial circular RNA. Circ Res 26. Haque S, Harries LW (2017) Circular RNAs (cir-
117(10):884–890 cRNAs) in health and disease. Genes (Basel)
11. Health Quality O (2005) Biventricular pacing (car- 8(12):353
diac resynchronization therapy): an evidence-based 27. Qu S, Yang X, Li X et al (2015) Circular RNA:
analysis. Ont Health Technol Assess Ser 5(13):1–60 a new star of noncoding RNAs. Cancer Lett
12. Khafaji HA, Bener AB, Rizk NM et al (2012) 365(2):141–148
Elevated serum leptin levels in patients with acute 28. Mazzuca MQ, Wlodek ME, Dragomir NM et al
myocardial infarction; correlation with coronary (2010) Uteroplacental insufficiency programs
angiographic and echocardiographic findings. BMC regional vascular dysfunction and alters arte-
Res Notes 5:262 rial stiffness in female offspring. J Physiol 588(Pt
13. Kiuchi MG, Andrea BR, da Silva GR et al (2015) 11):1997–2010
Pulmonary artery ablation to treat pulmonary arterial 29. Maass PG, Glazar P, Memczak S et al (2017) A map
hypertension: a case report. J Med Case Rep 9:284 of human circular RNAs in clinically relevant tis-
14. Fang S, Guo H, Cheng Y et al (2018) circHECTD1 sues. J Mol Med (Berl) 95(11):1179–1189
promotes the silica-induced pulmonary endothelial-­ 30. Liu C, Yao MD, Li CP et al (2017) Silencing of cir-
mesenchymal transition via HECTD1. Cell Death cular RNA-ZNF609 ameliorates vascular endothe-
Dis 9(3):396 lial dysfunction. Theranostics 7(11):2863–2877
15. Koide T, Nakanishi A, Ito I et al (1975) 31. Puro DG, Kohmoto R, Fujita Y et al (2016)
Hemodynamic and clinical significances of atrial Bioelectric impact of pathological angiogenesis
fibrillation, pulmonary vascular resistance and left on vascular function. Proc Natl Acad Sci U S A
ventricular function in rheumatic mitral stenosis. Jpn 113(35):9934–9939
Heart J 16(2):122–142 32. Carqueja IM, Sousa J, Mansilha A (2018) Vascular
16. Lindsay DC, Holdright DR, Clarke D et al (1996) malformations: classification, diagnosis and treat-
Endothelial control of lower limb blood flow in ment. Int Angiol 37(2):127. https://doi.org/10.23736/
chronic heart failure. Heart 75(5):469–476 S0392-9590.18.03961-5
17. Pluta A, Strozecki P, Krintus M et al (2015) Left 33. Yang P, Qiu Z, Jiang Y et al (2016) Silencing of
ventricular remodeling and arterial remodeling in cZNF292 circular RNA suppresses human glioma
patients with chronic kidney disease stage 1-3. Ren tube formation via the Wnt/beta-catenin signaling
Fail 37(7):1105–1110 pathway. Oncotarget 7(39):63449–63455
18. Radauskaite G, Vaskelyte J, Jurkevicius R (2007) 34. Laity JH, Lee BM, Wright PE (2001) Zinc finger
Progression of mitral regurgitation following proteins: new insights into structural and functional
ischemic mitral valve repair. Medicina (Kaunas) diversity. Curr Opin Struct Biol 11(1):39–46
43(8):619–622 35. Ding HX, Lv Z, Yuan Y et al (2018) The expres-
19. Rees W, Schuler S, Hummel M et al (1993) Heart sion of circRNAs as a promising biomarker in
transplantation in patients with muscular dystrophy the diagnosis and prognosis of human cancers: a
294 S. Ding et al.

systematic review and meta-analysis. Oncotarget dysfunction in spontaneously diabetic Torii rats.
9(14):11824–11836 BMC Complement Altern Med 17(1):497
36. Peng L, Chen G, Zhu Z et al (2017) Circular RNA 51. Su W, Dai DZ, Liu HR et al (2007) Upregulated
ZNF609 functions as a competitive endogenous endothelin system in diabetic vascular dysfunction
RNA to regulate AKT3 expression by sponging and early retinopathy is reversed by CPU0213 and
miR-150-5p in Hirschsprung’s disease. Oncotarget total triterpene acids from Fructus Corni. Clin Exp
8(1):808–818 Pharmacol Physiol 34(12):1228–1233
37. Sena CM, Pereira AM, Seica R (2013) Endothelial 52. Wondemagegn AT, Bizuayehu HM, Abie DD et al
dysfunction - a major mediator of diabetic vascular (2017) Undiagnosed diabetes mellitus and related
disease. Biochim Biophys Acta 1832(12):2216–2231 factors in east Gojjam (NW Ethiopia) in 2016:
38. Burd CE, Jeck WR, Liu Y et al (2010) Expression a community-based study. J Public Health Res
of linear and novel circular forms of an INK4/ARF-­ 6(1):834
associated non-coding RNA correlates with athero- 53. Herbreteau D, Brunereau L, Cottier J et al (1997)
sclerosis risk. PLoS Genet 6(12):e1001233 Hemangiomas and superficial vascular malforma-
39. Kumar L, Shamsuzzama HR et al (2017) Circular tions of the head and neck. Classification, diagnosis,
RNAs: the emerging class of non-coding RNAs and treatment. J Neuroradiol 24(4):274–290
their potential role in human neurodegenerative dis- 54. Shan K, Li CP, Liu C et al (2017) RNCR3: a regu-
eases. Mol Neurobiol 54(9):7224–7234 lator of diabetes mellitus-related retinal microvas-
40. Li Y, Dong Y, Huang Z et al (2017) Computational cular dysfunction. Biochem Biophys Res Commun
identifying and characterizing circular RNAs and 482(4):777–783
their associated genes in hepatocellular carcinoma. 55. Shan K, Liu C, Liu BH et al (2017) Circular
PLoS One 12(3):e0174436 noncoding RNA HIPK3 mediates retinal vascu-
41. Zhao L, Varghese Z, Moorhead JF et al (2018) CD36 lar dysfunction in diabetes mellitus. Circulation
and lipid metabolism in the evolution of atheroscle- 136(17):1629–1642
rosis. Br Med Bull 126:101. https://doi.org/10.1093/ 56. Chen G, Shi Y, Liu M et al (2018) circHIPK3 regu-
bmb/ldy006 lates cell proliferation and migration by sponging
42. Maiese K (2016) Disease onset and aging in the miR-124 and regulating AQP3 expression in hepato-
world of circular RNAs. J Transl Sci 2(6):327–329 cellular carcinoma. Cell Death Dis 9(2):175
43. Traylor M, Makela KM, Kilarski LL et al (2014) 57. Li Y, Zheng F, Xiao X et al (2017) CircHIPK3
A novel MMP12 locus is associated with large sponges miR-558 to suppress heparanase expression
artery atherosclerotic stroke using a genome-wide in bladder cancer cells. EMBO Rep 18(9):1646–1659
age-at-onset informed approach. PLoS Genet 58. Tian F, Wang Y, Xiao Z et al (2017) Circular RNA
10(7):e1004469 CircHIPK3 promotes NCI-H1299 and NCI-H2170
44. Holdt LM, Stahringer A, Sass K et al (2016) Circular cell proliferation through miR-379 and its target
non-coding RNA ANRIL modulates ribosomal RNA IGF1. Zhongguo Fei Ai Za Zhi 20(7):459–467
maturation and atherosclerosis in humans. Nat 59. Pesavento R, Prandoni P (2018) Prevention and
Commun 7:12429 treatment of the chronic thromboembolic pulmo-
45. Li H, Li K, Lai W et al (2018) Comprehensive cir- nary hypertension. Thromb Res 164:150. https://doi.
cular RNA profiles in plasma reveals that circular org/10.1016/j.thromres.2018.02.149
RNAs can be used as novel biomarkers for systemic 60. Aldabbous L, Abdul-Salam V, McKinnon T et al
lupus erythematosus. Clin Chim Acta 480:17–25 (2016) Neutrophil extracellular traps promote angio-
46. Holdt LM, Teupser D (2012) Recent studies of the genesis: evidence from vascular pathology in pulmo-
human chromosome 9p21 locus, which is associ- nary hypertension. Arterioscler Thromb Vasc Biol
ated with atherosclerosis in human populations. 36(10):2078–2087
Arterioscler Thromb Vasc Biol 32(2):196–206 61. Bonderman D, Jakowitsch J, Redwan B et al (2008)
47. Yang LP, Sun HL, Wu LM et al (2009) Baicalein Role for staphylococci in misguided thrombus
reduces inflammatory process in a rodent model resolution of chronic thromboembolic pulmo-
of diabetic retinopathy. Invest Ophthalmol Vis Sci nary hypertension. Arterioscler Thromb Vasc Biol
50(5):2319–2327 28(4):678–684
48. Wei L, Yin Z, Yuan Y et al (2010) A PKC-beta 62. Chen Z, Nakajima T, Tanabe N et al (2010)
inhibitor treatment reverses cardiac microvascular Susceptibility to chronic thromboembolic pulmo-
barrier dysfunction in diabetic rats. Microvasc Res nary hypertension may be conferred by miR-759 via
80(1):158–165 its targeted interaction with polymorphic fibrinogen
49. Takeda M, Mori F, Yoshida A et al (2001) Constitutive alpha gene. Hum Genet 128(4):443–452
nitric oxide synthase is associated with retinal vascu- 63. Pesavento R, Prandoni P (2015) Prevention and
lar permeability in early diabetic rats. Diabetologia treatment of the post-thrombotic syndrome and of
44(8):1043–1050 the chronic thromboembolic pulmonary hyperten-
50. Kim J, Jo K, Kim CS et al (2017) Aster koraiensis sion. Expert Rev Cardiovasc Ther 13(2):193–207
extract prevents diabetes-induced retinal vascular 64. Gu S, Li G, Zhang X et al (2015) Aberrant
expression of long noncoding RNAs in chronic
23 Circular RNAs in Vascular Functions and Diseases 295

t­hromboembolic pulmonary hypertension. Mol Med 78. Ding D, Zhang Y, Yang R et al (2016) miR-940 sup-
Rep 11(4):2631–2643 presses tumor cell invasion and migration via regula-
65. Guo L, Yang Y, Liu J et al (2014) Differentially tion of CXCR2 in hepatocellular carcinoma. Biomed
expressed plasma microRNAs and the potential reg- Res Int 2016:7618342
ulatory function of let-7b in chronic thromboembolic 79. Zhang EY, Cristofanilli M, Robertson F et al (2013)
pulmonary hypertension. PLoS One 9(6):e101055 Genome wide proteomics of ERBB2 and EGFR and
66. Mercier O, Sage E, Izziki M et al (2010) Endothelin other oncogenic pathways in inflammatory breast
a receptor blockade improves regression of flow-­ cancer. J Proteome Res 12(6):2805–2817
induced pulmonary vasculopathy in piglets. J Thorac 80. Jovanovic I, Zivkovic M, Jovanovic J et al (2014)
Cardiovasc Surg 140(3):677–683 The co-inertia approach in identification of specific
67. Miao R, Wang Y, Wan J et al (2017) Microarray microRNA in early and advanced atherosclerosis
expression profile of circular RNAs in chronic plaque. Med Hypotheses 83(1):11–15
thromboembolic pulmonary hypertension. Medicine 81. Rootman J, Heran MK, Graeb DA (2014) Vascular
(Baltimore) 96(27):e7354 malformations of the orbit: classification and the
68. Miyata M, Ito M, Sasajima T et al (2001) Effect role of imaging in diagnosis and treatment strate-
of a serotonin receptor antagonist on interleukin-­ gies*. Ophthal Plast Reconstr Surg 30(2):91–104
6-­induced pulmonary hypertension in rats. Chest 82. Raitoharju E, Lyytikainen LP, Levula M et al (2011)
119(2):554–561 miR-21, miR-210, miR-34a, and miR-146a/b are
69. Ogawa A, Firth AL, Yao W et al (2009) Inhibition up-regulated in human atherosclerotic plaques
of mTOR attenuates store-operated Ca2+ entry in in the tampere vascular study. Atherosclerosis
cells from endarterectomized tissues of patients 219(1):211–217
with chronic thromboembolic pulmonary hyper- 83. Bazan HA, Hatfield SA, Brug A et al (2017) Carotid
tension. Am J Physiol Lung Cell Mol Physiol plaque rupture is accompanied by an increase in the
297(4):L666–L676 ratio of serum circR-284 to miR-221 levels. Circ
70. Wang L, Guo LJ, Liu J et al (2013) MicroRNA Cardiovasc Genet 10(4):e001720
expression profile of pulmonary artery smooth mus- 84. Werner JA, Dunne AA, Folz BJ et al (2001) Current
cle cells and the effect of let-7d in chronic throm- concepts in the classification, diagnosis and treat-
boembolic pulmonary hypertension. Pulm Circ ment of hemangiomas and vascular malformations
3(3):654–664 of the head and neck. Eur Arch Otorhinolaryngol
71. Wynants M, Vengethasamy L, Ronisz A et al 258(3):141–149
(2013) NF-kappaB pathway is involved in CRP-­ 85. Altendorf-Hofmann A, Scheele J (2002) Indication
induced effects on pulmonary arterial endothe- and surgical outcome in hepatocellular carcinoma
lial cells in chronic thromboembolic pulmonary with infiltration of blood vessels, bile ducts and
hypertension. Am J Physiol Lung Cell Mol Physiol lymph nodes. Kongressbd Dtsch Ges Chir Kongr
305(12):L934–L942 119:635–641
72. Yang M, Deng C, Wu D et al (2016) The role of 86. Yamashita Y, Shirabe K, Aishima S et al (2015)
mononuclear cell tissue factor and inflammatory Predictors of microvascular invasion in hepatocel-
cytokines in patients with chronic thromboembolic lular carcinoma. Dig Dis 33(5):655–660
pulmonary hypertension. J Thromb Thrombolysis 87. Aravindhan V, Mohan V, Arunkumar N et al (2015)
42(1):38–45 Chronic endotoxemia in subjects with type-1 diabe-
73. Zhang YM, Liu WJ, Shi YZ et al (2008) Expressions tes is seen much before the onset of microvascular
of urotensin II and its receptor in pulmonary arter- complications. PLoS One 10(9):e0137618
ies in rats with chronic thromboembolic pulmonary 88. Xu L, Zhang M, Zheng X et al (2017) The circular
hypertension. Zhonghua Jie He He Hu Xi Za Zhi RNA ciRS-7 (Cdr1as) acts as a risk factor of hepatic
31(1):37–41 microvascular invasion in hepatocellular carcinoma.
74. Qu S, Zhong Y, Shang R et al (2017) The emerging J Cancer Res Clin Oncol 143(1):17–27
landscape of circular RNA in life processes. RNA 89. Shi Z, Chen T, Yao Q et al (2017) The circular
Biol 14(8):992–999 RNA ciRS-7 promotes APP and BACE1 degrada-
75. Fu C, Lv R, Xu G et al (2017) Circular RNA pro- tion in an NF-kappaB-dependent manner. FEBS
file of infantile hemangioma by microarray analysis. J 284(7):1096–1109
PLoS One 12(11):e0187581 90. Su C, Han Y, Zhang H et al (2018) CiRS-7 targeting
76. Zhou Z, Du D, Chen A et al (2018) Circular RNA miR-7 modulates the progression of non-small cell
expression profile of articular chondrocytes in an lung cancer in a manner dependent on NF-kappaB
IL-1beta-induced mouse model of osteoarthritis. signalling. J Cell Mol Med 22:3097. https://doi.
Gene 644:20–26 org/10.1111/jcmm.13587
77. Bhajun R, Guyon L, Pitaval A et al (2015) A statis- 91. Barrett SP, Parker KR, Horn C et al (2017) ciRS-7
tically inferred microRNA network identifies breast exonic sequence is embedded in a long non-coding
cancer target miR-940 as an actin cytoskeleton regu- RNA locus. PLoS Genet 13(12):e1007114
lator. Sci Rep 5:8336 92. Geng HH, Li R, Su YM et al (2016) The circular
RNA Cdr1as promotes myocardial infarction by
296 S. Ding et al.

mediating the regulation of miR-7a on its target highlight arterial-wall-specific mechanisms. Nat
genes expression. PLoS One 11(3):e0151753 Genet 49(7):1113–1119
93. Pan H, Li T, Jiang Y et al (2018) Overexpression of 108. Li C, Zhao L, Jiang W et al (2018) Correct microar-
circular RNA ciRS-7 abrogates the tumor suppres- ray analysis approaches in ‘Hsa-circRNA11783-2 in
sive effect of miR-7 on gastric cancer via PTEN/ peripheral blood is correlated with coronary artery
PI3K/AKT signaling pathway. J Cell Biochem disease and type 2 diabetes mellitus’. Diab Vasc Dis
119(1):440–446 Res 15(1):92–93
94. Peng L, Yuan XQ, Li GC (2015) The emerging land- 109. Bunnag SC (2006) Implications of microcirculation-­
scape of circular RNA ciRS-7 in cancer (review). research based information on prevention and treat-
Oncol Rep 33(6):2669–2674 ment of diabetes mellitus type 2: a perspective. Clin
95. Xu H, Guo S, Li W et al (2015) The circular RNA Hemorheol Microcirc 34(1–2):43–50
Cdr1as, via miR-7 and its targets, regulates insulin 110. Zhao Z, Li X, Li M (2018) Response to letter to
transcription and secretion in islet cells. Sci Rep the editor entitled ‘Correct microarray analysis
5:12453 approaches in “Hsa-circRNA11783-2 in peripheral
96. Woo KI, Kim YD (2014) Re: “vascular malforma- blood is correlated with coronary artery disease
tions of the orbit: classification and the role of imag- and type 2 diabetes mellitus”’. Diab Vasc Dis Res
ing in diagnosis and treatment strategies”. Ophthal 15(2):169–170
Plast Reconstr Surg 30(6):532 111. Bhavadharini B, Mahalakshmi MM, Maheswari K
97. Guo X, Chang Q, Pei H et al (2017) Long non-­ et al (2016) Use of capillary blood glucose for screen-
coding RNA-mRNA correlation analysis reveals the ing for gestational diabetes mellitus in resource-­
potential role of HOTAIR in pathogenesis of spo- constrained settings. Acta Diabetol 53(1):91–97
radic thoracic aortic aneurysm. Eur J Vasc Endovasc 112. Li CY, Ma L, Yu B (2017) Circular RNA hsa_
Surg 54(3):303–314 circ_0003575 regulates oxLDL induced vascular
98. He Q, Tan J, Yu B et al (2015) Long noncoding RNA endothelial cells proliferation and angiogenesis.
HIF1A-AS1A reduces apoptosis of vascular smooth Biomed Pharmacother 95:1514–1519
muscle cells: implications for the pathogenesis 113. Dai ZM, Kang HF, Zhang WG et al (2016) The
of thoracoabdominal aorta aneurysm. Pharmazie associations of single nucleotide polymorphisms in
70(5):310–315 miR196a2, miR-499, and miR-608 with breast can-
99. Li Y, Yang N (2018) Microarray expression profile cer susceptibility: a STROBE-compliant observa-
analysis of long non-coding RNAs in thoracic aortic tional study. Medicine (Baltimore) 95(7):e2826
aneurysm. Kaohsiung J Med Sci 34(1):34–42 114. Hashemi M, Sanaei S, Rezaei M et al (2016) miR-­
100. Liu G, Huang Y, Lu X et al (2010) Identification and 608 rs4919510 C>G polymorphism decreased the
characteristics of microRNAs with altered expres- risk of breast cancer in an Iranian subpopulation.
sion patterns in a rat model of abdominal aortic Exp Oncol 38(1):57–59
aneurysms. Tohoku J Exp Med 222(3):187–193 115. Kupcinskas J, Bruzaite I, Juzenas S et al (2014) Lack
101. Iyer V, Rowbotham S, Biros E et al (2017) A system- of association between miR-27a, miR-146a, miR-­
atic review investigating the association of microR- 196a-­2, miR-492 and miR-608 gene polymorphisms
NAs with human abdominal aortic aneurysms. and colorectal cancer. Sci Rep 4:5993
Atherosclerosis 261:78–89 116. Bao X, Zheng S, Mao S et al (2018) A potential risk
102. Joviliano EE, Ribeiro MS, Tenorio EJR (2017) factor of essential hypertension in case-control study:
MicroRNAs and current concepts on the pathogene- circular RNA hsa_circ_0037911. Biochem Biophys
sis of abdominal aortic aneurysm. Braz J Cardiovasc Res Commun 498:789. https://doi.org/10.1016/j.
Surg 32(3):215–224 bbrc.2018.03.059
103. Li Y, Maegdefessel L (2017) Non-coding RNA con- 117. Wise IA, Charchar FJ (2016) Epigenetic modi-
tribution to thoracic and abdominal aortic aneurysm fications in essential hypertension. Int J Mol Sci
disease development and progression. Front Physiol 17(4):451
8:429 118. Zhu HY, Wang SW, Martin LJ et al (2009) The role
104. Zheng C, Niu H, Li M et al (2015) Cyclic RNA of mitochondrial genome in essential hypertension
hsacirc000595 regulates apoptosis of aortic smooth in a Chinese Han population. Eur J Hum Genet
muscle cells. Mol Med Rep 12(5):6656–6662 17(11):1501–1506
105. Hu ZY, Luo JF, Zhong SL et al (2012) MicroRNAs 119. Dangwal S, Schimmel K, Foinquinos A et al (2017)
expression in normal and dissected aortic tissue. Noncoding RNAs in heart failure. Handb Exp
Zhonghua Xin Xue Guan Bing Za Zhi 40(5):406–410 Pharmacol 243:423–445
106. Kolopp-Sarda MN, Miossec P (2018) Cryoglobulins: 120. Hoebeeck LI, Rietzschel ER, Langlois M et al
an update on detection, mechanisms and clinical (2011) The relationship between diet and subclini-
contribution. Autoimmun Rev 17:457. https://doi. cal atherosclerosis: results from the Asklepios study.
org/10.1016/j.autrev.2017.11.035 Eur J Clin Nutr 65(5):606–613
107. Howson JMM, Zhao W, Barnes DR et al (2017) 121. Marques FZ, Booth SA, Charchar FJ (2015) The
Fifteen new risk loci for coronary artery disease emerging role of non-coding RNA in essential
23 Circular RNAs in Vascular Functions and Diseases 297

hypertension and blood pressure regulation. J Hum modulating expression of Cyclin Dependent Kinase
Hypertens 29(8):459–467 6 (CDK6) in pulmonary smooth muscle cells. Med
122. Marques FZ, Charchar FJ (2015) microRNAs in Sci Monit 22:4066–4072
essential hypertension and blood pressure regula- 131. Salem ESB, Fan GC (2017) Pathological effects of
tion. Adv Exp Med Biol 888:215–235 exosomes in mediating diabetic cardiomyopathy.
123. Marques FZ, Morris BJ (2012) Neurogenic hyper- Adv Exp Med Biol 998:113–138
tension: revelations from genome-wide gene expres- 132. Wu HJ, Zhang CY, Zhang S et al (2016) Microarray
sion profiling. Curr Hypertens Rep 14(6):485–491 expression profile of circular RNAs in heart tissue of
124. Zou M, Huang C, Li X et al (2017) Circular RNA mice with myocardial infarction-induced heart fail-
expression profile and potential function of hsa_cir- ure. Cell Physiol Biochem 39(1):205–216
cRNA_101238 in human thoracic aortic dissection. 133. Xu T, Wu J, Han P et al (2017) Circular RNA expres-
Oncotarget 8(47):81825–81837 sion profiles and features in human tissues: a study
125. Stokowy T, Wojtas B, Fujarewicz K et al (2014) using RNA-seq data. BMC Genomics 18(Suppl
miRNAs with the potential to distinguish follicular 6):680
thyroid carcinomas from benign follicular thyroid 134. Muller S, Appel B (2017) In vitro circularization of
tumors: results of a meta-analysis. Horm Metab Res RNA. RNA Biol 14(8):1018–1027
46(3):171–180 135. Zhang HD, Jiang LH, Sun DW et al (2018) CircRNA:
126. Xu RL, He W, Tang J et al (2018) Primate-specific a novel type of biomarker for cancer. Breast Cancer
miRNA-637 inhibited tumorigenesis in human pan- 25(1):1–7
creatic ductal adenocarcinoma cells by suppressing 136. Xin Z, Ma Q, Ren S et al (2017) The understanding
Akt1 expression. Exp Cell Res 363:310. https://doi. of circular RNAs as special triggers in carcinogen-
org/10.1016/j.yexcr.2018.01.026 esis. Brief Funct Genomics 16(2):80–86
127. Zhang JF, He ML, Fu WM et al (2011) Primate-­ 137. Memczak S, Jens M, Elefsinioti A et al (2013)
specific microRNA-637 inhibits tumorigenesis in Circular RNAs are a large class of animal RNAs
hepatocellular carcinoma by disrupting signal trans- with regulatory potency. Nature 495(7441):333–338
ducer and activator of transcription 3 signaling. 138. Du WW, Yang W, Chen Y et al (2017) Foxo3 circular
Hepatology 54(6):2137–2148 RNA promotes cardiac senescence by modulating
128. Li YM, Liu XY (2016) Molecular mechanisms multiple factors associated with stress and senes-
underlying application of serum procalcitonin and cence responses. Eur Heart J 38(18):1402–1412
stool miR-637 in prognosis of acute ischemic stroke. 139. Chen J, Cui L, Yuan J et al (2017) Circular RNA
Am J Transl Res 8(10):4242–4249 WDR77 target FGF-2 to regulate vascular smooth
129. Que T, Song Y, Liu Z et al (2015) Decreased miRNA- muscle cells proliferation and migration by spong-
­637 is an unfavorable prognosis marker and promotes ing miR-124. Biochem Biophys Res Commun
glioma cell growth, migration and invasion via direct 494(1–2):126–132
targeting Akt1. Oncogene 34(38):4952–4963 140. Li X, Zhao Z, Jian D et al (2017) Hsa-­
130. Sang HY, Jin YL, Zhang WQ et al (2016) circRNA11783-­2 in peripheral blood is correlated
Downregulation of microRNA-637 increases risk with coronary artery disease and type 2 diabetes
of hypoxia-induced pulmonary hypertension by mellitus. Diab Vasc Dis Res 14(6):510–515
 Functional Role of Circular RNA
in Regenerative Medicine 24
Richard Y. Cao, Qiying Dai, Qing Li, and Jian Yang

Abstract 1 Introduction
Every year, millions of people around the
world suffer from different forms of tissue In recent years, chronic diseases like cardiovas-
trauma. Regenerative medicine refers to ther- cular diseases, cancer, diabetes, and chronic lung
apy that replaces the injured organ or cells. diseases account for more than 70% of patient
Stem cells are the frontiers and hotspots of deaths in the world [1]. Regenerative medicine
current regenerative medicine research. refers to therapies that aim at creating replace-
Circular RNAs (circRNAs) are essential for ments for the tissue or organ function loss due to
the early development of many species. It was trauma or diseases. Researches on regenerative
found that they could guide stem cell differen- medicine mainly focus on human cells, including
tiation through interacting with certain somatic cells, stem cells, and embryo-derived
microRNAs (miRNAs). Based on this con- cells. Unlike conventional medical technology,
cept, it is meaningful to look into how cir- regenerative medicine avoids the immune
cRNAs influence stem cells and its role in response by using patient’s own cells [2].
regenerative medicine. In this chapter we will At present, various approaches of regenerative
discuss the functional roles of circRNAs in the medicine have been found to repair damaged tis-
prevention, repair, or progression of chronic sues or cells [3]. It brings hope for many incur-
diseases, through the communication between able diseases [4]. Besides, the application of
stem cells. regenerative medicine opens up a new era of
health care and promotes radical innovations in
Keywords patient management. One of the approaches is
Circular RNA · Regenerative medicine · Stem the use of embryonic stem cells (ESCs) in regen-
cell erative medicine [5]. ESCs play an important role
in the tissue regeneration in various diseases due
to their ability of self-renewal [6]. They are
R. Y. Cao · Q. Li · J. Yang (*) widely studied in stem cell therapy and tissue
Zhongshan-Xuhui Hospital, Fudan University/ engineering [7].
Shanghai Clinical Research Center, Chinese Circular RNA (circRNA) is a group of RNA
Academy of Sciences, Shanghai, China
which has a covalently closed structure. Usually
Q. Dai circRNAs do not code proteins, but they have
MetroWest Medical Center, Framingham, MA, USA
been found to have significant regulatory func-
Department of Cardiology, First Affiliated Hospital tion in protein expressions. In addition, recent
of Nanjing Medical University, Nanjing, China

© Springer Nature Singapore Pte Ltd. 2018 299

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_24
300 R. Y. Cao et al.

studies have found that they were essential for the cell therapy has achieved great success in recent
early growth of various species. With its interac- years [17, 18].
tion with certain microRNAs, somatic cells could Stem cells have long been studied to explore
be transformed back to their pluripotent state disease pathophysiology, drug screening, and
[8–10]. circRNAs are prevalently expressed in safety evaluation [17]. The application of stem
undifferentiated human embryonic stem cells cells in clinical practice has been increasing,
(hESCs) [11]. circRNAs regulate gene transcrip- and the benefit it brings is tremendous [18]. The
tion mainly by acting as a “microRNA sponge.” advantage of stem cells or ancestral/precursor
MicroRNAs are another noncoding RNA group cells is that they can be quickly generated and
that regulates gene expression. ESCs differenti- have the potential of differentiation [19].
ate under the interaction of microRNAs and cir- Combining bioengineering techniques and new
cRNAs [12]. Because of the abundant number of biomaterials, it is possible to create human
noncoding RNAs and the complexity in their organ [20, 21].
interactions, few circRNAs have been well Hematopoietic stem cell transplantation has
studied. been successfully used in treating leukemia,
In this chapter, we will discuss the impact of lymphoma, and multiple myeloma [22]. It
circRNAs on ESCs and their use in regenerative reveals the next frontier of medical treatment,
medicine. with the aim of achieving structural or functional
repair [23]. Indeed, hematopoietic stem cell
transplant can also be applied in other diseases
2  he Status of Regenerative
T including but not limited to metabolic storage
Medicine diseases and extracellular matrix disorders like
epidermolysis bullosa [24].
Human life has been greatly extended with the Apart from saving life, regenerative medicine
improvement of medical care in the past could improve patients’ quality of life [25, 26].
decades. Along with this privilege, aging-asso- Based on the continuous medical and social
ciated disease and aging itself have become a needs of regenerative science, there is an urgent
problem [13]. Aging-associated disease need to design, implement, and validate feasible
includes atherosclerosis, cardiovascular dis- models [27]. Regenerative medicine will benefit
ease, cancer, arthritis, etc. Aging is also an to both patients and stakeholders.
independent risk factor for many diseases.
These health problems greatly affect patients’
life quality. Each year, about 100 million peo- 2.1  he Definition and Source
T
ple suffer from different forms of tissue and of Stem Cells
organ damage in the world [14].
Transplantation has been the only therapy for Stem cells are a class of cells with the ability of
permanent organ damage. However, the source of self-renewal and multi-differentiation. They
organs hardly ever met the demand. Not to men- include embryonic stem cells isolated from
tion the huge cost of the transplantation, the other in vivo embryos, pluripotent stem cells induced
huge concern is the immune rejection it brings. in vitro, and adult stem cells [28]. Stem cells are
Patients have to take immunosuppressant after also used as “seed” cells for screening and treat-
the transplantation, which further increases other ing genetic diseases and the construction of
complications such as infection [15, 16]. The organs in vitro. They are of great value in regen-
appearance of regenerative medicine has brought erative medicine [29]. Stem cells can be further
significant changes in treating aging-associated divided into embryonic stem cells (ES cells) and
diseases. As part of regenerative medicine, stem somatic stem cells, depending on the stage of
development [30]. Stem cells such as neural stem
24 Functional Role of Circular RNA in Regenerative Medicine 301

cells, adipose stem cells, epidermal stem cells, 2.2.2 Regeneration Therapy Using
and mesenchymal stem cells could be found in In Vitro Cultured Stem Cell
adult organs [31]. These stem cells referred to Lines
adult stem cells which are as effective in treating Despite the remarkable achievements in clinical
various diseases [32]. application of hematopoietic stem cells, other
In 1998, Professor Thomson [33] of the types of stem cells have not yet been well studied
University of Wisconsin successfully obtained due to their rarity and the difficulty in obtaining
human embryonic stem cell lines from human them [45].
fetal germ cells. It is the first study of human One resolution is to establish stem cell lines
embryonic stem cells and has been recognized as by in vitro culture techniques. The other way is to
a milestone in the study of stem cells. At the same induce differentiation on ESCs [46]. In order to
time, Professor John Gilhart of Johns Hopkins prevent transplant rejection, patient’s own stem
University cultivated the first human embryonic cells are often used [47]. Studies found that stem
germ cell line and established a human pluripo- cells could also be obtained from somatic cells.
tent stem cell line [34]. These achievements In 2006, Professor Shinya Yamanaka invented a
marked the initiation of regenerative medicine. In new way to reprogram differentiated cells back to
theory, ESCs could provide as a source of trans- ESCs by transferring a combination of four tran-
plants. However, immunologic rejection to dif- scription factors (Oct4, Sox2, Klf4, and c-Myc)
ferentiated ESCs still exists [35]. into these differentiated somatic cells. These
ESC-like cells are named as iPSC. The discovery
of iPSC makes it possible to produce large num-
2.2  he Frontier Development
T bers of individual-specific, well-differentiated
of Stem Cell and Regenerative stem cells [48].
Medicine
2.2.3  mall Molecule Compound-­
S
2.2.1  se Stem Cells Derived
U Induced Cell Reprogramming
from the Body to Regenerate Small molecule compounds play a key role in the
Treatment development of medicine [49]. The advanced
The most commonly used stem cells are hemato- biotechnology enables rapid progressing in cel-
poietic stem cells, mesenchymal stem cells, etc. lular studies [50]. There are three major issues in
[36]. Hematopoietic stem cells are a group of the reprogramming of somatic cells, pluripotent
adult stem cells which give rise to blood cells cell differentiation, transdifferentiation, and plu-
[37]. Researches on hematopoietic stem cells ripotency of embryonic stem cells.
have been greatly prospered by improvement of The invention of the technology to induce plu-
mouse models. The application of advanced ripotent stem cell enables humans to obtain
genetic techniques has established different mod- autologous cell and do organ transplantation. It is
els for various studies [38]. With the help of a milestone in the history of regenerative medi-
single-­
cell technology, scientists successfully cine [51]. Compared to genetic therapy, small
analyzed the gene expression changes during the molecule has the advantages of no genome inte-
transformation from hematopoietic stem cell pre- gration, simple operation, easy-to-controllable
cursors to hematopoietic stem cells [39]. dose, and able to be reversed [52]. Studies have
Hematopoietic stem cells are the earliest type shown that small molecules can influence cell
of stem cells that have been used for clinical fate through epigenetic modifications, signaling
treatment [40–42], such as bone marrow trans- pathways, and metabolic kinases. These lead to
plantation [43]. Years of practice has made hema- improved reprogramming efficiency and replace-
topoietic stem cell transplantation to be the most ment of relevant transcription factors [53]. The
mature type of regenerative medicine [44]. use of small molecules to regulate the signaling
pathway can also significantly affect the
302 R. Y. Cao et al.

