Article

Targeting 14-3-3q-mediated TDP-43 pathology in

amyotrophic lateral sclerosis and frontotemporal
dementia mice
Graphical abstract Authors
Yazi D. Ke, Annika van Hummel,
Carol Au, ..., Ian P. Blair,
Fabien Delerue, Lars M. Ittner

Correspondence
[email protected] (Y.D.K.),
[email protected] (L.M.I.)

In brief
TDP-43 forms neuronal deposits in
amyotrophic lateral sclerosis (ALS) and
frontotemporal dementia (FTD) due to
unknown causes. Ke et al. identified the
TDP-43 interactor 14-3-3q, which
promotes TDP-43 deposition. They
devised a treatment harnessing 14-3-3q’s
affinity for TDP-43, mitigating deficits and
neurodegeneration in ALS/FTD models
with implications for future therapy.

Highlights
d 14-3-3q interacts preferentially with pathogenic TDP-43

d Increased neuronal levels of the cytoplasmic 14-3-3q

promote TDP-43 pathology

d Degron-mediated reduction of pathogenic TDP-43 improved

ALS/FTD models

Ke et al., 2024, Neuron 112, 1249–1264

April 17, 2024 ª 2024 Elsevier Inc.
https://doi.org/10.1016/j.neuron.2024.01.022 ll
 ll

Article
Targeting 14-3-3q-mediated TDP-43
pathology in amyotrophic lateral sclerosis
and frontotemporal dementia mice
Yazi D. Ke,1,6,* Annika van Hummel,1,6 Carol Au,1 Gabriella Chan,1 Wei Siang Lee,1 Julia van der Hoven,1
Magdalena Przybyla,1 Yuanyuan Deng,1 Miheer Sabale,1 Nicolle Morey,1 Josefine Bertz,1 Astrid Feiten,1 Stefania Ippati,1
Claire H. Stevens,2 Shu Yang,3 Amadeus Gladbach,1 Nikolas K. Haass,4 Jillian J. Kril,1,5 Ian P. Blair,3 Fabien Delerue,1
and Lars M. Ittner1,7,*
1Dementia Research Centre and Department of Biomedical Sciences, Faculty of Medicine and Health Sciences, Macquarie University,

Sydney, NSW 2109, Australia

2School of Chemistry and Molecular Bioscience, University of Wollongong and Illawarra Health and Medical Research Institute, Wollongong,

NSW 2522, Australia

3Centre for MND Research, Faculty of Medicine and Health Sciences, Macquarie University, Sydney, NSW 2109, Australia
4The University of Queensland Diamantina Institute, Translational Research Institute, The University of Queensland, Brisbane, QLD 4102,

Australia
5School of Medical Sciences, Faculty of Medicine and Health, University of Sydney, Sydney, NSW 2050, Australia
6These authors contributed equally
7Lead contact

*Correspondence: [email protected] (Y.D.K.), [email protected] (L.M.I.)

https://doi.org/10.1016/j.neuron.2024.01.022

SUMMARY

Amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) are characterized by cytoplasmic
deposition of the nuclear TAR-binding protein 43 (TDP-43). Although cytoplasmic re-localization of TDP-43
is a key event in the pathogenesis of ALS/FTD, the underlying mechanisms remain unknown. Here, we iden-
tified a non-canonical interaction between 14-3-3q and TDP-43, which regulates nuclear-cytoplasmic shut-
tling. Neuronal 14-3-3q levels were increased in sporadic ALS and FTD with TDP-43 pathology. Pathogenic
TDP-43 showed increased interaction with 14-3-3q, resulting in cytoplasmic accumulation, insolubility, phos-
phorylation, and fragmentation of TDP-43, resembling pathological changes in disease. Harnessing this
increased affinity of 14-3-3q for pathogenic TDP-43, we devised a gene therapy vector targeting TDP-43 pa-
thology, which mitigated functional deficits and neurodegeneration in different ALS/FTD mouse models ex-
pressing mutant or non-mutant TDP-43, including when already symptomatic at the time of treatment. Our
study identified 14-3-3q as a mediator of cytoplasmic TDP-43 localization with implications for ALS/FTD
pathogenesis and therapy.

INTRODUCTION functional RNA/DNA binding protein encoded by the TARDBP

gene.7 It harbors an N-terminal region that mediates functional
Amyotrophic lateral sclerosis (ALS) and frontotemporal dementia dimerization and oligomerization of TDP-43,8–10 followed by 2
(FTD) are rapidly progressive and fatal neurodegenerative dis- RNA recognition motifs (RRM1/2) and a large C-terminal
eases with significant clinical, genetic, and pathological over- glycine-rich domain (GRD) that mediates protein-protein interac-
lap.1 Both diseases are neuropathologically characterized by tions.11,12 The G-rich domain contains the vast majority of path-
deposition of TDP-43 in neurons.2 A prevailing theory is that ogenic TARDBP mutations in familial ALS13–15 and is understood
mis-localization of nuclear TDP-43 to the cytoplasm triggers to mediate the interaction of TDP-43 with partner proteins.16
toxic events including aberrant phosphorylation and fragmenta- However, little is known about the functional role of GRD-medi-
tion of TDP-43 (=gain of toxic function).3 Furthermore, lost phys- ated TDP-43 interactions in physiology and disease.
iological nuclear functions of TDP-43 contribute to deficits (=loss The 14-3-3 protein family comprises 7 genes and isoforms
of function).4,5 The molecular mechanisms that regulate physio- (YWHAB/14-3-3b, YWHAG/14-3-3g, YWHAE/14-3-3ε, YWHAH/
logical nucleo-cytoplasmic shuttling of TDP-43 during mRNA 14-3-3h, YWHAS/14-3-3s, YWHAZ/14-3-3z, and YWHAQ/14-3-
processing and drive its cytoplasmic accumulation and second- 3q/t) with both redundant and non-redundant functions.17 14-3-
ary modifications in disease remain unknown.6 TDP-43 is a multi- 3s form functional dimers.18 The helical structure of 14-3-3

Neuron 112, 1249–1264, April 17, 2024 ª 2024 Elsevier Inc. 1249
 ll
Article

monomers form a conserved concave amphipathic groove as tween 14-3-3q and TDP-43 with augmented complex formation
canonical binding site for conserved phosphorylated serine/ driving cytoplasmic localization of TDP-43 variants, including
threonine consensus binding motifs of hundreds of part- pathogenic and dysfunctional mutations.
ners.19–21 14-3-3s are ubiquitously expressed cytoplasmic pro- Interaction/co-localization of TDP-43 with 2 other 14-3-3 iso-
teins that change the conformation, stability, localization, and forms, 14-3-3s and 14-3-3h, had been previously reported.23,24
functions of interaction partners. 14-3-3 interactions have been We therefore tested all 14-3-3 isoforms for potential interaction
linked to a wide range of physiological processes and disease with TDP-43 by coIP. TDP-43 interacted with 14-3-3h, 14-3-
mechanisms, including in cancer, metabolic, and neurodegener- 3g, and 14-3-3s stronger than with 14-3-3q but showed no overt
ative diseases.17 14-3-3 proteins are abundant in the brain, and interaction with 14-3-3ε, 14-3-3z and 14-3-3b (Figure S4A). How-
different isoforms govern a variety of processes in CNS develop- ever, 14-3-3s and 14-3-3z are not abundant in neurons (see
ment and function.22 TDP-43 has been reported to interact with BioGPS and Allan Brain Atlas online tools). We note that recom-
14-3-3s under cell stress conditions, competing with a 14-3-3s/ binant 14-3-3b has been reported to interact with TDP-43
forkhead box O (FOXO) protein complex.23 Furthermore, 14-3- in vitro.28 More importantly, only 14-3-3q showed significantly
3h co-localized with TDP-43 in sporadic ALS spinal cord neu- stronger interaction with TDP-43-A315T and, in particular,
rons.24 14-3-3q mRNA was previously found to be upregulated TDP-43-L248A/I249A/I250A compared with TDP-43 wild type
in ALS,25 whereas 14-3-3z expression was found to be reduced (WT), whereas other interacting isoforms showed no enhanced
in ALS motor cortex.26 Despite these observations, a role of 14- interaction (e.g., 14-3-3s interacted less with TDP-43-L248A/
3-3 proteins in ALS and, possibly, TDP-43 pathology remains to I249A/I250A) (Figures S4B and S4C). Hence, only TDP-43’s
be shown. interaction with 14-3-3q significantly changed with pathogenic
Here, we used an unbiased interaction screen to study the in- and dysfunctional variants of TDP-43.
teractome of the TDP-43 GRD. We identified and validated a
novel TDP-43 interaction with a 14-3-3 family protein, 14-3-3q,
and demonstrate novel physiological and pathological roles of Non-canonical 14-3-3q/TDP-43 interaction is regulated
this complex as well as the therapeutic potential of using this by RNA
interaction as a targeted gene therapy for FTD and ALS. Interest- 14-3-3 dimers typically interact with phosphorylated forms of
ingly, we found a novel, non-canonical site on the surface of 14- their interaction partner.29 By contrast, we found that phosphor-
3-3q that is required for its interaction with TDP-43 and estab- ylation-mimicking variants of TDP-43 interact less with 14-3-3q
lished TDP-43-bound RNA as a regulator of this interaction. (Figures 2A and S2D), supporting a non-canonical-type interac-
tion. Interestingly, we found a significant increase of cell-free
RESULTS in vitro reconstituted 14-3-3q/TDP-43 complexes when samples
were treated with RNase but not DNase (Figure 2B), suggesting
The novel TDP-43 interactor 14-3-3q binds RNA bound to TDP-43 limits this interaction. Accordingly, the
preferentially pathogenic and dysfunctional variants of L248A/I249A/I250A variant of the RRM2 domain of TDP-43 virtu-
TDP-43 ally completely prevented RNA binding compared with WT TDP-
To identify novel interaction partners of TDP-43’s GRD, we per- 43 as shown via RNA immunoprecipitation (Figure S3A), which is
formed a bacterial two-hybrid screen (Figures 1A and S1A–S1C; consistent with a markedly increased TDP-43-L248A/I249A/
Table S1). The top candidate identified (11 of 65 hits) was 14-3- I250A/14-3-3q complex formation as compared with TDP-43-
3q, a member of the 14-3-3 protein family.17 Co-immunoprecip- WT/14-3-3q (Figures 1C and 1D). By contrast, TDP-43 variants
itation (coIP) from non-transfected murine N2a cells and mouse harboring pathogenic mutations enriched similar amounts of
brains confirmed interaction between endogenous 14-3-3q and RNA (Figure S3A), rather suggesting an increased affinity for
TDP-43 (Figures S2A and S2B). To test whether a 14-3-3q/ 14-3-3q due to conformational changes. Structurally, 14-3-3q
TDP-43 interaction is disease relevant, we expressed 14-3-3q harbors 9 a-helices (Figure 2C),18 with helices aC, aE, aG, and
together with TDP-43 variants in 293T HEK cells. Surprisingly, aI contributing to canonical partner binding of 14-3-3q di-
14-3-3q interacted significantly more with TDP-43 variants mers.30,31 To identify the interaction motif(s) in 14-3-3q that
harboring pathogenic mutations, including A315T (Figures 1B mediate binding of TDP-43, we truncated 14-3-3q stepwise,
and S2C). Interestingly, 14-3-3q showed even stronger interac- revealing that the interaction is mediated by the sixth a-helix of
tion with a RRM2 variant of TDP-43, TDP-43-L248A/I249A/ 14-3-3q (Figure 2C). Accordingly, expressing a-helix 6 of 14-3-
I250A, that is unable to bind mRNA (Figures 1C, S2D, and 3q (=amino acids 135–164; ‘‘Fx’’) alone co-precipitated TDP-43
S3A). Co-expression of 14-3-3q with TDP-43-A315T resulted in (Figure 2D), but failed to pull down known 14-3-3q interaction
marked cytoplasmic co-localization (Figures 1D and S3B). partners YES-associated protein (YAP)32 and FOXO133
Furthermore, TDP-43-L248A/I249A/I250A, and another RRM2 (Figures S5A and S5B), further supporting non-canonical 14-3-
variant (TDP-43-I239A/L243A), which are known to localize to 3q/TDP-43 interaction. The a-helix 6 harbors a 10 amino acid
the nucleus and form large nuclear aggregates when expressed motif that is unique to 14-3-3q over other 14-3-3 isoforms (Fig-
long-term in cells,27 was found almost exclusively in the cyto- ure 2E) that is presented on opposing surfaces of 14-3-3q dimers
plasm when co-transfected with 14-3-3q (Figures 1D and S3B). and different from the canonical interaction sites in the center of
At the same time, non-mutant TDP-43 showed cytoplasmic the molecule (Figure 2F), explaining its unconventional interac-
localization together with 14-3-3q in a smaller but significant sub- tion with TDP-43. Taken together, we identified a non-canonical
set of cells. Taken together, we identified a novel interaction be- specific interaction of 14-3-3q with TDP-43 that is mediated by a

1250 Neuron 112, 1249–1264, April 17, 2024

 ll
Article

A B

C D

Figure 1. 14-3-3q preferentially interacts with mutant TDP-43 and drives its cytoplasmic accumulation
(A) Schematic of TDP-43 including 2 RNA/DNA-binding motifs (RRM1/2), nuclear localization sequence (NLS), and a glycine (G)-rich region. Bacterial-2-Hybrid
performed with TDP-43’s G-rich region linked to lcl (lcl-DTDP). lcl-DTDP/RNAPa-14-3-3q interaction was evident by colonies on selective screening medium
(SSM), while non-SSM (NSSM) colonies confirmed co-transformation. Negative controls: lcl-DTDP/RNAPa and lcl/RNAPa-14-3-3q; positive control: lcl-pos/
RNAPa-pos.
(B) Co-expression in 293T cells shows significantly enhanced co-immunoprecipitation (coIP) of myc-tagged 14-3-3q with V5-tagged TDP-43 variants carrying
pathogenic mutations compared with its non-mutant form.
(C) Co-expression in 293T cells shows markedly enhanced coIP of 14-3-3q-myc with TDP-43-L248A/I249A/I250A-V5 compared with its non-mutant form.
(D) Co-expression of 14-3-3q confers cytoplasmic localization of non-mutant TDP-43, TDP-43-A315T, and more so TDP-43-L248A/I249A/I250A, which all
normally localize to the nucleus (+MOCK) in 293T cells. Scale bar, 50 mm.
For (B) to (D), ****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05. Data points obtained from independent experiments. Error bars indicate SEM.

