Research Article

Experimental and Translational Hepatology

Reprogramming of rhythmic liver metabolism by intestinal clock

Authors
Min Chen, Yanke Lin, Yongkang Dang, Yifei Xiao, Fugui Zhang, ., Jing Yang, Kaisheng Liu, Baojian Wu

Correspondence
[email protected] (B. Wu), [email protected] (K. Liu), [email protected] (J. Yang), [email protected] (Z. Qin).

Graphical abstract

Intestine clock Clock malfunction

Normal phenotype Hyperlipidemia Glucose intolerance

DNL GNG GNG DNL

GNG DNL DNL GNG

SREBP-1c SREBP-1c SREBP-1c SREBP-1c

Highlights Impact and implications

 An intestinal clock reprograms the rhythmic transcriptome Our findings establish the centrality of the intestinal clock
of the liver. among peripheral tissue clocks, and associate liver-related
pathologies with its malfunction. Clock modifiers in the intes-
 Deficiencies in the intestinal clock affect food entrainment of
tine are shown to modulate liver metabolism with improved
the liver clock.
metabolic parameters. Such knowledge will help clinicians
 Intestinal clock deficiency causes improper temporal di- improve the diagnosis and treatment of metabolic diseases by
versions between gluconeogenesis and lipogenesis. incorporating intestinal circadian factors.
 Transcriptional reprogramming involves disruption of he-
patic SREBP-1c rhythmicity.

https://doi.org/10.1016/j.jhep.2023.04.040
© 2023 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. J. Hepatol. 2023, 79, 741–757
Research Article
Experimental and Translational Hepatology

Reprogramming of rhythmic liver metabolism by

intestinal clock
Min Chen1,†, Yanke Lin1,†, Yongkang Dang1, Yifei Xiao1, Fugui Zhang1, Guanghui Sun1, Xuejun Jiang1, Li Zhang1, Jianhao Du1, Shuyi Duan2,
Xiaojian Zhang2, Zifei Qin2,*, Jing Yang2,*, Kaisheng Liu3,*, Baojian Wu1,*

Journal of Hepatology 2023. vol. 79 j 741–757

See Editorial, pages 589–591

Background & Aims: Temporal oscillations in intestinal nutrient processing and absorption are coordinated by the local clock,
which leads to the hypothesis that the intestinal clock has major impacts on shaping peripheral rhythms via diurnal nutritional
signals. Here, we investigate the role of the intestinal clock in controlling liver rhythmicity and metabolism.
Methods: Transcriptomic analysis, metabolomics, metabolic assays, histology, quantitative (q)PCR, and immunoblotting were
performed with Bmal1-intestine-specific knockout (iKO), Rev-erba-iKO, and control mice.
Results: Bmal1 iKO caused large-scale reprogramming of the rhythmic transcriptome of mouse liver with a limited effect on its
clock. In the absence of intestinal Bmal1, the liver clock was resistant to entrainment by inverted feeding and a high-fat diet.
Importantly, Bmal1 iKO remodelled diurnal hepatic metabolism by shifting to gluconeogenesis from lipogenesis during the dark
phase, leading to elevated glucose production (hyperglycaemia) and insulin insensitivity. Conversely, Rev-erba iKO caused a
diversion to lipogenesis from gluconeogenesis during the light phase, resulting in enhanced lipogenesis and an increased sus-
ceptibility to alcohol-related liver injury. These temporal diversions were attributed to disruption of hepatic SREBP-1c rhythmicity,
which was maintained via gut-derived polyunsaturated fatty acids produced by intestinal FADS1/2 under the control of a
local clock.
Conclusions: Our findings establish a pivotal role for the intestinal clock in dictating liver rhythmicity and diurnal metabolism, and
suggest targeting intestinal rhythms as a new avenue for improving metabolic health.
© 2023 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.

Introduction and peripheral clocks are known to respond differentially to

timing cues (zeitgebers). Light is a primary zeitgeber for the
Circadian clocks are an evolutionarily conserved and intrinsic
central clock and entrains it through direct retinal innervation,
timing system, which results in 24-h rhythms in behaviours
whereas feeding is a potent zeitgeber synchronizing peripheral
and physiology, allowing organisms to anticipate diurnal
clocks.4 Hierarchy also exists among the peripheral clocks with
changes in the environment (the dark–light cycle being a
respect to food entrainment: the liver clock occupies a higher-
prominent example) imposed by the rotation of the earth.1 In
level rank and controls the entrainment of the circadian tran-
mammals, the circadian clock is present in almost every
scriptome in tissues with lower-level ranks, such as the lung.5
tissue and every cell, and functions in a self-sustained (or
Metabolic homeostasis is intimately linked to the clock
cell-autonomous) manner and relies on the interlocked
system.6 The circadian control of metabolism is pervasive
transcriptional–translational feedback loops of various core
because 15% of all identified metabolites in human plasma
clock genes (e.g. Bmal1, Clock, Pers, Crys, Rev-erba/b, Rors,
and saliva oscillate in a circadian manner and most metabolic
Dbp, and E4bp4).2 In the main loop, the BMAL1/CLOCK het-
genes (e.g. those involved in metabolism of carbohydrates and
erodimer activates the transcription of Pers and Crys, the
lipids as well as biosynthesis of heme and ATP) are cyclically
protein products of which in turn oppose the action of BMAL1/
expressed.7 Circadian patterns of metabolic genes can be
CLOCK, a so-called ‘negative feedback mechanism’.2 Clocks
attributed to oscillations in their regulators, such as PPARa/c,
in body tissues are deemed to be hierarchal: the suprachias-
SREBP, and CHREBP, which are under direct or indirect con-
matic nucleus (SCN) of the brain harbours a central clock,
trol of the molecular components of the clock system.8–10
which coordinates the clocks in peripheral tissues.3 The central
Temporal organisation of metabolic pathways is essential

Keywords: Circadian rhythm; Intestine clock; SREBP; Metabolic diseases.

Received 20 September 2022; received in revised form 10 April 2023; accepted 27 April 2023; available online 23 May 2023
* Corresponding author. Address: Institute of Molecular Rhythm and Metabolism, Guangzhou University of Chinese Medicine, Guangzhou, China (B. Wu);
Guangdong Provincial Clinical Research Center for Geriatrics, Shenzhen Clinical Research Center for Geriatrics, Shenzhen People’s Hospital (The Second
Clinical Medical College, Jinan University; The First Affiliated Hospital, Southern University of Science and Technology), Shenzhen, Guangdong, China
(K. Liu); Department of Pharmacy, the First Affiliated Hospital of Zhengzhou University, Zhengzhou, China (J. Yang, Z. Qin).
E-mail addresses: [email protected] (B. Wu), [email protected] (K. Liu), [email protected] (J. Yang), [email protected] (Z. Qin).
†
These authors contributed equally to this work.
https://doi.org/10.1016/j.jhep.2023.04.040

Journal of Hepatology, September 2023. vol. 79 j 741–757

Intestinal clock reprograms rhythmic liver metabolism

for health. Chronic perturbance of their circadian rhythms can Results

lead to various metabolic diseases, such as obesity, insulin
resistance, hyperglycaemia, and dyslipidaemia.11,12 Notably, Intestinal Bmal1 remodels the rhythmic transcriptome
circadian metabolism can be remodelled by nutritional chal- of liver
lenges (e.g. fasting, high-fat diet [HFD], and ketogenic diet), To assess the potential impact of the intestinal clock on liver
involving both transcriptional and epigenetic mechanisms.13–15 rhythmicity, mice with Bmal1 specifically deleted in the intestine
Time-restricted feeding (to the activity phase) shows consistent (Bmal1-iKO mice, showing loss of clock function in the intes-
metabolic benefits in humans and animals, including reduced tine) were generated as described in previous report.26 We
body weight and fat mass as well as improved insulin sensitivity collected liver samples from Bmal1-iKO and control (Bmal1-
and blood pressure.16–18 This promotes extensive research flox) mice every 4 h throughout a light–dark cycle, and subse-
into, and development of, chrono-nutritional approaches to quently performed transcriptomic analyses. Rhythmic tran-
manage metabolic diseases. scripts and related parameters (i.e. phase and amplitude) were
The intestine is anatomically connected to the liver via the determined using the JTK_CYCLE algorithm. Numerous genes
portal vein, and has a profound effect on hepatic functions by (n = 1,420, 61%) lost their rhythmicity in mouse liver following
directly delivering gut-derived products (e.g. nutrients, hor- deletion of intestinal Bmal1 (Fig. 1A,B). Only a small portion of
mones, and metabolites) to the liver.19,20 Extensive bowel genes (n = 169, 7%) retained their rhythmicity (Fig. 1A,B). The
resection, gut barrier damage, and functional exclusion of the phases and amplitudes of these common rhythmic genes were
small intestine can all lead to severe liver dysfunction and slightly affected (Fig. 1C). Intriguingly, a considerable number of
even cirrhosis.21,22 Intestinal disorders, such as inflammatory genes (n = 766, 32%) gained rhythmicity in the liver of Bmal1-
bowel diseases, are associated with various liver diseases, iKO mice (Fig. 1A,B). The genes with altered rhythmicity
including hepatic steatosis and primary sclerosing chol- converged on metabolic pathways (n = 135) related to FAs,
angitis.23,24 Although its major function is to transport and carbohydrates, amino acids, and nucleotides (Fig. 1D), sug-
absorb nutrients, the intestine also has a role in the meta- gesting that hepatic metabolism is reprogrammed by the in-
bolism and biosynthesis of nutrients (and related compounds), testinal clock. These disrupted metabolic genes included Fasn
such as fatty acids (FAs) and triglycerides (TGs).25 Given that and Acaca (involved in FA metabolism), as well as Gck and Pklr
nutrient processing in the body occurs in the intestine initially, (involved in glucose metabolism) (Fig. 1E). Dhtkd1 and Aldh1b1
and some aspects of intestinal nutrient processing (e.g. were representative genes involved in amino acid metabolism
macronutrient transport, TG resynthesis, and lipoprotein pro- that were disrupted by Bmal1 iKO, whereas Uck1 and Pnp2
duction) have been shown to follow a circadian pattern under were archetypical examples of perturbed genes related to
the control of local clock, it is conceivable that gut-derived nucleotide metabolism (Fig. 1E, Fig. S1A, and Table S1).
circadian nutritional signals can be relayed to the liver to Dhtkd1 is involved in lysine and tryptophan catabolism, and its
remodel its clock and shape its rhythmicity. Thus, several mutations can result in alpha-aminoadipic and alpha-
major questions are raised regarding potential crosstalk be- ketoadipic aciduria (AMOXAD), an autosomal recessive inborn
tween the intestine and liver clocks: does the intestinal metabolic disorder.27 Uck1 encodes UCK1, which is a pyrimi-
clock respond to feeding cues similar to the liver clock? dine ribonucleoside kinase that catalyses the phosphorylation
What are the effects of the intestinal clock on the liver of uridine and cytidine.28 Bmal1 iKO did not affect food intake,
clock and on peripheral clock rhythmicity? How do intestinal feeding rhythm, diurnal locomotor activity, or energy expendi-
clock-controlled metabolic cues guide the hepatic circa- ture,26 excluding the possibility that remodelling of the liver
dian system? rhythmic transcriptome resulted from changes in behaviours.
Here, we aimed to clarify the effects of the intestinal clock We specifically analysed the expression of hepatic core clock
on liver rhythmicity, and to test the metabolic significance genes (Bmal1, Rev-erba/Nr1d1, Clock, Per1, Per2, Per3, Cry1,
thereof. We found that specific loss of intestinal Bmal1 (Bmal1 Dbp, E4bp4, Rora, and Rorg) in Bmal1-iKO mice vs. control
iKO) in mice induced large-scale reprogramming of liver mice. These clock genes were marginally affected except for
rhythmic transcriptome but with a limited effect on its clock. In Per2 and Per3, which showed mild shifts and dampened
the absence of intestinal Bmal1, the liver clock was resistant rhythms (Fig. 1F,G, Fig. S1B–D, and Table S1).
to the entrainment by inverted feeding and HFD. Importantly, We further extended our round-the-clock transcriptomic an-
Bmal1 iKO disrupted diurnal metabolic homeostasis involving alyses to white adipose tissue (WAT) and kidney, and observed
lipids and glucose in the liver, relying on modulation of hepatic loss and gain of rhythmicity for numerous genes in Bmal1-iKO
SREBP-1c expression by gut-delivered polyunsaturated fatty mice (Fig. S2). In the WAT, rhythmicity-altered genes were
acids (PUFAs). Similar to intestinal Bmal1, intestinal Rev-erba enriched for amino acid metabolism, propanoate metabolism,
controlled rhythmic liver metabolism, but in an opposite di- and metabolism of xenobiotics (Fig. S1A–C). In the kidney, the
rection, consistent with the negative feedback association in affected genes converged on spliceosome, circadian rhythm,
maintaining circadian homeostasis. Our findings reveal a and the ErbB signalling pathway (Figs S2D–F and S3). This
pivotal role of the intestinal clock in shaping liver rhythmicity signified a remodelling effect of the intestinal clock on the
and diurnal metabolism, and suggest targeting intestinal rhythmicity of other peripheral tissues in addition to the liver.
rhythms as a new avenue for improving metabolic health. However, the effects might be secondary to the changes in liver
rhythmicity because the liver is shown to control the rhythmicity
Materials and methods of other distant peripheral tissues, including WAT and kidney, by
modulating daily fluctuations in blood-borne metabolites, such
All materials and methods are detailed in supplementary as glucose,29 although other mechanisms cannot be excluded.
materials and methods.