r­esonance cell reprogramming process [54]. tion is delayed or failed, since the dead cardiac
Transdifferentiation of pluripotent cells into cells cells cannot be replaced by functional cells. For a
of certain lineages or cells of different lineages is long time, heart has been recognized as a nonre-
an important way to obtain functional cells generative organ. The emergence of regenerative
in vitro [55–59]. medicine makes it theoretically possible to keep
cardiac function from ischemic state. In a study,
bone marrow stem cells were injected directly
2.3  he Application of Stem Cell
T into the coronary arteries in six patients with
and Regenerative Medicine myocardial infarction. After 10 weeks, the infarct
size was reduced by nearly 1/3, and their cardiac
2.3.1  tem Cells and Diabetes
S function was improved as well [67]. The
Treatment REPAIR-AMI trial (reinfusion of enriched pro-
Diabetes is one of the most common debilitating genitor cells and infarcted remodeling in acute
diseases that threaten human health [60]. At pres- myocardial infarction) was the first randomized,
ent, about 150 million people worldwide suffer double-blind, clinical study that aimed to prove
from diabetes. Both type 1 diabetes and type 2 the efficacy of stem cell therapy in myocardial
diabetes result from absolute or relative defi- infarction [73].
ciency of insulin secretion. Insulin deficiency
will result in dysregulation of sugar, lipids, pro- 2.3.3  tem Cells and Nerve Injury
S
teins, and electrolytes [61]. Drug treatment and Diseases
long-term injection of exogenous insulin are the Like myocardiocytes, neuro cells are the other
main treatments for diabetes. However, these group of cells with extremely limited regenera-
methods do not cure the disease, nor do they pre- tive ability. Depending on the site of injury, neuro
vent the diabetic complications. Islet transplanta- cell damage could cause tremendous disability.
tion could potentially cure diabetes. In 1990, The most vulnerable side is the central nervous
Scharp et al. reported that the first human alloge- system. The repair of central nervous injury has
neic islet cell transplantation has been success- always been a huge challenge. Stem cells replace
fully used to treat type 1 diabetes [62]. So far, damaged cells to secrete neurotrophic factors that
about 1000 diabetic patients have been treated can promote regeneration, protect neurons, and
with islet cell transplantation. thus reduce secondary damage. The process of
Suberi et al. used human embryonic stem cells stem cell therapy includes forming bridge-guided
(Hes-H9) to produce cells with β-cell characteris- nerve regeneration in injured areas, digesting col-
tics [63]. Assady et al. [64] and Lumelsky et al. lagenous scars, removing cell debris, regulating
[65] reported that human ESCs could differenti- immune responses, and repairing spinal cord
ate into insulin-producing cells. Bone marrow nonnervous tissue. In theory, stem cell therapy is
mesenchymal cells are found to be effective in plausible in treatment of neuro-injury. However,
treating type 1 diabetic mice [66]. Using islet cell the human nervous system may be more compli-
transplantation is a promising treatment for dia- cated than we thought. How the implanted stem
betes, although it remains in its early stage with a cell will interact with surrounding cells remains
lot of unsolved problems. uncertain. Plus ethical challenges are daunting,
making large clinical trials very difficult.
2.3.2  tem Cells and Cardiomyocyte
S
Injury Treatment
Cardiovascular disease could induce ventricular 3 Discovery of circRNA
wall remodeling and thus decrease cardiac func-
tion. In acute setting, restoration of blood flow is circRNA was first discovered in the RNA virus in
the best way to prevent damage to the heart mus- the 1970s [68]. In 1990, scientists used two-­
cle. It became more complicated if revasculariza- dimensional gel electrophoresis and electron
24 Functional Role of Circular RNA in Regenerative Medicine 303

microscopy to observe the circRNA molecule in Zhang et al. [81] identified a group of cir-
a Saccharomyces fungus [69]. In the following cRNAs which are abundantly expressed in
decades, some circular RNA transcripts were nucleus and can guide gene transcription.
also found in other transcripts, such as the dele- Ci-ankyrin repeat domain 52 (ci-ankrd52) is one
tion in colorectal carcinoma gene (DCC) tran- of them. It is located near the transcription site
scripts [70] and human Ets-1 gene and can affect the elongation of the RNA poly-
(ETwenty-Six-1, Ets-1) transcript [71], sex-­ merase II complex. The positive regulator of the
determining region Y (SRY) transcript [72], cyto- complex exerts a cis-regulatory effect on its
chrome P450 2C24 gene transcript [73], circular maternal gene. Silent information regulator 7 (ci-­
INK4 gene block antisense noncoding RNA in sirt7) also has a similar function. The INK4/ARF
the INK4 locus (cANRIL) [74], etc. site-associated long-chain noncoding
In recent years, with the rapid development of RNAANRIL inhibits the transcription of the cod-
bioinformatic technology, great achievements ing gene INK4/ARF by binding to the PcG com-
have been made in exploring circular RNA func- plex. This site also encodes cANRIL which also
tion [75]. Salzman et al. [76] discovered a good has transcriptional regulatory function [74].
amount of circRNAs related to human gene Some circRNAs can bind to proteins. For
expression. Jeck et al. [77] detected up to 25,000 example, CDR1as and Sry can bind to the miRNA
circRNAs in human fibroblasts. Memczak et al. effector Argonaute (AGO) [86] and inhibit
[78] identified 1950 human circRNAs, 1903 miRNAs-­mRNA cleavage. CircRNA52 interacts
mouse circRNAs, and 724 nematode circRNAs with RNA polymerase II complexes and affects
from RNA sequencing data. Guo et al. [79] found gene transcription [87]. In addition, Bohjanen
7112 circRNAs in 39 human cell lines. The cir- et al. [88] designed a circRNA that specifically
cRNAs discovered so far can be divided into the binds to the transactivating regulatory protein
following three categories based on their origins (Tat), which inhibited the expression of human
and constituent sequences: exonic circRNAs [76, immunodeficiency virus type 1 (HIV-1) gene.
77, 80], intronic circular RNA (circRNA) [81], Most circRNAs are present in the cytoplasm
and retained-intron circRNA [82]. [76], suggesting that they can be loaded into the
ribosome and translated into peptides. Like many
linear mRNAs without a 5′ cap and 3′ poly (A)
4  he Main Function
T tail, circRNA lacks a valid structure to initiate
of circRNAs translation. Some circRNAs are able to code pro-
tein. Once an internal ribosome entry site (IRES)
Salmena et al. [83] proposed the famous com- is activated, circRNAs could be translatable. For
peting endogenous RNA (ceRNA) regulation example, the core of hepatitis D virus (HDV)
hypothesis in which the biological function of contains a negative single-stranded circRNA,
ceRNA is accomplished through miRNA which encodes the related protein HDV antigen
response element (MRE). ceRNAs have differ- (HDAg) and plays a role in disease progression.
ent numbers and types of MREs that can com- In human osteosarcoma cells, some circular
petitively bind miRNAs and reduce the RNAs are found to be able to code protein,
inhibitory effect of miRNAs. There are multiple although their translation efficiency is very low
miRNA complementary binding sites on the cir- [89]. In addition, Talhouarne et al. [90] found that
cRNA, which can absorb miRNAs, just like the oocyte nuclei contained a large number of cir-
sponges. On the other hand, circRNA lacks poly cRNAs, most of which are 1000 nt in length and
(A) tails and 5′ ends, which enables them to resistant to exonuclease RNase R. Their content
escape degradation [84]. As a result, even a is significantly higher than in the nucleus. During
small portion of circRNAs can inhibit a large the development of fertilized eggs, circRNAs can
number of miRNAs [85]. be transmitted to offspring, indicating that they
304 R. Y. Cao et al.

play an important role in RNA-mediated genetic sponding linear mRNA. Among the differentially
inheritance. expressed 61 circRNAs, 11 were specifically
As more and more ribosomal data is available, expressed in hESCs [96]. The authors also vali-
whether circRNA can be translated in other cell dated that in the induced pluripotent stem cell
types or species is worthy of further study [91]. (iPSC) system, all 11 circRNAs were enriched in
iPSCs [97]. The expression of circRNA is often
accompanied by the expression of the corre-
5 The Function of circRNAs sponding linear gene [96]. After the comparison
in Stem Cell Pluripotency analysis, the authors found that the expression
and Reprogramming pattern of the three circRNAs is not consistent
with their corresponding linear genes. These
circRNAs are a new class of noncoding RNAs, three circRNAs are circBIRC6, circMAN1A2,
which have been studied in the recent 20 years. and circILKAP. Further, the authors analyzed the
They are covalently closed, single-stranded, and correlation between the expression of circRNAs
found to be prevalent in eukaryotes [71]. and the pluripotent state of stem cells using
The development of high-throughput sequenc- Northern blot. The results indicated that there
ing technology facilitated the discovery of many was a strong correlation between the expression
different circRNAs. However, few studies have of three circRNAs, namely, circBIRC6, circ-
reported their functions in somatic reprogram- CORO1C, and circMAN1A2, and the pluripotent
ming or embryonic stem cells. Like other non- state.
coding RNAs, circRNAs work by regulating The above studies have demonstrated the cor-
gene expression. In addition, circRNAs also relation between the expression pattern of the
affect protein transcription by “sponging” certain three circRNAs and the pluripotency state of
microRNAs. It was found that circBIRC6 is cells. However, it remains unknown whether
involved in regulation of stem cell pluripotency these circRNAs are involved in the regulation of
[92]. The study reported that a specific cleavage stem cell pluripotency [98]. Therefore, an RNA
factor, ESRP1, regulated the formation of cir- interference experiment was designed. The
cBIRC6. CircBIRC6 binds to miR-34a and miR-­ results showed that the positive rate of AP stain-
145 and regulates stem cell pluripotency [93, 94]. ing was significantly reduced after interfering
ESRP1 is regulated by transcription factors Oct4 with circBIRC6 and circCORO1C, but the effect
and Nanog. This is the first study that reveals the after interfering with circMAN1A2 was not obvi-
functions of circRNAs in stem cell pluripotency. ous. Differentiated status-related transcription
factors and marker genes showed similar results.

5.1 How to Screen

for Pluripotency Related 5.2 c ircRNA, Not Linear RNA, Is
to circRNAs Involved in the Regulation
of Stem Cell Pluripotency
Stem cell pluripotency-related circRNAs are
selected by high-throughput RNA sequencing. circRNA is formed by the processing precursors
Differentially expressed circRNAs were ana- of mRNA [99]. In order to explore whether these
lyzed the during ESCs differentiation. precursors are involved, researchers designed a
Pluripotency-related circRNAs are recorded in type of shRNAs that specifically targets linear
the database published in 2013 [78, 95]. RNAs and common exonic regions. Targeting the
In another study, researchers compared the corresponding exon regions can simultaneously
expression changes of circRNAs before and after interfere with linear RNAs and circRNAs inde-
hESCs differentiated into embryoid bodies (EB pendently. The AP staining results showed that
spheres). It also measured the changes of corre- the linear RNAs interfere with BIRC6 and
24 Functional Role of Circular RNA in Regenerative Medicine 305

CORO1C alone did not affect the pluripotent References

state of the cells. Q-PCR and immunofluores-
cence assays were performed to analyze pluripo- 1. Mozaffarian D, Benjamin EJ, Go AS et al (2015)
Heart disease and stroke statistics–2015 update:
tency and differentiation-related genes. The a report from the American Heart Association.
results showed that the circRNAs are involved in Circulation 131(4):e29–e322
the regulation of cellular pluripotency, while lin- 2. Mason C, Dunnill P (2008) A brief definition of
ear mRNAs did not directly affect the cellular regenerative medicine. Regen Med 3(1):1–5
3. Keshtkar S, Azarpira N, Ghahremani MH (2018)
pluripotency state. Mesenchymal stem cell-derived extracellular vesi-
Interfering with circBIRC6 and circCORO1C cles: novel frontiers in regenerative medicine. Stem
can affect the pluripotent state of stem cells. Cell Res Ther 9(1):63
Overexpression of them can promote the transi- 4. Daley GQ, Scadden DT (2008) Prospects for stem
cell-based therapy. Cell 132(4):544–548
tion to a pluripotent state [100]. Reporter gene 5. Arjmand B, Goodarzi P, Mohamadi-Jahani F et al
experiments showed that the expression of these (2017) Personalized regenerative medicine. Acta
circRNAs alone is not sufficient to induce cell Med Iran 55(3):144–149
reprogramming. In contrast, in the classical four-­ 6. Dushnik-Levinson M, Benvenisty N (1995)
Embryogenesis in vitro: study of differentiation of
factor (Oct4, Sox2, Klf4, and Myc, OSKM)- embryonic stem cells. Biol Neonate 67(2):77–83
induced iPSC system, overexpression of 7. Fox IJ, Daley GQ, Goldman SA et al (2014) Stem
circBIRC6 and circCORO1C can significantly cell therapy. Use of differentiated pluripotent stem
promote cell reprogramming. Based on prelimi- cells as replacement therapy for treating disease.
Science 345(6199):1247391
nary screening and functional studies, the authors 8. Li F, Zhang L, Li W et al (2015) Circular RNA
claim that circBIRC6 and circCORO1C are ITCH has inhibitory effect on ESCC by suppress-
involved in the regulation of stem cell pluripo- ing the Wnt/beta-catenin pathway. Oncotarget
tency and overexpression of them can promote 6(8):6001–6013
9. Yang W, Du WW, Li X et al (2016) Foxo3 activity
reprogramming of somatic cells. promoted by non-coding effects of circular RNA and
Foxo3 pseudogene in the inhibition of tumor growth
and angiogenesis. Oncogene 35(30):3919–3931
6 Conclusions 10. Houbaviy HB, Murray MF, Sharp PA (2003)
Embryonic stem cell-specific MicroRNAs. Dev Cell
5(2):351–358
Different from linear RNAs, circRNAs are cova- 11. Itskovitz-Eldor J, Schuldiner M, Karsenti D et al
lently closed RNAs that regulate gene expres- (2000) Differentiation of human embryonic stem
sion. circRNAs are characterized by high cells into embryoid bodies compromising the three
embryonic germ layers. Mol Med 6(2):88–95
stability, high conservation, and tissue specificity. 12. Zheng Q, Bao C, Guo W et al (2016) Circular RNA
They can interact with miRNAs by sponging profiling reveals an abundant circHIPK3 that regu-
them and form a complicated complex to regulate lates cell growth by sponging multiple miRNAs. Nat
gene transcription. Rapid development of high-­ Commun 7:11215
13. Spector M (2018) Biomedical materials to meet the
throughput sequencing technology and bioinfor- challenges of the aging epidemic. Biomed Mater
matics provides us with great tools for the further 13(3):030201
study of circRNAs. More and more circRNAs 14. McMahon DS, Thorsteinsdottir H, Singer PA et al
have been discovered. We have already made (2010) Cultivating regenerative medicine innovation
in China. Regen Med 5(1):35–44
great progress in associating them with disease 15. Mironov V, Visconti RP, Markwald RR (2004) What
and exploring their clinical significance. Some of is regenerative medicine? Emergence of applied
them have already been put into clinical trials. stem cell and developmental biology. Expert Opin
However, due to the unclear circRNA biogenesis Biol Ther 4(6):773–781
16. Kemp P (2006) History of regenerative medicine:
and their complicated interactions with other looking backwards to move forwards. Regen Med
molecules, more studies need to be done to 1(5):653–669
explore this huge but fascinating network. 17. Buganim Y (2016) Back to basics: refined nuclear
reprogramming techniques yield higher-quality stem
cells. Science 352(6292):1401
Competing Financial Interests The authors declare no
competing financial interests.
306 R. Y. Cao et al.

18. Niwa H (2018) The principles that govern tran- germ line cell-derived pluripotent stem cells in vitro.
scription factor network functions in stem cells. Stem Cells Dev 19(8):1183–1194
Development 145(6):dev157420 36. Fagoonee S, Pellicano R, Altruda F (2010) Stem
19. Yu J, Vodyanik MA, Smuga-Otto K et al (2007) cell world 2010: crucial questions from clinicians to
Induced pluripotent stem cell lines derived from biologists. Panminerva Med 52(2):95–96
human somatic cells. Science 318(5858):1917–1920 37. Butler AE, Huang A, Rao PN et al (2007)
20. Thrall JH (2004) Personalized medicine. Radiology Hematopoietic stem cells derived from adult donors
231(3):613–616 are not a source of pancreatic beta-cells in adult non-
21. Murphy MJ, Shahriari N, Payette M et al (2016) diabetic humans. Diabetes 56(7):1810–1816
Development of a curriculum in molecular diagnos- 38. Alonso-Ferrero ME, van Til NP, Bartolovic K et al
tics, genomics and personalized medicine for derma- (2018) Enhancement of mouse hematopoietic stem/
tology trainees. J Cutan Pathol 43(10):858–865 progenitor cell function via transient gene delivery
22. Terzic A, Nelson TJ (2010) Regenerative medicine using integration-deficient lentiviral vectors. Exp
advancing health care 2020. J Am Coll Cardiol Hematol 57:21–29
55(20):2254–2257 39. Wang Y, Tian H, Cai W et al (2018) Tracking hema-
23. Das AK, Pal R (2010) Induced pluripotent stem cells topoietic precursor division ex vivo in real time.
(iPSCs): the emergence of a new champion in stem Stem Cell Res Ther 9(1):16
cell technology-driven biomedical applications. 40. Kemp CB, Knight MJ, Scharp DW et al (1973)
J Tissue Eng Regen Med 4(6):413–421 Transplantation of isolated pancreatic islets into the
24. Terzic A, Nelson TJ (2013) Regenerative medicine portal vein of diabetic rats. Nature 244(5416):447
primer. Mayo Clin Proc 88(7):766–775 41. Ricordi C, Lacy PE, Finke EH et al (1988)
25. Kordelas L, da Silva Nardi F, Wagner B, et al (2018) Automated method for isolation of human pancre-
Elevated soluble human leukocyte antigen G levels atic islets. Diabetes 37(4):413–420
in patients after allogeneic stem cell transplantation 42. Ryan EA, Paty BW, Senior PA et al (2005) Five-­
are associated with less severe acute and chronic year follow-up after clinical islet transplantation.
graft-versus-host disease. Bone Marrow Transplant: Diabetes 54(7):2060–2069
https://doi.org/10.1038/s41409-018-0145-1 43. Lacotte S, Berney T, Shapiro AJ et al (2011) Immune
26. Safinia N, Grageda N, Scotta C et al (2018) Cell monitoring of pancreatic islet graft: towards a better
therapy in organ transplantation: our experience on understanding, detection and treatment of harmful
the clinical translation of regulatory T cells. Front events. Expert Opin Biol Ther 11(1):55–66
Immunol 9:354 44. Copelan EA (2006) Hematopoietic stem-cell trans-
27. Lo Monaco M, Merckx G, Ratajczak J et al (2018) plantation. N Engl J Med 354(17):1813–1826
Stem cells for cartilage repair: preclinical studies 45. Corbett MS, Webster A, Hawkins R et al (2017)
and insights in translational animal models and out- Innovative regenerative medicines in the EU: a bet-
come measures. Stem Cells Int 2018:9079538 ter future in evidence? BMC Med 15(1):49
28. Esteban MA, Wang T, Qin B et al (2010) Vitamin 46. Nishiga M, Guo H, Wu JC (2018) Induced pluripo-
C enhances the generation of mouse and human tent stem cells as a biopharmaceutical factory for
induced pluripotent stem cells. Cell Stem Cell extracellular vesicles. Eur Heart J 39:1848. https://
6(1):71–79 doi.org/10.1093/eurheartj/ehy097
29. Onteniente B (2013) The multiple aspects of stroke 47. Giacalone JC, Sharma TP, Burnight ER et al (2018)
and stem cell therapy. Curr Mol Med 13(5):821–831 CRISPR-Cas9-based genome editing of human
30. Zhang M, Ngo J, Pirozzi F et al (2018) Highly effi- induced pluripotent stem cells. Curr Protoc Stem
cient methods to obtain homogeneous dorsal neural Cell Biol 44:5B 7 1–5B 7 22
progenitor cells from human and mouse embryonic 48. Moldovan L, Barnard A, Gil CH et al (2017)
stem cells and induced pluripotent stem cells. Stem iPSC-derived vascular cell spheroids as building
Cell Res Ther 9(1):67 blocks for scaffold-free biofabrication. Biotechnol
31. Evans MJ, Kaufman MH (1981) Establishment in J 12(12):1700444
culture of pluripotential cells from mouse embryos. 49. Sangeetha KN, Sujatha S, Muthusamy VS et al
Nature 292(5819):154–156 (2017) Current trends in small molecule discov-
32. Alison MR, Islam S (2009) Attributes of adult stem ery targeting key cellular signaling events towards
cells. J Pathol 217(2):144–160 the combined management of diabetes and obesity.
33. Thomson JA, Itskovitz-Eldor J, Shapiro SS et al Bioinformation 13(12):394–399
(1998) Embryonic stem cell lines derived from 50. Wolpaw AJ, Stockwell BR (2014) Multidimensional
human blastocysts. Science 282(5391):1145–1147 profiling in the investigation of small-molecule-­
34. Shamblott MJ, Axelman J, Wang S et al (1998) induced cell death. Methods Enzymol 545:265–302
Derivation of pluripotent stem cells from cultured 51. Yamanaka S, Takahashi K (2006) Induction of plu-
human primordial germ cells. Proc Natl Acad Sci U ripotent stem cells from mouse fibroblast cultures.
S A 95(23):13726–13731 Tanpakushitsu Kakusan Koso 51(15):2346–2351
35. Fagoonee S, Hobbs RM, De Chiara L et al (2010) 52. Yao Y, Li S, Cao J et al (2018) Development of small
Generation of functional hepatocytes from mouse molecule biosensors by coupling the recognition of
24 Functional Role of Circular RNA in Regenerative Medicine 307

the bacterial allosteric transcription factor with iso- 69. Matsumoto Y, Fishel R, Wickner RB (1990)
thermal strand displacement amplification. Chem Circular single-stranded RNA replicon in
Commun (Camb) 54:4774. https://doi.org/10.1039/ Saccharomyces cerevisiae. Proc Natl Acad Sci U S
C8CC01764F A 87(19):7628–7632
53. Li W, Li K, Wei W et al (2013) Chemical approaches 70. Pasman Z, Been MD, Garcia-Blanco MA (1996)
to stem cell biology and therapeutics. Cell Stem Cell Exon circularization in mammalian nuclear extracts.
13(3):270–283 RNA 2(6):603–610
54. Marson A, Foreman R, Chevalier B et al (2008) Wnt 71. Cocquerelle C, Daubersies P, Majerus MA et al
signaling promotes reprogramming of somatic cells (1992) Splicing with inverted order of exons occurs
to pluripotency. Cell Stem Cell 3(2):132–135 proximal to large introns. EMBO J 11(3):1095–1098
55. D’Amour KA, Agulnick AD, Eliazer S et al (2005) 72. Capel B, Swain A, Nicolis S et al (1993) Circular
Efficient differentiation of human embryonic transcripts of the testis-determining gene Sry in
stem cells to definitive endoderm. Nat Biotechnol adult mouse testis. Cell 73(5):1019–1030
23(12):1534–1541 73. Zaphiropoulos PG (1997) Exon skipping and circu-
56. Efe JA, Hilcove S, Kim J et al (2011) Conversion lar RNA formation in transcripts of the human cyto-
of mouse fibroblasts into cardiomyocytes using chrome P-450 2C18 gene in epidermis and of the rat
a direct reprogramming strategy. Nat Cell Biol androgen binding protein gene in testis. Mol Cell
13(3):215–222 Biol 17(6):2985–2993
57. Graf T (2011) Historical origins of transdiffer- 74. Burd CE, Jeck WR, Liu Y et al (2010) Expression
entiation and reprogramming. Cell Stem Cell of linear and novel circular forms of an INK4/ARF-­
9(6):504–516 associated non-coding RNA correlates with athero-
58. Sekiya S, Suzuki A (2011) Direct conversion of sclerosis risk. PLoS Genet 6(12):e1001233
mouse fibroblasts to hepatocyte-like cells by defined 75. Wang PL, Bao Y, Yee MC et al (2014) Circular RNA
factors. Nature 475(7356):390–393 is expressed across the eukaryotic tree of life. PLoS
59. Huang P, He Z, Ji S et al (2011) Induction of func- One 9(6):e90859
tional hepatocyte-like cells from mouse fibroblasts 76. Salzman J, Gawad C, Wang PL et al (2012) Circular
by defined factors. Nature 475(7356):386–389 RNAs are the predominant transcript isoform from
60. Saifan C, Nasr R, Mehta S et al (2013) Diabetes hundreds of human genes in diverse cell types. PLoS
insipidus: a challenging diagnosis with new drug One 7(2):e30733
therapies. ISRN Nephrol 2013:797620 77. Jeck WR, Sorrentino JA, Wang K et al (2013)
61. Devendra D, Liu E, Eisenbarth GS (2004) Circular RNAs are abundant, conserved, and associ-
Type 1 diabetes: recent developments. BMJ ated with ALU repeats. RNA 19(2):141–157
328(7442):750–754 78. Memczak S, Jens M, Elefsinioti A et al (2013)
62. Scharp DW, Lacy PE, Santiago JV et al (1990) Circular RNAs are a large class of animal RNAs
Insulin independence after islet transplantation into with regulatory potency. Nature 495(7441):333–338
type I diabetic patient. Diabetes 39(4):515–518 79. Guo JU, Agarwal V, Guo H et al (2014) Expanded
63. Suberi M, McKeever WF (1977) Differential right identification and characterization of mammalian
hemispheric memory storage of emotional and non-­ circular RNAs. Genome Biol 15(7):409
emotional faces. Neuropsychologia 15(6):757–768 80. Danan M, Schwartz S, Edelheit S et al (2012)
64. Assady S, Maor G, Amit M et al (2001) Insulin pro- Transcriptome-wide discovery of circular RNAs in
duction by human embryonic stem cells. Diabetes archaea. Nucleic Acids Res 40(7):3131–3142
50(8):1691–1697 81. Zhang Y, Zhang XO, Chen T et al (2013) Circular
65. Lumelsky N, Blondel O, Laeng P et al (2001) intronic long noncoding RNAs. Mol Cell
Differentiation of embryonic stem cells to insulin-­ 51(6):792–806
secreting structures similar to pancreatic islets. 82. Salzman J, Chen RE, Olsen MN et al (2013) Cell-­
Science 292(5520):1389–1394 type specific features of circular RNA expression.
66. Urban VS, Kiss J, Kovacs J et al (2008) PLoS Genet 9(9):e1003777
Mesenchymal stem cells cooperate with bone 83. Salmena L, Poliseno L, Tay Y et al (2011) A ceRNA
marrow cells in therapy of diabetes. Stem Cells hypothesis: the Rosetta Stone of a hidden RNA lan-
26(1):244–253 guage? Cell 146(3):353–358
67. Strauer BE, Brehm M, Zeus T et al (2001) 84. Huntzinger E, Izaurralde E (2011) Gene silencing
Intracoronary, human autologous stem cell trans- by microRNAs: contributions of translational repres-
plantation for myocardial regeneration following sion and mRNA decay. Nat Rev Genet 12(2):99–110
myocardial infarction. Dtsch Med Wochenschr 85. Hansen TB, Wiklund ED, Bramsen JB et al (2011)
126(34–35):932–938 miRNA-dependent gene silencing involving
68. Wu Q, Wang Y, Cao M et al (2012) Homology-­ Ago2-mediated cleavage of a circular antisense
independent discovery of replicating pathogenic RNA. EMBO J 30(21):4414–4422
circular RNAs by deep sequencing and a new com- 86. Peng Y, Song X, Zheng Y et al (2017) Circular RNA
putational algorithm. Proc Natl Acad Sci U S A profiling reveals that circCOL3A1-859267 regulate
109(10):3938–3943 type I collagen expression in photo aged human
308 R. Y. Cao et al.

dermal fibroblasts. Biochem Biophys Res Commun 94. Xu N, Papagiannakopoulos T, Pan G et al (2009)
486(2):277–284 MicroRNA-145 regulates OCT4, SOX2, and KLF4
87. Wilusz JE (2017) Circular RNAs: unexpected out- and represses pluripotency in human embryonic
puts of many protein-coding genes. RNA Biol stem cells. Cell 137(4):647–658
14(8):1007–1017 95. Wu CS, Yu CY, Chuang CY et al (2014) Integrative
88. Bohjanen PR, Colvin RA, Puttaraju M et al (1996) A transcriptome sequencing identifies trans-splicing
small circular TAR RNA decoy specifically inhibits events with important roles in human embryonic
tat-activated HIV-1 transcription. Nucleic Acids Res stem cell pluripotency. Genome Res 24(1):25–36
24(19):3733–3738 96. Suzuki H, Zuo Y, Wang J et al (2006) Characterization
89. Bolha L, Ravnik-Glavac M, Glavac D (2017) of RNase R-digested cellular RNA source that con-
Circular RNAs: biogenesis, function, and a role sists of lariat and circular RNAs from pre-mRNA
as possible Cancer biomarkers. Int J Genomics splicing. Nucleic Acids Res 34(8):e63
2017:6218353 97. Buganim Y, Markoulaki S, van Wietmarschen N
90. Talhouarne GJ, Gall JG (2014) Lariat intronic RNAs et al (2014) The developmental potential of iPSCs
in the cytoplasm of Xenopus tropicalis oocytes. is greatly influenced by reprogramming factor selec-
RNA 20(9):1476–1487 tion. Cell Stem Cell 15(3):295–309
91. Yang Y, Wang Z (2018) Constructing GFP-based 98. Gabut M, Samavarchi-Tehrani P, Wang X et al (2011)
reporter to study back splicing and translation of cir- An alternative splicing switch regulates embryonic
cular RNA. Methods Mol Biol 1724:107–118 stem cell pluripotency and reprogramming. Cell
92. Yu CY, Li TC, Wu YY et al (2017) The circular 147(1):132–146
RNA circBIRC6 participates in the molecular cir- 99. Ashwal-Fluss R, Meyer M, Pamudurti NR et al
cuitry controlling human pluripotency. Nat Commun (2014) circRNA biogenesis competes with pre-­
8(1):1149 mRNA splicing. Mol Cell 56(1):55–66
93. Choi YJ, Lin CP, Ho JJ et al (2011) miR-34 miRNAs 100. Kawamura T, Suzuki J, Wang YV et al (2009) Linking
provide a barrier for somatic cell reprogramming. the p53 tumour suppressor pathway to somatic cell
Nat Cell Biol 13(11):1353–1360 reprogramming. Nature 460(7259):1140–1144
 The Role of Circular RNAs
in Cerebral Ischemic Diseases: 25
Ischemic Stroke and Cerebral
Ischemia/Reperfusion Injury

Jian Yang, Mengli Chen, Richard Y. Cao, Qing Li,

and Fu Zhu

Abstract Keywords
Cerebral ischemic diseases including isch- Circular RNAs · Cerebral ischemic diseases ·
emic stroke and cerebral ischemia reperfusion Ischemic stroke · Cerebral ischemia/reperfu-
injury can result in serious dysfunction of the sion injury
brain, which leads to extremely high mortality
and disability. There are no effective therapeu-
tics for cerebral ischemic diseases to date. 1 Introduction
Circular RNAs are a kind of newly investi-
gated noncoding RNAs. It is reported that cir- The brain is the most sensitive organ to hypoxia.
cular RNAs are enriched in multiple organs, When cerebral blood flow is suddenly inter-
especially abundant in the brain, which indi- rupted, tissues are deprived of oxygen and glu-
cates that circular RNAs may be involved in cose, leading to dysfunction of brain as well as
cerebral physiological and pathological pro- other parts of body controlled by the cerebral
cesses. In this chapter, we will firstly review ischemic regions, which is called ischemic stroke
the pathophysiology, underlying mechanisms, or cerebral ischemia/reperfusion injury without
and current treatments of cerebral ischemic or with restoration of blood flow and oxygen [1].
diseases including ischemic stroke and cere- Ischemia elicits tissue anoxia which is the basis
bral ischemia/reperfusion injury. Secondly, of ischemic injury and primes the tissue for sub-
the characteristics and function of circular sequent reperfusion damage [2]. Ischemic stroke
RNAs will be outlined, and then we are going and cerebral ischemia/reperfusion injury can
to introduce the roles circular RNAs play in both result in serious dysfunction of the brain,
human diseases. Finally, we will summarize which leads to extremely high mortality and dis-
the function of circular RNAs in cerebral isch- ability [3, 4]. However, there are no effective
emic diseases. therapeutics against cerebral ischemic diseases.
Therefore, it is required to develop novel effec-
tive therapeutic strategies for treatment of cere-
J. Yang · R. Y. Cao · Q. Li · F. Zhu (*)
Zhongshan-Xuhui Hospital, Fudan University/ bral ischemic diseases. As a special type of
Shanghai Clinical Research Center, Chinese noncoding RNAs, circular RNAs play a vital
Academy of Sciences, Shanghai, China regulatory role in RNA metabolism [5]. It is
M. Chen known that circular RNAs are abundant in vari-
Department of Cardiology, The First Affiliated ous organs, especially highly enriched in
Hospital of Nanjing Medical University, ­synapses. The expression levels of certain circu-
Nanjing, China