unique and novel motif on the surface of the 14-3-3q molecule sequence modification. Importantly, immunoprecipitation of
and negatively regulated by the mRNA bound to TDP-43. TDP-43 from 14-3-3qzFx/zFx brain extracts showed the modifica-
tion disrupting endogenous 14-3-3q/TDP-43 complex formation
Binding to 14-3-3q is critical for nucleo-cytoplasmic in vivo (Figure 3C) in the absence of any overt phenotype in
shuttling of TDP-43 under stress young 14-3-3qzFx/zFx mice. We have previously shown that
To show that amino acids 135–164 of 14-3-3q harbor the TDP-43 stressing murine neurons with the proteasome inhibitor MG-
binding site in vivo and to study this interaction in a physiological 132 induced TDP-43 pathology.34 Therefore, we challenged pri-
environment with endogenous proteins, we abolished the 14-3- mary neurons from the 14-3-3qzFx/zFx line with abolished 14-3-
3q/TDP-43 interaction by replacing the 31aa-encoding 3q/TDP-43 interaction, and the respective wild-type controls,
sequence of Fx in Ywhaq with the corresponding sequence of with MG-132 (Figure 3D). Although MG-132-treated wild-type
Ywhaz that encodes 14-3-3z (Figure 3A), given that 14-3-3z neurons showed slow and incomplete re-distribution of TDP-
does not interact with TDP-43 (Figure S4). Expression levels of 43 from the nucleus to the cytoplasm, 14-3-3qzFx/zFx neurons
the altered protein 14-3-3qzFx were comparable to 14-3-3q in presented with rapid re-distribution of TDP-43 to the cytoplasm
wild-type samples when probed with an amino-terminal-specific with virtually complete clearance of nuclear TDP-43 in those
antibody to 14-3-3q (Figure 3B). Antibodies specific to the Fx cells. This was in the absence of increased cytotoxicity in 14-
sequence of 14-3-3q no longer detect 14-3-3qzFx in 14-3- 3-3qzFx/zFx neurons. Together, these data support a critical role
3qzFx/zFx mouse brain extracts, confirming the successful of 14-3-3q in nucleo-cytoplasmic shuttling of TDP-43 under

Neuron 112, 1249–1264, April 17, 2024 1251

 ll
Article

A B

Figure 2. Non-canonical, RNA-dependent interaction between 14-3-3q and TDP-43

(A) Co-expression of V5-tagged phosphorylation mimicking S/D TDP-43 variants in 293T cells disrupted coIP with myc-tagged 14-3-3q, while phosphorylation
preventing S409A/S410A and non-mutant TDP-43-V5 variants IP-ed with 14-3-3q-myc.
(B) Western blots of in vitro-reconstituted 14-3-3q/TDP-43 complexes that were markedly enhanced upon RNase but not DNase treatment.
(C) Western blots of coIP from 293T cell extracts transfected with N- and C-terminal truncation variants of 14-3-3q and TDP-43 showed complex formation except
when a-helix 6 (DF) of 14-3-3q was removed. Note the enhanced 14-3-3q/TDP-43 coIP in the absence of a-helices 7–9 (DG).
(D) Co-expression of 14-3-3q a-helix 6 alone (14-3-3q-Fx-V5) coIP-ed with TDP-43 similar to a GD variant of 14-3-3q from 293T cell extracts.
(E) Alignment of a-helix 6 of 14-3-3 isoforms corresponding to 14-3-3q-Fx. Red box, 10 aa sequence unique to 14-3-3q.
(F) Different views of 14-3-3q dimer 3D structure highlighting the unique 10 aa motif (see E) in red presented at the surface (arrows). Individual 14-3-3q monomers
are depicted in white and gray with canonical binding sites in the center (asterisks).
For (A) and (B), ****p < 0.0001; ***p < 0.001; **p < 0.01. Data points obtained from independent experiments. Error bars indicate SEM.

stress conditions and suggest a fine balance of available 14-3-3q ALS inclusions or the identification of mutations in disease.2,13–15
binding sites for TDP-43 to carry out its physiological function. Therefore, we stained brain and spinal cord sections from
different neurodegenerative diseases for 14-3-3q (Table S2).
Increased 14-3-3q/TDP-43 complexes in human FTD We found intensive neuronal 14-3-3q staining of hippocampal
and ALS neurons in frontotemporal lobar degeneration with TDP-43
Increased 14-3-3q mRNA levels have been previously reported pathology (FTLD-TDP). This was found together with cyto-
in ALS samples,25 preceding the discovery of TDP-43 in FTD/ plasmic and frequent nuclear depletion of TDP-43, without overt

1252 Neuron 112, 1249–1264, April 17, 2024

 ll
Article

A B C

Figure 3. 14-3-3q/TDP-43 interaction during nucleo-cytoplasmic shuttling

(A) Schematic illustrating the CRISPR modification strategy to exchange the 31aa sequence, qFx, in the murine 14-3-3q-encoding Ywhaq locus, spanning over 30
of exon 3 and 50 of exon 4 (red) with the corresponding sequence from 14-3-3z (green), resulting in 14-3-3qzFx/zFx mice.
(B) Western blots of brain extracts from wild-type C57BL/6, 14-3-3qzFx/zFx and iTDP-43A315T mice with a 14-3-3q antibody specific to its N terminus confirmed
indistinguishable levels of 14-3-3q in all samples, while an antibody specific to the Fx sequence (qFx-spec) did detect 14-3-3q in C57BL/6 and iTDP-43A315T but
not 14-3-3qzFx/zFx samples, confirming protein modification. Levels of TDP-43 were comparable in C57BL/6 and 14-3-3qzFx/zFx mice and increased in iTDP-
43A315T. Gapdh confirmed equal loading.
(C) IP with pan-specific TDP-43 antibodies co-precipitated endogenous 14-3-3q from C57BL/6 and iTDP-43A315T but not 14-3-3qzFx/zFx brains (arrowhead),
confirming disruption of the endogenous 14-3-3q/TDP-43 interaction in 14-3-3qzFx/zFx mice. *, IgG heavy- and light-chain bands are indicated.
(D) MG132 induces dose- and time-dependent cytoplasmic localization of TDP-43 in C57BL/6 and more so in 14-3-3qzFx/zFx neurons over 12 and 24 h inde-
pendent of cytotoxicity.
For (D), ****p < 0.0001; ***p < 0.001; **p < 0.01. Data points obtained from independent experiments. Error bars indicate SEM.

TDP-43 aggregates, but infrequent staining of cytoplasmic mutase type 1 (SOD1) (TDP-43 negative) showed control-like
phosphorylated TDP-43 (Figures 4A–4C and S6A). Using prox- 14-3-3q staining (Figures 4F and S6C). This indicates that 14-
imity ligation assay (PLA) together with antibodies to TDP-43 3-3q may contribute to TDP-43 pathology in FTLD-TDP and
and 14-3-3q, we found intensive cytoplasmic labeling of neurons SALS, while TDP-43 pathology in familial ALS with C9ORF72 ex-
in FTLD-TDP brains in contrast to only occasional signals in pansions is likely due to alternative mechanisms, such as RNA
neurologically normal controls (Figure 4D), consistent with foci and repeat peptide translation.35–38 However, it cannot be
increased cytoplasmic TDP-43/14-3-3q complexes in FTLD- fully excluded that TDP-43 interaction with 14-3-3q expressed
TDP. Brains from FTLD with tau pathology (FTLD-TAU), FTLD at endogenous levels also contributes to TDP-43 mis-localiza-
with ubiquitinated deposits (FTLD-U), and Alzheimer’s disease tion in familial ALS with C9ORF72 expansions. Conversely,
(AD) showed comparable 14-3-3q staining to sections from increased TDP-43 levels and pathology in iTDP-43A315T mice
neurologically normal controls (Figure S6B). Motor neurons in did not result in increased 14-3-3q mRNA expression (or of other
spinal cord sections from sporadic ALS (SALS) with TDP-43 pa- 14-3-3 isoforms) (Figure S6D), suggesting that increased 14-3-
thology also stained intensively for 14-3-3q when compared with 3q levels are not a result of aberrant TDP-43. iTDP-43A315T
control tissue (Figure 4E). By contrast, familial ALS with muta- mice express human (h) A315T mutant TDP-43 in CNS neurons
tions in C9ORF72 (with TDP-43 pathology) or superoxide dis- and develop FTD/ALS pathology with cytoplasmic localization,

Neuron 112, 1249–1264, April 17, 2024 1253

 ll
Article

A B C Figure 4. Increased neuronal 14-3-3q levels

in human FTLD-TDP and SALS
(A) Enhanced neuronal 14-3-3q staining (closed
arrow) and cytoplasmic TDP-43 or phosphorylated
TDP-43 (pTDP-43) (arrowheads) with frequent nu-
clear depletion (open arrowheads) in human fron-
totemporal lobar degeneration with TDP-43 pa-
thology (FTLD-TDP) compared with control (CTR;
open arrow) in CA4 of the hippocampus (n = 22
[FTLD-TDP] and n = 28 [CTR]). Insets: TDP-43/
D pTDP-43 inclusion (asterisk) and nuclear deple-
tion (open arrowhead) in the dentate gyrus in FTLD-
TDP.
(B) Quantification of optical staining density (OD) of
cytoplasmic 14-3-3q labeling relative to back-
ground in hippocampus CA4 of FTLD-TDP vs.
CTR.
(C) Number of cases with prominent generalized
neuronal 14-3-3q staining of all cases examined
with the indicated underlying hippocampal neuro-
pathology. (1) Indicate cases with single 14-3-3q
positive neurons; FTLD-TAU, FTLD with tau pa-
E thology; FTLD-U, FTLD with uncharacterized
ubiquitinated inclusions; AD, Alzheimer’s disease;
F CTR, controls without neuropathology.
(D) Intensive cytoplasmic labeling in FTLD-TDP
cases as compared with control (CTR) using
proximity ligation assay for direct detection of
TDP-43/14-3-3q interactions in situ.
(E) Increased neuronal 14-3-3q staining in human
sporadic amyotrophic lateral sclerosis (SALS;
closed arrow) compared with control (CTR;
open arrow) in motor neurons of the spinal cords (n = 17 [SALS] and n = 7 [CTR]). Insets: motor neurons with TDP-43 inclusions in SALS and nuclear TDP-43
in CTR.
(F) Number of cases with prominent 14-3-3q staining of all cases examined with the indicated underlying spinal cord neuropathology. FALS-C9ORF72, familial
ALS with C9ORF72 mutations; FALS-SOD1, FALS with SOD1 mutations; CTR, controls without neuropathology. yTwo cases had no motor neurons left to
assess.
For (B), (C), and (E), ****p < 0.0001; **p < 0.01. Error bars indicate SEM. Scale bars, 50 mm.

fragmentation, phosphorylation, ubiquitination, and insolubility TDP-43 insolubility compared with controls when challenged
of TDP-43 with associated neuron loss and functional deficits.39 with MG-132 (Figure 5B), in line with physiological 14-3-3q levels
stabilizing soluble TDP-43 during nucleo-cytoplasmic shuttling
Increased 14-3-3q levels drive TDP-43 pathology in cells processes.
and mice To study the effects of increased 14-3-3q levels in vivo, we
To understand how increased 14-3-3q expression affects TDP- next used adeno-associated viruses (AAVs) to express 14-3-3q
43, we overexpressed 14-3-3q in different models. Human SH- in the hippocampus of 3-week-old non-transgenic and iTDP-
SY5Y neuroblastoma cells express endogenous 14-3-3q and 43A315T mice. Increasing neuronal 14-3-3q levels resulted in
TDP-43. Stably overexpressing 14-3-3q in SH-SY5Y cells re- accumulation of insoluble TDP-43 fragments in non-transgenic
sulted in increased levels of insoluble endogenous TDP-43 as and markedly more in iTDP-43A315T mice (Figures 5C and
compared with control and 14-3-3q knockdown cells (Figure 5A). S7B). As a result, total amounts of insoluble TDP-43, comprising
Stressing 14-3-3q-overexpressing SH-SY5Y cells with MG-132 full-length and fragmented TDP-43 in the urea fraction, were
further increased insolubility of TDP-43 and resulted in the for- increased in iTDP-43A315T mice with neuronal expression of
mation of a 25 kDa fragment that was also significantly phos- 14-3-3q. Furthermore, AAV-14-3-3q-injected iTDP-43A315T
phorylated (Figure 5B). Interestingly, 14-3-3q-overexpressing mice showed substantial loss of hTDP-43-expressing hippo-
human SH-SY5Y cells also displayed abnormal Stathmin 2 campal neurons compared with controls (Figures 5D and S7C).
(STMN2) splicing (Figure S7A), similar to aberrant STMN2 By contrast, AAV-mediated 14-3-3q expression had limited toxic
splicing associated with TDP-43 pathology specifically in hu- effect in non-transgenic controls, with non-significant trends
mans (the murine Stmn2 locus lack the required alternative toward reduced hippocampal neuron numbers (Figures 5D
exon structure).40–42 Consistent with our previous findings in mu- and S7C). Taken together, increasing 14-3-3q levels in cells
rine neurons,34 MG-132 treatment of controls and 14-3-3q and mice caused disease-like neuronal loss, insolubility, and
knockdown cells resulted in comparably less TDP-43 insolubility fragmentation of endogenous and transgenic TDP-43 as well
in the absence of overt fragmentation. Interestingly, however, as exacerbating neuropathological phenotypes in iTDP-
14-3-3q knockdown cells showed a moderately augmented 43A315T mice.