742 Journal of Hepatology, September 2023. vol. 79 j 741–757

 Research Article

A Liver B Liver: rhythmic transcripts C Bmal1-flox vs. Bmal1-iKO

Bmal1-flox Bmal1-iKO 50 80
Bmal1-flox
40

No. of genes

No. of genes
60
30
1,420 169 766 40
20
10 20
Bmal1-iKO
RN

0 0
-10 -6 -2 2 6 10 -3 -2 -1 0 1 2 3
Δ Phase (h) Δ Log2 amplitude

D
Rhythmicity-altered transcripts Metabolic pathways
Glucagon signaling pathway (21) Fatty acid metabolism (10)

Metabolic pathways (135) Carbon metabolism (12)

Lysine degradation (13) Pyrimidine metabolism (9)

Small cell lung cancer (16) Pyruvate metabolism (8)

NR

2
p53 signaling pathway (13) Tryptophan metabolism (8)
1
0 1 2 3 4 5 0 5 10 15
0
-Log10 (p value) -Log10 (p value)
-1
RR

-2
ZT 2 6 10 14 18 22 2 6 10 14 18 22

E Fasn Acaca Gck Pklr F Per2

400 20 300 200 30
* * Bmal1-flox

300 * 15 * * 150
Bmal1-iKO
200 * 20
FPKM

200 * 10 * * 100
* * *
100 10
100 5 50
0 0 0 0 0
Dhtkd1 Aldh1b1 Uck1 Pnp2 Per3
40 80 50 3 15
* *
30 60 * 40 *
* 2 10
FPKM

* 30
20 * 40 * 20 1 5
10 20 *
10
0 0 0 0 0
ZT 0 12 24 0 12 24 0 12 24 0 12 24 0 12 24
G Bmal1 Rev-erbα Clock Per1 Cry1
5 50 5 20 30
4 40 4 15
20
FPKM

3 30 3
10
2 20 2 10
1 10 1 5
0 0 0 0 0
ZT 0 12 24 0 12 24 0 12 24 0 12 24 0 12 24

Fig. 1. Intestinal Bmal1 remodels the rhythmic transcriptome of liver. (A) Heatmaps showing the genes that are non-rhythmic in Bmal1-flox and rhythmic in Bmal1-
iKO (NR), the genes that are rhythmic in Bmal1-flox and non-rhythmic in Bmal1-iKO (RN) as well as the genes that are rhythmic in both (RR). Genes are considered
rhythmic over the circadian cycle if their permutation-based, p values (ADJ.P) was <0.01 (JTK_CYCLE) (n = 3). (B) Venn diagrams showing the overlap between
rhythmic genes in Bmal1-flox and in Bmal1-iKO mice. (C) Histograms representing the distribution of phase (left) and amplitude (right) differences between Bmal1-flox
and Bmal1-iKO mice of the common rhythmic genes. The dashed vertical lines signify the median phase shift (median: 0.27) and median distribution of amplitude
difference (median: -0.18). (D) KEGG pathway analysis of genes with altered rhythmicity. (E) Expression of genes related to fatty acid (Fasn and Acaca), carbohydrate
(Gck and Pklr), amino acid (Dhtkd1 and Aldh1b1) and nucleotides (Uck1 and Pnp2) metabolism in the liver of Bmal1-flox and Bmal1-iKO mice (n = 3). (F) Expression of
the core clock genes Per2 and Per3 in the liver of Bmal1-flox and Bmal1-iKO mice (n = 3). (G) Expression of the core clock genes Bmal1, Clock, Rev-erba, Per1, and
Cry1 in the liver of Bmal1-flox and Bmal1-iKO mice (n = 3). Data are presented as mean ± SD. *p <0.05 vs. control (two-way ANOVA with Bonferroni post hoc test). iKO,
intestinal knockout; KEGG, Kyoto Encyclopedia of Genes and Genomes; ZT, zeitgeber time. (This figure appears in color on the web.)

Intestinal Bmal1 affects the entrainment of liver clock by feeding; DF) or ad libitum (AL). After 30 days (sufficient to reach
feeding time and high-fat diet a stable entraining effect), samples were collected for qPCR
and/or transcriptomic analyses. DF induced an 12-h phase
Given that the rhythmic transcriptome of liver is regulated by
the intestinal clock, we investigated whether the latter affects shift in intestinal clock genes (e.g. Bmal1, Clock, Rev-erba,
entrainment of the liver clock by feeding. To test this, mice were Dbp, Per1, Per2, Cry1, and Cry2) in Bmal1-flox mice (Fig. 2A,
maintained under a standard light–dark cycle and fed exclu- Fig. S4, and Table S1), which was excepted given that food is
sively during the light phase (daytime feeding or inverted able to entrain the intestinal clock.30 In line with their mRNA
changes, BMAL1, REV-ERBa, and DBP in the intestine were

Journal of Hepatology, September 2023. vol. 79 j 741–757 743

Intestinal clock reprograms rhythmic liver metabolism

12-h phase-shifted (Fig. 2B). Likewise, phase inversion was (Fig. S5A–C). Thus, in Bmal1-flox mice, the liver rhythmic
observed for the same set of clock genes in the liver of Bmal1- transcriptome and its clock remained aligned with the feeding
flox mice in response to DF (Fig. 2C–E), suggesting a fine time.29 By contrast, the hepatic clock genes were resistant to a
entrainment of the liver clock by feeding. In correspondence phase shift in response to DF in Bmal1-iKO mice (Fig. 2E–G). In
with the phase inversion in the liver clock on DF, we observed addition, the rhythmic transcriptome of liver was marginally
on average an almost 12-h shift in the rhythmic transcriptome phase-shifted in the KO mice on DF (Fig. S5D–F). DF had no

A Intestine B Intestine
AL DF
AL DF
Bmal1 Clock Rev-erbα Dbp
2.0 2.0 10 3 ZT 2 6 10 14 18 22 2 6 10 14 18 22
* * * * * BMAL1
1.5 * 1.5 8
* * * 2 REV-ERBα
* * 6
1.0 1.0 * DBP
4 * * 1 *
0.5 0.5 2 * GAPDH

D
Rel mRNA

0.0 0.0 0 0
Per1 Per2 Cry1 Cry2
5 2.0 2.0 3 Liver AL DF
* * ZT 2 6 101418 22 2 6 1014 18 22
4 * * 1.5 * * * 1.5 *
3 * * * 2 BMAL1

Bmal1-flox
1.0 1.0 * *
2 REV-ERBα
1
1 0.5 0.5 DBP
0 0.0 0.0 0 GAPDH
ZT 0 12 24 0 12 24 0 12 24 0 12 24

C Bmal1-flox, liver E
AL DF
Bmal1 Rev-erbα Dbp
8 10 4 DF vs. AL
* * Bmal1-flox
6 8 3 * * Bmal1 Bmal1-iKO
* 6
4 * * * 2 * Rev-erbα
4
2 2 1
Rel mRNA

Dbp
0 0 0
Per1 Per2 Cry1 Cry1
6 * 6 4
* *
*
4 4 * 3 * * Per1
* *
* * * 2
Per2
2 2
1
-12 -8 -4 0 4 8 12
0 0 0 Δ phase (h)
ZT 0 12 24 0 12 24 0 12 24

F Bmal1-iKO, liver G
AL DF
Bmal1 Rev-erbα Dbp
10 10 10
8 * 8 * 8 *
6 * 6 6
4 * * 4 * 4 * * Liver AL DF
Rel mRNA

2 2 2 ZT 2 6 10 141822 2 6 10 14 18 22
0 0 0 BMAL1
Bmal1-iKO

Per1 Per2 Cry1

REV-ERBα
2.5 4 5
* * DBP
2.0 * 3 4
1.5 * * 3 GAPDH
2 * * *
1.0 * 2 *
1 *
0.5 1
0.0 0 0
ZT 0 12 24 0 12 24 0 12 24

Fig. 2. Intestinal Bmal1 affects the entrainment of liver clock by feeding. (A) Relative expression of core clock genes in IECs from Bmal1-flox mice fed either AL or
DF for 30 days (n = 8). (B) Relative expression of BMAL1, REV-ERBa, and DBP in IECs from Bmal1-flox mice fed either AL or DF for 30 days. (C) Relative expression of
core clock genes in the liver from Bmal1-flox mice fed either AL or DF for 30 days (n = 8). (D) Relative expression of BMAL1, REV-ERBa, and DBP in the liver from
Bmal1-flox mice fed either AL or DF for 30 days. (E) Phase differences in clock genes between DF and AL conditions in Bmal1-flox and Bmal1-iKO mice. (F) Relative
expression of core clock genes in the liver from Bmal1-iKO mice fed either AL or DF for 30 days (n = 8). (G) Relative expression of BMAL1, REV-ERBa and DBP proteins
in the liver from Bmal1-iKO mice fed either AL or DF for 30 days. Data are presented as mean ± SD. *p <0.05 vs. control (two-way ANOVA with Bonferroni post hoc test).
AL, ad libitum; DF, daytime feeding; IEC, intestinal epithelial cell; iKO, intestinal knockout; ZT, zeitgeber time.

744 Journal of Hepatology, September 2023. vol. 79 j 741–757

 Research Article

effects on intestinal clock genes (with constitutively low levels) We further observed no changes in b-hydroxybutyrate levels
in Bmal1-iKO mice (Fig. S6). Likewise, DF failed to shift the liver (Fig. 3G) or in expression of lipolytic genes (Atgl and Hsl) in
clock in Bmal1-iKO mice based on a different experimental WAT and skeletal muscle (Fig. S9D). In addition, the lipopro-
protocol (DF restricted to zeitgeber time (ZT) 2–6 for 1 week) tein lipase inhibitor tyloxapol did not affect the reduction of
(Fig. S7). Thus, DF uncoupled the intestinal clock from the liver plasma TG at ZT18 in Bmal1-iKO mice (Fig. S9E). Thus, the
clock. We concluded that entrainment of the liver clock by Bmal1-iKO-induced hypolipidaemic state at ZT18 resulted
feeding requires a normal function of the circadian clock in the from reduced hepatic lipogenesis rather than from diminished
intestine, supporting a crucial role of the intestinal clock in lipolysis or enhanced FA oxidation.
shaping the liver transcriptome. REV-ERBa (also known as NR1D1) is another component of
HFD is known to remodel the circadian transcriptome in the molecular clock that represses Bmal1 transcription in a circa-
liver.14 We asked whether the intestinal clock has a role in HFD- dian manner and helps maintain the oscillating robustness of
induced transcriptional remodelling. Consistent with a previous the clock syste.1,2 Loss of Rev-erba leads to upregulation of
report,14 3 days of HFD led to a considerably reduced ampli- Bmal1 in many tissues, including the intestine, with substantial
tude of Nampt and Upp2 oscillations, and to enhanced rhythms alterations during the light phase.26 We next investigated
of Pparg and its target gene Cidec in the liver of wild-type mice whether and how intestinal Rev-erba affects hepatic lipid
(Fig. S8). However, Bmal1-iKO mice fed HFD for 3 days had metabolism. To this end, mice with Rev-erba specifically
similar profiles of Nampt, Upp2, Pparg, and Cidec compared depleted in the intestine (named Rev-erba-iKO mice) were
with chow-fed Bmal1-flox controls (Fig. S8). Thus, the remod- generated as described in recent work,26 and the core clock
elling effects of HFD on hepatic gene transcription were largely and lipid metabolism-related genes in their livers were ana-
abolished by Bmal1 iKO. Again, the results underscore a role of lysed. Similar to Bmal1 iKO, Rev-erba iKO did not alter food
the intestinal clock in diet-induced reprogramming of the liver intake, feeding rhythm, or diurnal locomotor activity.26 Hepatic
rhythmic transcriptome. core clock genes were not significantly affected in Rev-erba-
iKO mice (Fig. S10). Hepatic lipogenic genes, such as Fasn,
Scd1, Acaca, and Elovl6, showed increases in their expression
Hepatic lipid metabolism in intestinal clock-deficient mice at ZT0–6 in Rev-erba-iKO mice, whereas other lipid-related
Our transcriptomic data suggested marked disruptions of genes remained largely unaffected (Fig. 3H and Fig. S9F,G).
rhythmic genes involved in hepatic lipid metabolism in Bmal1- In keeping with the enhanced expression of lipogenic genes,
iKO mice (Fig. 1). qPCR analyses confirmed that deficiency in these mice were hyperlipidaemic with higher levels of TGs and
intestinal Bmal1 caused substantial changes in several he- FFAs in both plasma and liver at ZT6 (Fig. 3I). Thus, Rev-erba-
patic genes involved in de novo lipogenesis (Fasn, Scd1, iKO mice had an opposite phenotype compared with Bmal1-
Acaca, and Elovl6), but no changes in genes related to lipid b- iKO mice with respect to the lipid level and timing of lipid
oxidation (Cpt1a, Cpt1c, Crat, Acsl1, Ehhadh, Acox1, change. Collectively, these results reveal bidirectional reprog-
Aldh3a2, Hadh, and Acaa1), lipid transport (Cd36, Vldlr, Apob, ramming of diurnal lipid metabolism in the liver by the intesti-
Apoe, and Apoc3), and cholesterol metabolism (Hmgcr, nal clock.
Hmgcs2, Acat2, Cyp7a1, Ldlr, Abcg5, and Abcg8) (Fig. 3A,B
and Fig. S9A). Hepatic Gpam was decreased in mice without
intestinal Bmal1, whereas other TG synthesis (Mogat1, Intestinal clock regulates the susceptibility of mouse liver
Mogat2, Dgat1, and Dgat2) and lipid sequestration (Plin2, to alcohol-induced injury
Cidec, G0s2, and Fitm1) genes were barely affected Given that hepatic lipogenesis is reprogrammed by the in-
(Fig. S9B). In general, lipogenic genes were downregulated testinal clock, we investigated whether the reprogramming
and their rhythms were dampened (Fig. 3A). Parallel changes has a pathophysiological consequence. To test this, Bmal1-
were also observed for their proteins (Fig. 3C). Meanwhile, iKO mice and controls were treated with acute gavage of
diurnal rhythms of G6pd2 and Me1, the genes involved in ethanol (three doses at 12-h intervals) to induce liver steatosis
the pentose phosphate pathway and pyruvate–malate cy- and injury (a model of acute alcoholic liver injury). The intes-
cle, respectively for generating NADPH, were blunted with tinal barrier (tight junctions) was unaffected by loss of intes-
reduced expression during the dark phase (Fig. S9C). These tinal Bmal1, as evidenced by no changes in ZO1, Occludin,
results prompted us to examine whether loss of intestinal and plasma lipopolysaccharide (LPS),32 and this held true
Bmal1 affected lipid homeostasis in the body. Given that body after binge ethanol exposure (Figs S11A and S12A). Thus, this
lipids display diurnal oscillations,31 we determined the lipid model allowed appropriate assessment of the effects of
levels in the plasma and liver at ZT0 (night–day transition), ZT6 metabolic reprogramming on alcohol-mediated liver injury
(midday), ZT12 (day–night transition), and ZT18 (midnight). without interference from a change in gut permeability (a po-
The knockouts had lower plasma levels of TGs and free FAs tential factor contributing to increased circulating endotoxin
(FFAs) at ZT18, but not at ZT0, ZT6, and ZT12 (Fig. 3D). Lower and liver injury).33 After ethanol gavage, Bmal1-iKO mice
levels of both TGs and FFAs were also noted in the liver at showed lower plasma levels of both alanine aminotransferase
ZT18 (Fig. 3e). By contrast, circulating and hepatic total (ALT) and aspartate aminotransferase (AST) compared with
cholesterol were not significantly affected (Fig. 3D,E). These controls (Fig. 4A). They also had lower TG and FFA levels in
findings agreed well with Bmal1-iKO-induced diurnal changes the plasma and liver (Fig. 4B,C), suggesting alleviated stea-
in lipogenic genes with striking downregulation during the tosis in these animals. This was confirmed by histology and Oil
dark phase (Fig. 3A). Concomitantly, the rate of hepatic de Red O staining (Fig. 4D,E and Fig. S13). Moreover, production
novo lipogenesis at ZT18 measured by tracing newly made of proinflammatory cytokines (such as Tnfa, Il6, Il1b, and
oleate was lower in Bmal1-iKO mice than in controls (Fig. 3F). Mcp1) was reduced in the ethanol-challenged knockouts