© Springer Nature Singapore Pte Ltd. 2018 309

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_25
310 J. Yang et al.

lar RNAs can be much higher than the canonical thorough understanding about the pathophysiol-
linear transcripts of their parent genes during the ogy of stroke is necessary.
central nervous system development, which indi-
cates that circular RNAs may be involved in cere- 2.1.2 Underlying Mechanisms
bral physiological and pathological processes The mechanisms of ischemic stroke are compli-
[6–8]. This chapter will be divided into three cated. Currently, it is acknowledged that the
parts. In the first part, we will outline the patho- mechanisms of ischemic stroke injury include
physiology, underlying mechanisms, and current metabolic disorders, inflammation, depolariza-
treatments of cerebral ischemic diseases includ- tion of penumbra, and cellular apoptosis
ing ischemic stroke and cerebral ischemia/reper- [14–17].
fusion injury. In the second part, we will review
the characteristics and function of circular RNAs Metabolic Disorders
and then briefly introduce the roles circular RNAs The brain is the biggest consumer of energy in
play in human diseases. In the final part, we will the human body and therefore very sensitive to
focus on talking about the function of circular reduction of oxygen and glucose. When ischemic
RNAs in cerebral ischemic diseases. stroke happens, blood supply to brain is blocked,
which results in lack of oxygen and glucose,
leading to dysfunction in energy metabolism and
2 Cerebral Ischemic Diseases substance metabolism in the brain [14]. It has
been demonstrated that combining NAD+ with
2.1 Ischemic Stroke NADPH can provide more neuroprotective
effects in both cellular and animal models of
2.1.1 Pathophysiology ischemic stroke [18]. Besides, the downregula-
Stroke is among the leading causes of deaths and tion of energy metabolism-related peroxisome
disability worldwide, the high morbidity and proliferator-activated receptor gamma (PPARγ)
mortality caused by stroke bring heavy financial gene was reported to prevent relapse of stroke
and mental burden to society and family [9, 10]. and other vascular disorders in patients with
Each year, 15 million people worldwide suffer stroke [19]. Apart from energy metabolism, sub-
from stroke. Ischemic stroke and hemorrhagic stance metabolism is also involved. It is reported
stroke are two main types recognized, and the that an increasing glucose range quartile was
former accounts for more than 80% of total cases positively associated with initial neurologic
[11]. Ischemic stroke is mainly caused by middle severity, and the association remained significant
cerebral artery occlusion by blood clot or plaque after dichotomization regarding glycated hemo-
formation which results in reduction of blood globin levels on admission [20]. Another research
supply to the brain tissues, leading to insufficient on pancreatic β-cell function and prognosis of
supply of oxygen and glucose to the affected nondiabetic patients with ischemic stroke dem-
brain area [12]. Ischemic stroke is a dynamic pro- onstrates the association between pancreatic
cess, during which neurons adjacent hypoper- β-cell function and an increased risk of 12-month
fused areas undergo necrosis since ischemia poor prognosis in nondiabetic patients with isch-
leads to disturbances in membrane integrity and emic stroke [21].
cell stability. It has been estimated via serial MR
brain imaging that each minute of ischemia leads Inflammation
to the death of 1.9 million neurons and the Ischemic stroke can induce the activation inflam-
destruction of 14 billion synapses in human matory cascade. The activation of innate and
patients with middle cerebral artery occlusion adaptive immune system induced by ischemic
[13]. The high mortality and disability caused by stroke causes a massive migration of peripheral
stroke makes it urgent for us to pay attention and leukocytes into the brain, which triggers focal
to search for solutions. In order to achieve this, inflammatory responses, promotes tissue death,
25 The Role of Circular RNAs in Cerebral Ischemic Diseases: Ischemic Stroke and Cerebral… 311

and deteriorates clinical outcome [22]. Increasing apoptosis following ischemic stroke and inhibi-
evidence suggests that antigen presentation cells tion of caspase-6 or caspase-8 can ameliorate the
(APC) are reduced in the periphery and increased neuronal damage resulting from ischemic stroke
in the ischemic brain both in rodent and human [37]. Besides, it is reported that administration of
stroke; the accumulation is associated with mesencephalic astrocyte-derived neurotrophic
expression of MHC class II molecules and the factor (MANF) can protect against ischemic
co-stimulatory molecules CD80 [23–26]. stroke-induced neuronal apoptosis by inhibiting
Besides, both CD4+ and CD8+ T cells are the cleavage of caspase-3 [38]. PARP-1 is another
involved in cerebral ischemic injury [27]. important protein in cell death regulation, which
Moreover, monocytes and macrophages play is cleaved by caspase-3 into two fragments and
important roles in cerebral ischemic stroke acts as a marker for apoptotic cell death [39]. A
because of their expression of the anti-­ report published on Stroke journal revealed that
inflammatory and pro-inflammatory mediators the high level of chelating zinc decreases brain
[15, 28]. Not only inflammatory cells but also damage and improves neurological functions via
inflammatory mediators participate in this patho- inhibition of PARP-1 [40]. In addition to neuro-
logical process. For instance, chemokines are nal apoptosis, platelet apoptosis is also critical in
considered pro-inflammatory because of their the pathogenesis of stroke. Platelet can induce
ability of mediating signals that induce leukocyte caspase-3 gene expression when stimulated with
recruitment to damaged tissues. They are com- ADP, which plays a key role in apoptosis signal-
prised of four subfamilies including CC, CXC, ing. Furthermore, compared to the control group,
XC, and CX3C chemokines [29–31]. However, it the expression level of platelet cytochrome-c and
has been reported that chemokines can induce caspase-3 is significantly increased in stroke
both neuroprotective and deleterious effects on patients [41, 42]. Collectively, cellular apoptosis
postischemic stroke patients, suggesting that is a significant part of ischemic stroke
more explorations are needed for the role of che- pathophysiology.
mokines in ischemic stroke and the chemokines
represent novel and potential therapeutic targets Penumbra
for prevention of ischemic stroke and reduction The penumbra was classically described as the
of consequent disability [32]. Previous studies tissue between the ischemic core and normal area
have suggested that the expression of IL-6 is of patients suffering from ischemic stroke; and
increased when ischemic stroke occurs and IL-6 the blood flow of penumbra is too low to main-
is believed to act as a robust early marker for out- tain electric activity but sufficient to preserve ion
come in acute ischemic stroke [33, 34]. channels [17]. Since 1977 when penumbra was
Collectively, inflammation cascade reaction, firstly defined, the physiologic characteristics of
including activation of different types of immune ischemic penumbra and the underlying mecha-
cells and initiation of various inflammatory nisms that mediate penumbra cell death have
mediators, plays a vital role in the pathophysiol- been deeply investigated [43]. The ultimate goal
ogy of cerebral ischemic stroke. of neuroprotective treatments is to save penum-
bra. If not treated in time, infarction will develop
Cellular Apoptosis inside penumbra due to consequent damage such
Increasing evidence indicates that cellular apop- as rapid depolarization, cell apoptosis, and
tosis contributes to ischemic stroke injury. Along inflammation like ischemic core [44]. Previous
with neuronal necrosis, cellular apoptosis is studies based on experimental cerebral ischemic
another important form of delayed neuronal stroke models have suggested that the depolariza-
death following ischemic stroke [35, 36]. The tions of penumbra are positively associated with
family of caspases plays vital role in the pathway the infarct volume, indicating that inhibition of
of apoptosis. Previous studies have shown that peri-infarct depolarization maybe a therapeutic
caspase-6 and caspase-8 are involved in neuronal strategy for improving the poor outcome of isch-
312 J. Yang et al.

emic stroke [45]. Besides, a reduction on protein Potential Therapeutic Strategies

synthesis in the penumbra is a sensitive meta- Since current treatment has a lot of limitation,
bolic result of cerebral ischemia. Moreover, the efforts have been made in searching for novel
low blood flow in penumbra is responsible for therapeutic strategies. Stem cells can exert multi-
reduction of adenosine 5′-triphosphate and fail- potency in treatment of cerebral ischemia. Bone
ure of Na/K pumps, which results in an increase marrow stromal cells (BMSCs) have the ability
of intracellular calcium and consequently induces of self-renewal and differentiation into neuronal
irreversible damage to brain tissues [46]. In a and glial lineages under certain conditions [59,
word, the penumbra plays an essential role in the 60]. Increasing evidence suggests that BMSCs
pathophysiology and diagnosis of ischemic transplantation can obviously improve the func-
stroke. tion and outcome of animal models of cerebral
ischemic stroke at different time points via differ-
2.1.3 Treatment ent transplantation routes [61]. More than 20% of
the miRNAs alter in the ischemic brain, suggest-
Clinical Treatment ing that miRNAs are key mediators in ischemic
The most effective treatment approach for isch- stroke biology [62–65]. miR-124 is mainly
emic stroke is to recover cerebral arteries blood expressed in the central nerve system neuronal
flow occluded by thrombus or embolism. The cells; and the expression of miR-124 in the brain
approach is called reperfusion therapy that is is 100 times higher than in other organs [66]. It
required as soon as possible [47–49]. There are has been reported that microRNA-124 exhibits
two types of strategies for recanalization of neuronal protective effects on rodent models of
blocked vessels: chemical and mechanical ways cerebral ischemic stroke via inhibiting cellular
[50]. Intravenous thrombolysis is the main type apoptosis [67]. Danshen is a Chinese medicinal
of chemical treatment for ischemic stroke. Until herb widely used for treatment of ischemic brain
now, recombinant tissue plasminogen activator and heart diseases. The most abundant and bioac-
(rtPA) is the only FDA-approved drug for patients tive component in danshen is salvianolic acid B,
following ischemic stroke. The narrow therapeu- which can attenuate apoptosis and inflammation
tic window that is from 3 h of the onset of a stroke in experimental stroke rats through activation of
until 4.5 h is usually a major limitation for its SIRT1 [68, 69]. Collectively, numerous experi-
clinical application [51–54]. Thrombectomy is mental studies on cerebral ischemic stroke have
the main mechanical treatment for ischemic provided multiple novel therapeutic strategies.
stroke. A tiny catheter is used to penetrate into
the blocked cerebral vessel along the natural line,
reach the focal lesion, and remove the thrombus 2.2 Cerebral Ischemia/
out of the vessel [55]. Despite of the efficacy of Reperfusion Injury
recanalization, there is a chance that rtPA may
increase the incidence of hemorrhage, which lead 2.2.1 Pathophysiology
to more serious outcomes [13]. As for thrombec- Stroke is the fifth leading cause of death and a
tomy, it is performed to a comprehensive stroke major cause of adult disability [4]. The most
center that limits to an artery with a diameter effective therapeutic method is immediate recov-
>2 mm [56, 57]. Early supportive care is another ering the blood flow via recanalization of the
clinical treatment, which is used to meet basic occluded arteries. Immediate restoration of the
demands for keeping patients alive [58]. blood supply can reduce more extensive brain tis-
Considering the limitation of recanalization in sue injury by salvaging a reversibly damage of
clinical application, novel therapeutic strategies penumbra [70, 71]. However, recanalization car-
are required for improved treatment of patients ries risks of further cellular necrosis and neural
following ischemic stroke. damage. Some patients may experience serious
25 The Role of Circular RNAs in Cerebral Ischemic Diseases: Ischemic Stroke and Cerebral… 313

deterioration in the form of fatal edema or intra- cerebral ischemia/reperfusion injury. The two
cranial hemorrhage following thrombolysis, common kinds of RNS, nitric oxide (NO) and
which is called cerebral ischemia/reperfusion peroxynitrite (ONOO−), are reported to partici-
injury [72]. Cerebral ischemia/reperfusion injury pate in cerebral ischemia/reperfusion injury. NO
could be defined as degeneration of ischemic but is generated together with superoxide, and they
salvageable brain tissue after reperfusion [73]. can interact with each other and produce ONOO−
Although thrombolysis and embolectomy can at a diffusion-limited rate. ONOO− inactivates
restore blood flow of the infarcted brain tissue, aconitase and superoxide dismutase (SOD) and
the treatment can also bring the risk of reperfu- mediates NO-induced damage of blood-brain
sion injury [50, 74]. In order to develop more barrier [86, 87]. The neurotoxic effect due to
effective and feasible strategies for reduction or overproduction of free radicals plays a significant
minimization of cerebral reperfusion injury, we part in cerebral ischemia/reperfusion injury;
need to understand the pathophysiology of cere- therefore, intervention of excessive free radicals
bral reperfusion injury. is critical for treatment of patients with cerebral
ischemia.
2.2.2 Underlying Mechanisms
Reperfusion can recover blood flow and the sup- Inflammation
ply of oxygen as well as many other energy mate- The inappropriate reperfusion of ischemic tissues
rials. However, this process bears risks of can produce numerous free radicals and many
worsening the original brain damage caused by other messengers, which trigger a chain of
ischemia through different mechanisms such as inflammatory reaction including activation of
release of free radicals, activation of inflamma- inflammatory cells and release and interaction of
tory cascade reaction, promotion of cellular inflammatory mediators [88]. The most impor-
apoptosis, calcium overload, and release of excit- tant inflammatory cells involved in cerebral isch-
atory amino acids, which is much more sophisti- emia/reperfusion injury are leukocytes, microglia,
cated than ischemia alone [75–78]. and astrocytes [89]. The signals received from
ischemic area can activate and induce leukocytes
Free Radicals to secret inflammatory cytokines which in turn
Free radicals are cytotoxic molecules that play recruit more leukocytes. Accumulating of leuko-
vital roles in the progress of cerebral ischemia/ cytes and increasing cytokines related conse-
reperfusion injury. There are two main types of quently deteriorate the vicious circle of
free radicals: reactive oxygen species (ROS) and inflammation [76]. Microglia are pro-­macrophage
reactive nitrogen species (RNS) [75, 79]. cells in the brain which are intimately associated
Physiologically, the level of ROS generated from with inflammation in cerebral ischemia/reperfu-
mitochondrial inner membrane is very low for sion injury. In the first few minutes of ischemia,
maintenance of cellular redox homeostasis [80– microglia are activated and transformed into
82]. When cerebral ischemia reperfusion hap- cerebral macrophages as soon as cellular apopto-
pens, blood flow and oxygen supply are sis occurs, resulting in neurotoxic effects in the
interrupted, which leads to accumulation of brain [90]. Astrocytes can promote the differen-
excessive ROS, resulting in tissue oxidative dam- tiation and proliferation of microglia, enhance
age in brain tissue suffered from ischemia reper- the phagocytic activity of macrophage, and
fusion. More and more evidences have shown induce generation of inflammatory mediators
that ROS accumulation can disturb the signal [91]. It has been demonstrated that interleukin-1β
transduction and induce lipid peroxidations in is able to reduce cerebral blood flow and increase
neural cells, which lead to cell death, activation the recruitment of neutrophils and generation of
of inflammation factors, and breakdown of blood-­ superoxide anion (O2−) in animal models of mid-
brain barrier [83–85]. On the other hand, RNS is dle cerebral artery occlusion (MCAO) [92].
also crucial for the pathophysiological process of Plenty of reports have indicated that inflamma-
314 J. Yang et al.

tory mediators including cytokines, chemokines, related to apoptosis including Bcl-2, FAS, and P53
and cell adhesion molecules can mediate the participate in the process of cellular apoptosis
pathophysiology of cerebral ischemia/reperfu- [105–107]. During cerebral ischemia reperfusion,
sion injury [93–96]. various signaling pathways, including MAPK sig-
naling pathway, PI3K/Akt/GSK-3β signaling
Calcium Overload pathway, JAK-­STAT signaling pathway, and tran-
Calcium is an important element in the body and scription factor NF-κB-involved signaling path-
takes a vital part in generation of bioelectricity and way, function in neurons and astrocytes
regulation of cellular function as well as metabo- [108–110].
lism. Recovery of blood flow in ischemic tissues
significantly increases the level of calcium which 2.2.3 Treatment
leads to cellular injury. This phenomenon is called
calcium overload [77, 97, 98]. Calcium overload- Clinical Treatment
induced cytotoxic effects cause neural damage As described above, the most effective treatment
through the following ways. Firstly, the augmented for cerebral ischemic diseases is restoration of
free calcium activates calmodulin, which binds to ischemic area. However, restoration carries the
free calcium and form a compound that cramp risks of further aggravating dysfunction induced
brain vessels and consequently aggravate blood by ischemia. Therefore, inhibition of reperfusion
and oxygen deficiency in cerebral ischemic area. injury is a crucial part of overall treatment for
Secondly, phospholipase C, phospholipase A2, cerebral ischemia. Hypothermia therapy for neu-
and nitric oxide synthase (NOS) as well as cal- roprotection was first proposed by Busto in 1987.
cium-dependent protease activated by calcium It has been proved that hypothermia therapy is
overload can catalyze the generation of oxygen able to decrease the cerebral metabolic rate,
free radicals. Thirdly, calcium overload enlarges decrease the production of inflammatory media-
the gaps between cerebral vascular endothelial tors, and inhibit the generation of free radicals
cells, resulting in high permeability of blood-brain and excitatory amino acids. In addition, hypo-
barrier and consequent cerebral edema. Finally, thermia therapy can be utilized in patients fol-
calcium overload can raise the production of excit- lowing cardiac arrest which apparently improve
atory amino acids which in turn aggravate calcium the function of the brain, especially when com-
overload, inducing the death of neurons [99–102]. bined with other therapeutics [111–114].
Collectively, the level of calcium greatly affects Moreover, cerebral ischemia/reperfusion injury
the cerebral pathophysiology, and this can be can be treated by reducing inflammation; antioxi-
another therapeutic target against cerebral isch- dants are effective agents for removing free radi-
emia/reperfusion injury. cals and inhibiting lipid peroxidation. Aspirin is a
validated antioxidative drug for brain tissue pro-
Cellular Apoptosis tection [23, 115]. Besides, vitamin C and vitamin
Cellular apoptosis is triggered in penumbra several E are regarded as preventive neuroprotective
days after cerebral ischemia [103]. When reperfu- agents due to their radical scavenging ability
sion injury appears, the damage can be exacer- [116, 117]. In addition, calcium antagonists can
bated. On one hand, deficiency of blood and reduce calcium influx by blocking calcium chan-
oxygen stimulates the expression of the apoptosis- nel and consequently alleviate brain damage
related genes. On the other hand, ROS, calcium [118]. However, current clinical treatments have
overload, NO, and energy metabolism disorders various limitations, and advanced novel thera-
will also induce neuronal apoptosis through differ- peutics are urgently required.
ent mechanisms. The caspases family is the initia-
tors and performers of apoptosis in mammal which Experimental Treatment
is deeply associated with cerebral ischemia/reper- The accumulation of glutamate in periphery area
fusion injury [104]. Besides, several other genes of ischemic region causes high levels of intracel-
25 The Role of Circular RNAs in Cerebral Ischemic Diseases: Ischemic Stroke and Cerebral… 315

lular sodium and calcium, resulting in neuronal closed loop structure [127]. In 1976, circular
injury. Lubeluzole is the antagonist of glutamate,RNAs were first identified by electron micros-
and pre-administration of lubeluzole can allevi- copy in a research associated with RNA viruses
ate neuronal damage in cerebral ischemia models [128]. At that time, biotechnology was not
[119]. It has been reported that tamoxifen could advanced enough to validate the abundant exis-
inhibit the release of excitatory amino acids in tence of circular RNAs, let alone their functions.
penumbra and improve the praxeology and histo- However, emerging evidences suggest that circu-
pathology results after reperfusion [120]. Anotherlar RNAs are abundant in mammal and more
research indicates that natural compounds from stable than linear RNAs because of their cova-
traditional herbal medicine can be applied in the lently closed loop structure which enables them
treatment of cerebral ischemia reperfusion injury to be resistant to RNA exonucleases [129]. In
[121]. In fact, Chinese medication and therapy addition, genes encoding circles in one species
such as acupuncture are suggested to have unique have great possibility of encoding circles in other
effects on treatment of cerebral ischemia reperfu-species, indicating that circular RNAs are evolu-
sion injury [122]. tionarily conserved among diverse species [130–
132]. There are three types of circular RNAs,
exonic circular RNAs, intronic circular RNAs,
3 Circular RNAs and retained-intron circular RNAs, which are
transcribed from pre-mRNA sequences by RNA
In the past few decades, protein-coding genes polymerase [133]. According to published docu-
and their transcripts in eukaryotes have been ments, most of circular RNAs show a sophisti-
investigated deeply [123]. However, the recent cated tissue- and cell type-specific expression
development of high-throughput RNA sequenc- pattern. Intriguingly, compared with other organs,
ing has identified that 98% of the whole-genome it seems that circular RNAs are more abundant in
transcripts are noncoding RNAs and the protein-­ the brain, which gives a clue of the potential
coding RNAs constitute only 2% [124]. function of circular RNAs in neurology [8,
Noncoding RNAs play vital roles in gene regula- 133–135].
tion. Currently, there are various noncoding
RNAs have been found including microRNAs,
long noncoding RNAs, small interfering RNAs, 3.2 Biological Functions
and small nuclear RNAs. In addition, circular of Circular RNAs
RNAs that recently catch our attention are sug-
gested with surprisingly strong function in regu- Circular RNAs have substantial impact on gene
lating gene expression [5, 125]. Due to their regulation. In this part, we will list some biologi-
specific structure, uncertain function, and low cal functions of circular RNAs as follows.
abundance, circular RNAs were ignored as trash
of genomes for decades [126]. Nowadays, circu- 3.2.1 MicroRNA Sponges
lar RNAs regain the attention of scientists MicroRNAs are an enriched class of noncoding
because more and more evidence has revealed RNAs and mediate regulation of mRNA tran-
their potency and prospect. scription through microRNA response elements
(MREs) [136]. Emerging studies reveal that sev-
eral circular RNAs can bind to MREs, acting as
3.1 Characteristics of Circular competitive endogenous (ce)RNAs for
RNAs microRNA-binding sites, which in turn modulate
the activity of microRNAs. For instance, the sex-­
Circular RNAs are structurally different from the determining region Y (SRY) is a circular RNA
other kinds of noncoding RNAs whose head 3′ found in mouse testis in 1993; and then it has
and tail 5′ ends combine covalently and form a been proved to be a microRNA-138 sponge by
316 J. Yang et al.

regulation microRNA-138-associated mRNA great deal of genes can play roles in biogenesis of
translation [135, 137, 138]. In addition to circ- circular RNAs; and the mechanism of splicing
SRY, cerebellar degeneration-related protein 1 control resembles a switch that flips between
(CDR1) transcript, called CIRS-7, is another cir- back-splicing or linear splicing [132, 141].
cular RNA discovered in mammal, which is
reported to act as a microRNA-7 sponge [139].
Besides, recent investigation suggests that circH- 3.3  ircular RNAs in Human
C
IPK3, derived from the HIPK3 gene Exon2, can Diseases
bind to 9 microRNAs with 18 potential binding
sites and particularly serve as a microRNA-124 Besides the diverse functions of circular RNAs
sponge [140]. physiological processes, the effects of circular
RNAs on pathological processes have also been
3.2.2 Transcription Regulation revealed. The change of circular RNA expression
The distribution of circular RNAs in cells pro- is involved with multiple types of diseases such
vides hints on their probable functions. According as cardiovascular diseases, cancers, infections,
to multiple researches, intronic circular RNAs and neurological diseases.
and retained-intron circular RNAs usually locate
in the nucleus, while exonic RNAs are primarily 3.3.1 Circular RNAs
situated in the cytoplasm [141]. Apart from act- and Cardiovascular Diseases
ing as microRNA sponges, several intronic circu- CircANRIL is a circular antisense noncoding
lar RNAs can exert posttranscriptional regulation RNA in the INK4 locus which has been reported
on gene expression. For example, c-sirt7, derived to be associated with atherosclerosis. CircANRIL
from lariats and interacting with the Pol II com- can bind to pescadillo homologue 1 (PES1) and
plex, can downregulate the gene expression of in turn impair ribosome biogenesis in vascular
the relevant ankyrin repeat domain 52 or sirtuin 7 smooth muscle cells, which is protective for
[132, 142]. Moreover, some retained-intron cir- arteries [146, 147]. Besides, a recent study has
cular RNAs can interact with the U1 component revealed that the well-known circular RNA
and recruit RNA polymerase II to the promoter Cdr1as can also serve as a microRNA-7a sponge
region of genes and upregulated the expression of in myocardial cells and modulate the function of
their target genes [143]. microRNA-7a in myocardial infarct injury [148].
Furthermore, circ-Foxo3 is found highly
3.2.3 Competition with Linear expressed in heart samples of aged patients and
Splicing mice, indicating that circFoxo3 is linked with
Back-splicing or splicing of pre-mRNA deter- cardiac senescence [149]. Another research sug-
mined the structure of RNA produced. Back-­ gests that the heart-related circular RNA (HRCR)
splicing, the way how circular RNAs generate, can regulate cardiomyocyte hypertrophy and
compete with splicing of pre-mRNA, leading to heart failure via acting as a microRNA-223
decreased production of linear mRNAs [144]. It sponge [150].
has been uncovered that the Muscleblind (MBL)
gene carries sequences which are able to form a 3.3.2 Circular RNAs and Cancer
circular RNA transcript. CircMBL can compete Considering the role of circular RNAs in regula-
with MBL pre-mRNA splicing, exhibiting nega- tion of cell cycles, cell proliferation, and cellular
tive regulatory effects on canonical splicing senescence, it is reasonable that circular RNA is
[133]. Besides, a circular RNA originated from implicated with cancer. For instance, ciRS-7 is a
SEPALLATA3 gene can form a R-loop and sus- ceRNA of microRNA-7 that can promote the ini-
pend the transcription. It also has an impact on tiation and progress of cancer by upregulating
recruitment of splicing factors for initiating tran- oncogenic EGRF and XIAP gene and downregu-
scription [145]. It has also been reported that a lating tumor-suppressed KLF4 [151, 152]. In
25 The Role of Circular RNAs in Cerebral Ischemic Diseases: Ischemic Stroke and Cerebral… 317

addition, has-circ-001988 has been proved to be injury are not satisfying. Despite current
involved in colorectal cancers and linked with the advanced biological and medical technologies,
differentiation and perineural invasion of cancer prognosis of patients with cerebral ischemic dis-
[153]. Moreover, it has been suggested that the eases is poor, which urges us to look for more
abundance of circular RNAs may be a potential effective interventions. During the past few
marker for cell proliferation in breast cancer decades, noncoding RNAs are hot topics in life
[154]. science. In this part, we briefly review the roles of
microRNAs and long noncoding RNAs by focus-
3.3.3 Circular RNAs and Preeclampsia ing on the roles of circular RNAs, a new star in
Emerging investigations provide hints on the role the family of noncoding RNAs, in cerebral isch-
of circular RNAs in preeclampsia. Based on emic diseases.
results from circular RNA microarray and previ-
ous studies, circular RNAs are hypothesized to
make contributions to the pathogenesis of pre- 4.1  oncoding RNAs in Cerebral
N
eclampsia via functioning as microRNA sponges Ischemic Diseases
[155]. It has also been reported that circular
RNAs in blood corpuscles can serve as predictor 4.1.1  icroRNAs in Cerebral Ischemic
M
of preeclampsia in early stage when combined Diseases
with plasma protein factor [156]. First, we talk about microRNAs. One publication
has revealed that increased microRNA-129-5p
3.3.4  ircular RNAs and Neurological
C levels exert protective effect against ischemia
Diseases reperfusion injury via alleviating neuronal
Circular RNAs are extremely abundant in the inflammation and blood-spinal cord barrier dam-
brain, particularly in neuropils and dendrites, age through regulation of high-mobility group
indicating the potential roles of circular RNAs in box-1 (HMGB1) and the Toll-like receptor
regulating synaptic function and neural plasticity (TLR)-3 pathway [162]. Besides, it has been
[8, 134]. Cdr1as has been demonstrated to be demonstrated that microRNA-130b can protect
associated with neurodegenerative diseases like the brain from ischemic stroke by targeting water
Alzheimer’s disease (AD). Cdr1as can increase channel protein aquaporin 4 [163]. In addition,
the AD-linked targets like ubiquitin protein ligase microRNA-431 also has a positive effect on rats
A via acting as a microRNA-7 sponge, increasing following cerebral ischemia/reperfusion injury
the clearance of amyloid peptides [157, 158]. by modulating the Rho/Rho-kinase signaling
Besides, Cdr1as causes high neural expression of pathway [164]. Similarly, microRNA-93 has
α-synuclein protein which is an essential compo- been reported to be protective for cerebral isch-
nent of Lewy bodies in Parkinson’s disease (PD) emia/reperfusion via suppressing inflammation
by sequestering microRNA-7 [159, 160]. Circular and cellular apoptosis [165].
RNAs play roles in memory as well. For exam-
ple, circPAIP2 can increase the expression of 4.1.2  ong Noncoding RNAs
L
memory-related gene PAIP2 through poly(A)- in Cerebral Ischemic Diseases
binding protein-associated signaling [161]. As to the role of long noncoding RNAs in cere-
bral ischemic diseases, there are plenty of corre-
sponding documents exist. It has been reported
4  ircular RNAs and Cerebral
C that Malat1 inhibits endothelial cell death and
Ischemic Diseases inflammation and consequently protects the cere-
bral microvasculature and parenchyma from
As previously mentioned, the valid therapies for cerebral ischemic insults [166]. Another study
cerebral ischemic diseases including acute isch- indicates that as a competing endogenous RNA
emic stroke and cerebral ischemia/reperfusion of microRNA-21, long noncoding RNA MEG3 is
318 J. Yang et al.

able to regulate ischemic neuronal death through microarrays and real-time PCR to identify the
targeting microRNA-21/PDCD4 signaling path- circular RNAs expression pattern at different
way [167]. Moreover, it has been found that the time points: 6, 12, and 24 h after reperfusion. The
silence of long noncoding RNA RMST can results suggested there are 283 circular RNAs
apparently reduce the brain infarct area and ame- altered compared with sham control, of which 16
liorate neurological function in mice following candidates are altered at all three time points of
cerebral ischemic insult [168]. A paper published reperfusion after ischemia. Furthermore, these 16
in 2017 suggested that long noncoding RNA H19 circular RNAs were validated to carry binding
can activate autophagy, thereby inducing cerebral sites for plenty of microRNAs. According to the
ischemia/reperfusion injury [169]. bioinformatics analysis, these identified circular
RNAs are functionally linked with pathophysiol-
ogy of stroke [171]. Another similar study was
4.2  ircular RNAs in Cerebral
C carried out based on a model of oxygen-glucose
Ischemic Diseases deprivation/reoxygenation (OGD/R) in HT22
cells by using circular RNA microarray for iden-
Actually, there are only a few papers about the tification of changes in circular RNAs expression
direct roles of circular RNAs in cerebral ischemic profiles. The results identified 2 upregulated cir-
diseases. In this part, we will first describe these cular RNAs and 13 downregulated circular RNAs
several literatures about the function of circular in the OGD/R model. Among those 15 circular
RNAs in cerebral ischemic diseases; and then we RNAs, the upregulation of mmu-circRNA-015947
will briefly review the roles of circular RNAs in was verified by quantitative real-time PCR and
atherosclerosis which represents a validated risk bioinformatics analysis, indicating that it can
factor for cerebral ischemia. interact with several target microRNAs. In addi-
tion, mmu-circRNA-015947 may be involved in
4.2.1  ircular RNAs and Cerebral
C apoptosis, metabolism, and immune-related
Ischemic Diseases pathways, suggesting that mmu-circRNA-015947
Ying Bai and his team published an article on the is likely to play a role in the pathophysiology of
Journal of Neuroscience in 2017. In this study, cerebral ischemia/reperfusion injury [172].
they used the plasma of acute ischemic stroke
patients including 13 females and 13 males as 4.2.2 Circular RNAs
well as a mouse stroke model for research. The and Atherosclerosis
results showed that circular RNA DLGAP4 Atherosclerosis is a chronic pathological process
(circDLGAP4) is increased both in the plasma of of arterial walls which is a critical risk factor of
patients and in rodent models. Besides, circDL- cerebrovascular diseases [173]. The major cause
GAP4 was shown to serve as a microRNA-143 of cerebral ischemic stroke is the blockage of
sponge that inhibits the expression of HECT cerebral vessels by embolism which is a common
domain E3 ubiquitin protein ligase 1. More consequence of atherosclerosis [174]. It has been
importantly, they found that overexpression of reported that the circular noncoding RNA ANRIL
circDLGAP4 can significantly reduce neurologi- can promote the progress of atherosclerosis via
cal deficits, diminish infarct area, and alleviate aggravating the inflammation of endothelial cells
damage of blood-brain barrier in mouse stroke and increasing the serum levels of lipids, TGs,
model following transient middle cerebral artery LDL, IL-1, IL-6, MMP-9, and CRP [175].
occlusion, indicating that circDLGAP4 is However, another study reported a contrary result
involved in cerebral ischemia which provide a that circular noncoding RNA ANRIL can protect
novel therapeutic target for treatment of cerebral atherosclerosis [146]. In addition, a screening
ischemic diseases [170]. Based on mouse model study identified that has-circ-0003575 was sig-
subjected to transient middle cerebral artery nificantly raised in oxLDL-induced human
occlusion, a research team used circular RNA umbilical vein endothelial cells (HUVECs) and
25 The Role of Circular RNAs in Cerebral Ischemic Diseases: Ischemic Stroke and Cerebral… 319

knockdown of has-circ-0003575 was able to References

advance the angiogenesis, indicating its role in
atheroprotection [176]. All of these observations 1. Turley KR, Toledo-Pereyra LH, Kothari RU (2005)
Molecular mechanisms in the pathogenesis and
suggest a potential role of circular RNAs in treatment of acute ischemic stroke. J Investig Surg
atherosclerosis-­related cerebrovascular diseases 18(4):207–218
such as cerebral ischemia. 2. Eltzschig HK, Eckle T (2011) Ischemia and reper-
fusion--from mechanism to translation. Nat Med
17(11):1391–1401
3. Ma Z, Xin Z, Di W et al (2017) Melatonin and
5 Conclusion and Perspective mitochondrial function during ischemia/reperfusion
injury. Cell Mol Life Sci 74(21):3989–3998
Cerebral ischemic diseases like ischemic stroke 4. Benjamin EJ, Virani SS, Callaway CW et al (2018)
Heart disease and stroke statistics-2018 update:
and cerebral ischemia/reperfusion injury can lead a report from the American Heart Association.
to extremely high mortality and disability due to Circulation 137:e67. https://doi.org/10.1161/
the lack of effective interventions. Although CIR.0000000000000558
thrombolysis and the other supportive medica- 5. Hentze MW, Preiss T (2013) Circular RNAs: splic-
ing’s enigma variations. EMBO J 32(7):923–925
tion can improve the prognosis of the patients, 6. Xie L, Mao M, Xiong K et al (2017) Circular RNAs:
the overall high mortality and disability has not a novel player in development and disease of the
been significantly ameliorated. The major mech- central nervous system. Front Cell Neurosci 11:354
anisms for the onset and development of cerebral 7. Earls LR, Westmoreland JJ, Zakharenko SS (2014)
Non-coding RNA regulation of synaptic plasticity
ischemic diseases include inflammation, oxida- and memory: implications for aging. Ageing Res
tive stress, cellular apoptosis, calcium overload, Rev 17:34–42
and the disorders of energy metabolism. There 8. Rybak-Wolf A, Stottmeister C, Glazar P et al (2015)
are lots of researches regarding the treatment and Circular RNAs in the mammalian brain are highly
abundant, conserved, and dynamically expressed.
underlying mechanisms of cerebral ischemic dis- Mol Cell 58(5):870–885
eases, in order to uncover potential therapeutics. 9. Lees KR, Bath PM, Naylor AR (2000) ABC of
Among these therapeutics, the newly studied cir- arterial and venous disease. Secondary preven-
cular RNAs hold the potential to improve the tion of transient ischaemic attack and stroke. BMJ
320(7240):991–994
quality of life of patients suffering from cerebral 10. Edwards JD, Kapral MK, Fang J et al (2017) Trends
ischemic diseases someday. Ying Bai and his in long-term mortality and morbidity in patients
team have demonstrated that upregulation of with no early complications after stroke and tran-
circDLGAP4 can significantly relieve the dam- sient ischemic attack. J Stroke Cerebrovasc Dis
26(7):1641–1645
age of blood-brain barrier in rodent model fol- 11. Grysiewicz RA, Thomas K, Pandey DK (2008)
lowing transient middle cerebral artery occlusion, Epidemiology of ischemic and hemorrhagic stroke:
which indicated the potential therapeutic func- incidence, prevalence, mortality, and risk factors.
tion of circular RNAs in cerebral ischemic dis- Neurol Clin 26(4):871–895 vii
12. Woodruff TM, Thundyil J, Tang SC et al (2011)
eases [170]. However, current investigations Pathophysiology, treatment, and animal and cel-
about the role of circular RNAs in cerebral isch- lular models of human ischemic stroke. Mol
emic diseases are insufficient in both quantity Neurodegener 6(1):11
and quality. In addition to searching for the exis- 13. Saver JL (2006) Time is brain–quantified. Stroke
37(1):263–266
tence of dysregulated circular RNAs in patients 14. Cobley JN, Fiorello ML, Bailey DM (2018) 13 rea-
and animal models of cerebral ischemic diseases, sons why the brain is susceptible to oxidative stress.
we should focus on the functions and underlying Redox Biol 15:490–503
mechanisms of circular RNAs in cerebral isch- 15. Jin R, Liu L, Zhang S et al (2013) Role of inflam-
mation and its mediators in acute ischemic stroke.
emic diseases in the future. J Cardiovasc Transl Res 6(5):834–851
16. Fricker M, Tolkovsky AM, Borutaite V et al (2018)
Competing Financial Interests The authors declare no Neuronal cell death. Physiol Rev 98(2):813–880
competing financial interests. 17. Ramos-Cabrer P, Campos F, Sobrino T et al (2011)
Targeting the ischemic penumbra. Stroke 42(1
Suppl):S7–S11
320 J. Yang et al.