1254 Neuron 112, 1249–1264, April 17, 2024

 ll
Article

A D

Figure 5. Increasing 14-3-3q levels caused disease-specific changes in TDP-43

(A) Presence of insoluble (‘‘U’’) endogenous TDP-43 in stable 14-3-3q-V5 overexpressing SH-SY5Y cells. TDP-43 was soluble (‘‘R’’) in cells with stable 14-3-3q
knockdown (14-3-3q-shR) and MOCK controls (ctr). Asterisk, V5-tagged 14-3-3q.
(B) Challenging SH-SY5Y cells with the proteasome inhibitor MG-132 enhanced insoluble TDP-43 in 14-3-3q-V5-expressing cells and caused TDP-43 frag-
mentation and phosphorylation (pTDP-43 (409/410); arrowheads). Asterisk, V5-tagged 14-3-3q.

(legend continued on next page)

Neuron 112, 1249–1264, April 17, 2024 1255

 ll
Article

Considering 14-3-3q expression promoted mis-localization of DD-induced degradation in neurons (Figure S8A). Co-expression
wild-type TDP-43 in cells (Figure 1D), we tested whether long- of DD-qFx with A315T-mutant hTDP-43 in neurons caused sig-
term AAV-mediated overexpression of 14-3-3q in spinal cords nificant reduction of hTDP-43 levels, consistent with induced
of naive C57BL/6 mice resulted in altered endogenous TDP-43 degradation (Figure S8B). Furthermore, AAV-mediated exp-
and functional deficits. Histopathological analysis of spinal cords ression of DD-qFx in brains of iTDP-43A315T mice from birth
after 10 months of 14-3-3q overexpression revealed significant (Figure 6A) resulted in markedly, 74.9% ± 5.2%-reduced and
cytoplasmic accumulation of TDP-43 in 14-3-3q overexpressing mutually exclusive cellular expression of transgenic hTDP-43
anterior horn motor neurons, while non-expressing or GFP con- to DD-qFx (Figures 6B and S9A) and significantly reduced
trol cells presented with virtually exclusively nuclear TDP-43 TDP-43 levels in brain extracts (Figure 6C). Residual hTDP-43
(Figures 5E and S7D). 8 months after AAV-14-3-3q injections, fe- staining in DD-qFx-expressing cells localized to their cytoplasm,
males showed normal motor performance and strength that possibly reflecting ongoing cytoplasmic hTDP-43/DD-qFx com-
deteriorated by 12 months of age, with significant motor deficits plex formation prior to degradation while transgenic hTDP-43
and reduced grip strength (Figure 5F). GFP expression had no continues (Figure 6B). Together, this suggests clearance of
comparable effects. For comparison, male mice showed a trend transgenic hTDP-43 mediated by DD-qFx degradation. Turnover
to reduced motor performance and muscle strength 8 months of DD-qFx was higher in ctr than iTDP-43A315T mice, likely due to
after AAV-14-3-3q that persisted over time. Female mice perform continued transgenic hTDP-43 expression and/or presence of
better during wire testing, likely due to their lower body weight. TDP-43 aggregates in iTDP-43A315T mice. Functional assess-
Finally, to determine the effect of increased neuronal 14-3-3q ment of DD-qFx-expressing iTDP-43A315T mice showed less
levels on endogenous pathogenic TDP-43, we used systemically disinhibition, less hyperactivity, reduced motor deficits, and
delivered neurotropic AAV.PHP to express 14-3-3q in the brains increased muscle strength as compared with control vector-in-
of CRISPR-modified mice carrying heterozygous N390D Tardbp jected iTDP-43A315T mice (Figures 6D–6G and S9B). There
mutation (Figure 5G). Heterozygous introduction of the N390D were no sex differences. Hence, DD-qFx induces clearance of
mutation in Tardbp has been shown to result in ALS-like pheno- pathogenic TDP-43 in neurons and prevented deficits in iTDP-
types in mice at advanced ages with motor neuron loss in the spi- 43A315T mice.
nal cord observed from 6 months of age.43 Although 14-3-3q Next, we tested the same approach in an additional mouse
expression did not alter numbers of TDP-43-positive neurons model, based on AAV-mediated expression of non-mutant
in the hippocampus of wild-type mice, we found significantly TDP-43 in CNS neurons (Figure 6H). We used the neurotropic
fewer 14-3-3q neurons in the hippocampus of N390D mice AAV9 serotype, AAV.PHP.B, to achieve uniform expression in
together with a loss of TDP-43-positive neurons. Taken together, CNS neurons44 by temporal vein injection in naive newborn
chronically increased 14-3-3q levels compromised endogenous C57BL/6 mice.45 This resulted in widespread neuronal expres-
TDP-43 localization in motor neurons with late-onset motor def- sion of human TDP-43 when co-delivered with AAV-ctr and
icits in wild-type mice and promoted neuronal loss in mice with much lower hTDP-43 expression when co-delivered with AAV-
an endogenous N390D mutant TDP-43. DD-qFx at birth (Figure 6I). Until 10 weeks of age, AAV-hTDP-43
and AAV-ctr showed comparable performance during fortnightly
14-3-3q interaction domain-mediated degradation of inverted wire testing (Figure 6J), suggesting comparable strength.
TDP-43 prevents deficits in FTD/ALS mice Thereafter, AAV-hTDP-43 progressively declined, but this was
The unique mode of interaction together with the preference for fully prevented when mice were co-treated with AAV-DD-qFx at
aberrant forms of TDP-43 prompted us to explore whether 14-3- birth. This was corroborated by direct assessment of grip
3q could be used to target pathological TDP-43 therapeutically. strength, which was significantly reduced in male AAV-hTDP-43
Therefore, we designed a degron construct comprising 14-3- mice as compared with AAV-ctr and prevented by DD-qFx and
3q-Fx (only the sixth a-helix of 14-3-3q) fused to a C-terminal showed a comparable trend in female mice (Figure 6K). Accord-
FK506 binding protein (FKBP)-L106P degradation domain (DD) ingly, muscle size and weights of 19-week-old female and male
and a N-terminal V5 tag for detection (hereby termed ‘‘DD- AAV-hTDP-43 were significantly reduced as compared with the
qFx’’). DD-qFx accumulated in primary neurons only in the pres- respective controls (Figure 6L). Hence, DD-qFx prevented deficits
ence of the stabilizing compound Shield1 confirming efficient induced by non-mutant TDP-43 expression in mice.

(C) AAV-mediated expression of 14-3-3q-V5 in the hippocampus resulted in insolubility and fragmentation (arrowheads/chevron) of TDP-43 in ctr and iTDP-
43A315T mice, respectively. Quantification of TDP-43 fragment levels from independent experiments. c-TDP-43, antibody to the C terminus of TDP-43.
(D) Staining brains for neuronal NeuN (green) and transgenic TDP-43 (blue) in 14-3-3q-V5 expressing 4-week-old iTDP-43A315T hippocampus showed loss of
hippocampal neurons (open arrowhead). Quantification of hTDP-43-expressing neurons in the CA1 of AAV-vec and AAV-14-3-3q-V5 injected iTDP-43A315T mice
and of NeuN-positive neurons in AAV-vec and AAV-14-3-3q-V5 injected control and iTDP-43A315T mice.
(E) Long-term AAV-mediated 14-3-3q-V5 (green; bottom right) expression in spinal cord motor neurons of naive C57BL/6 mice resulted in increased cytoplasmic
and reduced nuclear TDP-43 staining (open arrowheads), as compared with nuclear TDP-43 in neighboring non-transduced cells (arrowheads) or GFP-ex-
pressing AAV-vec controls. Quantification shows significantly increased percentage of 14-3-3q-expressing neurons with cytoplasmic TDP-43.
(F) Reduced motor performance during wire testing and decreased forearm peak grip strength after long-term expression of 14-3-3q-V5 compared with vector
controls (vec). Significant changes were observed in female cohorts.
(G) AAV-mediated expression of 14-3-3q-V5 in control and TardbpN390D mice, with significantly less 14-3-3q-V5 expression in the hippocampus of TardbpN390D
than control brains, while numbers of TDP-43-positive neurons remain comparable. Scale bar, 200 mm.
For (C), (D), (F), and (G), ***p < 0.001; **p < 0.01; *p < 0.05. Data points obtained from independent experiments. Error bars indicate SEM.

1256 Neuron 112, 1249–1264, April 17, 2024

 ll
Article

A B C

D E F G

H I J

K L

Figure 6. 14-3-3q interaction-mediated targeting of pathogenic TDP-43 prevented deficits in TDP-43 mouse models
(A) Intracerebral DD-qFx or control vector (vec) AAV delivery in newborn mice that were analyzed 3 months later.
(B) Predominantly nuclear hTDP-43 in the cortex of vec-injected iTDP-43A315T mice was significantly reduced in DD-qFx- compared with vec-expressing neurons.
Quantification of 100 neurons per section confirms significant reduction of hTDP-43 levels in DD-qFx-expressing cells.
(C) Western blots of brain extracts show reduced TDP-43 levels in iTDP-43A315T brains expressing DD-qFx from birth. mCherry and V5 confirmed AAV-mediated
expression. Note that DD-qFx expressed higher in iTDP-43A315T than ctr mice.
(D) Disinhibition (as reflected by increased open arm time in the elevated plus maze) of vec-treated iTDP-43A315T mice was significantly reduced upon DD-qFx
expression.
(E) Increased activity (as reflected by higher distance traveled in the open field) of vec-treated iTDP-43A315T mice was significantly reduced upon DD-qFx
expression (n = 8).
(F) Reduced motor performance (as reflected by short time to fall off rotating rod) of vec-treated iTDP-43A315T mice was comparable to ctr mice upon DD-qFx
expression.
(G) Improved grip strength of iTDP-43A315T mice expressing DD-qFx compared with vec-treated was observed.

(legend continued on next page)

Neuron 112, 1249–1264, April 17, 2024 1257

 ll
Article

rNLS mice express DNLS mutant human TDP-43 in neurons, proteasome marker proteasome subunit alpha type-1 (PSMA1)
resulting in rapid progressive neurodegeneration with progres- and found marked labeling of neurons with DD-qFx expression
sive limb paralysis and early death.46,47 We treated rNLS mice together with clearance of A315T mutant hTDP-43, while neu-
at birth with AAV-DD-qFx and AAV-ctr by i.v. delivery while sup- rons without DD-qFx expression did not show proteasome acti-
pressing DNLS-TDP-43 transgene expression with doxycycline vation and maintained transgenic hTDP-43 expression (Fig-
until weaning (days 21) to allow CNS maturation. After doxycy- ure 7E). This suggests neuron-specific proteasome activation
cline withdrawal, AAV-ctr rNLS developed first signs of motor by DD-qFx correlates with hTDP-43 clearance. At the same
impairment within 7–10 days (limb clasping and shaky gate), time, levels of mRNA encoding transgenic hTDP-43 or murine
rapidly progressing to paralysis (Figure S9C). By contrast, DD- TDP-43 remained unaffected by DD-qFx treatment, in line with
Fqx treatment significantly delayed onset of paralysis after post-transcriptional effects of the treatment (Figure S10A). Aber-
symptom onset (+63.9%) in this model and markedly improved rant hTDP-43 expression results in suppression of endogenous
survival (+38.5%) (Figures S9C and S9D), extending the protec- murine TDP-43 expression39,46 due to intrinsic control mecha-
tive efficacy of DD-qFx to a third independent model of ALS/FTD. nisms that have been linked to nuclear depletion of function
TDP-43 in neurons with cytoplasmic TDP-43 pathology in hu-
14-3-3q/TDP-43 interaction domain ameliorates deficits mans.2,27,48,49 This is associated with loss-of-physiologic-func-
in symptomatic FTD/ALS mice tions of TDP-43, such as compromised alternative mRNA
We have previously shown that short-term suppression of trans- splicing resulting in the presence of transcripts harboring cryptic
genic hTDP-43 expression ameliorates deficits of iTDP-43A315T exons in ALS/FTD50 and TDP-43-deficient mice.51 In our treated
mice.39 Here, we used AAV.PHP.B for systemic delivery of DD- mice, we found that concomitantly with proteasome activation
qFx (or controls) via tail vein injections for CNS neuron expres- and hTDP-43 removal, DD-qFx-expressing neurons recovered
sion in 3-month-old iTDP-43A315T mice with established deficits endogenous TDP-43 expression (Figures 7F and S10B), as well
(Figure 7A). Western blotting confirmed reduction of hTDP-43 in as a marked reduction of cryptic exon-containing Ap3b2
DD-qFx-expressing iTDP-43A315T mice compared with controls mRNA transcripts in the brains of AAV-DDqFx-treated iTDP-43
(Figure 7B). Notably, DD-qFx expression in control and iTDP- mice (Figure 7G), in line with their improved functional
43A315T mice was comparable without accumulation in the latter performance.
as seen after direct brain delivery, suggesting better therapeutic Taken together, neuronal DD-qFx expression decreased path-
dosing after systemic AAV delivery.44 We found comparable and ogenic TDP-43 levels and improved established functional defi-
reproducible DD-qFx and control vector expression patterns cits of iTDP-43A315T mice. To this end, both DD-qFx delivery at
throughout the CNS within 2 weeks, with the hippocampus of birth or in symptomatic TDP-43-expressing mice prevented
treated iTDP-43A315T mice shown here (Figure 7C). Importantly, and partially reversed functional deficits, respectively. Reduction
we found a 32.6±6.6% reduction in hTDP-43-expressing neu- of transgenic hTDP-43 coincided with neuron-specific protea-
rons in the absence of overt cell loss (Figure 7C). Furthermore, some activation and recovery of lost endogenous TDP-43
DD-qFx co-localized with hTDP-43 in remaining neurons, many expression and function. DD-qFx may have direct inhibitory ef-
of which showed only weak transgenic TDP-43 staining. Next, fects by binding pathogenic TDP-43 in addition to induced
we functionally assessed 3.5-month-old iTDP-43A315T mice degradation. Taken together, these data suggest that patholog-
(i.e., 2 weeks after DD-qFx AAV delivery). At this age, untreated ical TDP-43 can be targeted and cleared using specific interac-
iTDP-43A315T mice present with profound impairments.39 tion sequences, introducing a new avenue of treating ALS
Notably, DD-qFx expression improved disinhibition of iTDP- and FTD.
43A315T mice (Figure 7D), while hyperactivity and motor deficits
remained unchanged (Figures S9E–S9G). Previously, we DISCUSSION
observed comparable improvements of disinhibition, when
iTDP-43A315T mice were treated with doxycycline to completely In this study, we identified 14-3-3q as a novel interactor of TDP-
suppress transgene expression.39 Expression of control vectors 43. This cytoplasmic interaction is negatively regulated by RNA
had no effects on the deficits of iTDP-43A315T mice in these bound to TDP-43 and governed by a novel non-canonical site
tasks. DD-qFx left performance of control mice unaffected. To on the surface of 14-3-3q. Although the interaction of 14-3-3q
address the mechanism of therapeutic action, we stained brain with wild-type TDP-43 is weak and possibly transient in nature,
sections of iTDP-43 mice early after DD-qFx-treatment with the 14-3-3q interacts markedly stronger with both pathogenic and

(H) Intravenous (i.v.) AAV-hTDP-43 (AAV-hTDP) or control vector (vec) with DD-qFx or vec delivery via the temporal vein in newborn mice that were analyzed
3 months later.
(I) Widespread and strong neuronal expression of hTDP-43 in the cortex in vec co-delivered mice and markedly lower hTDP-43 expression in DD-qFx co-injected
mice. GFP indicates control AAV expression in neurons.
(J) Progressive decline in body strength (as reflected by reduce inverted wire times and corresponding linear regression slope differences) in vec-treated AAV-
hTDP-43 mice as comparable to AAV-DD-qFx-injected AAV-hTDP mice and controls (n = 10).
(K) Reduced grip strength in male vec-treated AAV-hTDP-43 mice as comparable to AAV-DD-qFx-injected AAV-hTDP mice.
(L) Atrophy of tibialis anterior (TA) muscle with reduced size and weight in female and male vec-treated AAV-hTDP-43 mice as comparable to AAV-DD-qFx-
injected AAV-hTDP mice.
For (B) to (G) and (J) to (L), ****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05; ns, not significant. Data points obtained from independent experiments. Error bars
indicate SEM.