Journal of Hepatology, September 2023. vol. 79 j 741–757 745

Intestinal clock reprograms rhythmic liver metabolism

compared with controls (Fig. 4F). We examined the expression chronic ethanol feeding (10 days) plus single-binge ethanol
of major lipogenic genes, and observed consistent reductions (another method inducing alcoholic liver injury). Similarly, in-
in these genes in Bmal1-iKO mice after binge ethanol expo- testinal permeability did not differ between Bmal1-iKO and
sure (Fig. S11B). Altogether, loss of intestinal Bmal1 protected Bmal1-flox mice (Figs S11C and S12B). Bmal1-iKO mice
mice from acute alcoholic liver injury, most likely because of showed lower plasma ALT and AST, as well as reduced
an effect of reprogrammed liver metabolism (reduced lipid circulating and hepatic TGs (Fig. S11D,E). Concurrently, they
biosynthesis). We additionally challenged the animals by had lower levels of genes encoding proinflammatory

A Fasn Scd1 Acaca

Bmal1-flox
Elovl6
Bmal1-iKO B 5 Bmal1-flox ZT18
Bmal1-iKO
8 5 4 6 4
*

Rel mRNA
*
4 * * * 3
Rel mRNA

6 * 3
* 3 * *
4 2 * 2
2
2 * 1 1
1
0 0 0 0 0

Hmgcs2
Aldh3a2
Ehhadh

Cyp7a1
Hmgcr
Acaa1

Apoc3
Acox1

Abcg5
Abcg8
Cpt1a
Cpt1c

Acat2
Acsl1

Hadh

Cd36

Apob
Apoe
Vldlr
Crat

Ldlr
ZT 0 12 24 0 12 24 0 12 24 0 12 24

C Bmal1-flox Bmal1-iKO
ZT 2 6 10 14 18 22 2 6 10 14 18 22 Lipid Lipid Cholesterol
FASN β-oxidation transport metabolism
SCD1
ACC1
ELOVL6
GAPDH

D
Plasma FFA (mg/dl)

Plasma cholesterol
Plasma TG (mg/dl)

150 100 100 Bmal1-flox

80 80 Bmal1-iKO
100

(mg/dl)
60 * 60
*
50 40 40
20 20
0 0 0
0 6 12 18 0 6 12 18 0 6 12 18
ZT ZT ZT

E F G
FFA (mg/g protein)

15 2.0 1.5 0.4 5 BHB

TG (mg/g protein)

Bmal1-flox
oleate (μg/mg liver)
(mg/g protein)

Newly synthesized

mmol/L or μmol/g
Bmal1-iKO
Cholesterol

1.5 0.3 4
10 1.0
1.0 3
5 * * 0.5 0.2
0.5 2
*
0 0.0 0.0 0.1 1
0 6 12 18 0 6 12 18 0 6 12 18
ZT ZT ZT 0.0 0
Ctrl B-iKO Plasma Liver

H Rev-erbα-flox Rev-erbα-iKO I
FFA (mg/gprotein) Plasma FFA (mg/dl)

Plasma cholesterol
Plasma TG (mg/dl)

5 Fasn 5 Scd1 250 Rev-erbα-flox Rev-erbα-iKO 60 200

4 4
* 200
*
Rel mRNA

* * * 40
* 150
(mg/dl)

3 * 3 * 150
100
2 2 100 20
1 1 50 50
0 0 0 0 0
0 6 12 18 0 6 12 18 0 6 12 18 0 6 12 18 0 6 12 18
4 Elovl6 4 Acaca 15 2.5 5
TG (mg/g protein)

(mg/g protein)

* 2.0 * 4
Cholesterol

*
Rel mRNA

3 * 3
* * 10 1.5 3
2 2
5 1.0 * 2
1 1 0.5 1
0 0 0 0.0 0
0 6 12 18 0 6 12 18 0 6 12 18 0 6 12 18 0 6 12 18
ZT ZT ZT ZT ZT

Fig. 3. Deficiency of intestinal clock affects hepatic lipid metabolism. (A,B) Relative expression of hepatic genes involved in (A) de novo lipogenesis (Acaca, Fasn,
Elovl6, and Scd1) and (B) lipid b-oxidation, lipid transport, and cholesterol metabolism in Bmal1-iKO and Bmal1-flox mice. (C) Immunoblotting of hepatic enzymes
involved in de novo lipogenesis in Bmal1-iKO and Bmal1-flox mice. (D) Plasma TGs, FFAs, and cholesterol in Bmal1-flox and Bmal1-iKO mice at different times of the
day. (F) Hepatic TGs, FFAs, and cholesterol in Bmal1-flox and Bmal1-iKO mice at different times of the day. (F) De novo lipogenesis rate measured in Bmal1-flox and
Bmal1-iKO mice by tracing hepatic 2H-oleate synthesised from 2H-labelled water. (G) Fasting b-hydroxybutyrate concentrations. (H) Relative expression of hepatic
genes involved in de novo lipogenesis (Acaca, Fasn, Elovl6, and Scd1) in Rev-erba-flox and Rev-erba-iKO mice. (I) Plasma and hepatic TGs, FFAs, and cholesterol
levels in Rev-erba-flox and Rev-erba-iKO mice at different times of the day. Data are presented as mean ± SD (n = 8). *p <0.05 vs. control (t test or two-way ANOVA with
Bonferroni post hoc test). BHB, b-hydroxybutyrate; B-iKO, Bmal1-iKO; FFA, free fatty acid; iKO, intestinal knockout; TG, trigylceride; ZT, zeitgeber time.

746 Journal of Hepatology, September 2023. vol. 79 j 741–757

 Research Article

A B

Plasma FFA (mg/dl)

150 150 300 80

Plasma TG (mg/dl)
Plasma AST (U/L)
Plasma ALT (U/L)
flox flox
iKO
* iKO
60 *
100 100 * 200 * *
* 40
*
50 50 100 * *
* * 20
0 0 0 0
Ctrl Alcohol Ctrl Alcohol Ctrl Alcohol Ctrl Alcohol Ctrl Alcohol Ctrl Alcohol Ctrl Alcohol Ctrl Alcohol
Bmal1 Rev-erbα Bmal1 Rev-erbα Bmal1 Rev-erbα Bmal1 Rev-erbα

C 100 150 D Bmal1-flox Bmal1-iKO E Bmal1-flox Bmal1-iKO

FFA (mg/g protein)

TG (mg/g protein)

flox * *
80 iKO

Ctrl

Ctrl
100
60 *

Oil Red
H&E
40
50
*

Alcohol

Alcohol
20
0 0
Ctrl Alcohol Ctrl Alcohol Ctrl Alcohol Ctrl Alcohol
Bmal1 Rev-erbα Bmal1 Rev-erbα

F 2.5 flox flox-alcohol 18 G Rev-erbα-flox Rev-erbα-iKO H Rev-erbα-flox Rev-erbα-iKO

iKO iKO-alcohol * * * * * * * *
2.0

Ctrl

Ctrl
Rel mRNA

* * * * * * * * 12
1.5

Oil Red
H&E
1.0

Alcohol

Alcohol
6
0.5
0.0 0
Tnfα Il-6 Il-1β Mcp1 Tnfα Il-6 Il-1β Mcp1
Bmal1 Rev-erbα

Fig. 4. Intestinal clock regulates the susceptibility of mouse liver to alcohol-induced injury. (A) Plasma ALT and AST levels in Bmal1-iKO, Rev-erba-iKO, and their
control mice gavaged with alcohol or saline (n = 8).(B) Plasma TG and FFA levels in Bmal1-iKO, Rev-erba-iKO, and their control mice gavaged with alcohol or saline (n = 8).
(C) Hepatic TG and FFA levels in Bmal1-iKO, Rev-erba-iKO, and their control mice gavaged with alcohol or saline (n = 8). (D) Haematoxylin and eosin (H&E) staining of livers
from Bmal1-iKO and Bmal1-flox mice gavaged with alcohol or saline. (E) Oil Red O staining of liver from Bmal1-iKO and Bmal1-flox mice gavaged with alcohol or saline. (F)
Relative expression of inflammatory factors in the liver from Bmal1-iKO, Rev-erba-iKO, and their control mice gavaged with alcohol or saline (n = 8). (G) H&E staining of
livers from Rev-erba-iKO and Rev-erba-flox mice gavaged with alcohol or saline. (H), Oil Red O staining of liver from Rev-erba-iKO and Rev-erba-flox mice gavaged with
alcohol or saline. Scale bars: 50 lm. Data are presented as mean ± SD. *p <0.05 vs. control (t test or two-way ANOVA with Bonferroni post hoc test). ALT, alanine
aminotransferase; AST, aspartate aminotransferase; FFA, free fatty acid; iKO, intestinal knockout; TG, trigylceride. (This figure appears in color on the web.)

cytokines, such as Tnfa, Il1b, Il6, and Mcp1 (Fig. S11F). Hepatic glucose output and insulin sensitivity in intestinal
Decreased sensitivity to alcoholic liver injury was further clock-deficient mice
confirmed by histology and Oil Red O staining (Fig. S11G). In line with the transcriptomic data, qPCR analyses showed
These findings supported a protective effect of Bmal1 iKO on that Bmal1 iKO caused considerable changes in several he-
alcoholic liver injury. patic genes involved in gluconeogenesis and glycolysis (G6pc,
Next, we examined the effects of Rev-erba iKO on alcoholic Pck1, Gck/Hk4, and Pklr) and in Pygl (encoding PYGL, a rate-
liver injury induced by acute gavage of ethanol. Given that Rev- limiting enzyme for glycogenolysis), but no changes in genes
erba iKO and Bmal1 iKO reprogrammed the hepatic lipid meta- related to glycogenesis and glucose transport (Gys2, Ugp2,
bolism in an opposite direction (Fig. 3), we predicted that Rev- Gbe1, and Glut2/Slc2a2) (Fig. 5A,B and Fig. S14A,B). The key
erba-iKO mice would have an increased susceptibility to alcoholic gluconeogenic genes G6pc and Pck1 were upregulated in the
liver injury. We confirmed that the intestinal barrier (tight junction) KO animals, whereas the glycolytic genes Gck and Pklr were
was largely unchanged in Rev-erba-iKO mice under both normal downregulated (Fig. 5A,B). Notably, their rhythms were all
and ethanol-exposed conditions (Figs S11H and S12A). After disrupted (Fig. 5A,B). The expression changes implicated
ethanol treatment, Rev-erba-iKO mice showed higher plasma altered hepatic glucose metabolism in mice without intestinal
levels of both ALT and AST compared with controls (Fig. 4A). They Bmal1. Intriguingly, based on glucose tolerance tests (GTTs),
also had higher TGs and FFAs in plasma and liver (Fig. 4B,C).
Bmal1-iKO mice showed ZT-dependent abnormalities in
Aggravated steatosis was further supported by histology and Oil glucose metabolism, with normal glucose tolerance at ZT6–8
Red O staining (Fig. 4G,H and Fig. S13). Moreover, genes and ZT12–14, mild impairment of glucose tolerance at ZT0–2,
encoding proinflammatory cytokines (Tnfa, Il6, Il1b, and Mcp1) and robust impairment at ZT18–20 (Fig. 5C and Fig. S14C).
were higher in the ethanol-treated knockouts than in controls They were less sensitive to insulin action at ZT18–20 but not at
(Fig. 4H). Elevations in the lipogenic genes persisted in Rev-erba- ZT6–8 according to insulin tolerance tests (ITTs) (Fig. 5D). In-
iKO mice compared with controls after binge ethanol exposure sulin clamp revealed a lower glucose infusion rate (GIR) in
(Fig. S11I). These data validated our prediction that mouse liver Bmal1-iKO mice, which was driven by a higher rate of hepatic
reprogrammed by Rev-erba iKO was more sensitive to alcohol- glucose production (HGP), but not by a low glucose disposal
induced injury. Taken together, hepatic metabolic reprogram- rate (Fig. 5E). These data suggested that impaired glucose
ming by intestinal clock modulates the susceptibility of mice to tolerance in mice without intestinal Bmal1 resulted from
alcoholic liver injury. increased glucose production. Supporting this, we found higher