18. Huang Q, Sun M, Li M et al (2017) Combination 34. Waje-Andreassen U, Krakenes J, Ulvestad E et al

of NAD(+) and NADPH offers greater neuroprotec- (2005) IL-6: an early marker for outcome in acute
tion in ischemic stroke models by relieving meta- ischemic stroke. Acta Neurol Scand 111(6):360–365
bolic stress. Mol Neurobiol 55:6063. https://doi. 35. Xu N, Zhang Y, Doycheva DM et al (2018)
org/10.1007/s12035-017-0809-7 Adiponectin attenuates neuronal apoptosis induced
19. Liu J, Wang LN (2015) Peroxisome proliferator-­ by hypoxia-ischemia via the activation of AdipoR1/
activated receptor gamma agonists for preventing APPL1/LKB1/AMPK pathway in neonatal rats.
recurrent stroke and other vascular events in patients Neuropharmacology 133:415–428
with stroke or transient ischaemic attack. Cochrane 36. Lossi L, Castagna C, Merighi A (2015) Neuronal
Database Syst Rev 8(10):CD010693. https://doi. cell death: an overview of its different forms in
org/10.1002/14651858.CD010693.pub3 central and peripheral neurons. Methods Mol Biol
20. Kim YS, Kim C, Jung KH et al (2017) Range of glu- 1254:1–18
cose as a glycemic variability and 3-month outcome 37. Shabanzadeh AP, D’Onofrio PM, Monnier PP et al
in diabetic patients with acute ischemic stroke. PLoS (2015) Targeting caspase-6 and caspase-8 to pro-
One 12(9):e0183894 mote neuronal survival following ischemic stroke.
21. Pan Y, Chen W, Jing J et al (2017) Pancreatic beta-­ Cell Death Dis 6:e1967
cell function and prognosis of nondiabetic patients 38. Yang W, Shen Y, Chen Y et al (2014) Mesencephalic
with ischemic stroke. Stroke 48(11):2999–3005 astrocyte-derived neurotrophic factor prevents neu-
22. Iadecola C, Anrather J (2011) The immunology of ron loss via inhibiting ischemia-induced apoptosis.
stroke: from mechanisms to translation. Nat Med J Neurol Sci 344(1–2):129–138
17(7):796–808 39. van Wijk SJ, Hageman GJ (2005) Poly(ADP-ribose)
23. Felger JC, Abe T, Kaunzner UW et al (2010) Brain polymerase-1 mediated caspase-independent cell
dendritic cells in ischemic stroke: time course, death after ischemia/reperfusion. Free Radic Biol
activation state, and origin. Brain Behav Immun Med 39(1):81–90
24(5):724–737 40. Zhao Y, Pan R, Li S et al (2014) Chelating intracel-
24. Kostulas N, Li HL, Xiao BG et al (2002) Dendritic lularly accumulated zinc decreased ischemic brain
cells are present in ischemic brain after permanent injury through reducing neuronal apoptotic death.
middle cerebral artery occlusion in the rat. Stroke Stroke 45(4):1139–1147
33(4):1129–1134 41. Rengan R, Ochs HD, Sweet LI et al (2000) Actin
25. Gelderblom M, Leypoldt F, Steinbach K et al (2009) cytoskeletal function is spared, but apoptosis is
Temporal and spatial dynamics of cerebral immune increased, in WAS patient hematopoietic cells.
cell accumulation in stroke. Stroke 40(5):1849–1857 Blood 95(4):1283–1292
26. Yilmaz A, Fuchs T, Dietel B et al (2009) Transient 42. Cevik O, Adiguzel Z, Baykal AT et al (2013) The
decrease in circulating dendritic cell precursors after apoptotic actions of platelets in acute ischemic
acute stroke: potential recruitment into the brain. stroke. Mol Biol Rep 40(12):6721–6727
Clin Sci (Lond) 118(2):147–157 43. Astrup J, Symon L, Branston NM et al (1977)
27. Yilmaz G, Arumugam TV, Stokes KY et al (2006) Cortical evoked potential and extracellular K+
Role of T lymphocytes and interferon-gamma in and H+ at critical levels of brain ischemia. Stroke
ischemic stroke. Circulation 113(17):2105–2112 8(1):51–57
28. Hallenbeck JM (1996) Significance of the inflam- 44. Liu R, Yuan H, Yuan F et al (2012) Neuroprotection
matory response in brain ischemia. Acta Neurochir targeting ischemic penumbra and beyond for
Suppl 66:27–31 the treatment of ischemic stroke. Neurol Res
29. Zhai SK, Volgina VV, Sethupathi P et al (2014) 34(4):331–337
Chemokine-mediated B cell trafficking dur- 45. Mies G, Iijima T, Hossmann KA (1993) Correlation
ing early rabbit GALT development. J Immunol between peri-infarct DC shifts and ischaemic neuro-
193(12):5951–5959 nal damage in rat. Neuroreport 4(6):709–711
30. Fernandez EJ, Lolis E (2002) Structure, function, 46. White SH, Brisson CD, Andrew RD (2012)
and inhibition of chemokines. Annu Rev Pharmacol Examining protection from anoxic depolarization by
Toxicol 42:469–499 the drugs dibucaine and carbetapentane using whole
31. Murphy PM (2002) International Union of cell recording from CA1 neurons. J Neurophysiol
Pharmacology. XXX. Update on chemokine recep- 107(8):2083–2095
tor nomenclature. Pharmacol Rev 54(2):227–229 47. Yoshie T, Ueda T, Takada T et al (2018) Effects of
32. Mirabelli-Badenier M, Braunersreuther V, Viviani pretreatment cerebral blood volume and time to
GL et al (2011) CC and CXC chemokines are piv- recanalization on clinical outcomes in endovascular
otal mediators of cerebral injury in ischaemic stroke. Thrombectomy for acute ischemic stroke. J Stroke
Thromb Haemost 105(3):409–420 Cerebrovasc Dis 27:1802. https://doi.org/10.1016/j.
33. Herrmann O, Tarabin V, Suzuki S et al (2003) jstrokecerebrovasdis.2018.02.009
Regulation of body temperature and neuroprotection 48. White CJ, Abou-Chebl A, Cates CU et al (2011)
by endogenous interleukin-6 in cerebral ischemia. Stroke intervention: catheter-based therapy for acute
J Cereb Blood Flow Metab 23(4):406–415 ischemic stroke. J Am Coll Cardiol 58(2):101–116
25 The Role of Circular RNAs in Cerebral Ischemic Diseases: Ischemic Stroke and Cerebral… 321

49. Powers WJ, Derdeyn CP, Biller J et al (2015) 2015 62. Wu P, Zuo X, Ji A (2012) Stroke-induced microR-
American Heart Association/American Stroke NAs: the potential therapeutic role for stroke. Exp
Association focused update of the 2013 guidelines Ther Med 3(4):571–576
for the early management of patients with acute 63. Bhalala OG, Srikanth M, Kessler JA (2013) The
ischemic stroke regarding endovascular treatment: emerging roles of microRNAs in CNS injuries. Nat
a guideline for healthcare professionals from the Rev Neurol 9(6):328–339
American Heart Association/American Stroke 64. Vemuganti R (2010) The MicroRNAs and stroke: no
Association. Stroke 46(10):3020–3035 need to be coded to be counted. Transl Stroke Res
50. Moussaddy A, Demchuk AM, Hill MD (2018) 1(3):158–160
Thrombolytic therapies for ischemic stroke: triumphs 65. Zheng HW, Wang YL, Lin JX et al (2012) Circulating
and future challenges. Neuropharmacology 134:272. MicroRNAs as potential risk biomarkers for hema-
https://doi.org/10.1016/j.neuropharm.2017.11.010 toma enlargement after intracerebral hemorrhage.
51. Del Zoppo GJ, Saver JL, Jauch EC et al (2009) CNS Neurosci Ther 18(12):1003–1011
Expansion of the time window for treatment of acute 66. Mishima T, Mizuguchi Y, Kawahigashi Y et al
ischemic stroke with intravenous tissue plasmino- (2007) RT-PCR-based analysis of microRNA
gen activator: a science advisory from the American (miR-1 and -124) expression in mouse CNS. Brain
Heart Association/American Stroke Association. Res 1131(1):37–43
Stroke 40(8):2945–2948 67. Sun Y, Gui H, Li Q et al (2013) MicroRNA-124 pro-
52. Wapshott T, Blum B, Kelsey W et al (2017) tects neurons against apoptosis in cerebral ischemic
Investigation of gender differences and exclusive cri- stroke. CNS Neurosci Ther 19(10):813–819
teria in a diabetic acute ischemic stroke population 68. Wang ZS, Luo P, Dai SH et al (2013) Salvianolic
treated with recombinant tissue-type plasminogen acid B induces apoptosis in human glioma U87 cells
activator (rtPA). J Vasc Interv Neurol 9(6):26–32 through p38-mediated ROS generation. Cell Mol
53. Guo Z, Yu S, Chen X et al (2018) Suppression of Neurobiol 33(7):921–928
NLRP3 attenuates hemorrhagic transformation after 69. Lv H, Wang L, Shen J et al (2015) Salvianolic acid
delayed rtPA treatment in thromboembolic stroke B attenuates apoptosis and inflammation via SIRT1
rats: involvement of neutrophil recruitment. Brain activation in experimental stroke rats. Brain Res Bull
Res Bull 137:229–240 115:30–36
54. Shere A, Goyal H (2017) Cannabis can augment 70. Li D, Shao Z, Vanden Hoek TL et al (2007)
thrombolytic properties of rtPA: intracranial hemor- Reperfusion accelerates acute neuronal death
rhage in a heavy cannabis user. Am J Emerg Med induced by simulated ischemia. Exp Neurol
35(12):1988 e1981–1988 e1982 206(2):280–287
55. Jovin TG, Chamorro A, Cobo E et al (2015) 71. Jung JE, Kim GS, Chen H et al (2010) Reperfusion
Thrombectomy within 8 hours after symptom onset and neurovascular dysfunction in stroke: from basic
in ischemic stroke. N Engl J Med 372(24):2296–2306 mechanisms to potential strategies for neuroprotec-
56. Berkhemer OA, Fransen PS, Beumer D et al (2015) tion. Mol Neurobiol 41(2–3):172–179
A randomized trial of intraarterial treatment for 72. The NINDS t-PA Stroke Study Group (1997)
acute ischemic stroke. N Engl J Med 372(1):11–20 Intracerebral hemorrhage after intravenous t-PA ther-
57. Campbell BC, Mitchell PJ, Kleinig TJ et al apy for ischemic stroke. Stroke 28(11):2109–2118
(2015) Endovascular therapy for ischemic stroke 73. Pan J, Konstas AA, Bateman B et al (2007)
with perfusion-imaging selection. N Engl J Med Reperfusion injury following cerebral ischemia:
372(11):1009–1018 pathophysiology, MR imaging, and potential thera-
58. Shafi N, Kasner SE (2011) Treatment of acute isch- pies. Neuroradiology 49(2):93–102
emic stroke: beyond thrombolysis and supportive 74. Gobin YP, Starkman S, Duckwiler GR et al
care. Neurotherapeutics 8(3):425–433 (2004) MERCI 1: a phase 1 study of mechani-
59. Thwaites JW, Reebye V, Mintz P et al (2012) Cellular cal embolus removal in cerebral ischemia. Stroke
replacement and regenerative medicine therapies in 35(12):2848–2854
ischemic stroke. Regen Med 7(3):387–395 75. Nanetti L, Raffaelli F, Vignini A et al (2011)
60. Deng W, Obrocka M, Fischer I et al (2001) In vitro Oxidative stress in ischaemic stroke. Eur J Clin
differentiation of human marrow stromal cells into Investig 41(12):1318–1322
early progenitors of neural cells by conditions that 76. Jin R, Yang G, Li G (2010) Inflammatory mecha-
increase intracellular cyclic AMP. Biochem Biophys nisms in ischemic stroke: role of inflammatory cells.
Res Commun 282(1):148–152 J Leukoc Biol 87(5):779–789
61. Yang Z, Zhu L, Li F et al (2014) Bone marrow stro- 77. Kristian T, Gido G, Kuroda S et al (1998) Calcium
mal cells as a therapeutic treatment for ischemic metabolism of focal and penumbral tissues in rats
stroke. Neurosci Bull 30(3):524–534 subjected to transient middle cerebral artery occlu-
sion. Exp Brain Res 120(4):503–509
322 J. Yang et al.

78. Wang GH, Lan R, Zhen XD et al (2014) An-Gong-­ 94. Wasserman JK, Yang H, Schlichter LC (2008) Glial
Niu-Huang Wan protects against cerebral isch- responses, neuron death and lesion resolution after
emia induced apoptosis in rats: up-regulation of intracerebral hemorrhage in young vs. aged rats. Eur
Bcl-2 and down-regulation of Bax and caspase-3. J Neurosci 28(7):1316–1328
J Ethnopharmacol 154(1):156–162 95. Berti R, Williams AJ, Moffett JR et al (2002)
79. Allen CL, Bayraktutan U (2009) Oxidative stress Quantitative real-time RT-PCR analysis of inflam-
and its role in the pathogenesis of ischaemic stroke. matory gene expression associated with ischemia-­
Int J Stroke 4(6):461–470 reperfusion brain injury. J Cereb Blood Flow Metab
80. Circu ML, Aw TY (2010) Reactive oxygen species, 22(9):1068–1079
cellular redox systems, and apoptosis. Free Radic 96. Caso JR, Moro MA, Lorenzo P et al (2007)
Biol Med 48(6):749–762 Involvement of IL-1beta in acute stress-induced
81. Jezek P, Hlavata L (2005) Mitochondria in homeo- worsening of cerebral ischaemia in rats. Eur
stasis of reactive oxygen species in cell, tissues, and Neuropsychopharmacol 17(9):600–607
organism. Int J Biochem Cell Biol 37(12):2478–2503 97. Baird GS (2011) Ionized calcium. Clin Chim Acta
82. Lee HL, Chen CL, Yeh ST et al (2012) Biphasic 412(9–10):696–701
modulation of the mitochondrial electron transport 98. Zheng J, Zeng X, Wang S (2015) Calcium ion as cel-
chain in myocardial ischemia and reperfusion. Am lular messenger. Sci China Life Sci 58(1):1–5
J Physiol Heart Circ Physiol 302(7):H1410–H1422 99. Costa C, Martella G, Picconi B et al (2006) Multiple
83. Crack PJ, Taylor JM (2005) Reactive oxygen species mechanisms underlying the neuroprotective effects
and the modulation of stroke. Free Radic Biol Med of antiepileptic drugs against in vitro ischemia.
38(11):1433–1444 Stroke 37(5):1319–1326
84. Chan PH (2001) Reactive oxygen radicals in signal- 100. Zhou X, Fan GC, Ren X et al (2007) Overexpression
ing and damage in the ischemic brain. J Cereb Blood of histidine-rich Ca-binding protein protects against
Flow Metab 21(1):2–14 ischemia/reperfusion-induced cardiac injury.
85. Heo JH, Han SW, Lee SK (2005) Free radicals as Cardiovasc Res 75(3):487–497
triggers of brain edema formation after stroke. Free 101. Lopez-Caamal F, Oyarzun DA, Middleton RH et al
Radic Biol Med 39(1):51–70 (2014) Spatial quantification of cytosolic Ca(2)(+)
86. Jian Liu K, Rosenberg GA (2005) Matrix metallo- accumulation in nonexcitable cells: an analytical
proteinases and free radicals in cerebral ischemia. study. IEEE/ACM Trans Comput Biol Bioinform
Free Radic Biol Med 39(1):71–80 11(3):592–603
87. Chen XM, Chen HS, Xu MJ et al (2013) Targeting 102. Li LH, Tian XR, Hu ZP (2015) The key target of
reactive nitrogen species: a promising therapeutic neuroprotection after the onset of ischemic stroke:
strategy for cerebral ischemia-reperfusion injury. secretory pathway Ca(2+)-ATPase 1. Neural Regen
Acta Pharmacol Sin 34(1):67–77 Res 10(8):1271–1278
88. Gyoneva S, Davalos D, Biswas D et al (2014) 103. Liu HJ, Yang JP, Wang CH et al (2009) Endoplasmic
Systemic inflammation regulates microg- reticulum in the penumbra following middle cere-
lial responses to tissue damage in vivo. Glia bral artery occlusion in the rabbit. Neurol Sci
62(8):1345–1360 30(3):227–232
89. Zaleska MM, Mercado ML, Chavez J et al (2009) 104. Butterick TA, Duffy CM, Lee RE et al (2014) Use of
The development of stroke therapeutics: prom- a caspase multiplexing assay to determine apoptosis
ising mechanisms and translational challenges. in a hypothalamic cell model. J Vis Exp 16. https://
Neuropharmacology 56(2):329–341 doi.org/10.3791/51305(86)
90. Ekdahl CT, Kokaia Z, Lindvall O (2009) Brain 105. Wawryk-Gawda E, Chylinska-Wrzos P, Lis-­
inflammation and adult neurogenesis: the dual role Sochocka M et al (2014) P53 protein in prolif-
of microglia. Neuroscience 158(3):1021–1029 eration, repair and apoptosis of cells. Protoplasma
91. Gleichman AJ, Carmichael ST (2014) Astrocytic 251(3):525–533
therapies for neuronal repair in stroke. Neurosci Lett 106. Feng L, Balakir R, Precht P et al (1999) Bcl-2 regu-
565:47–52 lates chondrocyte morphology and aggrecan gene
92. Fabian RH, Perez-Polo JR, Kent TA (2000) expression independent of caspase activation and
Electrochemical monitoring of superoxide anion full apoptosis. J Cell Biochem 74(4):576–586
production and cerebral blood flow: effect of inter- 107. Daehn IS, Varelias A, Rayner TE (2010)
leukin-­1 beta pretreatment in a model of focal isch- T-lymphocyte-induced, Fas-mediated apoptosis is
emia and reperfusion. J Neurosci Res 60(6):795–803 associated with early keratinocyte differentiation.
93. Yan YP, Sailor KA, Lang BT et al (2007) Monocyte Exp Dermatol 19(4):372–380
chemoattractant protein-1 plays a critical role in 108. Zhou J, Du T, Li B et al (2015) Crosstalk between
neuroblast migration after focal cerebral ischemia. MAPK/ERK and PI3K/AKT signal pathways
J Cereb Blood Flow Metab 27(6):1213–1224 during brain ischemia/reperfusion. ASN Neuro
7(5):175909141560246
25 The Role of Circular RNAs in Cerebral Ischemic Diseases: Ischemic Stroke and Cerebral… 323

109. Si Y, Bao H, Han L et al (2013) Dexmedetomidine 122. Xia WG, Zheng CJ, Zhang X et al (2017) Effects of
protects against renal ischemia and reperfusion “nourishing liver and kidney” acupuncture therapy
injury by inhibiting the JAK/STAT signaling activa- on expression of brain derived neurotrophic factor
tion. J Transl Med 11:141 and synaptophysin after cerebral ischemia reperfu-
110. Tabassum R, Vaibhav K, Shrivastava P et al (2015) sion in rats. J Huazhong Univ Sci Technolog Med
Perillyl alcohol improves functional and histologi- Sci 37(2):271–278
cal outcomes against ischemia-reperfusion injury 123. Esteller M (2011) Non-coding RNAs in human dis-
by attenuation of oxidative stress and repression of ease. Nat Rev Genet 12(12):861–874
COX-2, NOS-2 and NF-kappaB in middle cerebral 124. Consortium EP (2012) An integrated encyclopedia
artery occlusion rats. Eur J Pharmacol 747:190–199 of DNA elements in the human genome. Nature
111. Field JM, Hazinski MF, Sayre MR et al (2010) 489(7414):57–74
Part 1: executive summary: 2010 American Heart 125. Patil VS, Zhou R, Rana TM (2014) Gene regulation
Association guidelines for cardiopulmonary by non-coding RNAs. Crit Rev Biochem Mol Biol
resuscitation and emergency cardiovascular care. 49(1):16–32
Circulation 122(18 Suppl 3):S640–S656 126. Nigro JM, Cho KR, Fearon ER et al (1991)
112. International Liaison Committee on R (2005) 2005 Scrambled exons. Cell 64(3):607–613
International consensus on cardiopulmonary resus- 127. Guo JU, Agarwal V, Guo H et al (2014) Expanded
citation and emergency cardiovascular care science identification and characterization of mammalian
with treatment recommendations. Part 5: acute coro- circular RNAs. Genome Biol 15(7):409
nary syndromes. Resuscitation 67(2–3):249–269 128. Sanger HL, Klotz G, Riesner D et al (1976) Viroids
113. Hypothermia after Cardiac Arrest Study G (2002) are single-stranded covalently closed circular
Mild therapeutic hypothermia to improve the neu- RNA molecules existing as highly base-paired
rologic outcome after cardiac arrest. N Engl J Med rod-like structures. Proc Natl Acad Sci U S A
346(8):549–556 73(11):3852–3856
114. Bernard SA, Gray TW, Buist MD et al (2002) 129. Rong D, Sun H, Li Z et al (2017) An emerging func-
Treatment of comatose survivors of out-of-hospital tion of circRNA-miRNAs-mRNA axis in human dis-
cardiac arrest with induced hypothermia. N Engl eases. Oncotarget 8(42):73271–73281
J Med 346(8):557–563 130. Wang PL, Bao Y, Yee MC et al (2014) Circular RNA
115. Strbian D, Kovanen PT, Karjalainen-Lindsberg ML is expressed across the eukaryotic tree of life. PLoS
et al (2009) An emerging role of mast cells in cerebral One 9(6):e90859
ischemia and hemorrhage. Ann Med 41(6):438–450 131. Salzman J, Chen RE, Olsen MN et al (2013) Cell-­
116. Polyak E, Ostrovsky J, Peng M et al (2018) type specific features of circular RNA expression.
N-acetylcysteine and vitamin E rescue animal lon- PLoS Genet 9(9):e1003777
gevity and cellular oxidative stress in pre-clinical 132. Jeck WR, Sorrentino JA, Wang K et al (2013)
models of mitochondrial complex I disease. Mol Circular RNAs are abundant, conserved, and associ-
Genet Metab 123:449. https://doi.org/10.1016/j. ated with ALU repeats. RNA 19(2):141–157
ymgme.2018.02.013 133. Ashwal-Fluss R, Meyer M, Pamudurti NR et al
117. Kim JH, Kim YC, Nahm FS et al (2017) The ther- (2014) circRNA biogenesis competes with pre-­
apeutic effect of vitamin C in an animal model of mRNA splicing. Mol Cell 56(1):55–66
complex regional pain syndrome produced by pro- 134. You X, Vlatkovic I, Babic A et al (2015) Neural cir-
longed hindpaw ischemia-reperfusion in rats. Int cular RNAs are derived from synaptic genes and reg-
J Med Sci 14(1):97–101 ulated by development and plasticity. Nat Neurosci
118. Chan SL, Cipolla MJ (2017) Treatment with low dose 18(4):603–610
fasudil for acute ischemic stroke in chronic hyperten- 135. Memczak S, Jens M, Elefsinioti A et al (2013)
sion. J Cereb Blood Flow Metab 37(9):3262–3270 Circular RNAs are a large class of animal RNAs
119. Koinig H, Vornik V, Rueda C et al (2001) Lubeluzole with regulatory potency. Nature 495(7441):333–338
inhibits accumulation of extracellular glutamate in 136. Bartel DP (2009) MicroRNAs: target recognition
the hippocampus during transient global cerebral and regulatory functions. Cell 136(2):215–233
ischemia. Brain Res 898(2):297–302 137. Capel B, Swain A, Nicolis S et al (1993) Circular
120. Zhang Y, Jin Y, Behr MJ et al (2005) Behavioral transcripts of the testis-determining gene Sry in
and histological neuroprotection by tamoxifen after adult mouse testis. Cell 73(5):1019–1030
reversible focal cerebral ischemia. Exp Neurol 138. Hansen TB, Jensen TI, Clausen BH et al (2013)
196(1):41–46 Natural RNA circles function as efficient microRNA
121. Wu PF, Zhang Z, Wang F et al (2010) Natural com- sponges. Nature 495(7441):384–388
pounds from traditional medicinal herbs in the treat- 139. Xu H, Guo S, Li W et al (2015) The circular RNA
ment of cerebral ischemia/reperfusion injury. Acta Cdr1as, via miR-7 and its targets, regulates insulin
Pharmacol Sin 31(12):1523–1531 transcription and secretion in islet cells. Sci Rep
5:12453
324 J. Yang et al.

140. Zheng Q, Bao C, Guo W et al (2016) Circular RNA patients with preeclampsia. Cell Physiol Biochem
profiling reveals an abundant circHIPK3 that regu- 39(4):1380–1390
lates cell growth by sponging multiple miRNAs. Nat 156. Zhang YG, Yang HL, Long Y et al (2016) Circular
Commun 7:11215 RNA in blood corpuscles combined with plasma
141. Salzman J, Gawad C, Wang PL et al (2012) Circular protein factor for early prediction of pre-eclampsia.
RNAs are the predominant transcript isoform from BJOG 123(13):2113–2118
hundreds of human genes in diverse cell types. PLoS 157. Lukiw WJ (2013) Circular RNA (circRNA) in
One 7(2):e30733 Alzheimer’s disease (AD). Front Genet 4:307
142. Zhang Y, Zhang XO, Chen T et al (2013) Circular 158. Bingol B, Sheng M (2011) Deconstruction for recon-
intronic long noncoding RNAs. Mol Cell struction: the role of proteolysis in neural plasticity
51(6):792–806 and disease. Neuron 69(1):22–32
143. Li Z, Huang C, Bao C et al (2015) Exon-intron cir- 159. Junn E, Lee KW, Jeong BS et al (2009) Repression
cular RNAs regulate transcription in the nucleus. Nat of alpha-synuclein expression and toxic-
Struct Mol Biol 22(3):256–264 ity by microRNA-7. Proc Natl Acad Sci U S A
144. Zhang XO, Wang HB, Zhang Y et al (2014) 106(31):13052–13057
Complementary sequence-mediated exon circular- 160. Danan M, Schwartz S, Edelheit S et al (2012)
ization. Cell 159(1):134–147 Transcriptome-wide discovery of circular RNAs in
145. Conn VM, Hugouvieux V, Nayak A et al (2017) A archaea. Nucleic Acids Res 40(7):3131–3142
circRNA from SEPALLATA3 regulates splicing of 161. Khoutorsky A, Yanagiya A, Gkogkas CG et al
its cognate mRNA through R-loop formation. Nat (2013) Control of synaptic plasticity and memory
Plants 3:17053 via suppression of poly(A)-binding protein. Neuron
146. Holdt LM, Stahringer A, Sass K et al (2016) Circular 78(2):298–311
non-coding RNA ANRIL modulates ribosomal RNA 162. Li XQ, Chen FS, Tan WF et al (2017) Elevated
maturation and atherosclerosis in humans. Nat microRNA-129-5p level ameliorates neuroinflam-
Commun 7:12429 mation and blood-spinal cord barrier damage after
147. Burd CE, Jeck WR, Liu Y et al (2010) Expression ischemia-reperfusion by inhibiting HMGB1 and
of linear and novel circular forms of an INK4/ARF-­ the TLR3-cytokine pathway. J Neuroinflammation
associated non-coding RNA correlates with athero- 14(1):205
sclerosis risk. PLoS Genet 6(12):e1001233 163. Zheng Y, Wang L, Chen M et al (2017) Upregulation
148. Geng HH, Li R, Su YM et al (2016) The circular of miR-130b protects against cerebral ischemic
RNA Cdr1as promotes myocardial infarction by injury by targeting water channel protein aquaporin
mediating the regulation of miR-7a on its target 4 (AQP4). Am J Transl Res 9(7):3452–3461
genes expression. PLoS One 11(3):e0151753 164. Han XR, Wen X, Wang YJ et al (2017) Protective
149. Du WW, Yang W, Chen Y et al (2017) Foxo3 circular effects of microRNA-431 against cerebral ischemia-­
RNA promotes cardiac senescence by modulating reperfusion injury in rats by targeting the Rho/Rho-­
multiple factors associated with stress and senes- kinase signaling pathway. J Cell Physiol 233:5895.
cence responses. Eur Heart J 38(18):1402–1412 https://doi.org/10.1002/jcp.26394
150. Wang K, Long B, Liu F et al (2016) A circular RNA 165. Tian F, Yuan C, Hu L et al (2017) MicroRNA-93
protects the heart from pathological hypertrophy inhibits inflammatory responses and cell apopto-
and heart failure by targeting miR-223. Eur Heart sis after cerebral ischemia reperfusion by targeting
J 37(33):2602–2611 interleukin-1 receptor-associated kinase 4. Exp Ther
151. Peng L, Yuan XQ, Li GC (2015) The emerging land- Med 14(4):2903–2910
scape of circular RNA ciRS-7 in cancer (review). 166. Zhang X, Tang X, Liu K et al (2017) Long noncod-
Oncol Rep 33(6):2669–2674 ing RNA Malat1 regulates cerebrovascular patholo-
152. Hansen TB, Kjems J, Damgaard CK (2013) gies in ischemic stroke. J Neurosci 37(7):1797–1806
Circular RNA and miR-7 in cancer. Cancer Res 167. Yan H, Rao J, Yuan J et al (2017) Long non-coding
73(18):5609–5612 RNA MEG3 functions as a competing endogenous
153. Wang X, Zhang Y, Huang L et al (2015) Decreased RNA to regulate ischemic neuronal death by target-
expression of hsa_circ_001988 in colorectal cancer ing miR-21/PDCD4 signaling pathway. Cell Death
and its clinical significances. Int J Clin Exp Pathol Dis 8(12):3211
8(12):16020–16025 168. Hou XX, Cheng H (2018) Long non-coding RNA
154. Nair AA, Niu N, Tang X et al (2016) Circular RNAs RMST silencing protects against middle cerebral
and their associations with breast cancer subtypes. artery occlusion (MCAO)-induced ischemic stroke.
Oncotarget 7(49):80967–80979 Biochem Biophys Res Commun 495(4):2602–2608
155. Qian Y, Lu Y, Rui C et al (2016) Potential signifi- 169. Wang J, Cao B, Han D et al (2017) Long non-­
cance of circular RNA in human placental tissue for coding RNA H19 induces cerebral ischemia reperfu-
25 The Role of Circular RNAs in Cerebral Ischemic Diseases: Ischemic Stroke and Cerebral… 325

sion injury via activation of autophagy. Aging Dis 173. Yamauchi H, Nishii R, Higashi T et al (2011) Silent
8(1):71–84 cortical neuronal damage in atherosclerotic disease
170. Bai Y, Zhang Y, Han B et al (2018) Circular RNA of the major cerebral arteries. J Cereb Blood Flow
DLGAP4 ameliorates ischemic stroke outcomes Metab 31(3):953–961
by targeting miR-143 to regulate endothelial-­ 174. Schaller B, Graf R (2004) Cerebral ischemia and
mesenchymal transition associated with blood-brain reperfusion: the pathophysiologic concept as a basis
barrier integrity. J Neurosci 38(1):32–50 for clinical therapy. J Cereb Blood Flow Metab
171. Mehta SL, Pandi G, Vemuganti R (2017) Circular 24(4):351–371
RNA expression profiles Alter significantly in 175. Song CL, Wang JP, Xue X et al (2017) Effect of circu-
mouse brain after transient focal ischemia. Stroke lar ANRIL on the inflammatory response of vascular
48(9):2541–2548 endothelial cells in a rat model of coronary athero-
172. Lin SP, Ye S, Long Y et al (2016) Circular RNA sclerosis. Cell Physiol Biochem 42(3):1202–1212
expression alterations are involved in OGD/R-­ - 176. Li CY, Ma L, Yu B (2017) Circular RNA hsa_
induced neuron injury. Biochem Biophys Res circ_0003575 regulates oxLDL induced vascular
Commun 471(1):52–56 endothelial cells proliferation and angiogenesis.
Biomed Pharmacother 95:1514–1519
 Part VII
Circular RNAs in Plants and in Archaea
 CircRNAs in Plants
26
Xuelei Lai, Jérémie Bazin, Stuart Webb,
Martin Crespi, Chloe Zubieta, and Simon J. Conn