1258 Neuron 112, 1249–1264, April 17, 2024

 ll
Article

A B

E F G

Figure 7. DD-qFx improved deficits in symptomatic iTDP-43A315T mice

(A) CNS expression of DD-qFx (or mCherry) via systemic AAV-PHP.B delivery in 3-month-old mice that were analyzed 2 weeks later.
(B) Western blots show reduced transgenic hTDP-43 and panTDP43 (human and mouse TDP-43) levels in iTDP-43A315T mouse brain extracts expressing DD-qFx-
V5 as compared with mCherry.
(C) Staining of brain from vec- and DD-qFx-expressing mice showed reduced number of hTDP-43-positive cells in the hippocampus. Scale bar, 200 mm.
(D) Disinhibition of iTDP-43A315T mice was significantly ameliorated upon DD-qFx expression as shown by time in closed and open arms of the elevated
plus maze.
(E) Neurons with expression of DD-Fqx (V5; red) in the cortex of human TDP-43 (hTDP-43; green) expressing iTDP-43A315T mice displayed increased proteasome
subunit alpha 1 (PSMA1; purple) labeling together with clearance of pathogenic hTDP-43 (arrows), while proteasome levels remained low in neurons with
persisting hTDP-43 expression in the absence of DD-Fqx (open arrowheads). Note that the residual hTDP-43 present in the latter neurons shows prior presence of
pathological TDP-43 or continued transgene activity.
(F) Staining of cortical brain sections from DD-Fqx (V5; purple) treated iTDP-43A315T mice for human TDP-43 (hTDP-43; red) and murine TDP-43 (mTDP-43; green)
showed suppressed mTDP-43 expression in hTDP-43-positive, DD-Fqx-negative neurons (open arrowheads), DD-Fqx-positive neurons with already reduced

(legend continued on next page)

Neuron 112, 1249–1264, April 17, 2024 1259

 ll
Article

dysfunctional variants of TDP-43 and resulted in cytoplasmic for TDP-43 enables other processes that prevent its nuclear re-
accumulation (Figure 7H). Increasing 14-3-3q levels, as we found entry (e.g., aggregation or aberrant interaction with other pro-
in ALS and FTLD-TDP, promoted insolubility of TDP-43 and its teins) remains to be shown. Previous reports suggested that
hyper-phosphorylation and fragmentation, both in cells and TDP-43 carrying human pathogenic mutations compromises
in vivo. Exploiting the novel 14-3-3q interaction motif, we devel- protein folding and reduces RNA binding and transport,12,59,60
oped a gene therapy vector that facilitated reduction of patho- which therefore may promote the interaction with 14-3-3q. How-
logical TDP-43 together with increased neuronal proteasome ever, our enrichment assay did not reveal different RNA binding
activity and prevented or reverted deficits in different TDP-43- between non-mutant TDP-43 and pathogenic variants. We pro-
expressing mouse models of FTD/ALS. posed that conformational changes in pathogenic mutant TDP-
14-3-3 proteins are known to interact with hundreds of part- 43 contribute to the increased interaction with 14-3-3q, similar
ners, recognizing phosphorylated consensus motifs that bind to a small number of 14-3-3 interactions relying on conforma-
to the inner amphipathic groove (AG) at the center of 14-3-3 tional changes of their binding partners.29 Whether 14-3-3q con-
3D structure.52 The 14-3-3q/TDP-43 interaction is therefore un- tributes to nuclear exit of TDP-43, as reported for other nuclear
usual in several ways; firstly, the interaction is facilitated by a interaction partners, or whether 14-3-3q only stabilizes TDP-43
new motif in a-helix 6 which resides on the surface of 14-3-3q after release of RNA prior to nuclear re-entry remains to be
protein structure opposite to the AG that facilitates conventional shown. However, increased retention of TDP-43 in the cyto-
14-3-3q/partner interactions. Previous reports on interactions plasm due to more binding to 14-3-3q in disease would diminish
with 14-3-3 proteins involving binding outside the AG are sparse the presence of nuclear TDP-43 over time. Finally, a shared
and limited to a hydrophobic motif formed by a-helices 7 and 8 feature of the 14-3-3q/TDP-43 and other 14-3-3 interactions is
as an additional site for partners/complexes that simultaneously that binding changes the localization of the partner,17 with
bind to the AG via canonical phosphorylated motifs.53,54 By TDP-43’s accumulation in the cytoplasm in the presence of
contrast, we found that the 14-3-3q a-helix 6 alone is sufficient increased 14-3-3q levels in cells, mice, and humans.
to stably interact with TDP-43. Secondly, the 14-3-3q/TDP-43 In sporadic FTLD-TDP and ALS, we found increased neuronal
interaction is negatively regulated by RNA bound to TDP-43. 14-3-3q staining together with the presence of 14-3-3q/TDP-43
To our knowledge, this reflects an entirely novel mechanism of complexes and cytoplasmic TDP-43. This is consistent with
regulating 14-3-3 interactions with their partners. This may also the previous report of increased 14-3-3q mRNA levels in
explain why the transient 14-3-3q/TDP-43 interaction was iden- ALS.25 Furthermore, we found that increasing 14-3-3q levels in
tified so readily in our prokaryotic screening system lacking TDP- cells resulted in insolubility, fragmentation, and phosphorylation
43 target RNAs, in contrast to previous TDP-43 interaction dis- of TDP-43, resembling changes found in the disease.2 Increasing
covery studies in mammalian systems.55–57 The RNA-dependent 14-3-3q levels in iTDP-43 mice augmented TDP-43 insolubility
regulation of the 14-3-3q/TDP-43 interaction may explain the and fragmentation as well as neuronal loss, while increasing
markedly increased interaction with L248A/I249A/I250A mutant 14-3-3q levels in non-transgenic mice drove insolubility, frag-
variants of the RRM2 domain of TDP-43 and hence its abolished mentation, and mis-localization of endogenous TDP-43 together
RNA binding and promoting of 14-3-3q/TDP-43-L248A/I249A/ with progressive motor and muscular deficits. Conversely,
I250A complexes. Although this sequence was initially identified increased mutant TDP-43 expression in mice did not change
as a nuclear export-mediating sequence,27 recent findings from 14-3-3q mRNA levels. Similarly, spinal cord neurons in human fa-
cells suggest that TDP-43 leaves the nucleus rather by passive milial ALS (FALS) with C9ORF72 expansion and concomitant
diffusion.58 Our data support a physiological role of 14-3-3q in TDP-43 pathology did not show marked 14-3-3q staining,
limiting the cytoplasmic localization of TDP-43, in particular un- together suggesting that TDP-43 pathology per se is not the
der stress conditions, given the nuclear clearance and cyto- cause of increased 14-3-3q expression. In FALS-C9ORF72,
plasmic accumulation of TDP-43 in 14-3-3zFx/zFx neurons and TDP-43 pathology is likely due to alternative mechanisms
TDP-43 insolubility in 14-3-3q-depleted SH-SY5Y cells upon compared with sporadic disease, like RNA foci formation and
proteasome inhibition. Whether 14-3-3q itself facilitates nuclear repeat-associated non-AUG (RAN) translation of repeat pep-
re-entry of TDP-43 or the lack cytoplasmic 14-3-3q binding sites tides.35–38 Taken together, our findings propose an early,

hTDP-43 and recovering mTDP-43 expression (arrows), and DD-Fqx-positive neurons with virtually completely removed hTDP-43 and marked recovered mTDP-
43 expression (closed arrowheads).
(G) RT-PCR demonstrates significant amounts of a cryptic exon-containing Ap3b2 transcripts (arrowhead) in addition to its regularly cleaved form (chevron) in
vec-treated iTDP-43A315T mice that is virtually absent from DD-Fqx-treated iTDP-43A315T mice in RNA extracts from hippocampi of 3.5-month-old mice upon
2 weeks of treatment.
For (B) to (D), (G), ****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05; ns, not significant. Data points obtained from independent experiments. Error bars indi-
cate SEM.
(H) Proposed model for the role of 14-3-3q/TDP-43 interaction under physiological and disease conditions.
Physiologically, TDP-43 transiently binds to 14-3-3q in the cytoplasm after release of RNA and prior to re-localizing to the nucleus. TDP-43 harboring pathogenic
mutations (FALS) interacts significantly stronger with 14-3-3q in the cytoplasm reducing nuclear re-entry. Increased interaction with 14-3-3q promotes patho-
logical changes of TDP-43, including aggregation, fragmentation, hyper-phosphorylation and cytoplasmic accumulation, and consequently neuronal loss and
functional deficits. In sporadic ALS and FTLD-TDP, increased levels of 14-3-3q provide increased binding sites for TDP-43 while it resides in the cytoplasm during
nucleo-cytoplasmic shuttling. Prolonged binding to 14-3-3q promotes pathological changes of TDP-43, including aggregation, fragmentation, hyper-phos-
phorylation and cytoplasmic accumulation, and consequently neuronal loss and functional deficits. Created with BioRender.com.

1260 Neuron 112, 1249–1264, April 17, 2024

 ll
Article

disease-driving role of 14-3-3q in sporadic FTD and ALS where plexes are negatively regulated by RNA binding of TDP-43.
increased binding sites prolong the residing time of TDP-43 in Physiologically, a transient interaction with 14-3-3q may stabilize
the cytoplasm, eventually leading to TDP-43 pathology (Fig- TDP-43 while it resides in the cytoplasm during RNA shuttling.
ure 7H). This is in line with staining of phosphorylated TDP-43 Our data suggest that increased 14-3-3q interaction with TDP-
in a subset of neurons that show increased 14-3-3q and cyto- 43 (due to TDP-43 mutations or increased 14-3-3q levels) makes
plasmic TDP-43. Whether increased 14-3-3q levels contribute the latter prone to pathological modifications, providing a new
to selective vulnerability of distinct neuronal populations in FTD mechanism for sporadic FTD/ALS and familial ALS with TARDBP
and ALS or how 14-3-3q itself promotes insolubility, fragmenta- mutations. Finally, based on the novel 14-3-3q interaction motif,
tion, and phosphorylation of TDP-43 remains to be shown. we have developed a gene therapy vector with translational po-
Possible mechanisms include increased local concentrations tential for targeting pathological TDP-43 that prevented and re-
of TDP-43 beyond threshold for aggregation or scaffolding of verted functional deficits in different TDP-43-expressing mouse
kinases/proteases and their substrate TDP-43, similar to scaf- models of FTD and ALS.
folding of known functional transcriptional complexes by 14-
3-3.54 STAR+METHODS
Capitalizing on our discovery of a novel interaction sequence,
14-3-3q Fx, that preferentially binds to pathological forms of Detailed methods are provided in the online version of this paper
TDP-43, we have devised a gene therapy vector, DD-qFx, for and include the following:
neuronal delivery and treatment of TDP-43 pathology. DD-qFx
treatment prevented functional deficits and mitigated neuropa- d KEY RESOURCES TABLE
thology in 3 independent TDP-43-based ALS/FTD mouse d RESOURCE AVAILABILITY
models and reverted deficits when administered to symptomatic B Lead contact
mice. Neurons with TDP-43 pathology present with clearance of B Materials availability
nuclear TDP-43, consistent with a loss of physiological functions B Data and code availability
in RNA metabolism and aberrant alternative mRNA splicing.48 d EXPERIMENTAL MODEL AND STUDY PARTICIPANT
Interestingly, 14-3-3q overexpression in human cells that re- DETAILS
sulted in pathological modification and mis-localization of B Animals (mice)
endogenous TDP-43 also caused aberrant mRNA splicing of B Human samples
the physiological TDP-43 target STMN2, similar to what has pre- B Cell lines
viously been reported for cells expressing mutant TDP-43.41,42 B Primary cell cultures
Accordingly, reconstitution of suppressed endogenous TDP-43 d METHOD DETAILS
expression upon DD-qFx treatment of ALS/FTD mouse models B Bacterial two-hybrid screening
and therefore recovery of lost physiological function may at least B Cloning
in part explain the preventive and therapeutic effects of our gene B Adeno-associated viruses
therapy vector. In addition, we found specific activation of B Immunoprecipitation
neuronal proteasome activity by DD-qFx as a mechanism of ac- B Western blotting
tion that may have beneficial effects beyond clearance of patho- B Cell culture staining
logical TDP-43, overcoming impaired proteasome function B Microscopy
linked to TDP-43 pathology and ALS/FTD pathogenesis.34,61,62 B In vitro complex assay
Although our data provide new mechanistic insight into the B RNA precipitations
pathogenesis of ALS and FTD, and proof-of-concept for a poten- B Quantitative PCR
tial disease-modifying gene therapy, further studies are required B RT-PCR
toward clinical translation. For example, differences in disease B 3D modelling
course, transgene expression levels, and mutant gene use be- B Sequential extraction
tween rodent models and humans remain inherent limitations B Motor and Behavioral Testing
to translation.63 Furthermore, although the selectivity of DD- B Immunohistochemistry
qFx for transgenic human/pathological TDP-43 in ALS/FTD d QUANTIFICATION AND STATISTICAL ANALYSIS
mice holds therapeutic promise, future studies in human ALS/ B Quantification
FTD model systems, such as stem-cell-derived neurons, are B Statistical analysis
required to consolidate the differential effects of DD-qFx on
pathological and non-pathological TDP-43 and refine a thera- SUPPLEMENTAL INFORMATION
peutic vector.
In summary, we have identified 14-3-3q as a novel interaction Supplemental information can be found online at https://doi.org/10.1016/j.
neuron.2024.01.022.
partner of TDP-43 that contributes to aberrant cytoplasmic local-
ization of neuronal TDP-43 and possibly ALS/FTD pathogenesis.
ACKNOWLEDGMENTS
Whether this mechanism contributes to non-neuronal TDP-43
pathology in ALS/FTD remains to be shown. Although conven- We thank Dr. Gradinaru and Dr. Deverman (Caltech, US) for providing the AAV-
tional 14-3-3 interactions rely on specific phosphorylation of its PHP.B capsid, Dr. Lee for antibodies to murine TDP-43, Dr. Walker (University
binding partners,20 we show that the 14-3-3q/TDP-43 com- of Queensland, Australia) for inducible DNLS-TDP-43 mice, Dr. Emilija