Journal of Hepatology, September 2023. vol. 79 j 741–757 747

Intestinal clock reprograms rhythmic liver metabolism

A8 G6pc 3 Pck1 4 Gck 8

Bmal1-flox
Pklr
Bmal1-iKO
B Bmal1-flox Bmal1-iKO
ZT 2 6 10 14 18 22 2 6 10 14 18 22
* * * * *
Rel mRNA

6 3 6 G6PC
* * 2 *
* * PCK1
4 * 2 * 4
* *
2
1
1
* * * * * GCK
* 2 *
PKLR
0 0 0 0
0 12 24 0 12 24 0 12 24 0 12 24 GAPDH

C 800 ZT6-8 (GTT) 1,000 ZT18-20 (GTT) D 300 ZT6-8 (ITT) 250 ZT18-20 (ITT) Bmal1-flox
Glucose (mg/dl)

Glucose (mg/dl)

Glucose (mg/dl)

Glucose (mg/dl)
600 800 * * 200 * Bmal1-iKO

600 200 150

400 * * * *
400 * * 100 * *
100
200 200 50
0 0 0 0
0 30 60 90 120 0 30 60 90 120 0 30 60 90 120 0 30 60 90 120
Time (min) Time (min) Time (min) Time (min)
E 60 80 100 F 300 G 300 Bmal1-flox

Glucose (mg/dl)

Glucose (mg/dl)
*
HGP (mg/kg/min)
GIR (mg/kg/min)

Bmal1-iKO
Rd (mg/kg/min)

80 *
60 200 * 200 *
40 *
* 60
* 40 100 100
40
20
20 20 0 0
0 6 12 18 0 6 12 18
0 0 0 ZT ZT
Ctrl B-iKO Basal Clamp Ctrl B-iKO

H ZT6-8 (PTT) ZT18-20 (PTT)

I 1,000 Bmal1-flox
J 250 Bmal1-flox
K 1.2 Bmal1-flox
400 300
Glucose production
(nmol/h/mg protein)
Glucose (mg/dl)

Glucose (mg/dl)

Bmal1-iKO Bmal1-iKO Bmal1-iKO

*

Liver glycogen

Insulin (ng/ml)
200

(mg/g tissue)
300 750 *
200 * * 0.8
* 150
200 * * 500
100 100
100 0.4
250 50
0 0
0 30 60 90 120 0 30 60 90 120 0 0 0.0
Time (min) Time (min) Ctrl Lac/Pyr ZT6 ZT18 Fed Fasted

L M Primary hepatocytes N Glucose (mg/dl)

800 ZT6-8 (GTT) Rev-erbα-flox
Vehicle Insulin Bmal1-flox Bmal1-iKO * * Rev-erbα-iKO
600
Bmal1 flox iKO flox iKO Vehicle Insulin Vehicle Insulin
400 *
p-AKT p-AKT
AKT AKT 200

GAPDH GAPDH 0
0 30 60 90 120
Time (min)

O 300 P 400 Rev-erbα-flox Q 200 ZT6-8 (ITT) R 400 ZT6-8 (PTT) S 600 Rev-erbα-flox
Glucose (mg/dl)

Glucose (mg/dl)

Glucose (mg/dl)

Glucose (mg/dl)

* *
Glucose production
(nmol/h/mg protein)

Rev-erbα-iKO
Rev-erbα-iKO
200 300 150 * 300 *
* * 400
200 100 * 200 *
100
*
50 * 100
100
200
0 0 0 *
0
0 6 12 18 0 6 12 18 0 30 60 90 120 0 30 60 90 120
ZT ZT Time (min) Time (min) 0
Ctrl Lac/Pyr

Fig. 5. Deficiencies in the intestinal clock affect hepatic glucose output and insulin sensitivity. (A) Relative expression of hepatic genes involved in gluconeo-
genesis and glycolysis (G6pc, Pck1, Gck, and Pklr) in Bmal1-iKO and Bmal1-flox mice. (B) Immunoblotting of hepatic enzymes involved in gluconeogenesis and
glycolysis in Bmal1-iKO and Bmal1-flox mice. (C) GTT results examined at ZT6–8 and ZT18–20 with Bmal1-iKO and Bmal1-flox mice. (D) ITT results examined at ZT6–8
and ZT18–20 with Bmal1-iKO and Bmal1-flox mice. (E) Insulin clamp at ZT18 in Bmal1-iKO and Bmal1-flox mice. (F) Fasting plasma glucose levels in Bmal1-iKO and
Bmal1-flox mice. (G) Plasma glucose levels in fed Bmal1-iKO and Bmal1-flox mice. (H) PTT results examined at ZT6–8 and ZT18–20 with Bmal1-iKO and Bmal1-flox
mice. (I) Glucose production in primary hepatocytes isolated from Bmal1-iKO and Bmal1-flox mice. (J) Liver glycogen levels in Bmal1-iKO and Bmal1-flox mice. (K)
Plasma insulin levels in Bmal1-iKO and Bmal1-flox mice. (L) Immunoblotting of total and phosphorylated AKT in Bmal1-iKO and Bmal1-flox mice after injection of
insulin or saline. (M) Immunoblotting of total and phosphorylated AKT in primary mouse hepatocytes isolated from Bmal1-iKO and Bmal1-flox mice. (N) GTT results
examined at ZT6–8 with Rev-erba-iKO and Rev-erba-flox mice. (O) Fasting plasma glucose levels in Rev-erba-iKO and Rev-erba-flox mice. (P) Plasma glucose levels in
fed Rev-erba-iKO and Rev-erba-flox mice. (Q) ITT results examined at ZT6–8 with Rev-erba-iKO and Rev-erba-flox mice. (R) PTT results examined at ZT6–8 with Rev-
erba-iKO and Rev-erba-flox mice. (S) Glucose production in primary hepatocytes isolated from Rev-erba-iKO and Rev-erba-flox mice. Data are presented as mean ±
SD (n = 8). *p <0.05 vs. control (t test or two-way ANOVA with Bonferroni post hoc test). B-iKO, Bmal1-iKO; GIR, glucose infusion rate; GTT, glucose tolerance test;
HGP, hepatic glucose production; iKO, intestinal knockout; ITT, insulin tolerance test; PTT, pyruvate tolerance test; Rd, rate of disposal; ZT, zeitgeber time.

748 Journal of Hepatology, September 2023. vol. 79 j 741–757

 Research Article

blood glucose concentrations at ZT18 and ZT0 (but not at ZT6 similarity screening based on the ChEA3 database.34 SREBP-1
and ZT12) in KO animals under both fasted and fed conditions (also known as SREBF-1) was the most enriched TF bound
(Fig. 5F,G), aligning with the ZT-dependent glucose-tolerance near the metabolic genes the rhythms of which were disrupted
phenotype. Pyruvate tolerance tests (PTTs) confirmed that by Bmal1 iKO (Fig. 6A). SREBP-1a (ubiquitously expressed) and
Bmal1-iKO mice had more extensive gluconeogenesis at SREBP-1c (predominantly expressed in the liver) are two splice
ZT18–20, but not at ZT6–8 (Fig. 5H). By contrast, intestinal variants of SREBP-1 and differ only in the length of their N
gluconeogenesis did not differ between Bmal1-iKO mice and terminus.35 Although both SREBP-1 and SREBP-2 are involved
controls (Fig. S15). In addition, glucose production from pyru- in lipid metabolism, they mainly regulate hepatic lipogenesis
vate/lactate was higher in primary hepatocytes isolated from and cholesterol metabolism, respectively.35 Remarkably, we
mice without intestinal Bmal1 than in those from control mice found that the diurnal rhythm of hepatic Srebp-1c was negated
(Fig. 5I). Additionally, liver glycogen was higher in Bmal1-iKO by Bmal1 iKO and its expression was considerably down-
mice at ZT18 than in controls (Fig. 5J), which was probably regulated, whereas Srebp-1a and Srebp-2 remained unaffected
the result of lower expression of Pygl. Altogether, these (Fig. 6B). Concomitantly, total and nuclear SREBP-1 proteins
findings suggested that increased hepatic glucose production were reduced and their rhythms were blunted (Fig. 6B). Insig2
in mice without intestinal Bmal1 resulted from and Scap, two regulators of proteolytic processing and nuclear
enhanced gluconeogenesis. translocation of SREBP, were not affected in the liver of mice
We next asked whether insulin signalling has a role in lacking intestinal Bmal1 (Fig. S17A) and, thus, their contribu-
Bmal1-iKO-induced insulin insensitivity. Both fasted and fed tions to disrupted expression of nuclear SREBP-1c can be
Bmal1-iKO mice had a slightly higher (but without a significant excluded. The temporal change in SREBP-1c paralleled the
difference) blood insulin concentration at ZT18 compared with alteration profiles of the lipid- and glucose-related genes
controls (Fig. 5K and Fig. S16A). Phosphorylation of AKT in the Acaca, Fasn, Elovl6, Scd1, G6pc, Pck1, Gck, and Pklr, which
liver was unaffected as a result of Bmal1 iKO in response to are known SREBP-1c targets (Figs 3 and 5),35–37 suggesting
insulin stimulation (Fig. 5L). Likewise, we did not find a signifi- that reprogramming of these metabolic genes by Bmal1 iKO
cant change in AKT phosphorylation between insulin-treated was driven by SREBP-1c. By contrast, Bmal1 iKO did not affect
Bmal1-iKO and control primary hepatocytes (Fig. 5M). The other known regulators of lipid and glucose metabolism,
genes involved in insulin synthesis and signalling (such as including Ppara, Pparg, Pgc-1a, Sirt1, Crebh, Chrebp, Lxri, and
INS1/2, INSR, and IRS2) and pancreatic clock genes were Foxo1 (Fig. S17B). To further test the role of SREBP-1c in
unaffected in Bmal1-iKO mice (Fig. S16B,C). In addition, metabolic reprogramming, we performed ectopic expression of
glucose-stimulated insulin secretion and pancreatic insulin did SREBP-1c in Bmal1-iKO mice specifically in the liver using an
not differ between KO and control mice (Fig. S14D,E). Thus, adenovirus vector expressing SREBP-1c/TBG (a liver-specific
glucose intolerance in mice without intestinal Bmal1 was not thyroxine-binding globulin promoter) (Fig. S18A). By adjusting
the result of a change in insulin responsiveness. the vector dose, exogenous SREBP-1c at ZT18 could be
Lastly, we examined glucose tolerance and insulin sensi- controlled to be the level of endogenous SREBP-1c in wild-
tivity in Rev-erba-iKO mice. These mice showed improvement type liver (Fig. S18B). Ectopic SREBP-1c expression in
in glucose tolerance in a ZT-dependent manner, with markedly Bmal1-iKO liver normalised expression of the metabolic genes
improved glucose tolerance at ZT6–8 (Fig. 5N and Fig. S14F). Acaca, Fasn, Elovl6, Scd1, G6pc, Pck1, Gck, and Pklr (Fig. 6C),
They also had lower fasting and fed blood glucose concen- Concurrently, the homeostasis phenotype associated with
trations and were more sensitive to insulin action (Fig. 5O–Q), Bmal1 iKO was abrogated by ectopic SREBP-1c expression
which was not related to a change in circulating glucagon level (Fig. 6D–H). Collectedly, these results revealed that SREBP-1c
(Fig. S14G). Rev-erba iKO led to reduced gluconeogenesis in is crucial for Bmal1-iKO-induced metabolic reprogramming.
the liver (revealed by PTTs) and to decreased glucose pro- We next investigated whether hepatic SREBP-1c per se is
duction from pyruvate/lactate in primary hepatocytes reprogrammed by the nutrients and metabolites delivered from
(Fig. 5R,S). Similar to Bmal1 iKO, Rev-erba iKO did not affect the gut. We performed metabolomic profiling of the liver and
insulin secretion or insulin signalling (Fig. S14H–L). In line with a intestine from Bmal1-iKO and control mice at ZT14 and ZT18.
reduced glucose output, we found downregulation of gluco- Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis of
neogenic genes (G6pc and Pck1) and upregulation of glycolytic differential metabolites identified two common pathways
genes (Gck and Pklr), but no changes in other glucose deregulated in the liver of Bmal1-iKO mice, namely, biosyn-
metabolism-related genes in the liver of Rev-erba-iKO mice thesis of PUFAs and linoleic acid metabolism (Fig. 6I). Thus, we
(Fig. S14M,N). Therefore, Rev-erba iKO showed ZT-dependent examined major PUFAs (C18:2n6, C18:3n6, C20:2n6, C20:3n6,
improvement in glucose tolerance because of reduced hepatic C20:4n6, C18:4n3, C22:5n3, and C22:6n3) in the liver and small
glucose production. intestine around the clock by performing mass spectrometric
analysis. We found upregulation of these PUFAs and loss of
their rhythmicity in both liver and intestine of Bmal1-iKO mice
Intestinal clock-induced reprogramming of hepatic (Fig. 6J,K). Meanwhile, the KO animals had upregulated intes-
metabolism by SREBP-1c tinal expression and dampened rhythms of FADS1 and FADS2
To explore the mechanisms underlying rhythmic disruption in (two key enzymes responsible for generating PUFAs), but un-
metabolic genes in the liver caused by deficiency of the in- altered levels of Elovl2 and Elovl5, two additional contributors
testinal clock, we performed transcription factor (TF)-binding to PUFA formation (Fig. 7A,B). Reduced levels of PUFAs in