Abstract in model plant systems show distinct features

Circular RNAs (circRNAs) are covalently of plant circRNAs compared with those from
closed, single-stranded transcripts that are animals, including putative roles in stress
ubiquitously expressed in all eukaryotes and response, differences in expression patterns,
even prokaryotic archaea. Although once and novel biogenesis mechanisms. This pro-
regarded as splicing artifacts, circRNAs are a vides a great opportunity to broaden our
novel class of regulatory molecules with knowledge of circRNAs using plant model
diverse biological functions, including regula- systems, such as Arabidopsis and rice, which
tion of transcription, modulation of alternative are ideal for phenotypic characterization and
splicing, and binding of miRNAs and proteins. genetic studies. In this review, we summarize
The majority of studies of circRNAs have current knowledge of plant circRNAs, discuss
been performed in animals with a focus on the their identification and biogenesis, describe
biogenesis, function, and mechanistic charac- potential functions, and propose future per-
terization of these molecules. In contrast, the spectives for plant circRNA study.
study of circRNAs in plants is just emerging.
Interestingly, recent circRNA profiling studies Keywords
circRNAs · Plants · Transcriptomics ·
Genome-wide profiling
Author contributed equally with all other contributors.
Xuelei Lai and Jeremie Bazin
X. Lai · C. Zubieta (*)
Laboratoire de Physiologie Cellulaire et Végétale,
CNRS Univ. Grenoble Alpes, CEA, INRA, BIG 1 Introduction
Grenoble, Grenoble, France
e-mail: [email protected] CircRNAs are a novel type of endogenous and
J. Bazin · M. Crespi largely noncoding RNA. Unlike their cognate
Institute of Plant Sciences Paris-Saclay, IPS2, messenger RNA (mRNA), circRNAs possess a
CNRS-INRA-University of Paris Sud, Paris-Diderot
and Evry, University of Paris Saclay, covalent bond linking their 3′ and 5′ ends and, as
Gif sur Yvette, France such, are hyperstable RNA molecules [1]. While
S. Webb · S. J. Conn (*) human circRNAs were initially reported almost
Flinders Centre for Innovation in Cancer, College of three decades ago [2], they were regarded as tran-
Medicine & Public Health, Flinders University, scriptional noise derived from splicing errors [3].
Bedford Park, Australia
However, in recent years they have received
e-mail: [email protected]

© Springer Nature Singapore Pte Ltd. 2018 329

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_26
330 X. Lai et al.

increased attention partly due to advances in where [14, 19–24] and by accompanying chap-
next-generation sequencing techniques and ters of this book. First, we summarize profiling of
highly efficient bioinformatics algorithms for circRNAs from different plant species and their
their discrimination and mapping. This has distinct features. We then discuss biogenesis of
unequivocally shown that circRNAs are abundant plant circRNAs based on studies in animals.
and ubiquitous among eukaryotes [4], including Finally, we propose potential functions of plant
yeast, C. elegans, mouse, human and plants, and circRNAs and future perspectives of circRNA
even prokaryotic archaea [5], adding a novel studies in plants.
class of RNA molecules with diverse regulatory
functions to the extensive collection of noncod-
ing RNAs present in eukaryotes [6]. 2 Plant circRNA Profiling
The majority of circRNAs exhibit cell-type-,
tissue-, or developmental stage-specific expres- 2.1 Experimental circRNA
sion patterns [7–10] and are highly regulated Detection
[11], implying their deliberate production by the
cell. In addition, some circRNAs are syntenically Next-generation sequencing (NGS) has greatly
conserved. These findings are suggestive of expanded the diversity of detected circRNAs. A
potential functionality of circRNAs. Indeed, number of strategies exist to enrich for circRNAs
emerging evidence has shown that specific cir- from RNA extraction through to library prepara-
cRNAs are involved in regulation of gene expres- tion. CircRNAs are known to be depleted in
sion at the transcriptional and posttranscriptional poly(A) + libraries (mRNA-seq) because they
levels. For example, increasing numbers of cir- lack poly(A) tails. In addition, by their chemical
cRNAs have been shown to act as miRNA nature, they do not exhibit free termini.
sponges, able to sequester and prevent miRNAs Consequently, intact circRNAs are unavailable
from binding to their corresponding target genes for direct ligation of RNA adapters common
[9, 12], or in one particular case, stabilizing them among several RNA-seq library preparation pro-
to improve miRNA targeting [13]. Some cir- tocols. However, they are known to be present in
cRNAs contain multiple binding sites for RNA-­ transcriptomic datasets produced from rRNA-­
binding proteins (RBP) and were therefore depleted samples and RNA ligation-independent
proposed to act as RBP sponges [14]. More protocols, including total RNA-seq [25].
recently, several studies show that a subpopula- Fortuitously, many datasets are publicly available
tion of circRNAs retain an internal ribosomal in NGS repositories for a number of plants spe-
entry site (IRES) which enables ribosome occu- cies including model organisms such as
pancy and can result in their translation into pro- Arabidopsis and various crops. The first genome-­
teins/peptides [15–18]. However, no overarching wide identification and characterization of cir-
functionality has been ascribed to the majority of cRNAs in plants were conducted during the last
circRNAs, and the aforementioned functions few years and exploited these publicly available
may or may not operate in different organisms total RNA-seq datasets from Arabidopsis, rice,
such as plants. and barley [26, 27]. As RNA splicing in plants
Compared with the extensive studies of cir- differs from that of animals, dedicated predictive
cRNAs in animals, the systematic characteriza- algorithms needed to be designed for plants.
tion of circRNAs in plants has only recently Thousands of circRNA were identified in each of
begun and focused primarily on model plant sys- these species, with the vast majority of them
tems such as Arabidopsis and rice. In this review, deriving from mRNA-coding exons. Strikingly,
we concentrate on what is known with respect to the parent genes of more than 700 exonic rice and
the profiling, biogenesis, and function of cir- Arabidopsis circRNAs were orthologous between
cRNAs in plants. Studies of circRNAs from other the two species, suggesting a conservation of cir-
systems have been extensively reviewed else- cRNAs in plants, as had been shown in mammals
26 CircRNAs in Plants 331

[9, 27, 28]. However, most circRNAs are approach, algorithms are used to identify back-­
expressed at low levels, and detecting and quanti- splicing junctions from the mapping information
fying them in total RNA-seq are challenging. A of a multiple split-read alignment. This approach,
high sequencing depth is required to have a sig- termed “segmented read approach,” uses the
nificant number of reads supporting their expres- unmapped reads from aligner programs, such as
sion. An additional protocol has been developed Bowtie [36], extracts their two end segments, and
to greatly enrich for circRNAs, using the linear performs a second alignment. It then looks for
ribonuclease R (RNase R) prior to rRNA deple- cases where the two end segments are mapped
tion and RNA-seq library preparation. This within spliced exons in the opposite orientation,
approach was successfully used in plant species which indicates a putative back-splicing event.
and yielded high-depth interrogation of cir- Then, depending on the software and the dataset
cRNAs (and intron lariats and other paired used, the presence of a canonical splice junction,
RNAs) in Arabidopsis [29], rice [30], soybean the number of reads overlapping the splice junc-
[31], tomato [32], and kiwifruit [33]. tion, and the enrichment of sequences containing
the splice junction in RNase R-treated samples
are used as criteria to define high-confidence cir-
2.2 Bioinformatics Detection cRNAs. The latter tools have been most exten-
of Circular RNA sively used in plants where they have led to the
identification of thousands of circRNAs in mul-
As mentioned, exonic circRNAs are formed by tiple plant species. Most of the resources pro-
back-splicing reactions which consist of a cova- duced by these studies have been or will be
lent ligation event between a downstream 3′ implemented into web-based databases such as
donor splice site and an upstream 5′ acceptor PlantcircBase (http://ibi.zju.edu.cn/plantcir-
splice site from a linear pre-mRNA [1]. cbase/), in which users can browse over 90,000
Sequencing reads that span this backsplice junc- circRNAs detected among 8 plants species
tion are considered as chimeric and are routinely (Oryza sativa, Arabidopsis thaliana, Zea mays,
rejected by most standard read-mapping algo- Solanum lycopersicum, Glycine max, Gossypium
rithms. However, this feature can also be hirsutum, Triticum aestivum, Hordeum vulgare),
exploited to specifically detect exonic circular identify those with miRNA seed sites which
RNAs. Indeed, read fragments overlapping a could act as miRNA sponges, and probe compet-
known or de novo predicted splice junction with ing endogenous RNA networks (circRNA-­
opposite orientation are likely to arise from a miRNA-­mRNA) [37].
backsplice event. A number of tools have been
used to recover such features. These tools use
two basic approaches. The first one, called 2.3 Limitations and Challenges
“pseudo-reference based” relies on the recon-
struction of all possible circRNA sequences by All NGS approaches have inherent limitations,
shuffling the exon-exon junctions based on and one should use caution when interpreting this
genome annotation, prior to read mapping [34]. data to extrapolate common features of cir-
This requires an accurate genome annotation cRNAs. For instance, in a study reanalyzing plant
with a thorough knowledge of all alternatively RNA-seq datasets, Ye et al. [27] identified 12,037
spliced mRNA isoforms. One of the available and 6,012 circRNAs from rice root and
tools called KNIFE [34] has been very recently Arabidopsis leaf, respectively. In addition to
applied to publicly available and RNase R-treated exonic circRNAs, a large number of circRNAs
RNA-seq datasets from maize and detected 85 are derived from noncoding regions or spanning
high confidence circRNA candidates [35]. The two or more genes. However, they could not vali-
second group of algorithms does not directly rely date any of these experimentally, suggesting that
on prior knowledge of gene annotation. In this distinct genomic features, such as repetitive
332 X. Lai et al.

sequences or gene duplication that differ between cRNAs in some conditions, such as quaking
plants and animals, decrease the specificity of (QKI) [11], muscleblind (MBL) [42], adenosine
prediction algorithms that were originally devel- deaminase 1 (ADAR1) [45], FUS [46], and
oped for applications in animals. The develop- DHX9 [47]. Homolog proteins of these factors,
ment of dedicated tools for plant circRNA calling while identified in plants, have not yet been stud-
such as PcircRNA_finder [38] has at least par- ied, and their putative role in circRNA biogenesis
tially overcome such problems in the detection of remains to be investigated. Here we summarize
plant circRNAs. In addition, depending on the these homolog proteins in Arabidopsis
tool used and the cut-off applied, the number of (Table 26.2) and assess their potential functions
detected circRNAs can be very different and only in circRNA biogenesis in plants based on knowl-
partially overlap [39]. Indeed, benchmarking edge of their counterparts from animals.
analysis of five existing software on the same
RNA-seq datasets highlighted the large differ-
ence between tools, suggesting that several algo- 4  KI Homolog: KH Domain–
Q
rithms should ideally be combined to achieve Containing Proteins
reliable predictions [40]. A summary of plant cir- in Arabidopsis
cRNA profiling studies, including identification
methods, total circRNAs, stress treatments, and QKI regulates circRNA biogenesis during the
circRNAs with potential miRNA-binding sites, is human epithelial/mesenchymal transition [11]. It
given in Table 26.1. These data highlight the con- binds to the introns flanking the exon on the host
servation and distinct features of plant circRNAs RNA and homodimerizes; therefore, it could pro-
from different plant species. mote circRNA biogenesis by exon looping,
bringing the 3′ and 5′ termini into close proxim-
ity for circularization. At the protein level, QKI
3 Plant circRNA Biogenesis contains a STAR domain responsible for RNA
binding and homodimerization that is composed
CircRNAs in plants and animals can arise from of a central KH domain flanked by QUA1 and
exons, introns, and/or intergenic regions. In ani- QUA2 domains [48]. The two key features that
mals, circularized exons are typically bracketed are important for QKI’s function as a circRNA
by long introns that contain complementary regulator are its dimerization ability and RNA-­
sequences such as ALU elements [28, 41] and/or binding activity [11]. QKI dimerization involves
micro-repeat regions sufficient to promote intron the dimerization domain, QUA1 [49] and the
base-pairing [42, 43]. In contrast, yeast circRNAs RNA-binding domain, KH [36] which, together
could be generated through exon-containing lar- with the QUA2 domain, are responsible for RNA-­
iat precursors that lacked noticeable flanking binding activity [36]. The QKI dimer preferen-
intronic secondary structure or repeating tially binds to two A/U-rich motifs [50–52] that
sequences [44]. Unlike the animal circRNAs, can be on the same or separate RNA molecules
most plant circRNAs have limited repetitive and [49].
reverse complementary sequences in intronic QKI homolog proteins in Arabidopsis are
sequences flanking exons [27], suggesting that KH domain-containing proteins, which are
plants might favor different mechanisms of RNA-­binding proteins that have been shown to
­circRNA biogenesis compared with those of ani- participate in pre-mRNA processing [53] and to
mals. One such mechanism has been proposed in be important for various domains of plant devel-
maize, where transposons seem to play a role in opment and survival, including flowering regu-
the biogenesis of circRNAs [35]. lation [54, 55], stress responses [56–58], as well
Recently, several studies have shown that as hormone signaling [59]. There are 26 KH
RNA-binding proteins may serve as regulatory domain-­containing proteins in Arabidopsis, 5 of
activators or inhibitors in the formation of cir- which are highly similar to QKI (hereafter
Table 26.1 Identification and characteristics of circRNAs currently identified in plant species
Plant species Arabidopsis Kiwifruit Soybean Tomato Wheat Maize Rice Barley
Tissues Leaves Leaves Whole plants, Leaf, root, and Leaf, root, and Fruit Leaves of Seedling Root Mature leaf Leaves,
roots, stems, stem tissues stem wheat leaves and panicle grains
leaves, flowers, tissues seedlings tissues and grain
and siliques transfer
cells
Datasets or Reanalysis of RNA-­seq rRNA depletion rRNA depletion rRNA depletion rRNA depletion CircRNA RNase Reanalysis rRNA RNA-­seq
circRNA library published data (GEO and RNase and RNase R and RNase R and RNase enrichment R-seq, of published depletion (Genbank
prep RNA-seq data accession R treatment treatment treatment R treatment kit (cloud-seq 977 public RNA-seq data and RNase SRA297575)
(PRJNA218215) GSE43616) Inc.) maize (GenBank R treatment
RNA-­Seq PRJNA215013)
Bioinformatics BOWTIE2 MapSplice BOWTIE2 (v2.0.5), CIRI CIRI CIRI CIRI KNIFE, BOWTIE2 Self-­designed CIRI
program (v2.0.5) (version 2.0) find_circ program circ_finder, (v2.0.5) pipeline
(Memczak et al., CIRCexplorer2,
2013) CIRI
Treatment High-­light N/A Drought, salinity Pathogen N/A Chilling Dehydration N/A Pi-starvation N/A Micronutriants
conditions and heat infection stressed such as iron
(Pseudomonas and well- and zinc
syringae pv. watered
actinidiae) conditions
Total circRNA 6012 168 5861 (85.14% 3582 (21.44%/ 5372 (46.43% 854 (72.01%/ 88 (6.82%/ 5329 in total, 12,037 2354 (57.60%/ NA
count (exonic/ (85.70%/ (94.04%/NA /3.77%/ 14.55%/ /48.05%/ 3.63%/ 2.27%/60.22%/ 2804 high (50.46%/4.03% 2.55%/29.74%/
intronic/ 0.015% /5.95%/NA) 11.09%/NA); 64.01%/NA) 5.55%/NA) 24.35%/NA) 2.27%); 28.41% confident /5.86%/39.65%) NA);
intergenic/ /0.5%/13.76%) unique exon-intronic (mostly exonic) 1.66%
others) mitochondrial exon-intronic
and chloroplast
circRNAs
sere also
detected
circRNAs with 5.0% contain N/A 39 (0.67%) 9 have N/A 2134 (39.7%) 102 circRNAs Six of the circRNAs have 6.6% contain 235 exonic N/A
miRNA-binding potential more than 1 circRNAs with predicted differentially an average of potential circRNAs
sites miRNA-binding different contained miRNA-binding expressed 1.33 miRNA- miRNA-binding contain putative
sites miRNA-binding predicted sites for circRNAs binding sites sites miRNA-binding
sites binding sites 24 miRNAs have putative ranging from 1 sites, among
for 92 miRNAs. miRNA-binding to 9 which only
Of these sites (3 to 8 31 circRNAs
circRNAs, sites); had 2 to 6
only 352 had 26 miRNAs miRNA-binding
2 to 6 were predicted sites
miRNA-binding
sites
(continued)
Table 26.1 (continued)
Plant species Arabidopsis Kiwifruit Soybean Tomato Wheat Maize Rice Barley
Differential Temporal circRNAs showed Developmental The expression Tissue-specific 163 circRNAs 62 circRNAs N/A Temporal- and N/A Cellular level
expression expression developmental- stage- and of the majority expression: had significant showed stress-dependent alterations
pattern and specific expression tissue-specific of circRNAs 62.4% (484) of difference significant expression pattern across tissues
stress-dependent pattern in expression pattern was tissue-, leaf circRNAs, between the difference and in
expression Arabidopsis leaves, taxon-, and 83.5% (2647) of control and between the PEG response to
pattern; mutant e.g., circ- stage- specific root circRNAs chilling injury and control micronutrients
(tnr-1) AT1G29965 was pathogen and 72.2% group, 138 treatment groups, iron and zinc
background only expressed in infection (1563) of stem circRNAs containing 16
expression the early stage of dependent circRNAs, only upregulated, 25 upregulated
pattern leaf growth, while expression 2.7% (143) of circRNAs circRNAs and 46
circ-AT5G18590 the total downregulated downregulated
was only detected circRNAs were circRNAs
in the mature stage expressed in all
of leaf growth, and the tissues
circ-AT4G08300
was expressed only
in the senescence
process
Distinct A significantly N/A N/A Both exonic and 4451 (82.8%) N/A Predicted The flanking 27 rice exonic Parental genes Fluctuations of
features of positive intronic circRNAs were miRNA-binding intron length of circRNAs were with multiple circRNAs do
circRNAs correlation was circRNAs were generated from circRNAs the junctions of found to be exons are not correlate
observed for the significantly the paralogous contain more circRNAs is differentially preferentially with the levels
expression positively genes abundant significantly expressed under circularized; a of their
profiles of some correlated to miRNA-binding larger than that phosphate- large number of parental-loci
circRNAs and parent sites than those of linear sufficient and circRNA encoded linear
their parent genes protein-coding predicted from transcripts of starvation isoforms derived transcripts;
genes and other plant randomly conditions; a from alternative circRNA from
intronic species selected genes; significantly backsplicing mitochondria
circRNAs are a LINE1-like positive genome were
class of highly elements (LLEs) correlation was detected
remarkable and their reverse observed for the
regulators the complementary expression
parent genes pairs (LLERCPs) profiles of some
comparing to are significantly circRNAs and
that of exonic enriched in the their parent genes
circRNAs flanking regions
of circRNAs
Plant species Arabidopsis Kiwifruit Soybean Tomato Wheat Maize Rice Barley
Proposed N/A Potential N/A Regulate host Tissue Play roles in the Potential miRNA circRNAs are N/A A case study: Regulatory
function of posttranscriptional pathogen differentiation chilling injury sponges likely to be Overexpression role in
circRNAs regulators in the interactions in soybean regulation involved in the of micronutrient
senescence of modulation of Os08circ16564 homeostasis;
Arabidopsis leaves phenotypic reduces regulate both
variation by expression of its nuclear and
LLERCPs parental gene. organellar
Os08circ16564 gene
contain potential expression
binding sites of
miR172, a
miRNA plan
crucial role in
the development
of the rice
spikelet and
floral organ)
Reference [27] [64] [35] [33] [31] [32] [33] [35] [27] [30] [26]
336 X. Lai et al.

Table 26.2 Arabidopsis homologs of RBPs implicated in animal circRNA biogenesis

Features
important as
circRNA Putative/known
biogenesis function(s) of the
Binding sites/ circRNA biogenesis promoter/ Homolog/similar proteins homolog proteins
RBP motifs mechanism repressor in Arabidopsis in Arabidopsis
QKI [11] A/U-rich Promote circRNA Dimerization, KH domain-containing mRNA
motifs biogenesis by RNA binding proteins: AT4G26480, processing,
exon-looping AT5G56140, AT3G08620, flowering
AT1G09660, and regulation,
AT2G38610 stress response,
and hormone
signaling
MBL [42] MBL- Dimerization, No homologs in Putative
binding sites RNA binding Arabidopsis, closest activity of RNA
(expanded matches: Zinc finger binding or
CUG or CCCH domain-­containing nuclear acid
CCUG proteins: AT2G47850, binding
repeats) AT3G02830, AT3G06410,
[88,89] AT3G48440, AT5G16540,
AT5G18550, and
AT5G63260
FUS [46] FUS- RNA-binding No homologs in N/A
binding sites activity Arabidopsis, closest
matches: Glycine-rich
RNA-binding protein 3
(AT5G61030), TBP-
associated factor 15
(AT1G50300), and early
flowering 9 (AT5G16260)
DHX9 Inverted-­ Repress circRNA Nuclear RNA DExH-box ATP-­ ATP-dependent
[47] repeat Alu biogenesis by helicase activity, dependent RNA helicases: RNA helicase
elements acting as nuclear interact with AT2G35920, AT2G01130, activity
RNA resolvase ADAR, a AT1G77030, AT1G33390,
co-repressor AT1G48650, and
AT5G04895
ADAR1 N/A Repress circRNA RNA editing ADAR proteins can be N/A
[45] biogenesis with activity found in nearly all
putative metazoa but absent in all
mechanisms of (a) protozoa, yeast, and
compete with plants [90]
circRNA-
promoting factors,
such as MBL. (b)
Editing or
hyper-editing of
introns flanking
circRNAs

called Arabidopsis QKI-like proteins for sim- tures of both proteins in complex with RNA
plicity) (Table 26.2) [60]. Protein sequence were resolved, and key modules/residues
alignment suggests that Arabidopsis QKI-like responsible for homodimerization (α1 and α2 of
proteins share highly conserved amino acid res- QUA1 domain) and RNA binding (α3, α4 and
idues with those in the central STAR domains of β3 of KH domain, and α7 of QUA2 domain)
human QKI and GLD-1, a C. elegans QKI were identified [49] (Fig. 26.1). Interestingly,
homolog protein (Fig. 26.1). The crystal struc- all of these modules/residues are highly con-
26 CircRNAs in Plants 337

Fig. 26.1 Protein sequence alignment of Homo sapiens boxed and colored in white and red background; con-
QKI (Hs_QKI), C. elegans GLD-1 (Ce_GLD-1), and served residues are boxed and colored in red without
Arabidopsis QKI-like proteins. The STAR domain of QKI background; secondary structure elements are derived
and GLD-1 consists of QUA1, KH, and QUA2 subdo- from crystal structure of QKI (PDB: 4jvh) [49]; TT and
mains, which share highly conserved residues with that of T.T represent turns that connect defined secondary struc-
Arabidopsis QKI-like proteins. Identical residues are ture elements
338 X. Lai et al.

served in the Arabidopsis QKI-like proteins tion pattern was shown for a number of circRNA-­
(Fig. 26.1), suggesting that these proteins may mRNA-­miRNA pairs during M- to S- and G- to
share similar physiological functions with QKI, M-stages, suggesting a possible role for cir-
e.g., in circRNA biogenesis [11]. Although this cRNAs in leaf senescence in this network.
is an appealing hypothesis, no experimental Initial studies have attempted to identify cir-
data is available yet to support or refute this cRNAs from plants that may act as true miRNA
putative function. Extensive studies of QKI sponges, yet no circRNA with high density of
homologs in Arabidopsis would require charac- miRNA-binding sites as described for Sry and
terization of the RNA-binding motif in plants; ciRS-7 from mammals has been shown for a
mutant/overexpression lines of the Arabidopsis plant circRNA [12]. Further identification of cir-
QKI-like proteins; biochemical activity studies, cRNAs from plant species may reveal the pres-
such as homodimerization and RNA-binding ence of circRNAs that can act as miRNA sponges
activity; and biophysical studies, e.g., RNA- with multiple miRNA-binding sites, however, to
looping ability. date this has not been shown. Unlike in mam-
mals, miRNAs from plants exhibit much higher
sequence complementarity to their target site in
5 Plant circRNA Function order to cleave and/or negatively regulate mRNA
expression, facilitating the identification of
A number of distinct functional mechanisms for miRNA targets [65, 66]. The availability of bio-
animal circRNAs have been identified, and plant informatics prediction tools for miRNAs and
circRNAs may exhibit similar conserved func- their targets in Arabidopsis and the publicly
tions. These include miRNA sponging, transcrip- available databases such as miRTarBase (http://
tional modulation, translation of circRNAs into mirtarbase.mbc.nctu.edu.tw/) which collects val-
proteins/peptides, and altering protein function idated miRNA targets will facilitate scanning of
through direct protein binding. The consequences circRNAs for miRNA-binding sites [67, 68]. For
of these activities include altering cell cycle pro- a comparison of miRNA prediction tools for
gression [61], cell proliferation [62], and cell plants, see Srivastava, et al. [69]. These tools
migration [63], all of which are indispensable for have predicted numerous plant circRNAs which
plant growth and development. With this in mind, could bind miRNAs and impact miRNA-mRNA
we summarize the known plant circRNA func- regulatory networks; however, unlike their ani-
tions below, with scope for much greater func- mal counterparts, these putative circRNA-­
tional characterization of circRNAs in plants, miRNA interactions have yet to be experimentally
leveraging animal circRNA functional studies. validated. The development of new pipelines
incorporating prediction software, miRNA data-
bases, and new filtering procedures offers great
5.1 CircRNAs as miRNA Sponges promise for more robust prediction and valida-
tion of miRNA targets which will likely reveal
Based on the work in the mammalian field, cir- many new connections between miRNAs and cir-
cRNA function in plants has largely focused on cRNAs [68].
their potential role as miRNA sponges or as act- Some of the most powerful methods for iden-
ing in the miRNA pathway. Recent work by Liu tifying protein-RNA complexes are crosslinking
et al. [64] has investigated the putative circRNA-­ immune-precipitation sequencing (CLIP-seq)
miRNA-­mRNA network based on the hypothesis techniques. For example, to identify miRNA-­
that both mRNAs and circRNAs are targets of the circRNA species genome wide, AGO-CLIP tech-
same miRNA. Based on differentially expressed niques can be used [70]. Argonaute (AGO)
mRNAs, circRNAs, and miRNAs, an anticorrela- proteins are part of the RNA-induced silencing
26 CircRNAs in Plants 339

complex (RISC), and this complex recognizes role for circRNAs; however, the mechanism of
mRNA- or circRNA-containing sequences com- this remains to be elucidated.
plementary to the miRNA. These techniques
have been successfully performed in plants and
are one method to help identify circRNAs that 5.3  ircRNA in Gene Expression
C
may act as miRNA sponges [68]. CLIP experi- Regulation: A Mechanistic
ments targeting different RNA-binding proteins Study
(RBPs) involved in circRNA activity are also eas-
ily envisaged. While circRNAs exhibit differential expression
patterns in leaf senescence, photosynthesis, and
stress response, a mechanistic basis for the puta-
5.2 CircRNAs in Stress Response tive role of circRNAs in these processes has not
been described. To date, the only example of a
Plants, as sessile organisms, encounter various mechanism for circRNA function with respect to
environmental stresses, such as drought, heat, gene regulation was postulated by Conn et al.
salinity, cold, and pathogen infection [71]. To [75]. In this seminal study, the authors describe
tackle these environmental challenges, plants the differential expression of floral MADS gene-­
have evolved various sophisticated biological derived circRNAs under different temperature
pathways, in which massive gene expression conditions, demonstrate a positive correlation
reprogramming occurs. RNA molecules, such as with their parental splice variant for a few of the
microRNAs (miRNAs) and small interfering circRNAs, and propose a mechanism of R-loop-­
RNAs (siRNAs), have been shown to play critical mediated alternative splicing (AS), favoring gen-
roles in gene expression regulation under differ- eration of the circRNA and its associated mRNA
ent stress conditions in plants [72, 73]. construct. The SEPALLATA3 gene gives rise to
Interestingly, circRNAs have been shown to dis- the canonical mRNA that includes all exons, an
play stress-specific expression patterns in exon 6-skipped splice variant and a circRNA
Arabidopsis in the first report of plant circRNAs consisting of exon 6. Overexpression studies of
[27]; therefore, one intriguing question is whether SEP3 exon 6 circRNA and a control circRNA
or not circRNAs also play a role in stress demonstrated that the exon 6 circRNA was able
responses of plants or are the products of stress to alter the splicing balance, favoring its own bio-
responses on different RNA machineries such as genesis and the biogenesis of its associated
the spliceosome. mRNA at the cost of the canonical splice form.
Studies under different stress conditions These studies suggest the tantalizing possibility
including oxidative stress, drought, and nutrient that circRNAs may have general gene regulatory
deficiency have further revealed differential roles in AS via R-loop formation, although this
expression of circRNAs which in most cases remains to be proven on a wider scale. Indeed,
does not correlate with expression of their associ- over 60% of genes in Arabidopsis are alterna-
ated mRNA, further suggesting a functional role tively spliced, and the putative role of circRNAs
of circRNA in environmental and stress response. in splicing may provide a general role for these
For example, 62 circRNAs were identified that molecules. As tight control of mRNA and AS
exhibit differential expression under dehydration transcript abundance is a hallmark of many genes,
stress and are associated with photosynthesis and including key regulatory genes, profiling cir-
hormone signal pathways in wheat [74]. cRNAs in plant NGS datasets, examining the
Moreover, a number of circRNAs in rice and bar- potential role of specific circRNAs in modulating
ley have also been reported to respond to nutrient gene expression and AS, and determining the
depletion such as phosphate, iron, and zinc [26, effects of circRNAs at the phenotypic level are
27]. These studies suggest a posttranscriptional crucial challenges in the field.
340 X. Lai et al.