Neuron 112, 1249–1264, April 17, 2024 1261

 ll
Article
Robinson and Yijun Lin for AAV production, Andrea Chan for histology, the Functional and dynamic polymerization of the ALS-linked protein TDP-
team at the Macquarie Animal Research Services for animal husbandry, 43 antagonizes its pathologic aggregation. Nat. Commun. 8, 45.
Troy Butler for help with animal experiments, and Jessica Spathos for mouse 10. Tsoi, P.S., Choi, K.J., Leonard, P.G., Sizovs, A., Moosa, M.M., MacKenzie,
colony management. Funding was from NHMRC (1081916, 1123564, K.R., Ferreon, J.C., and Ferreon, A.C.M. (2017). The N-Terminal Domain of
1132524, 1136241, 1143848, and 1143978), ARC (DP1096674, ALS-Linked TDP-43 Assembles without Misfolding. Angew. Chem. Int. Ed.
DP170100781, DP170100843, and DP210101957), and MND Australia Engl. 56, 12590–12593.
(GIA1824). Y.K. and L.I. are inventors on a patent application, submitted by
11. Sun, Y., and Chakrabartty, A. (2017). Phase to Phase with TDP-43.
Macquarie University, that covers targeting 14-3-3q to prevent neuronal
Biochemistry 56, 809–823.
toxicity.
12. Buratti, E., and Baralle, F.E. (2001). Characterization and functional impli-
cations of the RNA binding properties of nuclear factor TDP-43, a novel
AUTHOR CONTRIBUTIONS
splicing regulator of CFTR exon 9. J. Biol. Chem. 276, 36337–36343.
Conceptualization, Y.D.K. and L.M.I.; methodology, Y.D.K., A.v.H., F.D., and 13. Sreedharan, J., Blair, I.P., Tripathi, V.B., Hu, X., Vance, C., Rogelj, B.,
L.M.I.; formal analysis, Y.D.K., A.v.H., C.A., G.C., Y.D., M.S., and L.M.I.; inves- Ackerley, S., Durnall, J.C., Williams, K.L., Buratti, E., et al. (2008). TDP-
tigation, Y.D.K., A.v.H., C.A., G.C., W.S.L., J.v.d.H, M.P., Y.D., M.S., N.M., 43 mutations in familial and sporadic amyotrophic lateral sclerosis.
J.B., A.F., S.I., C.H.S., S.Y., A.G., and L.M.I.; resources, N.K.H., J.J.K., and Science 319, 1668–1672.
I.P.B.; writing – original draft, Y.D.K. and L.M.I.; writing – review & editing,
14. Kabashi, E., Valdmanis, P.N., Dion, P., Spiegelman, D., McConkey, B.J.,
Y.D.K., A.v.H., C.A., G.C., W.S.L., J.v.d.H, M.P., Y.D., M.S., N.M., J.B., A.F.,
Vande Velde, C., Bouchard, J.P., Lacomblez, L., Pochigaeva, K.,
S.I., C.H.S., S.Y., A.G., N.K.H., J.J.K., I.P.B., F.D., and L.M.I.; visualization,
Salachas, F., et al. (2008). TARDBP mutations in individuals with sporadic
Y.D.K., A.v.H., and L.M.I.; supervision, Y.D.K., N.K.H., J.J.K., I.P.B., and
and familial amyotrophic lateral sclerosis. Nat. Genet. 40, 572–574.
L.M.I.; project administration, Y.D.K. and L.M.I.; funding acquisition, Y.D.K.,
15. Lattante, S., Rouleau, G.A., and Kabashi, E. (2013). TARDBP and FUS mu-
N.K.H., and L.M.I.
tations associated with amyotrophic lateral sclerosis: summary and up-
date. Hum. Mutat. 34, 812–826.
DECLARATION OF INTERESTS
16. Buratti, E., Brindisi, A., Giombi, M., Tisminetzky, S., Ayala, Y.M., and
Y.D.K. and L.M.I. are inventors, and A.v.H. and W.S.L. are contributors to pat- Baralle, F.E. (2005). TDP-43 binds heterogeneous nuclear ribonucleopro-
ents covering the gene therapy vectors presented here that have been tein A/B through its C-terminal tail: an important region for the inhibition of
licensed to Celosia Therapeutics (Australia), of which L.M.I. is a co-founder. cystic fibrosis transmembrane conductance regulator exon 9 splicing.
Celosia Therapeutics was not involved in design, data generation/analysis, J. Biol. Chem. 280, 37572–37584.
or funding of this study. 17. Bridges, D., and Moorhead, G.B. (2005). 14-3-3 proteins: a number of
functions for a numbered protein. Sci. STKE 2005, re10.
Received: August 8, 2023
18. Liu, D., Bienkowska, J., Petosa, C., Collier, R.J., Fu, H., and Liddington, R.
Revised: November 20, 2023
(1995). Crystal structure of the zeta isoform of the 14–3-3 protein. Nature
Accepted: January 22, 2024
376, 191–194.
Published: February 15, 2024
19. Yaffe, M.B., Rittinger, K., Volinia, S., Caron, P.R., Aitken, A., Leffers, H.,
REFERENCES Gamblin, S.J., Smerdon, S.J., and Cantley, L.C. (1997). The structural ba-
sis for 14–3-3:phosphopeptide binding specificity. Cell 91, 961–971.
1. Burrell, J.R., Halliday, G.M., Kril, J.J., Ittner, L.M., Götz, J., Kiernan, M.C., 20. Muslin, A.J., Tanner, J.W., Allen, P.M., and Shaw, A.S. (1996). Interaction
and Hodges, J.R. (2016). The frontotemporal dementia-motor neuron dis- of 14–3-3 with signaling proteins is mediated by the recognition of phos-
ease continuum. Lancet 388, 919–931. phoserine. Cell 84, 889–897.
2. Neumann, M., Sampathu, D.M., Kwong, L.K., Truax, A.C., Micsenyi, M.C., 21. Coblitz, B., Wu, M., Shikano, S., and Li, M. (2006). C-terminal binding: an
Chou, T.T., Bruce, J., Schuck, T., Grossman, M., Clark, C.M., et al. (2006). expanded repertoire and function of 14–3-3 proteins. FEBS Lett. 580,
Ubiquitinated TDP-43 in frontotemporal lobar degeneration and amyotro- 1531–1535.
phic lateral sclerosis. Science 314, 130–133. 22. Berg, D., Holzmann, C., and Riess, O. (2003). 14-3-3 proteins in the ner-
3. Shenouda, M., Zhang, A.B., Weichert, A., and Robertson, J. (2018). vous system. Nat. Rev. Neurosci. 4, 752–762.
Mechanisms Associated with TDP-43 Neurotoxicity in ALS/FTLD. Adv. 23. Zhang, T., Baldie, G., Periz, G., and Wang, J. (2014). RNA-processing pro-
Neurobiol. 20, 239–263. tein TDP-43 regulates FOXO-dependent protein quality control in stress
4. Yang, C., Wang, H., Qiao, T., Yang, B., Aliaga, L., Qiu, L., Tan, W., response. PLoS Genet. 10, e1004693.
Salameh, J., McKenna-Yasek, D.M., Smith, T., et al. (2014). Partial loss 24. Umahara, T., Uchihara, T., Shibata, N., Nakamura, A., and Hanyu, H.
of TDP-43 function causes phenotypes of amyotrophic lateral sclerosis. (2016). 14-3-3 eta isoform colocalizes TDP-43 on the coarse granules in
Proc. Natl. Acad. Sci. USA 111, E1121–E1129. the anterior horn cells of patients with sporadic amyotrophic lateral scle-
5. Ni, J., Ren, Y., Su, T., Zhou, J., Fu, C., Lu, Y., Li, D., Zhao, J., Li, Y., Zhang, rosis. Brain Res. 1646, 132–138.
Y., et al. (2023). Loss of TDP-43 function underlies hippocampal and 25. Malaspina, A., Kaushik, N., and de Belleroche, J. (2000). A 14-3-3 mRNA is
cortical synaptic deficits in TDP-43 proteinopathies. Mol. Psychiatry 28, up-regulated in amyotrophic lateral sclerosis spinal cord. J. Neurochem.
931–945. 75, 2511–2520.
6. Ederle, H., and Dormann, D. (2017). TDP-43 and FUS en route from the nu- 26. Dervishi, I., Gozutok, O., Murnan, K., Gautam, M., Heller, D., Bigio, E., and
cleus to the cytoplasm. FEBS Lett. 591, 1489–1507. Ozdinler, P.H. (2018). Protein-protein interactions reveal key canonical
7. Buratti, E., and Baralle, F.E. (2008). Multiple roles of TDP-43 in gene pathways, upstream regulators, interactome domains, and novel targets
expression, splicing regulation, and human disease. Front. Biosci. 13, in ALS. Sci. Rep. 8, 14732.
867–878. 27. Winton, M.J., Igaz, L.M., Wong, M.M., Kwong, L.K., Trojanowski, J.Q., and
8. Mompeán, M., Romano, V., Pantoja-Uceda, D., Stuani, C., Baralle, F.E., Lee, V.M. (2008). Disturbance of nuclear and cytoplasmic TAR DNA-bind-
Buratti, E., and Laurents, D.V. (2016). The TDP-43 N-terminal domain ing protein (TDP-43) induces disease-like redistribution, sequestration,
structure at high resolution. FEBS Journal 283, 1242–1260. and aggregate formation. J. Biol. Chem. 283, 13302–13309.
9. Afroz, T., Hock, E.M., Ernst, P., Foglieni, C., Jambeau, M., Gilhespy, 28. Volkening, K., Leystra-Lantz, C., Yang, W., Jaffee, H., and Strong, M.J.
€ckthun, A., Mittl, P., et al. (2017).
L.A.B., Laferriere, F., Maniecka, Z., Plu (2009). Tar DNA binding protein of 43 kDa (TDP-43), 14-3-3 proteins and