Journal of Hepatology, September 2023. vol. 79 j 741–757 749

Intestinal clock reprograms rhythmic liver metabolism

A Putative TF binding enriched at rhythm B 3

Srebp-1c
2
Srebp-1a
3
Srebp-2
disrupted metabolic genes * Bmal1-flox
Bmal1-iKO

Rel mRNA
60 *
2 2
Intergrated score

*
1
(mean rank)

40 1
* * 1

20 0 0 0
ZT 0 12 24 0 12 24 0 12 24
0 Total Nucleus
G
P1

D3

1
Bmal1-flox Bmal1-iKO Bmal1-flox Bmal1-iKO
2I

5A
AR
F
EB

E
GT

NR
ZB

ZT 2 6 10 14 18 22 2 6 10 14 18 22 2 6 10 14 18 22 2 6 10 14 18 22
PP
SR

SREBP-1
GAPDH
LaminA/C

C Bmal1-flox-ctrl D 200 100 E 10 2.0 F 200

Plasma FFA (mg/dl)

FFA (mg/g protein)

Plasma TG (mg/dl)

TG (mg/g protein)
#
Bmal1-iKO-ctrl

Glucose (mg/dl)
4 # # #
Bmal1-iKO-Ad-SREBP-1c 150 75 8 150 *
1.5 #
* *
Rel mRNA

3 6
100 * 50 * 1.0 100
2 # 4 *
# # # # # # # *
50 25 2 0.5 50
1 *
* * * * *
0 0 0 0.0 0
0
a
ns

ck
d1

k1

lr
l6

6p
ac

Pk
ov
Fa

G
Pc
Sc
Ac

G
El

G 1,000 ZT18-20 (GTT) H 200 ZT18-20 (ITT) I ZT14 Central carbon metabolism in cancer
Glucose (mg/dl)

Glucose (mg/dl)

# #
#
750 * #
#
150 * Pyrimidine metabolism
* *
500 * #
100 #
Mineral absorption
#
250 * 50 * Pantothenate and CoA biosynthesis

Lipid Linoleic acid metabolism ZT18 African trypanosomiasis

0 0 metabolism Biosynthesis of unsaturated fatty acids Asthma
0 30 60 90 120 0 30 60 90 120
Time (min) Time (min) ABC transporters Biosynthesis of unsaturated fatty acids
Lipid
Bmal1-flox-Ctrl β-alanine metabolism metabolism Linoleic acid metabolism
Bmal1-iKO-Ctrl
Bmal1-iKO-Ad-SREBP-1c 0 1 2 3 0 1 2 3
-Log10 (p value) -Log10 (p value)

J Linoleic γ-linolenic Eicosadienoic acid

K Linoleic γ-linolenic Eicosadienoic acid
acid (C18:2n6) acid (C18:3n6) (C20:2n6) acid (C18:2n6) acid (C18:3n6) (C20:2n6)
150k 6,000 400k 300k 5,000 40k
* * *
*
Intensity

Intensity

100k 4,000 * 200k *

200k 2,500 20k * *
50k 2,000 100k

0 0 0 0 0 0
ZT 0 6 12 18 0 6 12 18 0 6 12 18 ZT 0 6 12 18 0 6 12 18 0 6 12 18
Dihomo-γ-linolenic Arachidonic acid Stearidonic Dihomo-γ-linolenic Arachidonic acid Stearidonic
acid (C20:3n6) (C20:4n6) acid (C18:4n3) acid (C20:3n6) (C20:4n6) acid (C18:4n3)
50k 200k 200k 6,000 200k 30k
* * * * *
* *
Intensity

Intensity

4,000 * *
25k 100k 100k 100k * 15k
2,000

0 0 0 0 0 0
ZT 0 6 12 18 0 6 12 18 0 6 12 18 ZT 0 6 12 18 0 6 12 18 0 6 12 18
Docosapentaenoic Docosahexaenoic Docosapentaenoic Docosahexaenoic
acid (C22:5n3) acid (C22:6n3) acid (C22:5n3) acid (C22:6n3)
4,000 4,000 30,000 3,000
*
Intensity

Intensity

20,000 * 2,000 *
2,000 * 2,000
10,000 1,000

0 0 0 0
ZT 0 6 12 18 0 6 12 18 ZT 0 6 12 18 0 6 12 18

Fig. 6. Intestinal clock-induced reprogramming of hepatic metabolism by SREBP-1c. (A) Transcription factor-binding similarity screening of metabolic transcripts
based on ChEA3 database. (B) Relative expression of hepatic Srebp-1c, Srebp-1a, and Srebp-2 as well as total and nuclear SREBP-1 in Bmal1-iKO and Bmal1-flox
mice. (C) Effects of ectopic expression of SREBP-1c on hepatic genes involved in de novo lipogenesis, gluconeogenesis, and glycolysis in Bmal1-iKO mice (n = 8).
(D,E) Plasma TG and FFA levels (D) and hepatic TG and FFA levels (E) in Bmal1-iKO mice at 3 weeks after injection of Ad-SREBP-1c (n = 8). (F) Plasma glucose levels in
Bmal1-iKO mice at 3 weeks after injection of Ad-SREBP-1c (n = 8). (G,H) GTT results (G) and ITT results (H) examined at ZT18–20 in Bmal1-iKO mice at 3 weeks after
injection of Ad-SREBP-1c (n = 8). (I) KEGG pathway analysis of differential metabolites in the livers from Bmal1-iKO and Bmal1-flox mice at ZT14 and ZT18. (J,K)

750 Journal of Hepatology, September 2023. vol. 79 j 741–757

 Research Article

various tissues have been described in humans carrying intestinal BMAL1 to the Stat3 promoter, which was reduced in
FADS1/2 variants.38,39 Thus, elevated PUFAs in the liver of Bmal1-iKO mice, but enhanced in Rev-erba-iKO mice (Fig. 7J).
Bmal1-iKO mice resulted from enhanced production of PUFAs Bmal1-overexpressing plasmids resulted in increased activity
in the intestine. Supporting this, intestinal epithelial cell (IEC)- of the luciferase reporter driven by the Stat3 promoter in CT26
specific knockdown of Fads1/2 in mice led to reduced PUFAs cells (Fig. 7K). Therefore, intestinal BMAL1 promotes Stat3
and enhanced Srebp1c expression in the liver, generating transcription via direct binding to its promoter. We next
opposite metabolic phenotypes to those resulting from Bmal1 examined whether STAT3 is required for BMAL1 regulation of
iKO (Fig. S19). We also tested the PUFA changes in Rev-erba- Fads1 and Fads2 expression. We quantified Fads1 and Fads2
iKO mice. Compared with the observations in Bmal1-iKO mice, transcripts in the intestine from Stat3–/– mice and wild-type
PUFAs were decreased in both the liver and small intestine of littermates. Intestinal Fads1 and Fads2 expression was
Rev-erba-iKO mice, which agreed well with reduced intestinal increased to Bmal1-iKO levels in Stat3–/– mice (Fig. 7L). Addi-
expression of the PUFA-processing enzymes FADS1 and tionally, knockdown of Stat3 markedly attenuated the inhibitory
FADS2 (Fig. 7C,D and Fig. S20). PUFAs selectively suppress effects of BMAL1 on Fads1 and Fads2 expression in CT26 cells
the transcription and expression of Srebp-1c through proteo- (Fig. 7M). These findings indicate that BMAL1 regulation of
lytic processing and an autoloop regulatory circuit.40 Thus, Fads1 and Fads2 requires STAT3. In support of this, intestinal
reprogrammed SREBP-1c in the liver is most likely the result of ectopic expression of STAT3 normalised metabolic genes in
changes in PUFAs. the liver of Bmal1-iKO mice, and abrogated their metabolic
We then investigated the molecular mechanism by which phenotypes (Fig. S23).
BMAL1 regulates Fads1 and Fads2 in the intestine. Generally,
BMAL1 (a transcriptional activator) functions to induce gene
transcription and expression.1 Given that BMAL1 showed a Hepatic metabolism in mice with chemical-perturbed
negative regulatory effect on intestinal Fads1 and Fads2, an intestinal clock
intermediate regulator was necessary to mediate the action of Given that Bmal1 iKO (a genetic manipulation) reprograms the
BMAL1. Such a mediator should be a target of BMAL1 and a rhythmic metabolism in the liver, it was of considerable interest
negative regulator of both Fads1 and Fads2. Bioinformatic to test whether chemicals that disturb the intestinal clock have
analysis using the JASPAR database predicted the binding a similar effect. We used REV-ERBa ligands to modulate in-
sites in Fads1 and Fads2 promoters for TFs such as SREBP-1, testinal BMAL1 expression.26 Gavage of SR9009 (a REV-ERBa
PPARa, and STAT3 (Fig. 7E).41 SREBP-1 and PPARa are agonist) at a dose of 50 or 100 mg/kg once daily for 7
known activators of Fads1 and Fads2 expression.42 Hepatic consecutive days significantly downregulated BMAL1 expres-
Fads1 and Fads2 were downregulated in response to sup- sion in the intestine, but had a minimal effect on hepatic BMAL1
pression of SREBP-1c in the liver of Bmal1-iKO mice and other clock genes (Fig. S24A–C). By contrast, SR8278 (a
(Fig. S21A). Thus, we tested STAT3, which binds to the pro- REV-ERBa antagonist) at an oral dose of 50 or 100 mg/kg
moters of target genes and either activates or inhibits their upregulated intestinal BMAL1 expression, but had no effect on
transcription.43,44 Chromatin precipitation (ChIP) assays hepatic BMAL1 expression (Fig. S24D–F). Negligible effects of
revealed binding of intestinal STAT3 to Fads1 and Fads2 pro- SR9009 and SR8278 in the liver in vivo resulted from their poor
moters, which was reduced in the intestine of Bmal1-iKO mice oral bioavailability and exceedingly low liver exposure
but enhanced in Rev-erba-iKO mice (Fig. 7F). Co-transfection (Fig. S25). We then examined hepatic lipid metabolism in the
of a Stat3-encoding plasmid with a luciferase reporter driven chemical-treated vs. vehicle-treated (control) mice at ZT18
by a Fads1 or Fads2 promoter led to a decrease in luciferase when Bmal1 iKO has the largest effect (Fig. 3). SR9009-treated
activity (Fig. 7G). Mutation experiments confirmed that a mice showed lower plasma and hepatic levels of both TGs and
–1458/–1448-bp region in Fads1 and a –852/–842-bp region in FFAs compared with controls (Fig. 8A,B). Hepatic lipogenic
Fads2 were required for STAT3 actions (Fig. 7G). Moreover, genes, such as Acaca, Fasn, Elovl6, and Scd1, were also
DNA methyltransferase 1 (DNMT1, with a role in gene promoter downregulated by SR9009, whereas other lipid-related genes
methylation) was involved in STAT3 repression of Fads1/2 were unaffected (Fig. 8C,D). By contrast, SR8278-treated mice
(Fig. S22A,B). Therefore, STAT3 binds to Fads1 and Fads2 were hyperlipidaemic, with elevated lipid levels and enhanced
promoters and inhibits their transcription in the intestine. expression of lipogenic genes in the liver (Fig. 8E–H). In line
Lastly, we asked whether and how BMAL1 regulates STAT3 with the changes in hepatic lipogenesis, Srebp-1c expression
expression. We found that intestinal STAT3 expression was in the liver was decreased by SR9009, but increased by
downregulated in Bmal1-iKO mice, but upregulated in Rev- SR8278 (Fig. 8I).
erba-iKO mice (Fig. 7H and Fig. S22C). In line with this, over- We next examined whether chemical perturbance of intes-
expression of Bmal1 led to elevations in STAT3 mRNA and tinal BMAL1 altered hepatic glucose metabolism and insulin
protein in CT26 cells, whereas knockdown of Bmal1 by small sensitivity. SR9009-treated mice showed a higher blood
interfering (si)RNA resulted in decreased cellular expression of glucose concentration at ZT18 (corresponding to a robust
STAT3 (Fig. 7I). Sequence analysis identified several E-box (a glucose intolerance caused by Bmal1 iKO (Fig. 5 and 8J). They
binding motif for BMAL1) elements in the promoter of Stat3 were more intolerant of glucose and less sensitive to insulin
(Fig. S22D). ChIP assays demonstrated actual binding of action at ZT18–20 (Fig. 8K,L). This was associated with

Amounts of PUFAs in the (J) liver and (K) small intestine from Bmal1-iKO and Bmal1-flox mice (n = 8). Data are presented as mean ± SD. *#p <0.05 vs. control (one-way
or two-way ANOVA with Bonferroni post hoc test). FFA, free fatty acid; GTT, glucose tolerance test; iKO, intestinal knockout; ITT, insulin tolerance test; KEGG, Kyoto
Encyclopedia of Genes and Genomes; PUFAs, polyunsaturated fatty acids; TG, trigylceride; ZT, zeitgeber time.