5.4 CircRNA as Biomarkers concerted effort among laboratories will illumi-

nate the functions and interactions of circRNAs
CircRNAs are highly stable molecules and can be in plant development and also offer potential par-
enriched in many cell types, thus making them allels to new functional roles in other
good candidates as biomarkers [19, 23, 76, 77]. eukaryotes.
Over the last few years, many studies have shown
the promising relevance of circRNAs as potential Acknowledgments This work was supported by
molecular markers for disease diagnosis and Australian Research Council Future Fellowship
(FT160100318 to S.C.), Action Thématique et Incitative
treatment in humans, with particular emphasis on sur Programme (ATIP)-Avenir (to C.Z.), Agence
various types of cancers [78–84]. In plants, vari- Nationale de la Recherche (project FloPiNet to C.Z. and
ous biomarkers have been studied and proven be X.L.), Grenoble Alliance for Integrated Structural Cell
valuable tools for both fundamental research and Biology (ANR-10-LABX-49-01 to C.Z.), and the
“Laboratoire d’Excellence (LABEX)” Saclay Plant
applied practices in crop breeding. For example, Sciences (SPS; ANR-10-LABX-40) and the ANR grant
biomarker genes can be harnessed to assess the SPLISIL, France (to M.C.).
response of plants to varying nitrogen conditions
and therefore are potential agronomic tools to Competing Financial Interests The authors declare no
monitor and optimize nitrogen fertilizer usage competing financial interests.
[85]. Another study suggested that biomarker
genes can be used to identify and differentiate
microbial pathogens [86]. Furthermore, meta-
bolic biomarkers allow the prediction of crop References
product quality [87]. circRNAs as biomarkers in
1. Chen LL (2016) The biogenesis and emerging roles of
plants are still a new concept but there exists a circular RNAs. Nat Rev Mol Cell Biol 17(4):205–211
potential utility for these molecules due to their 2. Nigro JM, Cho KR, Fearon ER et al (1991) Scrambled
long half-lives, resistance to common degrada- exons. Cell 64(3):607–613
tion pathways, and the ease and specificity of 3. Cocquerelle C, Mascrez B, Hetuin D et al (1993)
Mis-splicing yields circular RNA molecules. FASEB
detection. Only recently, Conn et al. suggested J 7(1):155–160
that circRNAs could function as bona fide bio- 4. Wang PL, Bao Y, Yee MC et al (2014) Circular RNA
markers of AS event for MADS box genes in is expressed across the eukaryotic tree of life. PLoS
Arabidopsis [75]. Hence, in future research on One 9(4):e95116
5. Danan M, Schwartz S, Edelheit S et al (2012)
the function of circRNA in plants, biomarker Transcriptome-wide discovery of circular RNAs in
function is an attractive topic to investigate. Archaea. Nucleic Acids Res 40(7):3131–3142
6. Kopp F, Mendell JT (2018) Functional classification
and experimental dissection of long noncoding RNAs.
Cell 172(3):393–407
6 Perspectives 7. Salzman J, Chen RE, Olsen MN et al (2013) Cell-type
specific features of circular RNA expression. PLoS
CircRNAs are the most contemporary class of Genet 9(9):e1003777
regulatory noncoding RNAs, and a plethora of 8. Salzman J, Gawad C, Wang PL et al (2012) Circular
RNAs are the predominant transcript isoform from
roles for these molecules in animals is emerging. hundreds of human genes in diverse cell types. PLoS
We propose leveraging this information to eluci- One 7(2):e30733
date the functions of specific circRNAs among 9. Memczak S, Jens M, Elefsinioti A et al (2013)
the thousands already identified in numerous Circular RNAs are a large class of animal RNAs with
regulatory potency. Nature 495(7441):333–338
plant species. With the ability to study whole 10. Gao Y, Wang J, Zhao F (2015) CIRI: an efficient and
organism function in higher eukaryotes, plants unbiased algorithm for de novo circular RNA identifi-
offer a perfect model to investigate the functions cation. Genome Biol 16:4
of these molecules. With the breadth of charac- 11. Conn SJ, Pillman KA, Toubia J et al (2015) The RNA
binding protein quaking regulates formation of cir-
terized functions for circRNAs, the cellular con- cRNAs. Cell 160(6):1125–1134
sequences can be recapitulated in plants. A
26 CircRNAs in Plants 341

12. Hansen TB, Jensen TI, Clausen BH et al (2013) 32. Zuo J, Wang Q, Zhu B et al (2016) Deciphering
Natural RNA circles function as efficient microRNA the roles of circRNAs on chilling injury in tomato.
sponges. Nature 495(7441):384–388 Biochem Biophys Res Commun 479(2):132–138
13. Piwecka M, Glazar P, Hernandez-Miranda LR et al 33. Wang ZP, Liu YF, Li DW et al (2017) Identification of
(2017) Loss of a mammalian circular RNA locus circular RNAs in kiwifruit and their species-specific
causes miRNA deregulation and affects brain func- response to bacterial canker pathogen invasion. Front
tion. Science 357(6357):eaam8526 Plant Sci 8:413
14. Greene J, Baird AM, Brady L et al (2017) Circular 34. Szabo L, Morey R, Palpant NJ et al (2015)
RNAs: biogenesis, function and role in human dis- Statistically based splicing detection reveals neural
eases. Front Mol Biosci 4:38 enrichment and tissue-specific induction of circu-
15. Pamudurti NR, Bartok O, Jens M et al (2017) lar RNA during human fetal development. Genome
Translation of CircRNAs. Mol Cell 66(1):9–21 e27 Biol 16:126
16. Granados-Riveron JT, Aquino-Jarquin G (2016) The 35. Chen L, Zhang P, Fan Y et al (2018) Circular RNAs
complexity of the translation ability of circRNAs. mediated by transposons are associated with tran-
Biochim Biophys Acta 1859(10):1245–1251 scriptomic and phenotypic variation in maize. New
17. Abe N, Matsumoto K, Nishihara M et al (2015) Phytol 217(3):1292–1306
Rolling circle translation of circular RNA in living 36. Feracci M, Foot JN, Grellscheid SN et al (2016)
human cells. Sci Rep 5:16435 Structural basis of RNA recognition and dimeriza-
18. Yang Y, Fan X, Mao M et al (2017) Extensive tion by the STAR proteins T-STAR and Sam68. Nat
translation of circular RNAs driven by N(6)- Commun 7:10355
methyladenosine. Cell Res 27(5):626–641 37. Chu Q, Zhang X, Zhu X et al (2017) PlantcircBase:
19. Barrett SP, Salzman J (2016) Circular RNAs: analy- a database for plant circular RNAs. Mol Plant
sis, expression and potential functions. Development 10(8):1126–1128
143(11):1838–1847 38. Chen L, Yu Y, Zhang X et al (2016) PcircRNA_
20. Lasda E, Parker R (2014) Circular RNAs: diversity of finder: a software for circRNA prediction in plants.
form and function. RNA 20(12):1829–1842 Bioinformatics 32(22):3528–3529
21. Salzman J (2016) Circular RNA expression: its 39. Szabo L, Salzman J (2016) Detecting circular RNAs:
potential regulation and function. Trends Genet bioinformatic and experimental challenges. Nat Rev
32(5):309–316 Genet 17(11):679–692
22. Cortés-López M, Miura P (2016) Emerging functions 40. Hansen TB, Veno MT, Damgaard CK et al (2016)
of circular RNAs. Yale J Biol Med 89(4):527 Comparison of circular RNA prediction tools. Nucleic
23. Fischer JW, Leung AK (2017) CircRNAs: a regula- Acids Res 44(6):e58
tor of cellular stress. Crit Rev Biochem Mol Biol 41. Westholm JO, Miura P, Olson S et al (2014)
52(2):220–233 Genome-wide analysis of Drosophila circular RNAs
24. Patop IL, Kadener S (2018) circRNAs in Cancer. Curr reveals their structural and sequence properties
Opin Genet Dev 48:121–127 and age-dependent neural accumulation. Cell Rep
25. Toubia J, Conn VM, Conn SJ (2018) Don’t go in 9(5):1966–1980
circles: confounding factors in gene expression profil- 42. Ashwal-Fluss R, Meyer M, Pamudurti NR et al (2014)
ing. EMBO J 37(11):e97945 circRNA biogenesis competes with pre-mRNA splic-
26. Darbani B, Noeparvar S, Borg S (2016) Identification ing. Mol Cell 56(1):55–66
of circular RNAs from the parental genes involved 43. Liang D, Wilusz JE (2014) Short intronic repeat
in multiple aspects of cellular metabolism in barley. sequences facilitate circular RNA production. Genes
Front Plant Sci 7:776 Dev 28(20):2233–2247
27. Ye CY, Chen L, Liu C et al (2015) Widespread 44. Barrett SP, Wang PL, Salzman J (2015) Circular RNA
noncoding circular RNAs in plants. New Phytol biogenesis can proceed through an exon-containing
208(1):88–95 lariat precursor. Elife 4:e07540
28. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular 45. Rybak-Wolf A, Stottmeister C, Glazar P et al (2015)
RNAs are abundant, conserved, and associated with Circular RNAs in the mammalian brain are highly
ALU repeats. RNA 19(2):141–157 abundant, conserved, and dynamically expressed.
29. Chen G, Cui J, Wang L et al (2017) Genome-wide Mol Cell 58(5):870–885
identification of circular RNAs in Arabidopsis thali- 46. Errichelli L, Dini Modigliani S, Laneve P et al (2017)
ana. Front Plant Sci 8:1678 FUS affects circular RNA expression in murine
30. Lu T, Cui L, Zhou Y et al (2015) Transcriptome-­ embryonic stem cell-derived motor neurons. Nat
wide investigation of circular RNAs in rice. RNA Commun 8:14741
21(12):2076–2087 47. Aktas T, Avsar Ilik I, Maticzka D et al (2017) DHX9
31. Zhao W, Cheng Y, Zhang C et al (2017) Genome-wide suppresses RNA processing defects originating
identification and characterization of circular RNAs from the Alu invasion of the human genome. Nature
by high throughput sequencing in soybean. Sci Rep 544(7648):115–119
7(1):5636 48. Foot JN, Feracci M, Dominguez C (2014) Screening
protein--single stranded RNA complexes by NMR
342 X. Lai et al.

spectroscopy for structure determination. Methods 62. Zheng QP, Bao CY, Guo WJ et al (2016) Circular
65(3):288–301 RNA profiling reveals an abundant circHIPK3 that
49. Teplova M, Hafner M, Teplov D et al (2013) Structure-­ regulates cell growth by sponging multiple miRNAs.
function studies of STAR family quaking proteins Nat Commun 7:11215
bound to their in vivo RNA target sites. Genes Dev 63. Hsiao KY, Lin YC, Gupta SK et al (2017) Noncoding
27(8):928–940 effects of circular RNA CCDC66 promote Colon
50. Ryder SP, Williamson JR (2004) Specificity of the Cancer growth and metastasis. Cancer Res
STAR/GSG domain protein Qk1: implications for the 77(9):2339–2350
regulation of myelination. RNA 10(9):1449–1458 64. Liu T, Zhang L, Chen G et al (2017) Identifying and
51. Ryder SP, Frater LA, Abramovitz DL et al (2004) characterizing the circular RNAs during the lifespan
RNA target specificity of the STAR/GSG domain of Arabidopsis leaves. Front Plant Sci 8:1278
post-transcriptional regulatory protein GLD-1. Nat 65. Mendes ND, Freitas AT, Sagot MF (2009) Current
Struct Mol Biol 11(1):20–28 tools for the identification of miRNA genes and their
52. Hall MP, Nagel RJ, Fagg WS et al (2013) Quaking targets. Nucleic Acids Res 37(8):2419–2433
and PTB control overlapping splicing regulatory 66. Wang XJ, Reyes JL, Chua NH et al (2004)
networks during muscle cell differentiation. RNA Prediction and identification of Arabidopsis thali-
19(5):627–638 ana microRNAs and their mRNA targets. Genome
53. Cheng Y, Kato N, Wang W et al (2003) Two RNA Biol 5(9):R65
binding proteins, HEN4 and HUA1, act in the pro- 67. Bulow L, Bolivar JC, Ruhe J et al (2012) ‘MicroRNA
cessing of AGAMOUS pre-mRNA in Arabidopsis Targets’, a new AthaMap web-tool for genome-wide
thaliana. Dev Cell 4(1):53–66 identification of miRNA targets in Arabidopsis thali-
54. Mockler TC, Yu X, Shalitin D et al (2004) Regulation ana. BioData Min 5(1):7
of flowering time in Arabidopsis by K homol- 68. Bonnet E, He Y, Billiau K et al (2010) TAPIR, a
ogy domain proteins. Proc Natl Acad Sci USA web server for the prediction of plant microRNA
101(34):12759–12764 targets, including target mimics. Bioinformatics
55. Rodriguez-Cazorla E, Ripoll JJ, Andujar A et al 26(12):1566–1568
(2015) K-homology nuclear ribonucleoproteins 69. Srivastava PK, Moturu TR, Pandey P et al (2014) A
regulate floral organ identity and determinacy in comparison of performance of plant miRNA target
Arabidopsis. PLoS Genet 11(2):e1004983 prediction tools and the characterization of features
56. Jiang J, Wang B, Shen Y et al (2013) The Arabidopsis for genome-wide target prediction. BMC Genomics
RNA binding protein with K homology motifs, 15:348
SHINY1, interacts with the C-terminal domain phos- 70. Chi SW, Zang JB, Mele A et al (2009) Argonaute
phatase-like 1 (CPL1) to repress stress-inducible gene HITS-CLIP decodes microRNA-mRNA interaction
expression. PLoS Genet 9(7):e1003625 maps. Nature 460(7254):479–486
57. Guan Q, Wen C, Zeng H et al (2013) A KH domain-­ 71. Suzuki N, Rivero RM, Shulaev V et al (2014)
containing putative RNA-binding protein is critical Abiotic and biotic stress combinations. New Phytol
for heat stress-responsive gene regulation and ther- 203(1):32–43
motolerance in Arabidopsis. Mol Plant 6(2):386–395 72. Phillips JR, Dalmay T, Bartels D (2007) The
58. Jeong IS, Fukudome A, Aksoy E et al (2013) role of small RNAs in abiotic stress. FEBS Lett
Regulation of abiotic stress signalling by Arabidopsis 581(19):3592–3597
C-terminal domain phosphatase-like 1 requires inter- 73. Khraiwesh B, Zhu JK, Zhu J (2012) Role of miRNAs
action with a k-homology domain-containing protein. and siRNAs in biotic and abiotic stress responses of
PLoS One 8(11):e80509 plants. Biochim Biophys Acta 1819(2):137–148
59. Thatcher LF, Kamphuis LG, Hane JK et al (2015) The 74. Wang YX, Yang M, Wei SM et al (2017) Identification
Arabidopsis KH-domain RNA-binding protein ESR1 of circular RNAs and their targets in leaves of Triticum
functions in components of Jasmonate Signalling, aestivum L under dehydration stress. Front Plant Sci
unlinking growth restraint and resistance to stress. 7:2024
PLoS One 10(5):e0126978 75. Conn VM, Hugouvieux V, Nayak A et al (2017) A
60. Lorkovic ZJ, Barta A (2002) Genome analysis: circRNA from SEPALLATA3 regulates splicing of its
RNA recognition motif (RRM) and K homology cognate mRNA through R-loop formation. Nat Plants
(KH) domain RNA-binding proteins from the flow- 3(5):17053
ering plant Arabidopsis thaliana. Nucleic Acids Res 76. Chen Y, Li C, Tan C et al (2016) Circular RNAs: a
30(3):623–635 new frontier in the study of human diseases. J Med
61. Du WW, Yang W, Liu E et al (2016) Foxo3 circular Genet 53(6):359–365
RNA retards cell cycle progression via forming ter- 77. Huang S, Yang B, Chen BJ et al (2017) The emerg-
nary complexes with p21 and CDK2. Nucleic Acids ing role of circular RNAs in transcriptome regulation.
Res 44(6):2846–2858 Genomics 109(5–6):401–407
26 CircRNAs in Plants 343

78. Meng S, Zhou H, Feng Z et al (2017) CircRNA: func- 84. Zhu X, Wang X, Wei S et al (2017) hsa_circ_0013958:
tions and properties of a novel potential biomarker for a circular RNA and potential novel biomarker for lung
cancer. Mol Cancer 16(1):94 adenocarcinoma. FEBS J 284(14):2170–2182
79. Han YN, Xia SQ, Zhang YY et al (2017) Circular 85. Yang XS, Wu J, Ziegler TE et al (2011) Gene
RNAs: a novel type of biomarker and genetic tools in expression biomarkers provide sensitive indicators
cancer. Oncotarget 8(38):64551–64563 of in planta nitrogen status in maize. Plant Physiol
80. Zhang HD, Jiang LH, Sun DW et al (2018) CircRNA: 157(4):1841–1852
a novel type of biomarker for cancer. Breast Cancer 86. Yan X, Gurtler J, Fratamico P et al (2011)
25(1):1–7 Comprehensive approaches to molecu-
81. Huang YS, Jie N, Zou KJ et al (2017) Expression pro- lar biomarker discovery for detection and
file of circular RNAs in human gastric cancer tissues. identification of Cronobacter spp. (Enterobacter saka-
Mol Med Rep 16(3):2469–2476 zakii) and Salmonella spp. Appl Environ Microbiol
82. Lu LS, Sun J, Shi PY et al (2017) Identification of 77(5):1833–1843
circular RNAs as a promising new class of diagnos- 87. Steinfath M, Strehmel N, Peters R et al (2010)
tic biomarkers for human breast cancer. Oncotarget Discovering plant metabolic biomarkers for pheno-
8(27):44096–44107 type prediction using an untargeted approach. Plant
83. Shao Y, Li J, Lu R et al (2017) Global circular RNA Biotechnol J 8(8):900–911
expression profile of human gastric cancer and its
clinical significance. Cancer Med 6(6):1173–1180
 Circular RNAs and Plant Stress
Responses 27
Celso Gaspar Litholdo Jr.
and Guilherme Cordenonsi da Fonseca

Abstract expressed. In tomato, 163 circRNAs demon-

Circular RNAs (circRNAs) are a novel class strated chilling-responsive expression,
of noncoding RNAs that have been exten- with 102 containing miRNA-binding sites and
sively explored in the past few years. The are predicted to act as miRNA sponges.
advent of new high-throughput sequencing Additionally, Arabidopsis seedlings presented
technologies coupled with bioinformatics 1583 heat-specific circRNAs, and it was
tools revealed the presence of these molecules also reported that heat stress could increase
in the transcriptome of a wide range of organ- the quantity, length, and alternative circular-
isms. In animals, circRNAs can modulate ization events of circRNAs. Finally, wheat
gene expression and act as sponges of miR- seedlings under dehydration stress had 62 cir-
NAs to inhibit their activity. It has been dem- cRNAs differentially expressed, with 6 being
onstrated that they have the potential to be predicted as miRNA sponges. Although the
diagnostic biomarkers as their expression is role of plant circRNAs during the biotic and
closely associated to human diseases, such as abiotic stresses is still poorly characterised,
Alzheimer and cancer. However, in plants these molecules have the potential to expand
their function remains elusive. Recently, the the number of targets and tools in the
role of the circRNAs in plant stress responses biotechnology field.
has been studied. During the infection of
Pseudomonas syringae in kiwifruit plants,
584 circRNAs were differentially expressed
in leaf samples, and a group of them could be 1 Introduction
further associated with the stage of infection.
Under phosphate deficiency conditions, 27 rice 1.1  ircular RNAs: Biosynthesis
C
circRNAs were reported to be differentially and Function in Plants

Circular RNAs (circRNAs) are single-stranded

C. G. Litholdo Jr. (*)
Laboratoire Génome et Développement des Plantes, covalently closed loop RNA molecules that were
Centre National pour la Recherche Scientifique first described in 1976, when it was demonstrated
(CNRS), Perpignan, France that plant viroids are composed of single-stranded
e-mail: [email protected] RNA molecules lacking free 5′ and 3′ends [1].
G. C. da Fonseca Posteriorly, another group identified for the first
Centro de Biotecnologia, Universidade Federal do time the existence of endogenous circRNA
Rio Grande do Sul, Porto Alegre, RS, Brazil

© Springer Nature Singapore Pte Ltd. 2018 345

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_27
346 C. G. Litholdo Jr. and G. C. da Fonseca

t­ranscripts. Studying the candidate tumor sup- spliceosome formation, including RNA pairing,
pressor gene (DCC), it was observed that several by either repetitive elements [23] or complemen-
abnormal transcripts composed of exons were tary sequences in the flanking introns [24].
scrambled during the splicing process [2]. These However, circRNAs identified in plants pos-
molecules were found at relatively low levels in sessed few repetitive and reverse complementary
the nonpolyadenylated component of cytoplas- sequences in the flanking introns, when com-
mic RNA from human and rodent cells [2]. Over pared to animals [9]. In Arabidopsis, it was iden-
the following years, more endogenous circRNAs tified that circRNAs have at least two different
were discovered in eukaryotes [3, 4]; however, short complementary sequences that ranged from
they were long considered transcriptional by-­ 4 to 11 nucleotides near the splice sites in more
products or the result of aberrant RNA splicing than 33% of the cases, suggesting that multiple
without a clear function [5]. combination of these sequences can compensate
Until recently, the study of circRNAs has been for their short length [25]. It was demonstrated in
neglected due to the difficulties in the detection rice that more than 90% of the identified cir-
of these molecules through the most used meth- cRNAs were flanked by noncanonical splicing
ods of RNA analysis. Since circRNAs lack free signals, contrasting with humans circRNAs,
5′- or 3′- ends, they cannot be detected by tech- which most are flanked by the canonical GT/AG
niques that need polyadenylated free ends, such splicing sites [26]. Altogether, these findings indi-
as rapid amplification of cDNA ends (RACE) or cated that plants have specific mechanisms for
RNAseq from poly(A)-enriched samples, which the regulation of the circRNAs when compared
is the most commonly used strategy for transcrip- to other organisms [27].
tome analysis [6]. However, the improvement of The most striking function of the circRNA
the high-throughput sequencing technology, par- that has been discovered so far is their ability to
ticularly with the advent of the RibominusSeq modulate gene expression, acting as miRNA
technique (RNAseq from rRNA-depleted sam- sponges [28]. MiRNA sponges are transcripts
ples), coupled with the development of new bio- that have miRNA-binding sites and therefore are
informatics tools, allowed the identification of able to “sequest” miRNAs, to inhibit their activ-
hundreds of circRNAs in different organisms [7, ity [29]. These transcripts are also called compet-
8]. Over the past few years, circRNAs have been ing endogenous RNAs in animals or target
identified in many plant species, including mimicry in plants [30], and they have been used
Arabidopsis, rice, barley, wheat, tomato, cotton, as molecular biology tools to study the function
soybean, potato, kiwi, maize, and orange [9–18]. of miRNAs [29, 31]. The advantage of the cir-
The biosynthesis of the circRNAs (Fig. 27.1) cRNAs as miRNA sponges is their intracellular
is considered to be conserved in all eukaryotes high stability, with half-lives exceeding 48 h,
and, as protein-coding genes, is dependent of whereas their linear counterparts exhibited half-­
RNA polymerase II and a noncanonical splicing lives of less than 20 h [23]. This higher stability
mechanism of the pre-messenger RNAs (pre-­ is due to their resistance to RNA exonucleases,
mRNAs), termed backsplicing [19]. The backs- since circRNAs do not have 5′- or 3′- free ends
plicing occurs when the 5′-end of the upstream [5]. However, no evidence has been shown that
exon is linked to 3′-end of the downstream exon plant circRNAs act as miRNA sponges [27].
producing a head-to-tail splicing junction [20]. In Several studies show that plant circRNAs have
some cases, the circularization retains introns the potential to be miRNA sponges, but com-
between exons producing the exon-intron cir- pared to animals, they represent a smaller propor-
cRNAs or EIciRNAs [21]. circRNAs can be also tion of the total circRNAs (around 5%) and have
derived from lariat introns that escape debranch- lesser miRNA-binding sites [9, 12, 32].
ing and form a stable circular molecule, termed In plants, it has been demonstrated that cir-
intronic circRNA or ciRNA [22]. cRNAs can negatively modulate the expression
The regulation of the circularization events is of their parental genes. In rice, the overexpres-
dependent by cis and trans elements during the sion of a circRNA greatly reduces the expression
27 Circular RNAs and Plant Stress Responses 347

Fig. 27.1 Circular RNA biosynthesis some cases, the circularization retains introns between
The parental gene of the circRNA is transcribed by RNA exons producing the exon-intron circRNAs. When a lariat
polymerase II. The backsplicing can occur by RNA pair- intron escapes the debranching, it can form a stable circu-
ing of complementary sequences in the flanking introns. lar molecule termed intronic circRNA. The purple arrows
The result of the backsplicing is a circular molecule pro- indicate the splicing sites. The colored rectangles repre-
duced by a head-to-tail splicing junction. The circRNA sent the exons, while the black lines represent the introns
that contains only exons is termed exonic circRNA. In

levels of its parental gene in different tissues [33]. RibominusSeq libraries are the most commonly
In a more recent work, a circRNA derived from used method, these molecules can also be found
exon 6 of the SEPALLATA3 (SEP3) gene of in the traditional RNAseq libraries from poly(A)-
Arabidopsis thaliana was demonstrated to bind enriched samples, however with a lower effi-
to its cognate DNA locus through the formation ciency of detection [33]. Several works are also
of a RNA:DNA hybrid resulting in transcrip- using RNAseq libraries from rRNA-depleted
tional pausing and an decrease on the abundance samples further treated with RNAse R, an enzyme
of its sourced gene [34]. This circRNA also that degrades linear but not circular RNAs, to
increases the abundance of an exon-skipped improve the sensibility and decrease the number
alternative splicing variant that lacks the exon 6 of false positives for the detection of circRNAs
driving floral homeotic phenotypes [34]. These [15, 23, 35, 36].
findings provided new insights about the biologi- The identification of circRNAs relies mainly
cal roles of the plant circRNAs. by detection of sequence reads that map only in
the backsplicing junctions, but not in the refer-
ence genome, often termed backsplicing reads or
1.2  ethods to Identify and Study
M junction reads. To address that goal, several
circRNAs in Plants algorithms and programs were developed includ-
ing CIRI [37], find_circ [7], MapSplice [38],
High-throughput sequencing and bioinformatics CIRCexplorer [24], and testrealign [39]. The dif-
analysis have been the most used tools to char- ferences among them are the use of different
acterize and identify circRNAs. Although read aligners, the requirement of inputs as gene
348 C. G. Litholdo Jr. and G. C. da Fonseca

annotations, and the search for non-exonic cir- 2  otential Role of circRNAs
P
cRNAs [40]. A study compared five different During Plant Stress
algorithms and evaluated the levels of bona fide Responses
and false-­ positive circRNAs using data from
libraries treated with RNase R and untreated sam- Circular RNAs are new discovered players in the
ples [41]. These results showed that the false- RNA-mediated gene regulation, acting in several
positive rate ranged from 12% to 28%, but was biological processes, at the transcriptional and
especially high for highly expressed circRNAs, posttranscriptional levels. In animals, circRNAs
suggesting that the more reliable approach is the are abundant in transcriptomics data, and several
combination of at least two circRNA identifica- has been demonstrated to participate in the occur-
tion methods [41]. However, all these programs rence of human diseases, such as cancers and
were designed primarily for human or animal Alzheimer, which creates diagnostic and thera-
datasets, and considering the differences in the peutic targets based on circRNAs [45, 46]. In
organization of plant and mammal genomes, a plants, circRNA function is poorly characterized
software called PcircRNA_finder was developed although its presence in plant transcriptomics is
[42]. This software was able to predict, with more found widespread and highly abundant, since the
high-confidence, circRNAs from rice libraries first report in the plant model Arabidopsis [8].
when compared to CIRCexplorer and find_circ We will here compile all the information avail-
[42]. able in the literature to give us hints of the role
Over the past few years, an increasingly of circRNAs during plant stress responses, which
number of studies have allowed the develop- is also indicated in Table 27.1.
ment of plant circRNA databases. The
PlantcircBase [43] compiles the information of
almost 90,000 circRNAs from eight plant spe- 2.1 circRNAs and Biotic Stresses
cies. Detailed information is available for each
circRNA entry, such as ID, genomic position, The first report identifying circRNAs under a
annotation and experimental validation of the biotic stress condition was the study of the plant-­
backsplicing site. Additionally, splicing signals pathogen interaction between Arabidopsis and
of each circRNAs are provided, which is a par- Pseudomonas syringae pv. tomato. Leaf samples
ticularly useful information, considering that were sequenced in the Illumina HiSeq 2000 plat-
most of the plant circRNAs do not have the form, after removal of ribosomal and linear
canonical GT-AG splicing signals [26]. Another RNAs, by magnetic beads and RNase R, respec-
interesting feature found in this database is the tively, the latest a biochemical approach named
prediction of networks involving circRNA- CircleSeq [25]. 85 million reads per sample were
miRNA-mRNA, allowing the identification of obtained, and 803 circRNAs were identified
putative miRNA sponges. using 16 distinct RNA sequencing data sets from
The mostly used technique to validate cir- A. thaliana.
cRNAs from in silico analysis is the reverse tran- The authors used the following filter criteria
scription (RT)-PCR using convergent primers to for circRNA identification: (1) presence of align-
detect linear transcripts, as the parental mRNA, ment of the reads with mapping quality of >10;
and divergent primers to detect the circular RNA (2) presence of either a “GT” or “AG” signal
[9, 33]. Divergent primers are designed as such: within five nucleotides of the donor or acceptor
the forward primer aligns in the downstream site, respectively, or a back-splice site; (3) seg-
exon, and the reverse primer aligns in the ments from the same read must be in the same
upstream exon of the circRNA sequence. The strand with a reversed order; and (4) circRNA
reverse primer is also used in the cDNA synthesis sequence was presented in at least two
of the circular and linear transcripts [44]. Arabidopsis datasets [25]. 108 circRNAs out of
27 Circular RNAs and Plant Stress Responses 349

Table 27.1 Biotic and abiotic stresses currently known to deregulate plant circular RNAs expression
Number of differentially
Plant stress Plant species expressed circrnas Reference
Biotic Bacterial infection (Pseudomonas Actinidia sp. (kiwi 584 [10]
syringae pv. actinidiae) fruit)
Bacterial infection (Pseudomonas Arabidopsis thaliana 803 [25]
syringae pv. Tomato)
Abiotic Low- and high- light Arabidopsis thaliana 6012 [9]
Heating (38 °C) Arabidopsis thaliana 1583 [47]
Iron and zinc treatments Hordeum vulgare 62 [13]
(barley)
Phosphate-starvation Oryza sativa (rice) 27 [9]
Chilling (0 °C) Solanum lycopersicum 163 [12]
(tomato)
Dehydration (PEG) Triticum aestivum 62 [11]
(wheat)
The number of differentially expressed circRNAs identified during plant stress conditions is indicated