1262 Neuron 112, 1249–1264, April 17, 2024

 ll
Article
copper/zinc superoxide dismutase (SOD1) interact to modulate NFL 44. Deverman, B.E., Pravdo, P.L., Simpson, B.P., Kumar, S.R., Chan, K.Y.,
mRNA stability. Implications for altered RNA processing in amyotrophic Banerjee, A., Wu, W.L., Yang, B., Huber, N., Pasca, S.P., et al. (2016).
lateral sclerosis (ALS). Brain Res. 1305, 168–182. Cre-dependent selection yields AAV variants for widespread gene transfer
29. Dougherty, M.K., and Morrison, D.K. (2004). Unlocking the code of 14–3-3. to the adult brain. Nat. Biotechnol. 34, 204–209.
J. Cell Sci. 117, 1875–1884. 45. Gombash Lampe, S.E., Kaspar, B.K., and Foust, K.D. (2014). Intravenous
30. Rittinger, K., Budman, J., Xu, J., Volinia, S., Cantley, L.C., Smerdon, S.J., injections in neonatal mice. J. Vis. Exp. 93, e52037.
Gamblin, S.J., and Yaffe, M.B. (1999). Structural analysis of 14–3-3 phos- 46. Igaz, L.M., Kwong, L.K., Lee, E.B., Chen-Plotkin, A., Swanson, E., Unger,
phopeptide complexes identifies a dual role for the nuclear export signal of T., Malunda, J., Xu, Y., Winton, M.J., Trojanowski, J.Q., et al. (2011).
14–3-3 in ligand binding. Mol. Cell 4, 153–166. Dysregulation of the ALS-associated gene TDP-43 leads to neuronal death
31. Brunet, A., Kanai, F., Stehn, J., Xu, J., Sarbassova, D., Frangioni, J.V., and degeneration in mice. J. Clin. Invest. 121, 726–738.
Dalal, S.N., DeCaprio, J.A., Greenberg, M.E., and Yaffe, M.B. (2002). 14- 47. Walker, A.K., Spiller, K.J., Ge, G., Zheng, A., Xu, Y., Zhou, M., Tripathy, K.,
3-3 transits to the nucleus and participates in dynamic nucleocytoplasmic Kwong, L.K., Trojanowski, J.Q., and Lee, V.M. (2015). Functional recovery
transport. J. Cell Biol. 156, 817–828. in new mouse models of ALS/FTLD after clearance of pathological cyto-
32. Basu, S., Totty, N.F., Irwin, M.S., Sudol, M., and Downward, J. (2003). Akt plasmic TDP-43. Acta Neuropathol. 130, 643–660.
phosphorylates the Yes-associated protein, YAP, to induce interaction 48. Tziortzouda, P., Van Den Bosch, L., and Hirth, F. (2021). Triad of TDP43
with 14-3-3 and attenuation of p73-mediated apoptosis. Mol. Cell control in neurodegeneration: autoregulation, localization and aggrega-
11, 11–23. tion. Nat. Rev. Neurosci. 22, 197–208.
33. Boreddy, S.R., Pramanik, K.C., and Srivastava, S.K. (2011). Pancreatic tu- 49. Ayala, Y.M., De Conti, L., Avendaño-Vázquez, S.E., Dhir, A., Romano, M.,
mor suppression by benzyl isothiocyanate is associated with inhibition of D’Ambrogio, A., Tollervey, J., Ule, J., Baralle, M., Buratti, E., et al. (2011).
PI3K/AKT/FOXO pathway. Clin. Cancer Res. 17, 1784–1795. TDP-43 regulates its mRNA levels through a negative feedback loop.
34. van Eersel, J., Ke, Y.D., Gladbach, A., Bi, M., Götz, J., Kril, J.J., and Ittner, EMBO J. 30, 277–288.
L.M. (2011). Cytoplasmic accumulation and aggregation of TDP-43 upon 50. Ling, J.P., Pletnikova, O., Troncoso, J.C., and Wong, P.C. (2015). TDP-43
proteasome inhibition in cultured neurons. PLoS One 6, e22850. repression of nonconserved cryptic exons is compromised in ALS-FTD.
Science 349, 650–655.
35. Mori, K., Weng, S.M., Arzberger, T., May, S., Rentzsch, K., Kremmer, E.,
Schmid, B., Kretzschmar, H.A., Cruts, M., Van Broeckhoven, C., et al. 51. Jeong, Y.H., Ling, J.P., Lin, S.Z., Donde, A.N., Braunstein, K.E., Majounie,
(2013). The C9orf72 GGGGCC repeat is translated into aggregating dipep- E., Traynor, B.J., LaClair, K.D., Lloyd, T.E., and Wong, P.C. (2017). Tdp-43
tide-repeat proteins in FTLD/ALS. Science 339, 1335–1338. cryptic exons are highly variable between cell types. Mol. Neurodegener.
12, 13.
36. Chew, J., Gendron, T.F., Prudencio, M., Sasaguri, H., Zhang, Y.J.,
Castanedes-Casey, M., Lee, C.W., Jansen-West, K., Kurti, A., Murray, 52. Sluchanko, N.N., Beelen, S., Kulikova, A.A., Weeks, S.D., Antson, A.A.,
M.E., et al. (2015). Neurodegeneration. C9ORF72 repeat expansions in Gusev, N.B., and Strelkov, S.V. (2017). Structural Basis for the
mice cause TDP-43 pathology, neuronal loss, and behavioral deficits. Interaction of a Human Small Heat Shock Protein with the 14-3-
Science 348, 1151–1154. 3 Universal Signaling Regulator. Structure 25, 305–316.
37. O’Rourke, J.G., Bogdanik, L., Muhammad, A.K.M.G., Gendron, T.F., Kim, 53. Rezabkova, L., Man, P., Novak, P., Herman, P., Vecer, J., Obsilova, V., and
K.J., Austin, A., Cady, J., Liu, E.Y., Zarrow, J., Grant, S., et al. (2015). Obsil, T. (2011). Structural basis for the 14–3-3 protein-dependent inhibi-
C9orf72 BAC Transgenic Mice Display Typical Pathologic Features of tion of the regulator of G protein signaling 3 (RGS3) function. J. Biol. Chem.
ALS/FTD. Neuron 88, 892–901. 286, 43527–43536.
38. Peters, O.M., Cabrera, G.T., Tran, H., Gendron, T.F., McKeon, J.E., 54. Taoka, K., Ohki, I., Tsuji, H., Furuita, K., Hayashi, K., Yanase, T.,
Metterville, J., Weiss, A., Wightman, N., Salameh, J., Kim, J., et al. Yamaguchi, M., Nakashima, C., Purwestri, Y.A., Tamaki, S., et al. (2011).
(2015). Human C9ORF72 Hexanucleotide Expansion Reproduces RNA 14-3-3 proteins act as intracellular receptors for rice Hd3a florigen.
Foci and Dipeptide Repeat Proteins but Not Neurodegeneration in BAC Nature 476, 332–335.
Transgenic Mice. Neuron 88, 902–909. 55. Freibaum, B.D., Chitta, R.K., High, A.A., and Taylor, J.P. (2010). Global
39. Ke, Y.D., van Hummel, A., Stevens, C.H., Gladbach, A., Ippati, S., Bi, M., analysis of TDP-43 interacting proteins reveals strong association with
€ger, S., van der Hoven, J., Volkerling, A., et al. (2015). Short-
Lee, W.S., Kru RNA splicing and translation machinery. J. Proteome Res. 9, 1104–1120.
term suppression of A315T mutant human TDP-43 expression improves 56. Ling, S.C., Albuquerque, C.P., Han, J.S., Lagier-Tourenne, C., Tokunaga,
functional deficits in a novel inducible transgenic mouse model of FTLD- S., Zhou, H., and Cleveland, D.W. (2010). ALS-associated mutations in
TDP and ALS. Acta Neuropathol. 130, 661–678. TDP-43 increase its stability and promote TDP-43 complexes with FUS/
40. Prudencio, M., Humphrey, J., Pickles, S., Brown, A.L., Hill, S.E., TLS. Proc. Natl. Acad. Sci. USA 107, 13318–13323.
Kachergus, J.M., Shi, J., Heckman, M.G., Spiegel, M.R., Cook, C., et al. 57. Chou, C.C., Zhang, Y., Umoh, M.E., Vaughan, S.W., Lorenzini, I., Liu, F.,
(2020). Truncated stathmin-2 is a marker of TDP-43 pathology in fronto- Sayegh, M., Donlin-Asp, P.G., Chen, Y.H., Duong, D.M., et al. (2018).
temporal dementia. J. Clin. Invest. 130, 6080–6092. TDP-43 pathology disrupts nuclear pore complexes and nucleocytoplas-
41. Klim, J.R., Williams, L.A., Limone, F., Guerra San Juan, I., Davis- mic transport in ALS/FTD. Nat. Neurosci. 21, 228–239.
Dusenbery, B.N., Mordes, D.A., Burberry, A., Steinbaugh, M.J., 58. Ederle, H., Funk, C., Abou-Ajram, C., Hutten, S., Funk, E.B.E.,
Gamage, K.K., Kirchner, R., et al. (2019). ALS-implicated protein TDP-43 Kehlenbach, R.H., Bailer, S.M., and Dormann, D. (2018). Nuclear egress
sustains levels of STMN2, a mediator of motor neuron growth and repair. of TDP-43 and FUS occurs independently of Exportin-1/CRM1. Sci.
Nat. Neurosci. 22, 167–179. Rep. 8, 7084.
42. Melamed, Z., López-Erauskin, J., Baughn, M.W., Zhang, O., Drenner, K., 59. Alami, N.H., Smith, R.B., Carrasco, M.A., Williams, L.A., Winborn, C.S.,
Sun, Y., Freyermuth, F., McMahon, M.A., Beccari, M.S., Artates, J.W., Han, S.S.W., Kiskinis, E., Winborn, B., Freibaum, B.D., Kanagaraj, A.,
et al. (2019). Premature polyadenylation-mediated loss of stathmin-2 is a et al. (2014). Axonal transport of TDP-43 mRNA granules is impaired by
hallmark of TDP-43-dependent neurodegeneration. Nat. Neurosci. 22, ALS-causing mutations. Neuron 81, 536–543.
180–190. 60. Bhandare, V.V., and Ramaswamy, A. (2018). The proteinopathy of D169G
43. Huang, S.L., Wu, L.S., Lee, M., Chang, C.W., Cheng, W.C., Fang, Y.S., and K263E mutants at the RNA Recognition Motif (RRM) domain of tar
Chen, Y.R., Cheng, P.L., and Shen, C.-K.J. (2020). A robust TDP-43 DNA-binding protein (tdp43) causing neurological disorders: A computa-
knock-in mouse model of ALS. Acta Neuropathol. Commun. 8, 3. tional study. J. Biomol. Struct. Dyn. 36, 1075–1093.

Neuron 112, 1249–1264, April 17, 2024 1263

 ll
Article
61. Scotter, E.L., Chen, H.J., and Shaw, C.E. (2015). TDP-43 Proteinopathy 68. Bi, M., Gladbach, A., van Eersel, J., Ittner, A., Przybyla, M., van Hummel,
and ALS: Insights into Disease Mechanisms and Therapeutic Targets. A., Chua, S.W., van der Hoven, J., Lee, W.S., Mu €ller, J., et al. (2017). Tau
Neurotherapeutics 12, 352–363. exacerbates excitotoxic brain damage in an animal model of stroke. Nat.
62. Götzl, J.K., Lang, C.M., Haass, C., and Capell, A. (2016). Impaired protein Commun. 8, 473.
degradation in FTLD and related disorders. Ageing Res. Rev. 32, 122–139. 69. Ittner, L.M., Ke, Y.D., and Götz, J. (2009). Phosphorylated Tau Interacts
63. Ittner, L.M., Halliday, G.M., Kril, J.J., Götz, J., Hodges, J.R., and Kiernan, with c-Jun N-terminal Kinase-interacting Protein 1 (JIP1) in Alzheimer
M.C. (2015). FTD and ALS-translating mouse studies into clinical trials. Disease. J. Biol. Chem. 284, 20909–20916.
Nat. Rev. Neurol. 11, 360–366. 70. Ke, Y.D., Dramiga, J., Schu€tz, U., Kril, J.J., Ittner, L.M., Schröder, H., and
64. Fath, T., Ke, Y.D., Gunning, P., Götz, J., and Ittner, L.M. (2009). Primary Götz, J. (2012). Tau-mediated nuclear depletion and cytoplasmic accumu-
support cultures of hippocampal and substantia nigra neurons. Nat. lation of SFPQ in Alzheimer’s and Pick’s disease. PLoS One 7, e35678.
Protoc. 4, 78–85. 71. Xiao, B., Smerdon, S.J., Jones, D.H., Dodson, G.G., Soneji, Y., Aitken, A.,
65. von Jonquieres, G., Mersmann, N., Klugmann, C.B., Harasta, A.E., Lutz, and Gamblin, S.J. (1995). Structure of a 14-3-3 protein and implications for
B., Teahan, O., Housley, G.D., Fröhlich, D., Kra€mer-Albers, E.M., and coordination of multiple signalling pathways. Nature 376, 188–191.
Klugmann, M. (2013). Glial promoter selectivity following AAV-delivery to 72. van Hummel, A., Chan, G., van der Hoven, J., Morsch, M., Ippati, S., Suh,
the immature brain. PLoS One 8, e65646. L., Bi, M., Asih, P.R., Lee, W.S., Butler, T.A., et al. (2018). Selective
66. Delerue, F., White, M., and Ittner, L.M. (2014). Inducible, tightly regulated Spatiotemporal Vulnerability of Central Nervous System Neurons to
and non-leaky neuronal gene expression in mice. Transgenic Res. 23, Pathologic TAR DNA-Binding Protein 43 in Aged Transgenic Mice. Am.
225–233. J. Pathol. 188, 1447–1456.
67. Ittner, L.M., Koller, D., Muff, R., Fischer, J.A., and Born, W. (2005). The 73. van Eersel, J., Stevens, C.H., Przybyla, M., Gladbach, A., Stefanoska, K.,
N-terminal extracellular domain 23–60 of the calcitonin receptor-like re- Chan, C.K.-X., Ong, W.Y., Hodges, J.R., Sutherland, G.T., Kril, J.J., et al.
ceptor in chimeras with the parathyroid hormone receptor mediates asso- (2015). Early-onset axonal pathology in a novel P301S-Tau transgenic
ciation with receptor activity-modifying protein 1. Biochemistry 44, mouse model of frontotemporal lobar degeneration. Neuropathol. Appl.
5749–5754. Neurobiol. 41, 906–925.