Journal of Hepatology, September 2023. vol. 79 j 741–757 751

Intestinal clock reprograms rhythmic liver metabolism

A Rel mRNA
5 Fads1 * 6 Fads2 3 Elovl2 2.5 Elovl5
Bmal1-flox Bmal1-iKO

4 * 2.0
* * * * * *
3 * 4 * 2 1.5
*
2 2 1 1.0
1 0.5
0 0 0 0.0
ZT 0 12 24 0 12 24 0 12 24 0 12 24

C 6 Fads1 6 Fads2 10 Elovl2 2.5 Elovl5

Rev-erbα-flox Rev-erbα-iKO

*
* * 8 2.0
Rel mRNA

4 * * 4 6 1.5
* *
* *
2 2 4 1.0
2 0.5
0 0 0 0.0
ZT 0 12 24 0 12 24 0 12 24 0 12 24

B D E PPARα/RXR STAT3 SREBP-1

Bmal1-flox Bmal1-iKO Rev-erbα-flox Rev-erbα-iKO
ZT 2 6 10 14 18 22 2 6 10 14 18 22 ZT 2 6 10 14 18 22 2 6 10 14 18 22 -1,865 +235
FADS1 FADS1 Fads1
SREBP-1 PPARα/RXR STAT3
FADS2 FADS2
-2,017 +84
GAPDH GAPDH Fads2

F G
60 Fads1 4 Fads2 ZT6-IgG 1.5 Fads1-Luc 1.5 Fads2-Luc 1.5 Mutated-Luc Ctrl
* ZT6-STAT3 Stat3
* *
*
Input (%)

Input (%)

3 * ZT18-IgG
* 1.0 1.0 1.0
Rel RLU

Rel RLU

Rel RLU
* * * ZT18-STAT3
*
*
* *
* * * * *
* 2
* * * *
* * * * * * 0.5 0.5 0.5
1 * *
*
*
0 0 0.0 0.0 0.0
ng ng
O

O
O

50

0
0

0
0
0

s1

2
rl

rl
rl

rl

rl

rl

10
20

10
20
30

ds
Ct

Ct
Ct

Ct

Ct

Ct
iK

-iK
-iK

-iK

d
Fa

Fa
1-

al1
bα

bα

Stat3 Stat3
al

er

er
Bm

Bm
v-

v-
Re

Re

H I J
600 4 Stat3
* Ctrl ZT6-IgG
Bmal1-flox Bmal1-iKO Bmal1
Rel mRNA

400 * * ZT6-BMAL1
siNC
Input (%)

Stat3 siBmal1 3 * ZT18-IgG

2.5 200 * ZT18-BMAL1
4 * * * *
Rel mRNA

2.0 * Bmal1-flox Bmal1-iKO * 2 *

* 2 *
1.5 * ZT 2 6 10 14 18 22 2 6 10 14 18 22 * * *
0 * 1 * *
* STAT3 Bmal1 Stat3
1.0
0.5 GAPDH Ctrl Bmal1 siNC siBmal1 0
rl

rl

O
0.0 BMAL1
Ct

Ct
iK

-iK
ZT 0 12 24
1-

bα

STAT3
al

er
Bm

v-

GAPDH
Re

K 5 Stat3-Luc
L 6
M 2.0 Fads1-Luc 2.0 Fads2-Luc
Ctrl
Stat3-/-
4 * * * Bmal1-flox
1.5 * * 1.5 * *
Rel mRNA

Bmal1-iKO
4
Rel RLU

Rel RLU
Rel RLU

3
*
* * 1.0 1.0
2 * 2
1 0.5 0.5

0 0 0.0 0.0
ng Fads1 Fads2
rl
50

0
0

Bm rl
siS l1
t3

al Bm rl
siS l1
t3
Ct

10
20

Ct

Ct
a

a
ta

ta

Bmal1
+

+
1

1
al
Bm

Bm

Fig. 7. Intestinal Bmal1 negatively regulates Fads1 and Fads2 via STAT3. (A) Relative intestinal epithelial expression of Fads1, Fads2, Elovl2, and Elovl5 in Bmal1-
iKO and Bmal1-flox mice (n = 8). (B) Immunoblotting of intestinal epithelial FADS1 and FADS2 proteins in Bmal1-iKO and Bmal1-flox mice. (C) Relative intestinal
epithelial expression of Fads1, Fads2, Elovl2, and Elovl5 in Rev-erba-iKO and Rev-erba-flox mice (n = 8). (D) Immunoblotting of intestinal epithelial FADS1 and FADS2
proteins in Rev-erba-iKO and Rev-erba-flox mice. (E) Schematic of the transcription factor-binding sites in the Fads1 and Fads2 promoters obtained from JASPAR
database. (F) ChIP assays showing recruitment of intestinal epithelial STAT3 to Fads1 and Fads2 promoters in mice (n = 8). (G) Luciferase reporter assays showing
dose-dependent inhibition of Fads1 and Fads2 transcription by Stat3 (n = 8). (H) Relative intestinal epithelial expression of STAT3 in Bmal1-iKO and Bmal1-flox mice. (I)
Effects of overexpression or knockdown of Bmal1 on STAT3 expression (n = 8). (J) ChIP assays showing recruitment of intestinal epithelial BMAL1 to Stat3 promoter in
mice (n = 8). (K) Luciferase reporter assays showing dose-dependent induction of Stat3 transcription by Bmal1 (n = 8). (L) Relative intestinal expression of Fads1 and

752 Journal of Hepatology, September 2023. vol. 79 j 741–757

 Research Article

A 150
B 15
C

Plasma FFA (mg/dl)

25 2.5

FFA (mg/g protein)

TG (mg/g protein)
Plasma TG (mg/dl) 1.5 Vehicle
* 20 2.0

Rel mRNA
100 10 * SR9009-50
* 15 * * 1.5 1.0
* * SR9009-100
* * * *
50 10 5 1.0 * *
0.5 *
5 0.5
0 0 0 0.0 0.0
SR 900 icle

SR 900 icle

SR 900 icle

SR 900 icle

sn
09 0

09 0

09 0

09 0
00

00

00

00

l6

d1
ac
90 9-5

90 9-5

90 9-5

90 9-5

ov
Fa
-1

-1

-1

-1

Sc
SR Veh

SR Veh

SR Veh

SR Veh

Ac

El
D E 200
F 12 5

Plasma FFA (mg/dl)

3 80

FFA (mg/g protein)

TG (mg/g protein)
Plasma TG (mg/dl)
Vehicle
SR9009-50 * * * 4 *
* 60
Rel mRNA

2 SR9009-100 * 8 *
3
100 40 *
2
1 4
20 1
0 0 0 0 0

SR 827 icle

SR 827 icle

SR 827 icle

SR 827 icle
78 0

78 0

78 0

78 0
00

00

00

00
dh

dh
a

oe

s2
36

5
Eh c

82 8-5

82 8-5

82 8-5

82 8-5
dl
t1

t1

cg

-1

-1

-1

-1
ha

Ha

gc
Ap

oc
Cd

SR Veh

SR Veh

SR Veh

SR Veh
Vl
Cp

Cp

Ab
Hm
Ap

G H I J

Rel Srebp-1c mRNA

8 Vehicle 3 Vehicle 4 Vehicle 300 Vehicle

Glucose (mg/dl)
SR8278-50 SR8278-50
6 * 3
50
* 50
Rel mRNA

Rel mRNA

SR8278-100 SR8278-100 100 100

2 200 *
* * 2 *
4 * *
* 100
**
2 * * 1
1 **
0 0 0 0
a

sn

l6

d1

09

78

09

78
dh

dh

Hm 3
a

oe

s2

5
Eh c

r
ac

dl

oc
t1

t1

cg
ov
Fa

Sc

90

82

90

82
gc
ha

Ha

Ap
Vl
Cp

Cp
Ac

Ap

Ab
El

SR

SR

SR

SR
K 600
GTT L 400
ITT M 250 *
PTT N 8 Vehicle
Glucose (mg/dl)

Glucose (mg/dl)

Glucose (mg/dl)

* 6
SR9009-50 *
*

Rel mRNA
* 300 * 200 * SR9009-100
400 * *
* 150 *
200 4 *
* * 100 * *
200 2 *
Vehicle 100 50
SR9009
** **
0 0 0 0
0 30 60 90 120 0 30 60 90 120 0 30 60 90 120 G6pc Pck1 Pklr Gck
Time (min) Time (min) Time (min)
O 500
GTT P 200
ITT Q 200
PTT R 5 Vehicle
*
Glucose (mg/dl)

Glucose (mg/dl)

Glucose (mg/dl)

Vehicle 4 SR8278-50
Rel mRNA

400 * SR8278 150 * 150 SR8278-100

* * * * 3 *
300 * * *
* 100 * 100 * 2
200 * * *
100 50 50 1
** **
0 0 0 0
0 30 60 90 120 0 30 60 90 120 0 30 60 90 120 G6pc Pck1 Pklr Gck
Time (min) Time (min) Time (min)

Fig. 8. Chemical-perturbed intestinal clock causes reprogramming of hepatic metabolism. (A–D) Effects of oral SR9009 on (A) plasma TG and FFA levels, (B)
hepatic TG and FFA levels, (C) hepatic Acaca, Fasn, Elovl6, and Scd1, and (D) hepatic genes involved in lipid b-oxidation, lipid transport, and cholesterol metabolism in
mice. (E–H) Effects of oral SR8278 on (E) plasma TG and FFA levels, (F) hepatic TG and FFA levels, (G) hepatic Acaca, Fasn, Elovl6, and Scd1, and (H) hepatic genes
involved in lipid b-oxidation, lipid transport and cholesterol metabolism in mice. (I,J) Effects of oral SR9009 and SR8278 on (I) hepatic Srebp-1c and (J) plasma glucose
levels in mice. (K–M) (K) GTT, (L) ITT, and (M) PTT results, examined at ZT18–20 in mice after gavage of SR9009 for 7 days. (N) Effects of oral SR9009 on hepatic G6pc,
Pck1, Pklr, and Gck in mice. (O–P) (O) GTT, (P) ITT, and (Q) PTT results examined at ZT18–20 in mice after gavage of SR8278 for 7 days. (R) Effects of oral SR8278 on
hepatic G6pc, Pck1, Pklr, and Gck in mice. Data are presented as mean ± SD (n = 8). *p <0.05 vs. control (one-way or two-way ANOVA with Bonferroni post hoc test).
FFA, free fatty acid; GTT, glucose tolerance test; iKO, intestinal knockout; ITT, insulin tolerance test; TG, trigylceride; ZT, zeitgeber time.

enhanced gluconeogenesis as revealed by PTT, upregulation of hypoglycaemic and had improved glucose tolerance and in-
G6pc and Pck1, as well as downregulation of Gck and Pklr sulin hypersensitivity (Fig. 8J,O,P), accompanied by reduced
(Fig. 8M,N). By contrast, SR8278-treated mice were hepatic gluconeogenesis (Fig. 8Q,R). Altogether, chemical

Fads2 in Stat3–/–, Bmal1-iKO and control mice (n = 8). (M) Effects of Bmal1 on Fads1 and Fads2 promoter activities in Stat3-deficient CT26 cells (n = 8). Data are
presented as mean ± SD. *p <0.05 vs. control (t test or one-way or two-way ANOVA with Bonferroni post hoc test). ChIP, chromatin precipitation; iKO, intestinal
knockout; ZT, zeitgeber time.