803 overlapped with the exons, showing that A total of 584 circRNAs were differentially
complementary sequences are not exclusive of expressed during P. syringae infection, and the
introns, but found in the flanking sequences. In expression of particular circRNAs was further
addition, the chloroplast genome showed a higher associated with the stage of infection [10].
source of circRNA than the mitochondria; 6% of Moreover, exonic and intronic circRNAs showed
identified circRNA were found in the chloroplast a positive correlation with the originating protein-­
compared to 1% of circRNA in mitochondria. coding genes, and a weighted gene co-expression
Furthermore, the results revealed that both gen- network analysis (WGCNA) for searching poten-
eral and alternative splicing mechanisms gener- tial biological process that are associated with
ate circRNAs [25]. Although this work identified genes of interest identified circRNAs that were
hundreds of circRNAs in Arabidopsis under closely associated with respiratory burst, MAP
pathogenic interaction and contributes to under- kinase activity, and intracellular signal transduc-
standing this endogenous class of RNA, the bio- tion. Therefore, this study represents the first
logical function of circRNAs under biotic stresses report on the potential biological role of cir-
remains to be revealed. cRNAs in plant defense response against bacte-
More recently, circRNAs were also found to rial pathogen infection [10].
respond to bacterial invasion in kiwifruit plants
[10]. Leaf samples of three species belonging to
the genus Actinidia have their ribosomal RNA 2.2 circRNA and Abiotic Stresses
removed and sequenced using the Illumina HiSeq
2000 sequencer. These samples were previously Previous reports have shown that circRNAs are
infected with Pseudomonas syringae pv. actinid- differentially expressed under abiotic stresses,
iae, the causal agent of kiwi canker disease. In such as phosphate, zinc, and iron imbalance, low
addition, the authors identified that circRNAs and high light, chilling, and drought. However,
expression have tissue-specific patterns. A total the biological significance underlying circRNA
of 2884 million paired-end reads were sequenced, regulatory function during these condi-
and 2914 circRNAs from the bacterial infected tions remains to be elucidated.
samples were identified after CircRNA Identifier Computational analysis, using two publicly
(CIRI) tool and manual filtering. Among these available ribosomal RNA-depleted RNA
circRNAs, 64% were intergenic, 21% were sequencing data from different tissues and
exonic, and 14.55% were intronic circRNAs [10]. phosphate-­starvation and light treatments, identi-
350 C. G. Litholdo Jr. and G. C. da Fonseca

fied 12037 and 6012 circRNAs in Oryza sativa kinase [13]. Interestingly, the authors also found
roots and A. thaliana leaves, respectively [9]. cir- noncoding protein transcripts as parental source
cRNAs were selected based on back-spliced of circRNAs, such as microRNA and long non-
reads, and 50.5% and 85.7% represented exonic coding RNAs after iron and zinc treatments.
circRNAs in rice and Arabidopsis, respectively. Another abiotic stress that has been shown to
Approximately, 700 exonic circRNAs were origi- alter circRNAs expression is extreme tempera-
nated from homologous genes between both ture, either heat or chilling stress; therefore, cir-
plant species, suggesting a conservation in plants cRNAs have possibly also a role on
[9]. The authors found 27 differentially expressed temperature-related stress responses. Low-­
circRNAs under phosphate deficiency condition temperature stress was studied in tomato plants,
in rice, 6 upregulated and 21 downregulated; as chilled tomatoes are susceptible to develop
moreover, several circRNAs presented positive symptoms in this condition [12]. Tomato fruits
correlation with their parental genes. These were kept under 0° C for 72 h, and rRNA-­
results suggest that a regulatory role between cir- removed RNA samples were sequenced by
cRNAs and circRNAs-originating genes also Illumina platform. Using CIRI tools and differen-
exists in plants, and additionally, the differen- tial expression analysis, the authors identified
tially expressed circRNAs have a putative role in 3019 chilling-responsive circRNAs, with 770
response to phosphate deficiency [9]. unique circRNAs that were only identified under
Another recent study revealed the expres- the stress condition when compared to the con-
sional changes of circRNAs in response to nutri- trol sample. Stress-induced exonic circRNAs
ents [13]. This study was carried out across barley represented 72% of the total parental genes,
tissues and in response to iron and zinc treatment including genes that encode to enzymes involved
on both foliar and seed tissues. The authors used in, cellular redox homoeostasis network, cell
a RNAseq dataset available in the GenBank and wall degradation, membrane lipid peroxidation,
performed sequencing using the Illumina plat- arginine and polyamine metabolism, energy
form HiSeq 2000 and TruSeq technology. A total metabolism, and jasmonic acid and abscisic
of 262 putative circRNAs were identified by acid signalling. circRNAs related to genes that
these approaches, and after manual filtering, 62 encode to temperature-induced transcription fac-
circRNAs were selected. The majority of selected tors, such as the core-binding factor (CBF) and
circRNAs presented a weak and negative tran- WRKY superfamily, were also identified [12]. In
scriptional correlation with the parental genes. addition, the authors predicted 102 circRNAs
Moreover, when comparing treatments in seeds that can potentially act as miRNA sponges, due
and leaves, an opposite trend was observed. For to the predicted miRNA-binding site for 24 dis-
instance, CAX2 gene, which is an H+/cation anti- tinct mature miRNAs. Further analysis showed
porter and sequesters of calcium, iron, and zinc, that 163 circRNAs presented differential expres-
had its transcription level inverted correlated with sion changes between control and chilled plants.
the CAX2 circRNA, in an tissue-dependent man- The vast majority of the deregulated circRNAs
ner [13]. showed upregulation under chilling treatment
In addition, several identified circRNAs were [12].
derived from mitochondrial and nuclear genes Similar pattern of expressional changes were
previously known to be involved with cellular observed in A. thaliana plants under heat
functions that are related to respiratory complex, ­treatment [47]. Arabidopsis seedlings were
ion homeostasis, intracellular protein transport, treated with 38°C for 3 h, and rRNA-depleted
amino acid biosynthesis, transcription and trans- and circRNAs-­enriched RNA samples were sub-
lation, and hormonal signaling. These genes mitted to Illumina HiSeq 4000 sequencer. The
include COX1 subunit of the cytochrome c oxi- authors discovered 1599 novel circRNAs, with
dase, apocytochrome b, NADH dehydrogenase 1583 showing heat-specific accumulation. When
NAD9, and CTR1-like serine/threonine-protein comparing control samples, heat stress treatment
27 Circular RNAs and Plant Stress Responses 351

triggered a significant increase in circRNA quan- circRNAs in the control of transcriptional and
tity, size, and alternative circularization events. posttranscriptional gene expression is uncovered;
Interestingly, different from the other reports on however, the equivalent function in plant species
circRNAs and plant stress responses, the produc- still lacks consistent and experimental validation.
tion of circRNAs was originated mainly from Further, the already known biological role of ani-
nuclear genes rather than chloroplast and mito- mal circRNAs during development and progress
chondria, and moreover, compared to chilling of diseases represents the chance to target cir-
treatment [12], the majority of circRNAs under cRNAs for therapeutic and diagnostic purposes.
high temperature treatment showed a positive Despite the progress in understanding the role
correlation with their parental genes [47]. and action of animal circRNAs, there is little
Additionally, the authors predicted the miRNAs information about plant circRNAs and even less
that could be sequestered by heat-induced cir- about its biological role during plant develop-
cRNAs and built an interacting network compris- ment and stress responses.
ing the miRNA target genes. The predicted Based on high-throughput sequencing tech-
miRNA sponge-like circRNAs can potentially nologies and in silico analyses, there are indica-
influence the expression of genes related to heat tions that circRNAs-mediated gene regulation
stress, hydrogen peroxide, and phytohormones, might play functional roles in plant immune
by sequestering, for instance, the miR156, response (Table 27.1), through different mecha-
miR157, miR159, miR395, miR172, and nisms by acting directly on mRNAs, or as
miR857. microRNA sponges. Another recently raised
Our last herein mentioned plant stress that has question is the putative translation of exonic cir-
shown to regulate circRNAs is drought. Wheat cRNAs into functional peptides [35]; the major-
seedlings treated with polyethylene glycol (PEG), ity of identified plant circRNAs are derived
which artificially mimics dehydration stress were from exons, but the existance of such peptides
analyzed by high-throughput sequencing and still demands further additional research. cir-
real-time PCR [11]. The authors identified 62 cRNAs represent, therefore, a new and intrigu-
candidate circRNAs that showed a differential ing class of noncoding RNAs still poorly
expression under drought-like stress, with 16 cir- explored in the field of plant science; their puta-
cRNAs being upregulated and 46 downregu- tive regulatory role on functional coding and
lated at this condition. This work also predicted noncoding transcripts associated with bacterial,
the sponge action of circRNAs by showing that 6 temperature, nutrient starvation and toxicity,
out of the 62 circRNAs have predicted miRNA-­ and drought resistances needs to be fully
binding site that can potentially regulate 26 dis- addressed. Future investigations on the biologi-
tinct wheat miRNAs [11]. Further analysis, using cal significance of circRNAs during biotic and
Gene Ontology (GO) and Kyoto Encyclopedia of abiotic stresses can potentially expand the avail-
Genes and Genomes (KEGG), demonstrated able tools and targets for applied crop
an involvement of the identified circRNAs in sev- biotechnology.
eral cellular processes, such as photosynthesis,
porphyrin and chlorophyll metabolism, oxidative
phosphorylation, amino acid-related process, and References
phytohormone signaling.
1. Sanger HL, Klotz G, Riesner D et al (1976) Viroids
are single-stranded covalently closed circular RNA
molecules existing as highly base-paired rod-like
3 Concluding Remarks structures. Proc Natl Acad Sci USA 73:3852–3856
2. Nigro JM, Cho KR, Fearon ER et al (1991) Scrambled
Circular RNAs are a class of noncoding RNAs exons. Cell 64:607–613
3. Cocquerelle C, Daubersies P, Majérus MA et al
that have been increasingly studied in plants (1992) Splicing with inverted order of exons occurs
during the past 5 years. In animals, the role of proximal to large introns. EMBO J 11:1095–1098
352 C. G. Litholdo Jr. and G. C. da Fonseca

4. Capel B, Swain A, Nicolis S et al (1993) Circular 21. Li Z, Huang C, Bao C et al (2015) Exon-intron cir-
transcripts of the testis-determining gene Sry in adult cular RNAs regulate transcription in the nucleus. Nat
mouse testis. Cell 73:1019–1030 Struct Mol Biol 22:256–264
5. Cocquerelle C, Mascrez B, Hétuin D et al (1993) 22. Zhang Y, Zhang X-O, Chen T et al (2013) Circular
Mis-splicing yields circular RNA molecules. FASEB intronic long noncoding RNAs. Mol Cell 51:792–806
J 7:155–160 23. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular
6. Jeck WR, Sharpless NE (2014) Detecting and charac- RNAs are abundant, conserved, and associated with
terizing circular RNAs. Nat Biotechnol 32:453–461 ALU repeats. RNA 19:141–157
7. Memczak S, Jens M, Elefsinioti A et al (2013) 24. Zhang X-O, Wang H-B, Zhang Y et al (2014)
Circular RNAs are a large class of animal RNAs with Complementary sequence-mediated exon circulariza-
regulatory potency. Nature 495:333–338 tion. Cell 159:134–147
8. Wang PL, Bao Y, Yee M-C et al (2014) Circular RNA 25. Sun X, Wang L, Ding J et al (2016) Integrative
is expressed across the eukaryotic tree of life. PLoS analysis of Arabidopsis thaliana transcriptomics
ONE 9(3):e90859 reveals intuitive splicing mechanism for circu-
9. Ye CY, Chen L, Liu C et al (2015) Widespread noncod- lar RNA. FEBS Lett 590:3510–3516. https://doi.
ing circular RNAs in plants. New Phytol 208:88–95 org/10.1002/1873-3468.12440
10. Wang Z, Liu Y, Li D et al (2017) Identification of 26. Ye C-Y, Zhang X, Chu Q et al (2017) Full-length
circular RNAs in Kiwifruit and their species-specific sequence assembly reveals circular RNAs with
response to bacterial canker pathogen invasion. Front diverse non-GT/AG splicing signals in rice. RNA Biol
Plant Sci 8:413 14:1055–1063
11. Wang Y, Yang M, Wei S et al (2017) Identification of 27. Li Q-F, Zhang Y-C, Chen Y-Q et al (2017) Circular
circular RNAs and their targets in leaves of Triticum RNAs roll into the regulatory network of plants.
aestivum L. under dehydration stress. Front Plant Sci Biochem Biophys Res Commun 488:382–386
7:2024 28. Hansen TB, Jensen TI, Clausen BH et al (2013)
12. Zuo J, Wang Q, Zhu B et al (2016) Deciphering Natural RNA circles function as efficient microRNA
the roles of circRNAs on chilling injury in tomato. sponges. Nature 495:384–388
Biochem Biophys Res Commun 479:132–138. 29. Ebert MS, Neilson JR, Sharp PA (2007) MicroRNA
https://doi.org/10.1016/j.bbrc.2016.07.032 sponges: competitive inhibitors of small RNAs in
13. Darbani B, Noeparvar S, Borg S (2016) Identification mammalian cells. Nat Methods 4:721–726
of circular RNAs from the parental genes involved 30. Franco-Zorrilla JM, Valli A, Todesco M et al (2007)
in multiple aspects of cellular metabolism in Barley. Target mimicry provides a new mechanism for regula-
Front Plant Sci 7:776 tion of microRNA activity. Nat Genet 39:1033–1037
14. Zhao T, Wang L, Li S et al (2017) Characterization of 31. Todesco M, Rubio-Somoza I, Paz-Ares J et al (2010)
conserved circular RNA in polyploid Gossypium spe- A collection of target mimics for comprehensive anal-
cies and their ancestors. FEBS Lett 591:3660–3669 ysis of microRNA function in Arabidopsis thaliana.
15. Zhao W, Cheng Y, Zhang C et al (2017) Genome-wide PLoS Genet 6:e1001031
identification and characterization of circular RNAs 32. Wang K, Long B, Liu F et al (2016) A circular RNA
by high throughput sequencing in soybean. Sci Rep protects the heart from pathological hypertrophy
7:5636 and heart failure by targeting miR-223. Eur Heart
16. Zhou R, Zhu Y, Zhao J et al (2017) Transcriptome-­ J 37:2602–2611
wide identification and characterization of potato 33. Lu T, Cui L, Zhou Y et al (2015) Transcriptome-­
circular RNAs in response to Pectobacterium caroto- wide investigation of circular RNAs in rice. RNA
vorum subspecies brasiliense infection. Int J Mol Sci 21:2076–2087
19:E71 34. Conn VM, Hugouvieux V, Nayak A et al (2017) A
17. Chen L, Zhang P, Fan Y et al (2018) Circular RNAs circRNA from SEPALLATA3 regulates splicing of its
mediated by transposons are associated with tran- cognate mRNA through R-loop formation. Nat Plants
scriptomic and phenotypic variation in maize. New 3:17053
Phytol 217:1292–1306 35. Chen G, Cui J, Wang L et al (2017) Genome-Wide
18. Zeng R-F, Zhou J-J, Hu C-G et al (2018) Identification of Circular RNAs in Arabidopsis thali-
Transcriptome-wide identification and functional pre- ana. Front Plant Sci 8:1678
diction of novel and flowering-related circular RNAs 36. Tan J, Zhou Z, Niu Y et al (2017) Identification and
from trifoliate orange (Poncirus trifoliata L. Raf.). functional characterization of tomato CircRNAs
Planta 247:1191–1202 derived from genes involved in fruit pigment accumu-
19. Ashwal-Fluss R, Meyer M, Pamudurti NR et al (2014) lation. Sci Rep 7:8594
circRNA biogenesis competes with pre-mRNA splic- 37. Gao Y, Wang J, Zhao F (2015) CIRI: an efficient and
ing. Mol Cell 56:55–66 unbiased algorithm for de novo circular RNA identifi-
20. Salzman J, Gawad C, Wang PL et al (2012) Circular cation. Genome Biol 16:4
RNAs are the predominant transcript isoform from 38. Wang K, Singh D, Zeng Z et al (2010) MapSplice:
hundreds of human genes in diverse cell types. PLoS accurate mapping of RNA-seq reads for splice junc-
One 7:e30733 tion discovery. Nucleic Acids Res 38:e178
27 Circular RNAs and Plant Stress Responses 353

39. Hoffmann S, Otto C, Doose G et al (2014) A multi-­ 43. Chu Q, Zhang X, Zhu X et al (2017) PlantcircBase:
split mapping algorithm for circular RNA, splicing, a database for plant circular RNAs. Mol Plant
trans-splicing and fusion detection. Genome Biol 10:1126–1128
15:R34 44. Kulcheski FR, Christoff AP, Margis R (2016) Circular
40. Gaffo E, Bonizzato A, Kronnie G et al (2017) RNAs are miRNA sponges and can be used as a new
CirComPara: a multi-method comparative bioinfor- class of biomarker. J Biotechnol 238:42–51
matics pipeline to detect and study circRNAs from 45. Ghosal S, Das S, Sen R et al (2013) Circ2Traits: a
RNA-seq data. Non-Coding RNA 3:8 comprehensive database for circular RNA potentially
41. Hansen TB, Venø MT, Damgaard CK et al (2016) associated with disease and traits. Front Genet 4:283
Comparison of circular RNA prediction tools. Nucleic 46. Chen LL (2016) The biogenesis and emerging roles of
Acids Res 44:e58 circular RNAs. Nat Rev Mol Cell Biol 17:205–211
42. Chen L, Yu Y, Zhang X et al (2016) PcircRNA_ 47. Pan T, Sun X, Liu Y et al (2017) Heat stress
finder: a software for circRNA prediction in plants. alters genome-wide profiles of circular RNAs in
Bioinformatics 32:3528–3529 Arabidopsis. Plant Mol Biol 96:217–229
 Part VIII
Future Prospects
 Prospective Advances in Circular
RNA Investigation 28
Siti Aishah Sulaiman, Nor Azian Abdul Murad,
Ezanee Azlina Mohamad Hanif, Nadiah Abu,
and Rahman Jamal

Abstract 1 Introduction
circRNAs have emerged as one of the key
regulators in many cellular mechanisms and The emergence of circulating RNAs (circRNAs)
pathogenesis of diseases. However, with the was reported in the 1970s, where they were
limited knowledge and current technologies observed as circular structures captured by elec-
for circRNA investigations, there are several tron microscopy in the eukaryotic cytoplasm.
challenges that need to be addressed for. These Interest in this noncoding RNA did not last long
include challenges in understanding the regu- due to the belief then that it was low in abun-
lation of circRNA biogenesis, experimental dance. circRNAs are often assumed to be splicing
designs, and sample preparations to character- noise or byproducts of splicing [1–4], and that
ize the circRNAs in diseases as well as the microRNAs (miRNAs) may outcompete their
bioinformatics pipelines and algorithms. In activation [5]. However, circRNAs have
this chapter, we discussed the above chal- reemerged and have become a hot topic in many
lenges and possible strategies to overcome areas such as agriculture [6, 7] and medicine with
those limitations. We also addressed the dif- the discovery of their roles in gene regulation [1,
ferences between the existing applications and 8–11]. circRNAs have been identified through
technologies to study the circRNAs in dis- advances in RNA sequencing (RNA-seq) and
eases. By addressing these challenges, further bioinformatics analyses, where data have been
understanding of circRNAs roles and regula- successfully pooled to establish databases for
tions as well as the discovery of novel cir- predicting circRNA downstream targets. The
cRNAs could be achieved. establishment of circRNA databases provides
guidelines for the prediction and further evalua-
Keywords tion of circRNAs to further understand their bio-
circRNA · Challenges · Identification and genesis and roles.
analysis · Future directions

S. A. Sulaiman · N. A. Abdul Murad (*)

E. A. Mohamad Hanif · N. Abu (*) · R. Jamal
Universiti Kebangsaan Malaysia (UKM)
Medical Molecular Biology Institute (UMBI),
Kuala Lumpur, Malaysia
e-mail: [email protected];
[email protected]

© Springer Nature Singapore Pte Ltd. 2018 357

J. Xiao (ed.), Circular RNAs, Advances in Experimental Medicine and Biology 1087,
https://doi.org/10.1007/978-981-13-1426-1_28
358 S. A. Sulaiman et al.

2 Current Knowledge including regulating the processes of cellular

of circRNA Functions energy, cell death and senescence, genome stabil-
ity, angiogenesis, invasion and metastasis, inflam-
2.1 Bona Fide Disease Biomarkers mation, and metabolism [17]. In a study of
and Therapeutic Targets lead-induced neurotoxicity, circ-Rar1 acted as a
sponge of miR-671 and promoted apoptosis by
circRNA expression is tissue specific, indicating inducing caspase-8 and p38 expression [18].
that they harbor biological functions and can be Moreover, circRNAs also regulate the epithelial-­
used as biomarkers to distinguish cell origin and to-­mesenchymal transition (EMT). For example,
also tumor types [4, 12]. circRNAs can adopt a circHIPK3 played a crucial role in miR-558
protective tumor-suppressive role and can also sponging, leading to inhibition of migration, inva-
exhibit oncogenic properties. Currently, only a sion, and angiogenesis [19]. circHIPK3 is highly
few circRNAs and their intersection molecules expressed in liver cancer, and its depletion has pro-
and/or pathways have been identified, namely, liferative effects on liver cancer cells [20].
cZNF292 in glioma [13], circ-ITCH in lung can- Likewise, circPVT1 upregulation in gastric cancer
cer [14], and CDR1 (or ciRS-7) in the ischemic (GC) is postulated to have an effect on GC tissue
myocardium (MI) [15]. Circ-ITCH downregula- proliferation, acting as a sponge of miR-­125 [21].
tion has been observed in esophageal squamous A similar effect has been observed in other cancer
carcinoma (ESC). Circ-ITCH overexpression cells; for example, Huh7, HCT116, and HeLa
suppresses tumor growth via miR-7, miR-17, and cell proliferation was significantly reduced upon
miR-124 sponging to activate ITCH mRNA, knockdown of circHIPK3 [20]. circRNA profiling
which then undergoes degradation and ubiquiti- has yielded positive insights in the classification of
nation by phosphorylated Dvl2, ultimately circRNA biomarkers in cancer.
repressing the Wnt/β-catenin pathway [16]. In Expanding on the potential role of circRNAs
lung cancer, circ-ITCH expression in tumor tis- as disease biomarkers, several studies have
sues was significantly decreased compared to the explored the function and implication of cir-
adjacent non-tumor tissues (n = 78) [14]. Ectopic cRNAs in metabolic and non-cancer diseases.
expression of circ-ITCH results in increased One such circRNA is ciRS-7, a well-identified
expression of the ITCH gene, its parental cancer-­ circRNA that acts as a miR-7 a/b sponge or inhib-
suppressive gene. This subsequently leads to itor in brain tissues or islet cells [22]. Hypoxia
inhibition of lung cancer cell proliferation. circ-­ treatment led to increased expression of Cdr1as
ITCH is a sponge of the oncogenic miR-7 and and miR-7a in mice with myocardial infarction
miR-214, which are important for increasing (MI), causing an increase in the cardiac infarct
ITCH expression and suppressing Wnt/β-catenin size [15]. Cdr1as overexpression in mouse car-
signaling activation [14]. cZNF292 is crucial for diomyocytes (MCM) promotes apoptosis and is
tube formation in glioma, and cZNF292 silenc- reversed by miR-7a overexpression. Sp1 tran-
ing reduced cell proliferation and cell cycle pro- scription factor (SP1) and poly (ADP-ribose)
gression in human glioma U87MG and U251 cell polymerase (PARP) were both targeted by miR-
lines [13]. Here, the S/G2/M phase in cell cycle ­7a and induced apoptosis under hypoxia treat-
progression was arrested through the Wnt/β-­- ment [15]. Myosin VIIA and Rab-interacting
catenin signaling pathway and several genes, protein (Myrip) and paired box 6 (Pax6) activa-
including PRR11 (proline rich 11), CCNA (cyclin tion were observed as the downstream targets of
A), p-CDK2 (phosphorylated cyclin-dependent miR-7 [23], indicating that ciRS-7 is a potential
kinase 2), VEGFR1/2 (vascular endothelial alternative diagnostic strategy in metabolic dis-
growth factor 1/2), p-VEGFR1/2, and EGFR ease, namely, diabetes. However, further studies
(epidermal growth factor receptor) [13]. are required to rule out the functional roles of
In addition to their roles as biomarkers, cir- ciRS-7 and to prove whether it truly has any
cRNAs are involved in many cellular processes, implications in the advancement of diabetes.
28 Prospective Advances in Circular RNA Investigation 359

2.2 circRNAs in Agriculture linear mRNA. circRNAs are commonly config-

ured from pre-mRNA. They are organized by the
In humans, circRNA biogenesis is dependent on joining of the 3′ end of a downstream exon of a
RNA polymerase II transcription and the back-­ gene and then back-spliced to the 5′ end of the
splicing reaction of the precursor mRNA (pre-­ first exon, forming the covalently circularized
mRNA). In plants, circRNAs are localized in the RNA [1].
nucleus or other organelles, such as the mito- circRNA back-splicing is generated by the
chondria and chloroplast, or outside the cells, presence of inverted repeat sequences or Alu
such as with viroids [6]. The functional assess- repeat elements in the introns flanking the exons
ments of circRNAs in plants are still in their that bring the spliced sites of exons at the 3′ end
infancy. However, circRNAs may be involved in to the 5′ end of the first exon in close proximity
plant developmental processes and stress to one another via base-pairing [12]. circRNAs
responses (abiotic versus biotic stress), and are mainly derived from exonic coding regions
similar to humans, they also serve as miRNA and can appear in three different forms: exonic
sponges. circRNAs are tissue organelle depen- circRNAs (e-circRNAs), intronic circRNAs (ciR-
dent, indicating that they play a role in plant NAs), and retained-intron circRNAs
organelle development [7, 24] and can be distin- (ElcircRNAs).
guished via RNA-seq datasets. There is evidence circRNAs have become a hot topic for
that circRNAs also contribute to malnutrition in research with an enormous amount of research
Oryza sativa (rice) and barley [25, 26]. The being performed to understand their mecha-
described regulatory pathways indicate that a nisms. Solid evidence has emerged by in vivo
number of circRNAs may be involved in cellular and in vitro studies. circRNAs are regulated by
metabolism, hormonal signaling, intracellular multiple mechanisms and serve various func-
protein sorting, respiration, and carbohydrate tions. To date, four functional roles have been
metabolism [25]. An RNA-seq dataset of reported for circRNAs. First, circRNAs act as
Arabidopsis thaliana showed that circRNAs are miRNA sponges or competing endogenous
clustered in the chloroplast, suggesting that mRNAs. Introducing circRNAs may have spong-
circRNAs may play a vital role in plant photo- ing effects on multiple miRNAs, altogether
synthesis [27]. The composition of circRNAs has abrogating miRNA binding to the target
been proven to be an important element in the mRNA(s) [29]. Second, circRNAs bind and
field of plant RNAs. In-depth studies aimed at sequester proteins. For example, circularized
understanding the mechanism behind plant muscleblind (MBL1) is highly abundant in
development are crucial, as they may benefit the Drosophila and humans [4, 30, 31]. The start site
future agriculture sector. coding sequence present in circMbl and the
binding sites of MBL1 mRNA signifies the
recruitment binding of MBL1, indicating the
3 Challenges to Current capacity of circRNAs to sequester proteins
circRNA Investigation through MBL1 ribosomes to initiate translation
in Drosophila [31]. It has been suggested that
3.1 Challenges in Regulating circRNA protein binding can also inhibit transla-
circRNA Biogenesis tion by acting as a competing counterpart, as
observed in cytoplasmic circ-FOXO3 (forkhead
It is accepted that circRNAs act as miRNA and box O3), where interactions with ID1 (inhibitor
RNA-binding protein (RBP) sponges in addition of DNA binding 1, HLH protein) and E2F1 (E2F
to roles in regulating gene transcription and transcription factor 1) [32] deplete activation in
expression, as well as protein/peptide translators the cytoplasm; it has also been implicated in the
[28]. circRNAs are formed in a circular tail-less promotion of cell cycle arrest in NIH3T3 mouse
mode at the 5′ end, as also seen in the canonical embryo fibroblast cells [33]. The formation of
360 S. A. Sulaiman et al.

circRNA–mRNA complexes plays an inhibitory emerging circRNA–miRNA–mRNA interactions

role in cellular and biological functions. require further validation. In addition, the studies
However, there is a crucial need for in-depth of circRNAs in cancer have not focused on the
exploration of the physical interactions of cir- mechanisms of initiation, progression, and
cRNAs with multiple target proteins to under- metastasis. In cancers, most of the studies
stand their actual mechanisms and pathways. involved identification of circRNAs, including
Third, circRNAs regulate mRNA splicing or lung, colorectal, breast, and bladder cancers as
transcription. ElcircRNAs are localized in the well as glioma [4]. Currently, several circRNAs
nucleus, implying that they will undergo alterna- have been identified to be upregulated or down-
tive splicing and gene transcription [12]. Unlike regulated in cancers. The analysis of circRNAs
the canonical linear mRNAs, transcripts contain- poses several challenges including in terms of
ing introns in the nucleus will be exported from experimental design, techniques to be used, and
the nucleus and be degraded in the cytoplasm most importantly, the bioinformatics tools to be
[34]. Intron-containing circRNAs (ElcircRNAs used. The current knowledge regarding the regu-
and ciRNAs) remain in the nucleus, indicating lation of circRNA biogenesis is limited, which
their involvement in mRNA transcriptional regu- renders the studies on circRNAs more difficult.
lation [4]. Nevertheless, ciRNAs do not have
defined roles or biological processes. Lastly, cir-
cRNAs can be translated. circRNA transforma- 3.2 Challenges in Experimental
tion by N6-­methyladenosine (m6A) gives rise to Design and Technique
the initiation of protein translation [35].
Translating ribosomes have been associated with CircRNAs can be identified using several
circRNAs known as ribo-circRNAs via ribo- approaches, including RNA-seq, microarray, and
some footprinting. Ribo-circRNAs harbor a con- reverse transcription (RT)-PCR [17, 40, 41]. As
served termination codon at which translation circRNAs are novel products of regulated alter-
initiation occurs by utilizing the start codon of native splicing processes, detecting splicing
the target mRNA bound to membrane-associated through RNA-seq is the most favorable method.
ribosomes [3]. Important evidence from in vitro However, there are several limitations, where cir-
and in vivo studies has demonstrated the transla- cRNA cannot be detected due to several RNA
tional activity of circRNAs. The protein encoded library preparation steps, including RNA purifi-
by Drosophila mbl has been detected via mass cation, RNA or cDNA size selection, and RNA
spectrometry, indicating that the translational fragmentation. For purification, rRNA-depleted
mechanism occurs in normal settings [3]. cir- libraries is much better compared to poly(A)
cRNAs have also been implicated as driver selected or depleted since they retain the cir-
mechanisms of protein synthesis. An internal cRNAs in a particular sample. circRNAs lack a
ribosome entry site (IRES) was stably integrated poly(A) tail; however, poly(A) enrichment step
in a circRNA minigene to induce expression will significantly reduce circRNAs, thus making
in vitro, and western blotting and flow cytometry them difficult to be detected by RNA-seq. For
showed that it activated a number of putative size selection, random priming is preferable as
proteins. This is further strong evidence that cir- this method is not biased where small sizes of cir-
cRNAs may also function as mRNAs to drive cRNAs can also be detected. The RNA needs to
protein synthesis [10]. be fragmented prior to adaptor ligation or prim-
Studies on the involvement of circRNAs in ing in order to distinguish between the small
various diseases are still in the infancy stage, RNA and circRNAs. In addition, RNA-seq proto-
where expression has been identified but further cols can introduce technical artifacts during
functional and population validation are still ­ligation and RT. The ligation step can produce
lacking [15, 36–39]. Some key functions of chimeric complementary DNAs (cDNAs), thus
circRNAs have been identified; however, the might generate low levels of artefactual cir-
28 Prospective Advances in Circular RNA Investigation 361

cRNAs [42]. On the other hand, reverse tran- cRNAs with low positive rates. However, these
scriptase can lead to extensive template-switching algorithms took 2–3 days to complete individual
artifacts, where two different RNA molecules can dataset predictions [40]. Hence, identifying cir-
join together and affect the discovery of novel cRNAs with high specificity and sensitivity using
RNA isoforms [43, 44]. RT can also lead to strand a single algorithm is challenging. The usage of
displacement, which subsequently leads to dif- more algorithms for circRNA identification is
ferences in circRNA quantitative measurement advisable [41]. More importantly, individual vali-
[45]. RT is highly dependent on good RNA dation is crucial for novel circRNAs with regard
quality, and most importantly, the right library to the type of algorithms used, as false-positive
preparation method should be selected [46]. results are common [40]. The technical chal-
Other techniques that can be used to detect cir- lenges to studying circRNAs are discussed in the
cRNAs include the microarray and RT-quantitative next section.
(Q) PCR techniques [17]. Most genome-wide
studies use microarray for detecting circRNAs,
but it is performed in limited numbers of cancer 4 Comparisons
types and their adjacent normal tissue samples. Between Current circRNA
However, microarray cannot identify novel cir- Investigation Methods
cRNAs. RT-QPCR is commonly used to validate
the circRNAs identified in the discovery phase in 4.1 Comparison Between RNA
a larger sample size. Currently, most circRNA Preparation Methods
studies have been performed retrospectively; thus for Identification of circRNAs
conducting a large prospective clinical trial is
important for confirming circRNAs as disease As circRNAs are biologically closed loops of
biomarkers. The details regarding the challenges RNA and lack open ends, it is quite difficult to
in bioinformatics are discussed in the next identify and validate their presence. To date,
section. there are three main methods of screening for the
presence of circRNAs in biological samples:
microarray, RNA-seq, and bioinformatics analy-
3.3 Challenges in Bioinformatics ses of readily available data (Table 28.1) [48, 49].
Analysis Nevertheless, before proceeding with any analy-
sis, it is important to enrich the circRNA popula-
CircRNA identification in single-end (SE) RNA-­ tion and remove any other linear RNAs that may
seq data is performed by aligning the reads to the skew the population [24, 50]. Total cellular RNA
back-splice junctions. Many available algorithms contains a wide array of RNA species and arti-
can be used to identify circRNAs from the facts, and to avoid false positives, samples are
sequence data; however, their performance usually treated with RNAse R [24] to digest all
differs [47]. Five circRNA prediction algorithms, other linear RNAs with >7 nucleotides, which
i.e., circRNA_finder, CIRCexplorer, find_circ, will exclude circRNAs because they lack the
CIRI, and MapSplice, have different false-­ required 3′ end [24]. For microarray, only one
positive and false-negative circRNA detection available platform is Arraystar. This platform uti-
rates [40]. The ability of the algorithms to iden- lizes circRNA information from six published
tify a circRNA based on the splice site distance databases [24, 51–55]. Currently, this is the most
also varies. CIRI and circRNA_finder can iden- straightforward approach for identifying cir-
tify circRNAs with very low proximal splice sites cRNAs from human samples because it does
(<100 bp) compared to other algorithms; how- not require extensive bioinformatics analysis. A
ever, most of the circRNAs found were false pos- more comprehensive approach to identifying
itives [40]. On the other hand, CIRCexplorer and circRNAs is by sequencing the transcriptome
MapSplice produce the most reliable list of cir- population. Current RNA-seq library preparations
362 S. A. Sulaiman et al.