1264 Neuron 112, 1249–1264, April 17, 2024

 ll
Article

STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER

Antibodies
Mouse monoclonal anti-TDP-43 ProteinTech Cat# 60019-2-IG; RRID: AB_2200520
Rabbit polyclonal anti-c-terminal TDP-43 ProteinTech Cat# 12892-1-AP; RRID: AB_2200505
Rabbit polyclonal anti-TDP-43 ProteinTech Cat# 10782-2-AP; RRID: AB_615042
Mouse monoclonal anti-14-3-3q Abcam Cat# AB10439; RRID: AB_297180
Rabbit polyclonal anti-qFc-specific 14-3-3q ProteinTech Cat# 14503-1-AP; RRID: AB_2218096
Mouse monoclonal anti-V5 Invitrogen/Life Technologies Cat# R960CUS; RRID: AB_2792973
Rabbit polyclonal anti-V5 Sigma Cat# V8137; RRID:AB_261889
Mouse monoclonal anti-myc Life Technologies Cat# R950-25; RRID: AB_2556560
Mouse monoclonal anti-mCherry Abcam Cat# AB125096; RRID: AB_11133266
Rabbit polyclonal anti-mCherry Abcam Cat# AB167453; RRID: AB_2571870
(; Cat. No.; 1:800
Mouse monoclonal anti- CosmoBio Cat# CAC-TIP-BTD-M01
phospho-TDP-43 S409/410
Rabbit polyclonal anti- CosmoBio Cat# CAC-TIP-PTD-P02
phospho-TDP-43 S409/410
Mouse monoclonal anti-GAPDH Merck-Millipore Cat# MAB374; RRID: AB_2107445
Rabbit polyclonal anti-mouse TDP-43 Igaz et al. 46; generous gift NA
from Dr. Virginia Lee, UPenn
Mouse monoclonal anti-MAP2 Sigma Cat# M1406; RRID: AB_477171
Chicken polyclonal anti-NeuN Merck Millipore Cat# ABN91; RRID: AB_11205760
Goat-polyclonal anti-GFP Abcam Cat# AB6673; RRID: AB_305643
Rabbit monoclonal anti-PSMA1 Invitrogen Cat# MA5-32806; RRID: AB_2810082
Goat polyclonal anti-mouse IgG, Jackson ImmunoResearch Cat# 111-0350166
horseradish peroxidase coupled
Goat polyclonal anti-rabbit IgG, Jackson ImmunoResearch Cat# 111-035-144
horseradish peroxidase coupled
Goat polyclonal anti-mouse IgG, Life Technologies Cat# A21422; RRID: AB_2535844
Alexa Fluor 555 coupled
Goat polyclonal anti-rabbit IgG, Life Technologies Cat# A11008; RRID:AB_143165
Alexa Fluor 488 coupled
Donkey polyclonal anti-mouse IgG, Life Technologies Cat# A31571; RRID: AB_162542
Alexa Fluor 647 coupled
Donkey polyclonal anti-rabbit IgG, Life Technologies Cat# A21026
Alexa Fluor 488 coupled
Donkey polyclonal anti-goat IgG, Life Technologies Cat# A21206; RRID: AB_2535792
Alexa Fluor 555 coupled
Goat polyclonal anti-mouse IgG, Vector Cat# BA-9200; RRID: AB_2336171
biotinylated
Goat polyclonal anti-rabbit IgG, Vector Cat# BA-1000; RRID:AB_2313606
biotinylated
Bacterial and virus strains
E.coli strain for BacterioMatch: XL1-Blue Chem-Agilent Cat# 240065
MRF0 Kan (Genotype: D(mcrA)183
D(mcrCB-hsdSMR-mrr)173 endA1
supE44 thi-1 recA1 gyrA96 relA1 lac
[F0 proAB lacIq Z DM15 Tn5 (Kan r)]
(Continued on next page)

Neuron 112, 1249–1264.e1–e8, April 17, 2024 e1

 ll
Article

Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
44
Adeno-associated virus 9 variant Deverman et al. NA
PHP.B (AAV9.PHP.B)
Adeno-associated virus 9 (AAV9) University of Pennsylvania NA
Biological samples
human entorhinal cortex and spinal cord New South Wales Brain NA
tissue samples, Table S2 Tissue Resource Centre
Chemicals, peptides, and recombinant proteins
4,6-diamidino-2-phenylindole (DAPI) Molecular Probes Cat# D1306
Dulbecco’s Modified Eagle Medium (DMEM) Thermo Fisher Sci Cat# 11995065
Dulbecco’s Modified Eagle Thermo Fisher Sci Cat# 11320033
Medium/Nutrient Mixture F-12 (DMEM/F-12)
Fetal bovine serum (FBS) Hyclone Cat# SH30084.03 and Lot# DE27854269
Puromycin Sigma Cat# P8833
Neurobasal Invitrogen Cat# 21103049
B27 supplement Invitrogen Cat# 17504044
GlutaMAX Invitrogen Cat# 35050061
Poly-D-lysine Sigma Cat# P0899
2,3,5- triphenyltetrazolium chloride (TTC) Sigma Cat# 1.08380
MG132 (proteasome inhibitor) Sigma Cat# M8699
Complete protease inhibitor cocktail Roche Cat# 4693132001
RNase Out Life Technologies Cat# 10777019
vanadyl ribonucleoside complexes New England BioLabs Cat# S1402S
TRIzol Reagent Life Technologies Cat# 15596018
Critical commercial assays
BacterioMatch II two hybrid Chem-Agilent Cat# 240065
Duolink PLA Brightfield kit Millipore Sigma Cat# DUO92012
RNeasy Mini Kit Qiagen Cat# 74104
Fast SYBR green reaction mix Invitrogen Cat# 4385612
Bradford assay BioRad Cat# 5000203
Deposited data
BacterioMatch results shown in Table S1 This study NA
Experimental models: Cell lines
293T HEK cells ATCC Cat# CRL-11268
SH-SY5Y cells ATCC Cat# CRL-2266
E16.5 primary mouse neurons In house protocol; Fath et al.64 NA
Experimental models: Organisms/strains
iTDP-43A315T mice Ke et al.39 NA
rNLS mice Walker et al.47 NA
14-3-3qzFx/zFx mice This study NA
TardbpN390D mice This study NA
C57Bl/6 ARC, Perth, Australia JAX #000664
Oligonucleotides
sgRNA for the generation of 14-3-3qzFx/zFx This study NA
mice: 5’-acgtaagtatatttgaccca-3’
sgRNA for the generation of 14-3-3qzFx/zFx This study NA
mice: 5’-ttgtttctgtgagttgaaaa-3’
(Continued on next page)

e2 Neuron 112, 1249–1264.e1–e8, April 17, 2024

 ll
Article

Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
Oligomer for the generation This study NA
of 14-3-3qzFx/zFx mice: 5’-aaatgaagggag
attatttccggtatcttgctgaagtagcttgtggcgatg
ataagaaaggaattgtggaccagtcacagcaagcc
taccaagaggcgtttgatataagcaagaaggaga-3’
sgRNA for the generation of TardbpN390D This study NA
mice: 5’-tgggggtcagcatcaaatgc-3’
sgRNA for the generation of TardbpN390D This study NA
mice: 5’-gtcagcatcaaatgcaggat-3’
Oligomer for the generation This study NA
of TardbpN390D mice: 5’-aaataattccta
cagtggttctaattctggtgccccccttggttggg
ggtcagcatcagatgcaggatcgggcagtggtt
ttaatgggggctttggctcgagcatggattctaa
gtcttc-3’
qPCR primers for Ywhaq (14-3-3q): Macrogen, Korea NA
5’-GCTAAAACGGCTTTTGATGAGG-3’
and 5,-GTGCCCTGGATGCCTTTAGTT-3’
qPCR primers for Ywhab (14-3-3b): Macrogen, Korea NA
5’-CTCCAGTCCTCCGCGAAAAT-3’
and 5’-GAGAGTTCGTGTCCCTGCTC-3’
qPCR primers for Ywhag (14-3-3g) Macrogen, Korea NA
5’-GGCGGTCTTCGGTTTCCTTC-3’
and 5’-GTTCAGCTCGGTCACGTTCTT-3’
qPCR primers for Ywhae (14-3-3ε) Macrogen, Korea NA
5’-CGCACCCCATTCGTTTAGG-3’
and 5’-ATTCTGCTCTTCACCATCACC-3’
qPCR primers for Ywhaz (14-3-3z) Macrogen, Korea NA
5’-CTACGATCACGTCCAACCCG-3’
and 5’-GTCAAACGCTTCTGGCTGC-3’
qPCR primers for Ywhas (14-3-3s) Macrogen, Korea NA
5’- ACAACCTGACACTGTGGACG-3’
and 5’-CCTTTGGAGCAAGAACAGCG-3’
qPCR primers for human TARDBP Macrogen, Korea NA
5’- CTGCGGGAGTTCTTCTCTCA-3’
and 5’-CGCAATCTGATCATCTGCAA-3’
qPCR primers for murine Tardbp Macrogen, Korea NA
5’-AATCAGGGTGGGTTTGGTAACA-3’
and 5’-GCTGGGTTAATGCTAAAAGCAC-3’
qPCR primers for Gapdh Macrogen, Korea NA
5’-GTGAAGGTCGGTGTGAAC-3’
and 5’-ATCTCCACTTTGCCACTGCAA-3’
Recombinant DNA
pBT vector Chem-Agilent Cat# 240067
human brain cDNA pTRG plasmid library Chem-Agilent, Stratagene Cat# 982262
14-3-3q MISSION shRNA vector Sigma-Aldrich Cat# SHCLND
pLenti6/Ubc vector Life Technologies Cat# V499–10
pcDNA3.1/myc vector Life Technologies Cat# V855-20
pcDNA3.1/V5 vector Life Technologies Cat# V81020
pAAV-CAG-WPRE-bHHpA vector von Jonquieres et al.65; NA
gift from Dr. Matthias
Klugmann
PTuner-LVX vector Clontech (Takara Bio) Cat# 632173
(Continued on next page)

Neuron 112, 1249–1264.e1–e8, April 17, 2024 e3

 ll
Article

Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
pAAV-hSyn-EGFP vector Addgene gift from Bryan Roth
(Addgene plasmid # 50465;
http://n2t.net/addgene:50465;
RRID: Addgene_50465)
Software and algorithms
Fiji (Image J), Version 2.14.0/1.54f NIH https://imagej.net/software/fiji/downloads
ImageLab software, Version 6.0.1 BioRad https://www.bio-rad.com/en-au/product/
image-lab-software?ID=KRE6P5E8Z
ZEN, Version 3.1 Zeiss https://www.zeiss.com/microscopy/
de/produkte/software/zeiss-zen.html
Prism 10, Version 10.1.0 GraphPad https://www.graphpad.com/
BioRender BioRender.com https://www.biorender.com/
Illustrator CC 2022, Version 26.3.1 Adobe https://www.adobe.com/
products/illustrator.html
Other
Protein G beads Life Technologies Cat# 20422
FluoroMount Southern Biotech Cat# 0100-01
TALON resin Clontech (Takara Bio) Cat# 635502
Dynabeads Protein G Life Technologies Cat# 10003D
RNeasy spin columns Qiagen Cat# 74104

RESOURCE AVAILABILITY

Lead contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Lars M.
Ittner ([email protected]).

Materials availability
Plasmids, cell, and mouse lines generated in this study will be made available on request, but we may require payment and/or a
completed Materials Transfer Agreement if there is potential for commercial application.

Data and code availability

d Proteomics data from bacterial 2-hybroid match is provided in Table S2.
d No custom codes were used in this study.
d Any additional information required to reanalyze the data reported in this work paper is available from the Lead Contact upon
request.

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Animals (mice)
iTDP-43A315T mice have been previously described.39 These mice constitutively express human A315T mutant TDP-43 under control
of a doxycycline-controllable (Tet-OFF) promoter in CNS neurons. rNLS mice have previously been described.47 These mice express
human DNLS mutant TDP-43 in CNS nurons controlled by the Tet-OFF system, using a Thy-OFF line.66 rNLS mice were maintained
on doxycycline till day 21, when being weaned from their mothers. 14-3-3qzFx/zFx mice were generated by electroporation of fertilized
C57Bl/6 oozytes using spCas9 protein, sgRNAs (5’-acgtaagtatatttgaccca-3’ and 5’-ttgtttctgtgagttgaaaa-3’) and oligomer (5’-aaat
gaagggagattatttccggtatcttgctgaagtagcttgtggcgatgataagaaaggaattgtggaccagtcacagcaagcctaccaagaggcgtttgatataagcaagaagg
aga-3’) followed by implantation in foster females. Successful modification was confirmed by sequencing and offspring were sub-
sequently identified by genotyping PCR with primers 5’-tcaggtccaggtttgaatga-3’ and 5’-cgtaggctgcatctccttct-3’, producing a
281bp mutant amplicon. 14-3-3qzFx/zFx mice have no overt phenotype. TardbpN390D mice were generated using the same approach
with the sgRNAs (5’-tgggggtcagcatcaaatgc-3’ and 5’-gtcagcatcaaatgcaggat-3’) and oligomer (5’-aaataattcctacagtggttctaatt
ctggtgccccccttggttgggggtcagcatcagatgcaggatcgggcagtggttttaatgggggctttggctcgagcatggattctaagtcttc-3’) and genotyped by next
generation sequencing (iSeq, Illumina). Homozygous TardbpN390D mice have no overt phenotype. C57Bl/6 mice were obtained

e4 Neuron 112, 1249–1264.e1–e8, April 17, 2024

 ll
Article

from a commercial supplier (ARC Perth, Australia). Mice were group housed on a 12h light/dark cycle with ad libidum access to food
and water. Mice of both genders were used with no observed influence of sex. For tissue analysis, mice were terminated at indicated
time points either by cervical dislocation for rapid tissue extraction (RNA and protein analysis) or by transcardial perfusion with phos-
phate buffered saline followed by 10 minutes of cool 4% paraformaldehyde (PFA) and emersion fixation for 12 hours after brains were
retrieved before histological processing. Tibialis anterior muscles were removed and weight with an ultra-fine balance. All animal ex-
periments have been approved by the Macquarie University Animal Ethics Committee.

Human samples
Paraffin embedded human entorhinal cortex and spinal cord tissue samples were received from the New South Wales Brain Tissue
Resource Centre at the University of Sydney and the Sydney Brain Bank at Neuroscience Research Australia. Research reported in
this publication was supported by the National Institute on Alcohol Abuse and Alcoholism of the National Institutes of Health. Details
on human cases have been provided in the Table S2. Human cases of both genders were used with no observed differences. Brains
were removed at autopsy, fixed in 15% formalin for 14 days and then weighed. The cerebrum was embedded in 3% agarose and
sliced at 3-mm intervals. Tissue was sectioned at 5mm on a standard microtome (Thermo) and mounted on Superfrost plus glass
slides (Menzel) for subsequent staining procedures. Use of human samples has been approved by the Human Ethics Committees
of University of Sydney and Macquarie University.

Cell lines
All immunoprecipitation (IP) experiments were carried out in 293T HEK cells (ATCC; CRL-11268) maintained in DMEM containing
10% fetal bovine serum (FBS) as per standard protocols. SH-SY5Y cells (ATCC; CRL-2266) were maintained in DMEM/F-12 contain-
ing 10% FBS and used for 14-3-3q overexpression or knockdown. Stable overexpression and knockdown of 14-3-3q were achieved
through lentiviral transduction and media supplemented with puromycin.