Journal of Hepatology, September 2023. vol. 79 j 741–757 753

Intestinal clock reprograms rhythmic liver metabolism

perturbance of intestinal BMAL1 remodels lipid and glucose the clinical observations that patients with intestinal diseases,
metabolism in the liver, mimicking the effects of genetic defi- such as ulcerative colitis, are at a higher risk for developing liver
ciency of intestinal clock. disorders, such as fatty livers.54 This is because ulcerative
colitis is associated with local clock disruption; in particular,
Rev-erba is markedly downregulated with a flat expression
Discussion pattern.55 As revealed by the current study, suppression of
A previous study by Manella et al. established the liver as a key intestinal Rev-erba can result in enhanced lipid biosynthesis
organ coordinating the rhythmicity of peripheral tissues (such and fat accumulation in the liver because of elevated expres-
as the heart, lung, kidney, WAT, and quadriceps) in response to sion of SREBP-1c (Fig. 3).
feeding.29 Thus, it is proposed that the liver occupies the top Based on epidemiologic data, hepatosteatosis (hepatic
level among peripheral tissues with respect to food entrain- lipid accumulation) is highly correlated with insulin resistance,
ment.29 Our findings that the intestinal clock has an essential and they can contribute pathologically to each other.56
role in shaping liver rhythmicity (along with those of WAT and Overwhelming lipid accumulation can generate harmful lipid
kidney) and that entrainment of the liver clock by inverted species, such as diacylglycerols, which can activate several
feeding or HFD requires the presence of intestinal Bmal1 sug- protein kinases (e.g. PKCε) and inflammatory responses,
gest that, of the peripheral clocks, the intestinal clock occupies thereby disrupting insulin signalling and resulting in insulin
the highest rank in terms of entrainment responses to meal resistance.57 By contrast, insulin resistance can increase he-
timing. Importantly, we showed that disruptions of intestinal patic uptake of carbohydrates and FFAs into the liver, pro-
rhythms by loss of either Bmal1 (a positive element of clock) or moting lipogenesis and TG synthesis.58 In sharp contrast,
Rev-erba (a negative element of clock) were associated with insulin resistance caused by Bmal1 iKO engenders a hypo-
disturbances in temporal lipid and glucose metabolism in the lipidaemic state, whereas hepatic lipid accumulation (hyper-
liver (Fig. S26). Therefore, we could conclude that the circadian lipidaemia) induced by Rev-erba iKO is accompanied by
system and metabolic homeostasis in the liver are dependent insulin hypersensitivity (Figs 3 and 5). Coexistence of hep-
on the intestinal clock. Three biological replicates per condition atosteatosis and improved insulin sensitivity has been also
were used in the RNA-sequencing experiments, which was a noted in mice with liver-specific depletion of Hdac3 and in
study limitation because low sample size can compromise the CGI58-silenced mice,59 and concurrent insulin resistance and
sensitivity and accuracy of differential expression measure- hypolipidaemia are seen in SREBP1-silenced mice.60,61 In
ments.45 It is acknowledged that the mechanisms underlying addition, dissociation between fatty liver and insulin resis-
retained patterns of hepatic clock genes in the DF regime in tance is observed in humans carrying a variant of PNPLA3.62
Bmal1-iKO mice are not resolved here, but could be related to Thus, fatty liver and insulin resistance do not necessarily have
altered metabolic signals, such as FFAs.46 a cause-and-effect relationship. This is also supported by
The intestine is considerably influenced by circadian other studies.63,64 For instance, overexpression of DGAT2 in
rhythms. Many aspects of intestinal activities (such as gastric the liver leads to hepatic lipid accumulation without alterations
emptying, intestinal motility, epithelial cell renewal, DNA syn- in glucose or insulin resistance.64 Insulin hypersensitivity in
thesis, and nutrient processing) oscillate in a circadian mice lacking hepatic Pten is sufficient to cause hep-
pattern.47–49 Gastrointestinal disorders (e.g. irritable bowel atosteatosis.63 In fact, the coexistence of hepatic steatosis
syndrome, gastroesophageal reflex disease, and peptic ulcer and insulin resistance is generally promoted by pathological
disease) are the major complaints of shift workers and jet- overnutrition (e.g. HFD), which provides surplus nutrients and
lagged travellers, both groups of which show circadian metabolic precursors to the liver to oversaturate and
misalignment.50–52 Thus, the impact of intestinal rhythms on the constantly activate both lipogenesis and gluconeogenesis.63
rhythmicity of peripheral tissues could have been under- Hepatic glucose production and lipid biosynthesis are
estimated previously. Our study suggests that some, if not all, interconnected (through shared metabolic precursors, such as
metabolic rhythms can be transmitted from the intestine to the acetyl-CoA) and temporally coordinated, and both display a
liver using nutrients and metabolites as a transmitter. PUFAs circadian rhythm. In nocturnal mice, hepatic gluconeogenesis
serve as one such transmitter and convey the rhythms of in- dominates during the daytime (rest phase with less feeding) to
testinal FADS1/2 to hepatic SREBP-1c, leading to diurnal maintain normoglycaemia, whereas lipid synthesis is more
rhythms in lipid biosynthesis and glucose production in the liver extensive during the night (active and feeding period).65,66 This
(Figs 3 and 5). circadian shift between hepatic lipogenesis and gluconeogen-
The gut-to-liver rhythmicity transmission extends the esis is essential for metabolic homeostasis and health benefits.
concept of gut–liver communication (or gut–liver axis, having a Our findings establish the intestinal clock as a key regulator of
major impact on the diagnosis and treatment of liver diseases), rhythmic diversion between hepatic glucose production and
and underscores the role of the intestinal clock in maintaining lipid biosynthesis. Bmal1 iKO led to a shift from lipid biosyn-
liver and metabolic health. Rhythmicity transmission has a thesis to gluconeogenesis during the night, resulting in
limited effect on the liver clock, which is known to be sensitive enhanced glucose production (hyperglycaemia) and insulin
to energy deficiency as well as food-stimulated hormones, insensitivity (Fig. S26). Conversely, Rev-erba iKO resulted in a
such as insulin and glucagon.5,53 This is supported by the fact shift from gluconeogenesis to lipid biosynthesis in the daytime,
that chow diet-fed Bmal1-iKO mice had a normal metabolic resulting in enhanced lipogenesis and increased susceptibility
phenotype (along with unchanged body weight) and, thus, a to alcoholic liver injury (Fig. S26). These improper shifts are
similar nutritional state compared with wild-type mice and that attributed to a disruption of the hepatic SREBP-1c rhythm,
the insulin and glucagon signalling pathways are unaltered by which is normally maintained via gut-derived PUFAs produced
intestinal clock malfunction.26 Our findings could help explain by intestinal FADS1/2 under the control of local clock (Figs 6

754 Journal of Hepatology, September 2023. vol. 79 j 741–757

 Research Article

and 7). Thus, we highlight here the circadian nature of SREBP- whether the gut microbiome has a role in driving the observed
1c in the temporal coordination of lipid and glucose metabolism metabolic shifts in the liver. We found that loss of microbiota
by the intestinal clock. (germ-free, GF) did not affect the temporal alterations in lipid
The lipogenic genes Fasn, Acaca, Scd1, and Elovl6, as well and glucose metabolism in Bmal1-iKO and Rev-erba-iKO mice
as the glycolytic genes Gck and Pklr are bona fide target genes (Fig. S27A–J). Likewise, iKO-induced changes in intestinal
of SREBP-1c,35,37 which promotes their expression using a STAT3 and FADS1/2 as well as hepatic PUFAs, SREBP-1c, and
transcriptional activation mechanism.35 Consistent with prior its target genes were independent of gut microbiome (Figs
reports,61 SREBP-1c negatively regulates the gluconeogenic S27K–L and S28). Consistently, in previous reports, hepatic
genes G6pc and Pck1. Such regulation involves an indirect SREBP-1c and its target genes were unaffected in GF mice
mechanism: SREBP-1c inhibits HNF4 transactivation of the compared with conventionally raised SPF mice.71,72 Thus, we
gluconeogenic genes by interfering with promoter recruitment can exclude a role of the microbiome in intestinal clock-gated
of the coactivator PGC-1.67 SREBP-1 is also shown to sup- rhythmic liver metabolism with respect to lipid biosynthesis
press Pck1 expression by interfering with the stimulatory effect and glucose production. We demonstrated that circadian
of Sp1 at a sterol regulatory element.68 We found that hepatic dyssynchrony in the intestine can lead to aberrant liver meta-
Acaca rather than Acacb was downregulated with a blunted bolism, corroborating the role of circadian rhythms in the
rhythm in Bmal1-iKO mice because of suppression of SREBP- pathogenesis of fatty liver diseases, such as alcoholic liver
1c (Fig. 3A and Fig. S21B). This agrees well with a previous disease and nonalcoholic fatty liver disease.78 Hence, cor-
report showing that hepatic Acacb expression is uniquely recting intestinal circadian dyssynchrony could provide bene-
dependent on SREBP-1a and that SREBP-1c fails to substitute fits for the treatment of fatty liver diseases.
for SREBP-1a in activating Acacb because of the more potent In sum, the intestinal clock carries a prominent remodelling
activation domain in SREBP-1a.69,70 In fact, Acacb is not a effect on the rhythmic transcriptome of the liver, irrespective of
simple lipogenic gene required for lipid biosynthesis, but has a its clock. Importantly, the intestinal clock controls rhythmic lipid
crucial role in regulating FA partitioning into cytosolic TG or and glucose metabolism in the liver, relying on modulation of
mitochondria for oxidation.69 Thus, the unique alteration in hepatic SREBP-1c expression by temporally regulated gut-
Acaca corroborates the conclusion that the remodelling effects delivered PUFAs. Malfunction of the intestinal clock results in
of the intestinal clock are specific to SREBP-1c in the either insulin resistance or hepatic lipid accumulation with an
liver (Fig. 6B). increased susceptibility to alcoholic liver injury. These findings
Mounting evidence indicates the gut microbiome as a reveal a pivotal role of the intestinal clock in shaping liver
key modifier of the intestinal and hepatic rhythmic tran- rhythmicity and diurnal metabolism, and suggest targeting in-
scriptome.71–75 Microbially derived bile acids and short-chain testinal rhythms as a new avenue for improving meta-
FAs (SCFAs) are suggested to be entraining agents to periph- bolic health.
eral circadian rhythms.76,77 This prompted us to examine

Affiliations
1
Institute of Molecular Rhythm and Metabolism, Guangzhou University of Chinese Medicine, Guangzhou, China; 2Department of Pharmacy, the First Affiliated Hospital
of Zhengzhou University, Zhengzhou, China; 3Guangdong Provincial Clinical Research Center for Geriatrics, Shenzhen Clinical Research Center for Geriatrics, Shenzhen
People’s Hospital (The Second Clinical Medical College, Jinan University; The First Affiliated Hospital, Southern University of Science and Technology), Shenzhen,
Guangdong, China

Abbreviations Authors’ contributions

ALT, alanine aminotransferase; AST, aspartate aminotransferase; B-iKO, Bmal1- Participated in research design: MC, YL, ZQ, JY, KL, BW. Conducted experi-
iKO; BHB, b-hydroxybutyrate; ChIP, chromatin immunoprecipitation; DF, daytime ments: MC, YL, YD, YX, FZ, GS, XJ, LZ, JD, SD. Performed data analysis: MC, YL,
feeding; DNMT1, DNA methyltransferase 1; GTT, glucose tolerance test; FFA, free YD, YX, FZ, GS, XJ, XZ, BW. Wrote or contributed to the writing of the manu-
fatty acid; GC-MS, gas chromatography-mass spectrometry; GIR, glucose infu- script: MC, BW.
sion rate; H&E, haematoxylin and eosin; HFD, high-fat diet; HGP, hepatic glucose
production; ITT, insulin tolerance test; KEGG, Kyoto Encyclopedia of Genes and
Genomes; KO, knockout; PTT, pyruvate tolerance test; PUFAs, polyunsaturated Data availability statement
fatty acids; Rd, rate of disposal; SCN, suprachiasmatic nucleus; TG, triglycerides;
WAT: white adipose tissue; ZT, zeitgeber time. All data associated with this study are present in the manuscript or the supple-
mentary data online.

Financial support
This work was supported by the project for young Qihuang scholars of the Supplementary data
state administration of traditional Chinese medicine, the National Key R&D Supplementary data to this article can be found online at https://doi.org/10.1016/
Program of China (2020YFC2008300 and 2020YFC2008304), Young Elite j.jhep.2023.04.040.
Scientists Sponsorship Program by Henan Association for Science and Tech-
nology (2022HYTP045), the National Natural Science Foundation of China References
(82273040), and the Science and Technology Foundation of Shenz-
hen (JCYJ20200109144410181). Author names in bold designate shared co-first authorship.

[1] Ralph MR, Foster RG, Davis FC, Menaker M. Transplanted suprachiasmatic
Conflicts of interests nucleus determines circadian period. Science 1990;247:975–978.
The authors declare no conflicts of interest that pertain to this work. [2] Lu D, Zhao M, Chen M, Wu B. Circadian clock-controlled drug metabolism:
Please refer to the accompanying ICMJE disclosure forms for further details. implications for chonotherapeutics. Drug Metab Dispos 2020;48:395–406.