Table 28.1 Pros and cons of different methods to identify and validate circular RNAs
Method Pros Cons
Identification
1. Microarray Cost-­effective Only cover reported species of circular RNA
Some targets have been wet Cannot be used for new discovery of circular
lab-validated RNAs
Known targets, easier to validate Human, mouse, rat targets only [78]
Require a larger amount of starting material
Most targets are exonic and intronics circular
RNAs
2. RNA-Seq Less starting material is required Library preparations can be biased against
than microarray circular RNAs
Can discover new species circular High cost
RNAs
Can cover both intronic, intergenic, Bioinformatics pipeline is not well established yet
and exonic circular RNAs [66]
3. Bioinformatics Can reuse the data for other Needs more validation
approach of readily analyses Bioinformatics pipeline is not well established yet
available data Most of the library preparation protocol does not
enrich for circular RNA population [66]
Require higher expertise in bioinformatics
Validation
1. qRT-PCR Cost-­effective False positives can happen
Readily available in most labs Can amplify linear splicing events also [60]
Rolling RT effects [60]
2. DdPCR Less false positives as it eliminates High cost
rolling RT effects [60] Requires special machine/equipment to perform
3. Northern Blot Specific target [61] Low sensitivity [95]
Less false positives [47] Laborious and time-consuming [62]
4. FISH/ISH Less false positives Low sensitivity [96]

vary greatly, particularly in terms of biochemical of the circRNAs and identify the junction span-
preparation [41]. RNA purification is among the ning 5′ and 3′ ends [58]. From there, divergent
first steps in RNA-seq and can greatly influence primers can be designed to flank the junction
circRNA population [56]. For example, the poly sequence and use sequencing [58, 59]. This is to
(A) selection in RNA-seq library preparation can avoid the primers amplifying random linear
heavily deplete the circRNA population, as cir- splicing events [59]. ddPCR is a new technology
cRNAs do not contain poly (A) tails [41]. The that enables higher sensitivity in nucleic acids
size selection generally used in RNA-seq quantification [60]. It is more reliable and accu-
excludes nucleotides of <200 bp [24]. This would rate than qRT-­PCR for determining the quantity
exclude smaller circRNAs (even if present) of circRNAs, as it eliminates the effect of rolling
unless a small RNA-seq library kit is used [24]. RT products [60]. Northern blotting is another
Therefore, the most reliable, cost- and time-­ commonly used method for detecting circRNAs
effective method for identifying circRNAs at the [61, 62]. Similar to qRT-PCR, probes are
moment is microarray. The qRT-PCR approach is designed to target the junction sequence and are
commonly used to validate the presence of cir- detected on blots. Sometimes, northern blotting
cRNAs [57]. Other methods such as droplet digi- is preferred because the species targeted is
tal PCR (ddPCR), fluorescence in situ known, thus reducing false positives [47, 62].
hybridization, and northern blotting are also used Nevertheless, it remains an unpopular choice due
[57]. It is imperative to determine the sequence to its time-­consuming and laborious procedure
28 Prospective Advances in Circular RNA Investigation 363

[62]. Fluorescence in situ hybridization can also alignment to a reference genome without gene
be used to detect circRNAs [63, 64]. It utilizes annotation. There is a slight difference between
fluorescence-­coupled probes targeting the junc- the tools (Table 28.2) in terms of executing this
tional sequence [63, 64]. All four methods have approach. Segemehl [68], MapSplice [69],
been proven to be able to detect circRNAs, but CIRCexplorer [53], circRNA_finder [74], and
some are preferred over others. Table 28.1 com- double conjugated clustering (DCC) [75] create
pares the advantages and disadvantages of each an algorithm that aligns the unmapped reads to
method. spliced sites for detecting BSJ sequences. By
contrast, UROBORUS [66] and find_circ [51]
remove mapped reads and extract the first and
4.2 Comparison last 20-bp anchor sequences from the unmapped
Between Bioinformatics Tools reads before identifying BSJ sequences from the
for Identifying circRNAs mapping of these anchors. CIRI [71, 73] per-
forms local alignment using BWA-MEM to iden-
CircRNA identification in RNA-seq data analysis tify paired chiastic clipping (PCC) for de novo
was not performed until the use of splicing analy- detection of BSJ sequences and filters systemati-
sis [41, 65]. Therefore, numerous bioinformatics cally to eliminate false positives. The second
pipelines and algorithms have been developed for method is a pseudo-reference approach that cre-
circRNA detection [51, 53, 66–75], including SE ates a pseudo-sequence database or de novo pre-
or paired-end (PE) RNA-seq data (Table 28.2). diction of BSJs by combining the reference
These tools or algorithms use different and genome with the corresponding gene annotation
unique combinations of identification strategies to identify circRNAs (Table 28.2) [76].
based on the reference genome and annotations, Posttranscriptional exon shuffling (PTES)Finder
mapper of choice, back-splicing junction (BSJ) [67] and non-co-linear transcripts (NCL)Scan
identification methods, and other filtering options [70] create putative circRNA sequences based on
[41, 76]. Typically, the flow of the bioinformatics mapping information of the exon-to-exon junc-
analysis starts from the removal of mapped reads, tion sequences after alignment to the reference
followed by the mapping of splice junction sites genome. However, Known and Novel IsoForm
from the unmapped reads to identify the cir- Explorer (KNIFE) [72] creates all potential exon-­
cRNAs [76, 77]. Among these tools, the majority, to-­exon junction sequences from gene annota-
except segemehl [68], use external aligners such tions before the alignment. Additionally, KNIFE
as Bowtie [78, 79], Burrows-Wheeler Aligner uses mRNA forward-spliced junction reads to
(BWA) [23, 80], or Spliced Transcripts Alignment construct a forward-spliced sites database [72]
to a Reference (STAR) [81] to filter out the and also for further removal of candidate reads
sequencing reads aligned to the reference genome aligned to both databases to improve accuracy.
(mapped reads) [41, 76]. Previous comparison studies have shown that
To identify circRNAs from unmapped reads, most circRNA algorithms have little overlap
two methods or algorithms are used [76]. The between the prediction results, and no single
first method uses the split-alignment approach gold-standard algorithm is sensitive and specific
that focuses on BSJ reads to identify circRNAs. enough for all data [40, 77, 82, 83]. Using rRNA-­
The algorithm is based on the fact that BSJ reads depleted libraries with or without RNAse R treat-
are aligned to the reference sequences in reverse, ment, up to 76% and 28% of false positives were
particularly to the established GU-AG splicing detected in rRNA-depleted libraries and rRNA-­
sequences flanking the splice sites [76]. The depleted RNAse R–treated libraries, respectively
split-alignment approach can be gene annotation [40]. Testing these algorithms with simulated
dependent, where the BSJs are identified in data improved sensitivity and specificity,
between annotated exons, or gene annotation although there was a trade-off between sensitiv-
independent, where the BSJs are identified from ity and specificity, in which the algorithms with
 364

Table 28.2 Comparison between circRNA identification and prediction tools

Sequencing Mappers Splice-­ Annotation CircRNAs Analysis by previous studies
Computational read types aware detection method Precision Precision False positives False positives
tools (HeLa, %) (Hs68, %) (HeLa, n) (Hs68, n) References
CIRCexplorer PE and SE TopHat/STAR Yes Yes Split-alignment 50.09 68.54 1388 1856 [53]
circRNA_ PE and SE STAR Yes No Split-alignment 46.32 58.54 1597 2094 [74]
finder
CIRI PE and SE BWA-MEM No No Split-alignment 54.20 69.49 3210 3400 [71, 73]
DCC PE and SE STAR Yes Yes Split-alignment 45.22 63.08 1760 2107 [75]
find_circ SE Bowtie2 No No Split-alignment 36.99 59.75 2092 2377 [51]
KNIFE PE and SE Bowtie, No Yes Pseudo-­ 44.26 69.49 2055 2359 [72]
Bowtie2 reference
MapSplice PE and SE Bowtie No Yes Split-alignment 54.21 76.33 1765 1854 [69]
NCLscan PE and SE BWA, BLAT, Mixed Yes Pseudo-­ 45.06 64.73 954 892 [70]
Novoalign reference
PTESFinder PE and SE Bowtie, No Yes Pseudo-­ 35.65 63.29 2054 2474 [67]
Bowtie2 reference
UROBORUS PE and SE Bowtie, Mixed Yes Split-alignment 31.00 19.73 761 279 [66]
Bowtie2,
TopHat
Segemehl PE and SE Per Se No No Split-alignment 14.32 8.78 2506 3094 [68]
Summary of the 11 known circRNA identification tools including their precision analysis done by previous study [76, 77]
Note: MapSplice is a splice-aware aligner but uses Bowtie read mapper
PE paired-end, SE single-end
S. A. Sulaiman et al.
28 Prospective Advances in Circular RNA Investigation 365

the highest specificity also had the lowest sensi- little information for detection [41, 76]; thus
tivity, and vice versa [40]. Therefore, none of algorithms such as CIRI [71, 73] can detect
these tools share an exact pipeline of analysis and unbalanced BSJ reads due to its dynamic align-
adapt to a combination of selection criteria and ment algorithm, which can identify balanced BSJ
filtering. sequences to control the FDR. The detection of
One such difference is the implementation of unbalanced BSJs using balanced BSJ reads may
gene annotation, which is often preferred for its require an additional step to improve the accu-
improved accuracy and sensitivity for BSJ detec- racy, especially for low-abundance circRNAs.
tion in comparison to pseudo-reference or de The usage of multiple seed matching improves
novo prediction [82, 83]. Annotation-dependent the recovery of unbalanced BSJs [73]. In this
algorithms only detect uniquely mapped reads study [73], CIRI created two putative genomic
and require canonical splice signals [41, 76]; thus regions from short segments corresponding to
this can cause a blind spot for detecting circRNA BSJs and forward-spliced sequences. Then, it
isoforms (non-canonical signals). In other words, divided the short segments or potential unbal-
the pseudo-reference approach offers more anced BSJ reads into multiple short seeds and
flexibility for detecting non-canonical splice sig- counted the rate those seeds were matched to the
nals, but at the cost of high false positives. Even genomic region [73]. By comparing the matching
so, most of the non-canonical signals are small; counts and seed locations, the exact position of
thus many pseudo-reference algorithms choose the unbalanced BSJs or short segments could be
to implement GU-AG splicing sequences as a determined [73]. This evidence therefore indi-
default filter to control the false discovery rate cates that CIRI is a more preferable algorithm for
(FDR) [41, 76], therefore potentially undermin- those who need to identify low-abundance
ing the detection of non-canonical splice signals. circRNAs.
Importantly, this dependency on annotation may
create a blind spot in the algorithms for detecting
novel circRNAs. 4.3 Comparison of circRNA
Another important point is the type of mapper Databases and Biological
used in the sequence mapping. find_circ [51], Function Tools
CIRI [71, 73], KNIFE [72], and PTESFinder [67]
use the most standard reference aligners (Bowtie Besides circRNA identification tools, other com-
and BWA), which are not specifically designed to putational tools have been developed to investi-
note splice sites. Excluding KNIFE [72], these gate the biological function of circRNAs by
algorithms implement the split-alignment focusing on databases, differential expression,
approach or 20-bp anchor filters to control sensi- quantification, visualization, and length assem-
tivity, FDR, and specificity. Meanwhile, splice-­ bly. Currently, the available circRNA databases
aware mappers such as TopHat [84, 85], STAR are circBase (http://www.circbase.org/) [86], cir-
[81], and Novoalign have been optimized and cRNADb (http://reprod.njmu.edu.cn/circrnadb)
updated with the current reference genome, thus [14], circ2Traits (http://gyanxet-beta.com/
making them more convenient. One disadvantage circdb/) [87], and CircNet (http://circnet.mbc.
is that splice-aware mappers have more restricted nctu.edu.tw/) [88]; they are all freely accessible
reporting modes, filtering, and arguments and online databases. circBase is a database of
may not be able to detect circRNAs in non-­ eukaryote circRNAs with evidence of their
canonical signals [41, 76]. expression [86]. It supports the genome assem-
Besides that, the sensitivity and accuracy of blies of Homo sapiens (hg19), Mus musculus
these algorithms also depend on the recovery of (mm9), Caenorhabditis elegans (ce6), Latimeria
unbalanced BSJ reads [41, 76]. Standard 20-bp chalumnae and L. menadoensis (latCha1), and D.
anchor sequence filtering cannot capture the very melanogaster (dm3) [86]. circRNADb [14] con-
short reads of unbalanced BSJs because there is tains 32,914 annotated e-circRNAs with informa-
366 S. A. Sulaiman et al.

tion on the circRNA genetic sequences, exon needed to further explore circRNAs, their bio-
splicing, IRES, open reading frames (ORF), and genesis, and functions.
references [14]. circ2Traits [87] reports the asso-
ciations of circRNAs and disease by estimating
the likelihood of a circRNA being associated 5 Suggestions on Methods
with miRNA in the disease and the network pre- to Investigate circRNAs
diction of these associations between circRNAs,
miRNAs, protein-coding genes, and long non-­ 5.1 Suggestions for Improvement
coding RNAs [87]. CircNet accounts for cir-
cRNA isoforms and incorporates a novel naming For identifying circRNAs, technology providers
system for antisense circRNAs or ciRNAs and should create more microarray platforms. This
the location of BSJ sites on well-annotated exons will help reduce bias and create more opportuni-
[88]. CircNet also provides information on novel ties to include other types of newly discovered
circRNAs, integrated miRNA–circRNA net- circRNAs. For RNA-seq, the biochemical steps
works, expression profiles, genomic annotations, in the library preparation protocol should be
and circRNA isoform sequences [88]. established and validated [41]. A proper guide-
A few application tools have been developed line on enriching circRNAs and removing arti-
for investigating the biological functions of cir- facts and linear splicing events should be
cRNAs. Both Sailfish-cir [39] and CircTest [75] implemented. For circRNA validation, both
have been used to identify circRNA differential microarray and RNA-seq have their advantages
expression or quantification. Sailfish-cir [39] and disadvantages, but perhaps two or three
estimates circRNA expression based on BSJ methods can be used in combination to confirm
reads and other related reads from the data, while the presence of circRNAs. There should also be
CircTest [75] estimates circRNA expression minimal requirements or criteria for publishing
based on the BSJ read counts distributed in the circRNA results, for example, whether the cir-
samples. An additional algorithm in CIRI, cRNAs were enriched using enzymatic reaction
CIRI-AS, which refers to alternative splicing, can and how the primers/probes targeting the junc-
be used for investigating the internal structure tional sequences were designed and validated.
and alternative splicing of circRNAs [89]. In terms of the bioinformatics pipelines for
Similarly, three other tools have been developed circRNA identification and prediction, there is no
to characterize and visualize circRNA function single perfect and accurate algorithm or analysis
and structure. Full Circular RNA Characterization pipeline that has been developed so far [41, 76].
from RNA-seq (FUCHS) [90] was developed to circRNA detection in RNA-seq reads requires
detect the known exon-skipping events within careful consideration, as most tools depend on
circRNAs, whereas CircPro [91] can be used to BSJ reads that have significant read density bias,
identify the protein-coding potential of cir- and some of the sequence reads at the exon
cRNAs. CircView [92] can be used to aid visual- boundaries are usually associated with sequenc-
ization and understanding of circRNA functions, ing errors [2, 10, 37]. Therefore, these can lead to
particularly regulatory elements such as miRNA false-positive alignments at BSJ reads which will
response elements (MREs) and RBP binding not be able to represent a truly expressed
sites. For understanding circRNA interaction circRNA.
with proteins and miRNAs, CircInteractome is a Due to the high rate of inconsistency between
user-friendly, web-based tool for exploring cir- the bioinformatics tools discussed above, a vali-
cRNA interactions based on their potential bind- dation approach is needed to assess the accuracy
ing sites [93]. However, more such tools are and performance of these algorithms. One such
28 Prospective Advances in Circular RNA Investigation 367

method is using simulated data to assess the sen- most reliable algorithm. Therefore, developing
sitivity, specificity, and limitations [41] and to a reference-free algorithm with greater sensitiv-
identify the trade-off values, false positives, and ity and accuracy is the current challenge to
false negatives so that the user can identify the improving circRNA investigation.
tool that best suits their needs. An example of
simulated data is Benchmarker for Evaluating the
Effectiveness of RNA-Seq Software (BEERS) 6 Conclusions
[94], which simulates human and mouse PE and Perspectives
RNA-seq data (Illumina) with differential expres-
sion of genes, splicing events, sequencing errors, It is apparent that circRNAs are now coming of
variants, and insertions/deletions (indels). It is age and are being rediscovered as important reg-
important to note that experiment-generated data ulators of the cellular system. The biological
are more complex than simulated data, and it is functions of circRNAs include miRNA sponging,
unknown how similar simulated data are in com- protein binding and sequestration, regulation of
parison to experimental data [41]. Following that, mRNA splicing, and transcription of circRNAs
a study evaluating these algorithms found that that can be translated. circRNAs are involved in
CIRI [71, 73], CIRCexplorer [53], and KNIFE many cellular processes leading to cancer and
[72] showed more balanced performance but that non-cancer diseases. In cancer, circRNAs are
their sensitivity, specificity, and cost had a few important for regulating cell death and senes-
shortcomings [77]. To improve these algorithms, cence and the proliferation, migration, and inva-
a recent study suggested that combining predic- sion and inflammation processes. In non-cancer
tion algorithms and selecting the commonly diseases, such as diabetes, circRNAs are impor-
identified circRNAs resulted in reduced false tant players in the metabolic and inflammation
positives and improved the accuracy [82], partic- pathways. Nevertheless, there is still much
ularly for CIRI, find_circ, and UROBORUS, uncharted territory when it comes to understand-
which benefited the most. An improved algorithm ing the role of circRNAs. Most current circRNA
that can eliminate the need for strict post-filtering investigations are generally dependent on RNA-­
is required to provide more flexibility in explor- seq data, and many experimental designs and
ing novel circRNAs. library preparations are tailored for this purpose.
Rapid identification of circRNAs has Thus, this may introduce potential bias in pooling
revealed that the existing circRNAs are more circRNA populations for identification and pre-
diverse than expected. This could mean that diction analysis. Moreover, most if not all the
some circRNAs may have been overlooked. algorithms that have been developed for identify-
Identifying these novel circRNAs is by nature ing and predicting circRNA have little agree-
difficult due to the many confounding factors, ment, and their sensitivity and specificity require
and a reference- or annotation-free approach improvement. Given the widespread existence of
algorithm would be able to identify such cir- circRNAs in the human genome, it is possible
cRNAs, but at the expense of a higher FDR [41, that current technologies in circRNA investiga-
76]. This is due to the fact that most circRNAs tion may overlook other circRNA isoforms or
are derived from novel and unannotated splice low-abundance circRNAs that could be biologi-
sites and are often less likely to be true positives cally meaningful. In conclusion, circRNAs have
and with reference-free algorithms having a promising potential as disease biomarkers and
reduced accuracy of 13–27% [82]. Interestingly, regulators; however, understanding of their bio-
the performance of CIRI, a reference-­free algo- logical functions and interaction with other
rithm, is comparable to that of annotation algo- known regulators is currently limited and needs
rithms [82], suggesting that it is currently the further studies.
368 S. A. Sulaiman et al.

References apoptotic genes promotes lead-induced neuronal cell

apoptosis. Arch Toxicol 91(4):1671–1684
19. Li Y, Zheng F, Xiao X, et al (2017) CircHIPK3
1. Haque S, Harries LW (2017) Circular RNAs
sponges miR-558 to suppress heparanase
(CircRNAs) in health and disease. Genes 8(12):1–17
expression in bladder cancer cells. EMBO
2. Jeck WR, Sharpless NE (2014) Detecting and charac-
Rep:e201643581-e201643581.
terizing circular RNAs. Nat Biotechnol 32(5):453–461
20. Zheng Q, Bao C, Guo W et al (2016) Circular
3. Pamudurti NR, Bartok O, Jens M et al (2017)
RNA profiling reveals an abundant circHIPK3 that
Translation of CircRNAs. Mol Cell 66(1):9–21 e27
­regulates cell growth by sponging multiple miRNAs.
4. Qu S, Liu Z, Yang X et al (2018) The emerging func-
Nat Commun 7:1–13
tions and roles of circular RNAs in cancer. Cancer
21. Chen J (2016) Circular RNA profile identifies cir-
Lett 414(2018):301–309
cPVT1 as a proliferative factor and prognostic marker
5. Tay Y, Rinn J, Pandolfi PP (2014) The multilayered
in gastric cancer. (December):2016–2016
complexity of ceRNA crosstalk and competition.
22. Hansen TB, Jensen TI, Clausen BH et al (2013)
Nature 505(7483):344–352
Natural RNA circles function as efficient microRNA
6. Lee SM, Kong HG, Ryu CM (2017) Are circu-
sponges. Nature 495:384
lar RNAs new kids on the block? Trends Plant Sci
23. Wang Y, Liu J, Liu C et al (2013) MicroRNA-7 regu-
22(5):357–360
lates the mTOR pathway and proliferation in adult
7. Liu T, Zhang L, Chen G et al (2017) Identifying and
pancreatic β-cells. Diabetes 62(3):887–895
characterizing the circular RNAs during the lifespan
24. Salzman J, Chen RE, Olsen MN et al (2013) Cell-type
of Arabidopsis leaves. Front Plant Sci 8(July):1–9
specific features of circular RNA expression. PLoS
8. Dou Y, Cha DJ, Franklin JL et al (2016) Circular
Genet 9(9):e1003777
RNAs are down-regulated in KRAS mutant colon
25. Darbani B, Noeparvar S, Borg S (2016) Identification
cancer cells and can be transferred to exosomes. Sci
of circular RNAs from the parental genes involved
Rep 6(October 2015):1–11
in multiple aspects of cellular metabolism in barley.
9. Meng S, Zhou H, Feng Z et al (2017) CircRNA: func-
Front Plant Sci 7(June):1–8
tions and properties of a novel potential biomarker for
26. Ye CY, Chen L, Liu C et al (2015) Widespread
cancer. Mol Cancer 16(1):1–8
noncoding circular RNAs in plants. New Phytol
10. Wang Y, Wang Z (2015) Efficient backsplic-
208(1):88–95
ing produces translatable circular mRNAs. RNA
27. Sun X, Wang L, Ding J et al (2016) Integrative analy-
21(2):172–179
sis of Arabidopsis thaliana transcriptomics reveals
11. Zhong Z, Lv M, Chen J (2016) Screening differential
intuitive splicing mechanism for circular RNA. FEBS
circular RNA expression profiles reveals the regula-
Lett 590(20):3510–3516
tory role of circTCF25-miR-103a-3p/miR-107-CDK6
28. Han Y-N, Xia S-Q, Zhang Y-Y et al (2017) Circular
pathway in bladder carcinoma. Sci Rep 6(November
RNAs: a novel type of biomarker and genetic tools in
2015):1–12
cancer. Oncotarget 8(38):64551–64563
12. Zhang ZC, Guo XL, Li X (2018) The novel roles
29. Yang W, Du WW, Li X et al (2016) Foxo3 activity
of circular RNAs in metabolic organs. Genes Dis
promoted by non-coding effects of circular RNA and
5(1):16–23
Foxo3 pseudogene in the inhibition of tumor growth
13. Yang P, Qiu Z, Jiang Y et al (2016) Silencing of
and angiogenesis. Oncogene 35(30):3919–3931
cZNF292 circular RNA suppresses human glioma
30. Ashwal-Fluss R, Meyer M, Pamudurti NR et al (2014)
tube formation via the Wnt/β-catenin signaling path-
CircRNA biogenesis competes with pre-mRNA splic-
way. Oncotarget 7(39):63449–63455
ing. Mol Cell 56(1):55–66
14. Wan L, Zhang L, Fan K et al (2016) Circular RNA-­
31. Kramer MC, Liang D, Tatomer DC et al (2015)
ITCH suppresses lung cancer proliferation via inhib-
Combinatorial control of Drosophila circular RNA
iting the Wnt/β-Catenin pathway. Biomed Res Int
expression by intronic repeats, hnRNPs, and SR pro-
2016:1579490
teins. Genes Dev 29(20):2168–2182
15. Geng H-H, Li R, Su Y-M et al (2016) The circular
32. Du WW, Yang W, Chen Y et al (2017) Foxo3 circu-
RNA Cdr1as promotes myocardial infarction by
lar RNApromotes cardiac senescence by modulating
mediating the regulation of miR-7a on its target genes
multiple factors associated with stress and senescence
expression. PLoS One 11(3):e0151753
responses. Eur Heart J 38(18):1402–1412
16. Li F, Zhang L, Li W et al (2015) Circular RNA ITCH
33. Du WW, Yang W, Liu E et al (2016) Foxo3 circular
has inhibitory effect on ESCC by suppressing the
RNA retards cell cycle progression via forming ter-
Wnt/β-catenin pathway. Oncotarget 6(8):6001–6013
nary complexes with p21 and CDK2. Nucleic Acids
17. Kristensen LS, Hansen TB, Venø MT et al (2018)
Res 44(6):2846–2858
Circular RNAs in cancer: opportunities and chal-
34. Braunschweig U, Barbosa-Morais NL, Pan Q et al
lenges in the field. Oncogene 37(5):555–565
(2014) Widespread intron retention in mammals
18. Nan A, Chen L, Zhang N et al (2017) A novel regula-
functionally tunes transcriptomes. Genome Res
tory network among LncRpa, CircRar1, MiR-671 and
24(11):1774–1786
28 Prospective Advances in Circular RNA Investigation 369

35. Yang Y, Fan X, Mao M et al (2017) Extensive transla- 51. Memczak S, Jens M, Elefsinioti A et al (2013)
tion of circular RNAs driven by N 6 -methyladenosine. Circular RNAs are a large class of animal RNAs with
Cell Res 27(5):626–641 regulatory potency. Nature 495:333
36. Cui X, Niu W, Kong L et al (2016) hsa_cir- 52. Zhang Y, Zhang X-O, Chen T et al (2013)
cRNA_103636: potential novel diagnostic and Circular intronic long noncoding RNAs. Mol Cell
therapeutic biomarker in major depressive disorder. 51(6):792–806
Biomark Med 10(9):943–952 53. Zhang X-O, Wang H-B, Zhang Y et al (2014)
37. Guo X-Y, Chen J-N, Sun F et al (2017) cir- Complementary sequence-mediated exon circulariza-
cRNA_0046367 prevents Hepatoxicity of lipid per- tion. Cell 159(1):134–147
oxidation: an inhibitory role against hepatic steatosis. 54. Guo JU, Agarwal V, Guo H et al (2014) Expanded
Oxid Med Cell Longev 2017:3960197 identification and characterization of mammalian cir-
38. Hansen TB, Kjems J, Damgaard CK (2013) cular RNAs. Genome Biol 15(7):409
Circular RNA and miR-7 in Cancer. Cancer Res 55. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular
73(18):5609–5612 RNAs are abundant, conserved, and associated with
39. Tang C-M, Zhang M, Huang L et al (2017) ALU repeats. RNA 19(2):141–157
CircRNA_000203 enhances the expression of 56. Cooper DA, Cortés-López M, Miura P (2018)
fibrosis-­associated genes by derepressing targets of Genome-wide circRNA profiling from RNA-seq data.
miR-26b-5p, Col1a2 and CTGF, in cardiac fibro- In: Dieterich C, Papantonis A (eds) Circular RNAs:
blasts. Sci Rep 7:40342 methods and protocols. Springer, New York, pp 27–41.
40. Hansen TB, Venø MT, Damgaard CK et al (2016) https://doi.org/10.1007/978-1-4939-7562-4_3
Comparison of circular RNA prediction tools. Nucleic 57. Zhang Y, Liang W, Zhang P et al (2017) Circular
Acids Res 44(6):e58–e58 RNAs: emerging cancer biomarkers and targets. J Exp
41. Szabo L, Salzman J (2016) Detecting circular RNAs: Clin Cancer Res 36(1):152
Bioinformatic and experimental challenges. Nat Rev 58. Heumüller AW, Boeckel J-N (2018) Characterization
Genet 17:679–692 and validation of circular RNA and their host gene
42. Quail MA, Kozarewa I, Smith F et al (2008) A mRNA expression using PCR. In: Dieterich C,
large genome centre’s improvements to the Illumina Papantonis A (eds) Circular RNAs: methods and pro-
sequencing system. Nat Methods 5(12):1005–1010 tocols. Springer, New York, pp 57–67. https://doi.
43. Houseley J, Tollervey D (2010) Apparent non-­ org/10.1007/978-1-4939-7562-4_5
canonical trans-splicing is generated by reverse tran- 59. Panda AC, Abdelmohsen K, Gorospe M (2017)
scriptase in vitro. PLoS One 5(8):e12271 RT-qPCR detection of senescence-associated cir-
44. Luo GX, Taylor J (1990) Template switching by cular RNAs. Methods Mol Biol (Clifton, N.J.)
reverse transcriptase during DNA synthesis. J Virol 1534:79–87
64(9):4321–4328 60. Chen D-F, Zhang L-J, Tan K et al (2018) Application
45. Kelleher CD, Champoux JJ (1998) Characterization of droplet digital PCR in quantitative detection of the
of RNA strand displacement synthesis by Moloney cell-free circulating circRNAs. Biotechnol Biotechnol
murine leukemia virus reverse transcriptase. J Biol Equip 32(1):116–123
Chem 273(16):9976–9986 61. Capel B, Swain A, Nicolis S et al (1993) Circular
46. Panda AC, De S, Grammatikakis I et al (2017) High-­ transcripts of the testis-determining gene Sry in adult
purity circular RNA isolation method (RPAD) reveals mouse testis. Cell 73(5):1019–1030
vast collection of intronic circRNAs. Nucleic Acids 62. Schneider T, Schreiner S, Preußer C et al
Res 45(12):e116–e116 (2018) Northern blot analysis of circular RNAs.
47. Barrett SP, Salzman J (2016) Circular RNAs: analy- In: Dieterich C, Papantonis A (eds) Circular
sis, expression and potential functions. Development RNAs: methods and protocols. Springer
143(11):1838 New York, New York, pp 119–133. https://doi.
48. Zheng Y, Zhao F (2018) Detection and reconstruc- org/10.1007/978-1-4939-7562-4_10
tion of circular RNAs from transcriptomic data. In: 63. Kocks C, Boltengagen A, Piwecka M et al (2018)
Dieterich C, Papantonis A (eds) Circular RNAs: Single-molecule fluorescence in situ hybridiza-
methods and protocols. Springer, New York, pp 1–8. tion (FISH) of circular RNA CDR1as. In: Dieterich
https://doi.org/10.1007/978-1-4939-7562-4_1 C, Papantonis A (eds) Circular RNAs: methods and
49. Jakobi T, Dieterich C (2018) Deep computational cir- protocols. Springer New York, New York, pp 77–96.
cular RNA analytics from RNA-seq data. In: Dieterich https://doi.org/10.1007/978-1-4939-7562-4_7
C, Papantonis A (eds) Circular RNAs: methods and 64. Zirkel A, Papantonis A (2018) Detecting circular
protocols. Springer, New York, pp 9–25. https://doi. RNAs by RNA fluorescence in situ hybridization. In:
org/10.1007/978-1-4939-7562-4_2 Dieterich C, Papantonis A (eds) Circular RNAs: meth-
50. Bachmayr-Heyda A, Reiner AT, Auer K et al (2015) ods and protocols. Springer, New York, pp 69–75.
Correlation of circular RNA abundance with prolif- https://doi.org/10.1007/978-1-4939-7562-4_6
eration – exemplified with colorectal and ovarian 65. Huang M-S, Zhu T, Li L et al (2018) LncRNAs and
cancer, idiopathic lung fibrosis, and normal human CircRNAs from the same gene: masterpieces of RNA
tissues. Sci Rep 5:8057 splicing. Cancer Lett 415:49–57
370 S. A. Sulaiman et al.

66. Song X, Zhang N, Han P et al (2016) Circular RNA 82. Hansen TB (2018) Improved circRNA identification
profile in gliomas revealed by identification tool by combining prediction algorithms. Front Cell Dev
UROBORUS. Nucleic Acids Res 44(9):e87–e87 Biol 6:20
67. Izuogu OG, Alhasan AA, Alafghani HM et al (2016) 83. Baruzzo G, Hayer KE, Kim EJ et al (2017) Simulation-­
PTESFinder: a computational method to identify based comprehensive benchmarking of RNA-seq
post-transcriptional exon shuffling (PTES) events. aligners. Nat Methods 14(2):135–139
BMC Bioinformatics 17(1):31 84. Kim D, Pertea G, Trapnell C et al (2013) TopHat2:
68. Hoffmann S, Otto C, Doose G et al (2014) A multi-­ accurate alignment of transcriptomes in the presence
split mapping algorithm for circular RNA, splicing, of insertions, deletions and gene fusions. Genome
trans-splicing and fusion detection. Genome Biol Biol 14(4):R36
15(2):R34 85. Trapnell C, Pachter L, Salzberg SL (2009) TopHat:
69. Wang K, Singh D, Zeng Z et al (2010) MapSplice: discovering splice junctions with RNA-Seq.
accurate mapping of RNA-seq reads for splice junc- Bioinformatics 25(9):1105–1111
tion discovery. Nucleic Acids Res 38(18):e178–e178 86. Glažar P, Papavasileiou P, Rajewsky N (2014)
70. Chuang T-J, Wu C-S, Chen C-Y et al (2016) NCLscan: circBase: a database for circular RNAs. RNA
accurate identification of non-co-linear transcripts 20(11):1666–1670
(fusion, trans-splicing and circular RNA) with a good 87. Ghosal S, Das S, Sen R et al (2013) Circ2Traits:
balance between sensitivity and precision. Nucleic a comprehensive database for circular RNA poten-
Acids Res 44(3):e29–e29 tially associated with disease and traits. Front Genet
71. Gao Y, Wang J, Zhao F (2015) CIRI: an efficient and 4:283
unbiased algorithm for de novo circular RNA identifi- 88. Liu Y-C, Li J-R, Sun C-H et al (2016) CircNet: a data-
cation. Genome Biol 16(1):4 base of circular RNAs derived from transcriptome
72. Szabo L, Morey R, Palpant NJ et al (2015) Statistically sequencing data. Nucleic Acids Res 44(Database
based splicing detection reveals neural enrichment issue):D209–D215
and tissue-specific induction of circular RNA during 89. Gao Y, Wang J, Zheng Y et al (2016) Comprehensive
human fetal development. Genome Biol 16(1):126 identification of internal structure and alternative
73. Gao Y, Zhang J, Zhao F (2017) Circular RNA iden- splicing events in circular RNAs. Nat Commun
tification based on multiple seed matching. Brief 7:12060
Bioinform: bbx014–bbx014 90. Metge F, Czaja-Hasse LF, Reinhardt R et al (2017)
74. Westholm JO, Miura P, Olson S et al (2014) FUCHS—towards full circular RNA characterization
Genomewide analysis of Drosophila circular RNAs using RNAseq. PeerJ 5:e2934
reveals their structural and sequence properties 91. Meng X, Chen Q, Zhang P et al (2017) CircPro: an
and age-dependent neural accumulation. Cell Rep integrated tool for the identification of circRNAs
9(5):1966–1980 with protein-coding potential. Bioinformatics
75. Cheng J, Metge F, Dieterich C (2016) Specific iden- 33(20):3314–3316
tification and quantification of circular RNAs from 92. Feng J, Xiang Y, Xia S, et al (2017) CircView: a visu-
sequencing data. Bioinformatics 32(7):1094–1096 alization and exploration tool for circular RNAs. Brief
76. Gao Y, Zhao F (2018) Computational strategies for Bioinform:bbx070-bbx070
exploring circular RNAs. Trends Genet 34:389 93. Dudekula DB, Panda AC, Grammatikakis I et al
77. Zeng X, Lin W, Guo M et al (2017) A comprehensive (2016) CircInteractome: a web tool for exploring
overview and evaluation of circular RNA detection circular RNAs and their interacting proteins and
tools. PLoS Comput Biol 13(6):e1005420 microRNAs. RNA Biol 13(1):34–42
78. Langmead B, Salzberg SL (2012) Fast gapped-read 94. Grant GR, Farkas MH, Pizarro AD et al (2011)
alignment with bowtie 2. Nat Methods 9(4):357–359 Comparative analysis of RNA-Seq alignment algo-
79. Langmead B, Trapnell C, Pop M et al (2009) rithms and the RNA-Seq unified mapper (RUM).
Ultrafast and memory-efficient alignment of short Bioinformatics 27(18):2518–2528
DNA sequences to the human genome. Genome Biol 95. Streit S, Michalski CW, Erkan M et al (2008) Northern
10(3):R25–R25 blot analysis for detection and quantification of RNA
80. Li H, Durbin R (2009) Fast and accurate short in pancreatic cancer cells and tissues. Nat Protoc 4:37
read alignment with burrows–wheeler transform. 96. PCvdC M, Roeland WD, PMvG R et al (1998)
Bioinformatics 25(14):1754–1760 Sensitive mRNA detection by fluorescence in situ
81. Dobin A, Davis CA, Schlesinger F et al (2013) STAR: hybridization using horseradish peroxidase-labeled
ultrafast universal RNA-seq aligner. Bioinformatics oligodeoxynucleotides and tyramide signal amplifica-
29(1):15–21 tion. J Histochem Cytochem 46(11):1249–1259