Journal of Hepatology, September 2023. vol. 79 j 741–757 755

Intestinal clock reprograms rhythmic liver metabolism

[3] Dibner C, Schibler U, Albrecht U. The mammalian circadian timing system: [31] Sprenger RR, Hermansson M, Neess D, Becciolini LS, Sørensen SB,
organization and coordination of central and peripheral clocks. Annu Rev Fagerberg R, et al. Lipid molecular timeline profiling reveals diurnal crosstalk
Physiol 2010;72:517–549. between the liver and circulation. Cell Rep 2021;34:108710.
[4] Challet E, Caldelas I, Graff C, Pévet P. Synchronization of the molecular [32] Yu F, Zhang T, Zhou C, Xu H, Guo L, Chen M, et al. The circadian clock gene
clockwork by light- and food-related cues in mammals. Biol Chem Bmal1 controls intestinal exporter MRP2 and drug disposition. Theranostics
2003;384:711–719. 2019;9:2754–2767.
[5] Stokkan KA, Yamazaki S, Tei H, Sakaki Y, Menaker M. Entrainment of the [33] Xu J, Liu X, Gao B, Karin M, Tsukamoto H, Brenner D, et al. New approaches
circadian clock in the liver by feeding. Science 2001;291:490–493. for studying alcoholic liver disease. Curr Pathobiol Rep 2014;2:171–183.
[6] Asher G, Schibler U. Crosstalk between components of circadian and [34] Keenan AB, Torre D, Lachmann A, Leong AK, Wojciechowicz ML, Utti V,
metabolic cycles in mammals. Cell Metab 2011;13:125–137. et al. ChEA3: transcription factor enrichment analysis by orthogonal omics
[7] Dallmann R, Viola AU, Tarokh L, Cajochen C, Brown SA. The human circa- integration. Nucleic Acids Res 2019;47:W212–W224.
dian metabolome. Proc Natl Acad Sci U S A 2012;109:2625–2629. [35] Horton JD, Goldstein JL, Brown MS. SREBPs: activators of the complete
[8] Chen L, Yang G. PPARs integrate the mammalian clock and energy meta- program of cholesterol and fatty acid synthesis in the liver. J Clin Invest
bolism. PPAR Res 2014;2014:653017. 2002;109:1125–1131.
[9] Brewer M, Lange D, Baler R, Anzulovich A. SREBP1 as a transcriptional [36] Bécard D, Hainault I, Azzout-Marniche D, Bertry-Coussot L, Ferré P,
integrator of circadian and nutritional cues in the liver. J Biol Rhythms Foufelle F. Adenovirus-mediated overexpression of sterol regulatory element
2005;20:195–205. binding protein-1c mimics insulin effects on hepatic gene expression and
[10] Iizuka K. Recent progress on the role of ChREBP in glucose and lipid glucose homeostasis in diabetic mice. Diabetes 2001;50:2425–2430.
metabolism. Endocr J 2013;60:543–555. [37] Foretz M, Pacot C, Dugail I, Lemarchand P, Guichard C, Le Lièpvre X, et al.
[11] Shi SQ, Ansari TS, McGuinness OP, Wasserman DH, Johnson CH. Circadian ADD1/SREBP-1c is required in the activation of hepatic lipogenic gene
disruption leads to insulin resistance and obesity. Curr Biol expression by glucose. Mol Cel Biol 1999;19:3760–3768.
2013;23:372–381. [38] Schaeffer L, Gohlke H, Müller M, Heid IM, Palmer LJ, Kompauer I, et al.
[12] Han H, Dou J, Hou Q, Wang H. Role of circadian rhythm and impact of Common genetic variants of the FADS1 FADS2 gene cluster and their
circadian rhythm disturbance on the metabolism and disease. J Cardiovasc reconstructed haplotypes are associated with the fatty acid composition in
Pharmacol 2022;79:254–263. phospholipids. Hum Mol Genet 2006;15:1745–1756.
[13] Hatori M, Vollmers C, Zarrinpar A, DiTacchio L, Bushong EA, Gill S, et al. [39] Lattka E, Illig T, Koletzko B, Heinrich J. Genetic variants of the FADS1 FADS2
Time-restricted feeding without reducing caloric intake prevents metabolic gene cluster as related to essential fatty acid metabolism. Curr Opin Lipidol
diseases in mice fed a high-fat diet. Cel Metab 2012;15:848–860. 2010;21:64–69.
[14] Eckel-Mahan KL, Patel VR, de Mateo S, Orozco-Solis R, Ceglia NJ, Sahar S, [40] Nakakuki M, Kawano H, Notsu T, Imada K, Mizuguchi K, Shimano H. A novel
et al. Reprogramming of the circadian clock by nutritional challenge. Cell processing system of sterol regulatory element-binding protein-1c regulated
2013;155:1464–1478. by polyunsaturated fatty acid. J Biochem 2014;155:301–313.
[15] Murakami M, Tognini P, Liu Y, Eckel-Mahan KL, Baldi P, Sassone-Corsi P. [41] Fornes O, Castro-Mondragon JA, Khan A, van der Lee R, Zhang X,
Gut microbiota directs PPARc-driven reprogramming of the liver circadian Richmond PA, et al. Jaspar 2020: update of the open-access database of
clock by nutritional challenge. EMBO Rep 2016;17:1292–1303. transcription factor binding profiles. Nucleic Acids Res 2020;48:D87–D92.
[16] Garaulet M, Gómez-Abellán P, Alburquerque-Béjar JJ, Lee YC, Ordovás JM, [42] Matsuzaka T, Shimano H, Yahagi N, Amemiya-Kudo M, Yoshikawa T,
Scheer FA. Timing of food intake predicts weight loss effectiveness. Int J Hasty AH, et al. Dual regulation of mouse Delta(5)- and Delta(6)-desaturase
Obes (Lond) 2013;37:604–611. gene expression by SREBP1 and PPARalpha. J Lipid Res 2002;43:107–114.
[17] Sutton EF, Beyl R, Early KS, Cefalu WT, Ravussin E, Peterson CM. Early [43] Zhang Q, Wang HY, Woetmann A, Raghunath PN, Odum N, Wasik MA.
time-restricted feeding improves insulin sensitivity, blood pressure, and STAT3 induces transcription of the DNA methyltransferase 1 gene (DNMT1)
oxidative stress even without weight loss in men with prediabetes. Cel Metab in malignant T lymphocytes. Blood 2006;108:1058–1064.
2018;27:1212–1221. [44] Niu G, Wright KL, Ma Y, Wright GM, Huang M, Irby R, et al. Role of Stat3 in
[18] Gill S, Panda S. A smartphone app reveals erratic diurnal eating patterns in regulating p53 expression and function. Mol Cel Biol 2005;25:7432–7440.
humans that can be modulated for health benefits. Cel Metab [45] Schurch NJ, Schofield P, Gierlin  ski M, Cole C, Sherstnev A, Singh V, et al.
2015;22:789–798. How many biological replicates are needed in an RNA-seq experiment and
[19] Albillos A, de Gottardi A, Rescigno M. The gut-liver axis in liver disease: which differential expression tool should you use? RNA 2016;22:839–851.
pathophysiological basis for therapy. J Hepatol 2020;72:558–577. [46] Mukherji A, Kobiita A, Chambon P. Shifting the feeding of mice to the rest
[20] Mells JE, Anania FA. The role of gastrointestinal hormones in hepatic lipid phase creates metabolic alterations, which, on their own, shift the peripheral
metabolism. Semin Liver Dis 2013;33:343–357. circadian clocks by 12 hours. Proc Natl Acad Sci U S A
[21] Plauth M, Raible A, Gregor M, Hartmann F. Inter-organ communication be- 2015;112:E6683–E6690.
tween intestine and liver in vivo and in vitro. Semin Cel Biol 1993;4:231–237. [47] Segers A, Depoortere I. Circadian clocks in the digestive system. Nat Rev
[22] Abu-Wasel B, Molinari M. Liver disease secondary to intestinal failure. Bio- Gastroenterol Hepatol 2021;18:239–251.
med Res Int 2014;2014:968357. [48] Okazaki F, Matsunaga N, Okazaki H, Azuma H, Hamamura K, Tsuruta A,
[23] Spagnuolo R, Abenavoli L, Corea A, Larussa T, Mancina RM, Cosco C, et al. et al. Circadian clock in a mouse colon tumor regulates intracellular iron
Multifaceted pathogenesis of liver steatosis in inflammatory bowel disease: a levels to promote tumor progression. J Biol Chem 2016;291:7017–7028.
systematic review. Eur Rev Med Pharmacol Sci 2021;25:5818–5825. [49] Pácha J, Sumová A. Circadian regulation of epithelial functions in the in-
[24] Palmela C, Peerani F, Castaneda D, Torres J, Itzkowitz SH. Inflammatory testine. Acta Physiol (Oxf) 2013;208:11–24.
bowel disease and primary sclerosing cholangitis: a review of the phenotype [50] Voigt RM, Forsyth CB, Keshavarzian A. Circadian rhythms: a regulator of
and associated specific features. Gut Liver 2018;12:17–29. gastrointestinal health and dysfunction. Expert Rev Gastroenterol Hepatol
[25] Pan X, Hussain MM. Gut triglyceride production. Biochim Biophys Acta 2019;13:411–424.
2012;1821:727–735. [51] Chrobak AA, Nowakowski J, Zwolin ska-Wcisło M, Cibor D, Przybylska-
[26] Yu F, Wang Z, Zhang T, Chen X, Xu H, Wang F, et al. Deficiency of intestinal Felus M, Ochyra K, et al. Associations between chronotype, sleep distur-
Bmal1 prevents obesity induced by high-fat feeding. Nat Com- bances and seasonality with fatigue and inflammatory bowel disease
mun 2021;12:5323. symptoms. Chronobiol Int 2018;35:1142–1152.
[27] Danhauser K, Sauer SW, Haack TB, Wieland T, Staufner C, Graf E, et al. [52] Segawa K, Nakazawa S, Tsukamoto Y, Kurita Y, Goto H, Fukui A, et al.
DHTKD1 mutations cause 2-aminoadipic and 2-oxoadipic aciduria. Am J Peptic ulcer is prevalent among shift workers. Dig Dis Sci 1987;32:449–453.
Hum Genet 2012;91:1082–1087. [53] Oosterman JE, Kalsbeek A, la Fleur SE, Belsham DD. Impact of nutrients on
[28] Van Rompay AR, Norda A, Lindén K, Johansson M, Karlsson A. Phos- circadian rhythmicity. Am J Physiol Regul Integr Comp Physiol
phorylation of uridine and cytidine nucleoside analogs by two human uridine- 2015;308:R337–R350.
cytidine kinases. Mol Pharmacol 2001;59:1181–1186. [54] Chao CY, Battat R, Al Khoury A, Restellini S, Sebastiani G, Bessissow T.
[29] Manella G, Sabath E, Aviram R, Dandavate V, Ezagouri S, Golik M, et al. The Co-existence of non-alcoholic fatty liver disease and inflammatory bowel
liver-clock coordinates rhythmicity of peripheral tissues in response to disease: a review article. World J Gastroenterol 2016;22:7727–7734.
feeding. Nat Metab 2021;3:829–842. [55] Wang S, Lin Y, Yuan X, Li F, Guo L, Wu B. REV-ERBa integrates colon clock
[30] Pan X, Hussain MM. Clock is important for food and circadian regulation of with experimental colitis through regulation of NF-jB/NLRP3 axis. Nat
macronutrient absorption in mice. J Lipid Res 2009;50:1800–1813. Commun 2018;9:4246.

756 Journal of Hepatology, September 2023. vol. 79 j 741–757

 Research Article

[56] Clark JM. The epidemiology of nonalcoholic fatty liver disease in adults. J [68] Chakravarty K, Wu SY, Chiang CM, Samols D, Hanson RW. SREBP-1c and
Clin Gastroenterol 2006;40:S5–S10. Sp1 interact to regulate transcription of the gene for phosphoenolpyruvate
[57] Glass CK, Olefsky JM. Inflammation and lipid signaling in the etiology of carboxykinase (GTP) in the liver. J Biol Chem 2004;279:15385–15395.
insulin resistance. Cel Metab 2012;15(5):635–645. [69] Im SS, Hammond LE, Yousef L, Nugas-Selby C, Shin DJ, Seo YK, et al.
[58] Sanyal AJ. Mechanisms of Disease: pathogenesis of nonalcoholic fatty liver Sterol regulatory element binding protein 1a regulates hepatic fatty acid
disease. Nat Clin Pract Gastroenterol Hepatol 2005;2:46–53. partitioning by activating acetyl coenzyme A carboxylase 2. Mol Cel Biol
[59] Crooke R, Shulman GI, Xue B, Shi H, Yu L. CGI-58 knockdown in mice 2009;29:4864–4872.
causes hepatic steatosis but prevents diet-induced obesity and glucose [70] Horton JD, Shah NA, Warrington JA, Anderson NN, Park SW, Brown MS,
intolerance. J Lipid Res 2010;51:3306–3315. et al. Combined analysis of oligonucleotide microarray data from transgenic
[60] Jideonwo V, Hou Y, Ahn M, Surendran S, Morral N. Impact of silencing and knockout mice identifies direct SREBP target genes. Proc Natl Acad Sci
hepatic SREBP1 on insulin signaling. PLoS ONE 2018;13:e0196704. U S A 2003;100:12027–12032.
[61] Ruiz R, Jideonwo V, Ahn M, Surendran S, Tagliabracci VS, Hou Y, et al. [71] Leone V, Gibbons SM, Martinez K, Hutchison AL, Huang EY, Cham CM,
Sterol regulatory element-binding protein-1 (SREBP1) is required to regulate et al. Effects of diurnal variation of gut microbes and high-fat feeding on host
glycogen synthesis and gluconeogenic gene expression in mouse liver. J circadian clock function and metabolism. Cell Host Microbe
Biol Chem 2014;289:5510–5517. 2015;17:681–689.
[62] Franko A, Merkel D, Kovarova M, Hoene M, Jaghutriz BA, Heni M, et al. [72] Weger BD, Gobet C, Yeung J, Martin E, Jimenez S, Betrisey B, et al. The
Dissociation of fatty liver and insulin resistance in I148M PNPLA3 carriers: mouse microbiome is required for sex-specific diurnal rhythms of gene
differences in diacylglycerol (DAG) FA18:1 lipid species as a possible expression and metabolism. Cel Metab 2019;29:362–382.
explanation. Nutrients 2018;10:1314. [73] Mukherji A, Kobiita A, Ye T, Chambon P. Homeostasis in intestinal epithelium
[63] Sun Z, Miller RA, Patel RT, Chen J, Dhir R, Wang H, et al. Hepatic Hdac3 is orchestrated by the circadian clock and microbiota cues transduced by
promotes gluconeogenesis by repressing lipid synthesis and sequestration. TLRs. Cell 2013;153:812–827.
Nat Med 2012;18:934–942. [74] Joyce SA, MacSharry J, Casey PG, Kinsella M, Murphy EF, Shanahan F,
[64] Monetti M, Levin MC, Watt MJ, Sajan MP, Marmor S, Hubbard BK, et al. et al. Regulation of host weight gain and lipid metabolism by
Dissociation of hepatic steatosis and insulin resistance in mice over- bacterial bile acid modification in the gut. Proc Natl Acad Sci U S A
expressing DGAT in the liver. Cel Metab 2007;6:69–78. 2014;111:7421–7426.
[65] Takeda Y, Kang HS, Freudenberg J, DeGraff LM, Jothi R, Jetten AM. Reti- [75] Larsson E, Tremaroli V, Lee YS, Koren O, Nookaew I, Fricker A, et al.
noic acid-related orphan receptor c (RORc): a novel participant in the diurnal Analysis of gut microbial regulation of host gene expression along the length
regulation of hepatic gluconeogenesis and insulin sensitivity. PLoS Genet of the gut and regulation of gut microbial ecology through MyD88. Gut
2014;10:e1004331. 2012;61:1124–1131.
[66] Guan D, Xiong Y, Trinh TM, Xiao Y, Hu W, Jiang C, et al. The hepatocyte [76] Dantas Machado AC, Brown SD, Lingaraju A, Sivaganesh V, Martino C,
clock and feeding control chronophysiology of multiple liver cell types. Chaix A, et al. Diet and feeding pattern modulate diurnal dynamics of the ileal
Science 2020;369:1388–1394. microbiome and transcriptome. Cel Rep 2022;40:111008.
[67] Yamamoto T, Shimano H, Nakagawa Y, Ide T, Yahagi N, Matsuzaka T, et al. [77] Frazier K, Chang EB. Intersection of the gut microbiome and circadian
SREBP1 interacts with hepatocyte nuclear factor-4 alpha and interferes with rhythms in metabolism. Trends Endocrinol Metab 2020;31:25–36.
PGC-1 recruitment to suppress hepatic gluconeogenic genes. J Biol Chem [78] Saran AR, Dave S, Zarrinpar A. Circadian rhythms in the pathogenesis and
2004;279:12027–12035. treatment of fatty liver disease. Gastroenterology 2020;158:1948–1966.

Journal of Hepatology, September 2023. vol. 79 j 741–757 757