Primary cell cultures

Primary mouse neurons were cultured from mouse embryo brains as established by us before.64 Briefly, brains were removed from
mouse embryos 16.5 days after time mating and placed in culture medium on ice for careful removal of meninges. Sub-dissected
hippocampi and cortices were dissociated by titruation, cell suspensions passed through cell strainers and seeded at 22,000
cells/cm2 on poly-D-lysine coated glass coverslips. Cells were cultured in Neurobasal medium supplemented with B27 and
GlutaMAX at 37 C and 5% CO2 until use for experimentation.

METHOD DETAILS

Bacterial two-hybrid screening

BacterioMatch II two hybrid system was done following the manufacturer’s instructions (Chem-Agilent). Briefly, the carboxyl-terminal
part of human TDP-43 (corresponding to aa 259-415) was cloned into the pBT (bait) vector and used to identify interacting partners
from a human brain cDNA pTRG plasmid library. Detection of protein-protein interaction partners is based on transcriptional activa-
tion of the HIS3 reporter gene and positives are further verified by a secondary streptomycin resistance reporter. All propagation and
transformation were done using the chemically competent cells provided within the kit. Colonies were visualized for images by incu-
bation of growth plates with 2% TTC/PBS (Sigma) for 10 min at 37 C.

Cloning
Point mutations and truncation variants were generated by standard site-directed mutagenesis.67 Knockdown of 14-3-3q was done
using the 14-3-3q MISSION shRNA lentiviruses (Sigma-Aldrich) with sequence CCGGCAGTTGCTTAGAGACAACCTACTCG
AGTAGGTTGTCTCTAAGCAACTGTTTTTG. Stable overexpression of 14-3-3q (C-terminal V5-tag) in SH-SY5Y cells was achieved
using lentiviruses (cloning vector pLenti6/Ubc; Life Technologies). All 14-3-3 isoforms were cloned with a C-terminal myc tag into
pcDNA3.1/myc (Life Technologies) and TDP-43 wild-type/mutants were cloned with a C-terminal V5 tag into pcDNA3.1/V5 (Life
Technologies) for IP. Truncation variants of 14-3-3q were generated by site directed mutagenesis as previously described.67

Adeno-associated viruses
14-3-3q (V5-tagged) was cloned into a rAAV vector under the CMV early enhancer fused to modified chicken b-actin promoter using
the plasmid pAAV-CAG-WPRE-bHHpA (gift from Matthias Klugmann) as backbone and removing EGFP.65 The same vector with
mCherry or EGFP expression were used as control. To generate pAAV-CAG-DD-qFx the coding sequence for amino acids 135 to
164 of 14-3-3q were cloned downstream of a FKBP-L106P degradation domain from the PTuner-LVX vector (Clontech) and upstream
of a V5 tag sequence. Enhanced GFP was expressed from pAAV-hSyn-EGFP. pAAV-hSyn-EGFP was a gift from Bryan Roth (Addg-
ene plasmid # 50465). Human TDP-43 and human A315T-mutant TDP-43 were cloned into pAAV-hSyn-EGFP by replacing EGFP
cDNA. Packaging of rAAV9 vectors were performed as previously described68 using the capsid AAV9.PHP.B.44 2 ml of rAAV (1 x
1013 viral genomes/ml) was injected into the hippocampus (-1.94mm AP, 1.6mmML, 1.8mm DV from lambda) of 3 weeks old
wild-type or iTDP-43A315T mice.39 For spinal cord injections, 1 ul of rAAV (1 x 1013 viral genomes/ml) was injected directly into the

Neuron 112, 1249–1264.e1–e8, April 17, 2024 e5

 ll
Article

spinal cord of cryo-anaesthetized neonatal pups (P0-2). For direct brain injections (i.c.v.), 1 ul of rAAV (2-3 x 1013 viral genomes/ml)
was injected directly into brain bilaterally into the cortex, midbrain and unilaterally into the cerebellum of cryo-anaesthetized neonatal
pups (P0-2).68 For systemic administration, 100ml of rAAV (5 x 1013 viral genomes/ml) were injected intravenously via the tail vein in
adult mice or 30ml of rAAV (5 x 1012 viral genomes/ml) were injected intravenously via the temporal vein in newborn mice. All animal
experiments have been approved by the Macquarie University Animal Ethics Committee.

Immunoprecipitation
Immunoprecipitation (IP) was done as previously described.69 Briefly, 293T HEK cells were co-transfected with variants of TDP-43 in
pcDNA3.1/V5 and/or 14-3-3 isoforms/variants. Cells were lysed in RIPA buffer. Equal amounts of proteins were incubated overnight
with 1 ul of V5 antibody (Life Technologies) and precipitated using magnetic Protein G beads (Life Technologies). Co-IP was further
confirmed by Western blotting using myc antibodies.

Western blotting
Western blotting has been done as previously described.70 Primary antibodies used for immunoblotting were human TDP-43
(ProteinTech; Cat. No. 60019-2-IG; 1:2,000), c-terminal TDP-43 (ProteinTech; Cat. No. 12892-1-AP; 1:1,000), pan-TDP-43
(ProteinTech; Cat. No. 10782-2-AP; 1:2,000), pan 14-3-3q (Abcam; Cat. No. AB10439; 1:1,000), qFc-specific 14-3-3q (ProteinTech;
Cat. No. 14503-1-AP; 1:1,000), V5 (Invitrogen; R960CUS; 1:5,000), myc (Life Technologies; R950-25; 1:5,000), mCherry (Abcam;
AB125096; 1:2,000), phospho-TDP-43 S409/410 (Cosmobio; CAC-TIP-BTD-M01; 1:4,000), GAPDH (Merck-Millipore; MAB374;
1:5,000) and mouse TDP-43 (generous gift from Dr. Virginia Lee; 1:1,000).

Cell culture staining

For immunocytochemistry, cells were fixed in 4% PFA at indicated time points and blocked with 3% heat-inactivated goat serum/2%
BSA. Antibodies, V5 (Sigma; V8137; 1:2,000), myc (Life Technologies; R950-25; 1:2,000), MAP2 (Sigma; M1406; 1:1,000) and sec-
ondary Alexa Fluor 488, 555 (Life Technologies) were used. Nuclei were staining with DAPI (Molecular Probes). Coverslips were
mounted in FluoroMount (Southern Biotech). Primary neurons cultured for 14 days were treated with the proteasome inhibitor
MG132 at indicated times and doses prior to staining.

Microscopy
All cell culture fluorescence images were taken with either a BX51 epifluorescence or a confocal FV10i microscopes (Olympus). All
brain and spinal cord fluorescence images were taken with a BX51 (monochrome camera; Olympus) or entire slides were scanned
with slide scanner AxioScan (Zeiss).

In vitro complex assay

HEK293T cells were transfected with full length WT TDP-43 (c-terminally V5-tagged) or 14-3-3q (with poly-histidine-tag). Cells trans-
fected with TDP-43 constructs were lysed in IP buffer (IPB) (20 mM of Tris-HCl (pH 7.8), 150 mM of NaCl, 0.1 % (v/v) of NP-40) sup-
plemented with EDTA-free Complete protease inhibitor cocktail (Roche) and V5-tagged TDP-43 was IP-ed (as above) with mouse
anti-v5 antibody (Life Technologies). Subsequently, the lysate was washed twice with IPB, twice with DNase/RNase buffer (DRB)
(10mM Tris- HCl (pH 7.6), 2.5 mM of MgCl2 and 0.5 mM of CaCl2) and reconstituted in DRB. The 14-3-3q-HIS transfected cells
were lysed in IPB and purified via TALON resin (Clontech Laboratories). Briefly, the lysates were incubated with TALON resins
for 2 h at 4  C, washed thrice with IPB and eluted in in vitro interaction buffer (IVIB) (20 mM Tris-HCl (pH 7.8), 0.1 M of NaCl, 20%
(v/v) of glycerol, 5 mM of MgCl2, 5 mM of CaCl2, 0.1 % (v/v) of NP-40, 1 mM of EDTA, 0.1 mM DTT and 0.2 mM of PMSF). For
RNA and DNA digestions, the magnetic bead suspensions containing bound V5-tagged TDP-43 were incubated with either DNase,
RNase or buffer (control) for 10 min at 37  C. After digestion, all the reaction mixtures were washed once with ice-cold IPB and re-
suspended in IVIB. The purified TDP-43 and 14-3-3q were subsequently incubated for 2 h at 4  C for in vitro interaction. The reaction
was washed as per regular IP and eluted in 4x sample buffer for Western blotting.

RNA precipitations
10 cm dish of HEK293T cells was transfected with V5-tagged hTDP-43 variants as described above. At 48 hours post-transfection, cells
were rinsed with PBS and lysed using polysomal lysis buffer (100 mM KCl, 5 mM MgCl2, 10 mM HEPES pH 7.0, 0.5% Nonidet P-40, 1 mM
DTT, 13 protease inhibitors) supplemented with 100 units ml1 RNase Out (Life Technologies) and 400 mM vanadyl
ribonucleoside complexes (Promega). Lysates were centrifuged at 16,0003g for 15 min at 4  C and supernatants were diluted 2-fold
in NT2 buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM MgCl2, 0.05% Nonidet P-40, 20 mM DTT, 20 mM EDTA, 13 protease
inhibitors) supplemented with 200 units ml1 RNase Out and 400 mM vanadyl ribonucleoside complexes. Lysates were then incubated
with mouse anti-V5 monoclonal antibody (Life Technologies, Cat No. R960CUS) at 4  C for 16 h with gentle mixing. Immunoprecipitation
with Dynabeads Protein G (Life Technologies) occurred for 1 h at 4  C. After washing with ice-cold NT2 buffer, one-tenth volume of the
beads were collected for Western blot analysis. In parallel, RNA was eluted from the remaining beads using TRIzol Reagent (Life
Technologies) and chloroform extraction. RNA was further purified using the RNeasy spin columns (Qiagen) and quantified on the Nano-
drop (ThermoFisher).

e6 Neuron 112, 1249–1264.e1–e8, April 17, 2024

 ll
Article

Quantitative PCR
RNA purification and quantitative PCR was performed as previously described.68 Briefly, RNA was extracted from mouse cortical
brain tissue or SH-SY5Y cells using RNeasy Mini Kit (Qiagen), following the manufacturer’s instructions. To remove contaminating
genomic DNA, an on-column DNA-digest was performed with RNase-free DNase I (Qiagen). cDNA was synthesized from 2.5 mg
of total RNA with the second strand cDNA-synthesis kit (Invitrogen). mRNA levels were determined by quantitative PCR, using a
Fast SYBR green reaction mix (Invitrogen) and gene-specific primer pairs, using a Mx3000 real-time PCR cycler (Stratagene). Levels
were expressed as a fold change of the housekeeping gene Gapdh and converted to fold difference relative to control tissue. These
primers were used (5’>3’): 14-3-3q forward (F): GCTAAAACGGCTTTTGATGAGG, reverse (R): GTGCCCTGGATGCCTTTAGTT; 14-3-
3b F: CTCCAGTCCTCCGCGAAAAT, R: GAGAGTTCGTGTCCCTGCTC; 14-3-3g F: GGCGGTCTTCGGTTTCCTTC, R: GTTCAGCTC
GGTCACGTTCTT; 14-3-3ε F: CGCACCCCATTCGTTTAGG, R: ATTCTGCTCTTCACCATCACC; 14-3-3z F: CTACGATCACG
TCCAACCCG, R: GTCAAACGCTTCTGGCTGC; 14-3-3s F: ACAACCTGACACTGTGGACG, R: CCTTTGGAGCAAGAACAGCG; hu-
man TARDBP F: CTGCGGGAGTTCTTCTCTCA; human TARDBP R: CGCAATCTGATCATCTGCAA; murine Tardbp F: AATCAG
GGTGGGTTTGGTAACA; murine Tardbp R: GCTGGGTTAATGCTAAAAGCAC; Gapdh F: GTGAAGGTCGGTGTGAAC; R: ATCT
CCACTTTGCCACTGCAA. Human STMN2 and reference primer sequences were previously published.42

RT-PCR
RNA was extracted from mouse brains and cDNA synthesized as outlined above. Primers and PCR conditions were adopted from
Jeong YH et al.51 Bands were separated by agarose gel electrophoresis and visualized using SYBR safe gel stain (Thermo) and a
GelDoc imaging system (BioRad). Band intensities of cryptic Ap3b2 amplicons were quantified with Fiji Image J (NIH) and normalized
to Gapdh band intensities.

3D modelling
The putative TDP-43 interaction site was visualized on a 3D structure of 14-3-3q (Protein Data Bank Europe entry #5iqp)71 using the
Swiss-Pdb Viewer version 3.7.

Sequential extraction
Sequential extraction with RIPA followed by urea buffers of cell and brain extracts has been done as described previously.39 50–
100 mg of cortical tissue was solubilized in modified RIPA buffer (50 mM Tris–Hcl
(pH 8), 150 mM NaCl, 1 % Triton X, 0.5 % Na-deoxycholate, 1 % sodium dodecyl sulfate, 3 mM EDTA and protease inhibitor
(Roche)) using a motorized tissue douncer. The homogenate is sonicated at 20% amplitude for 30 short pulses, incubated at 4 C
for 30 mins while rotating, followed by high-speed centrifugation of the supernatant at 50,0003g, 4 C for 1 hour. The resulting su-
pernatant constitutes the RIPA-soluble fraction. Pellets were resuspended in RIPA buffer homogenised and high-speed centrifuged
as above for 2 more rounds to clear any remaining RIPA-soluble proteins before the final pellet is resuspended in UREA buffer (7 M
UREA, 2 M thiourea, 4 % Chaps 3 mM Tris, pH 8.5). The UREA homogenate is sonicated at 20% amplitude for 30 pulses and high
speed centrifuged as above. The resulting supernatant constitutes the RIPA-insoluble/UREA soluble fraction. The protein concen-
trations were determined using the Bradford assay (BioRad). Samples were aliquoted, snap frozen and stored at -80 C until further
use.

Motor and Behavioral Testing

Wire test was performed as previously described.72 Briefly, mice were placed on a wire mesh and allowed to hang upside down,
latency to fall off was recorded. Rotarod testing was done as previously described.39 Grip Strength was performed as previously
described72 using a grip strength meter to measure peak forearm strength (Chatillon, AMETEK). Disinhibition and anxiety were as-
sessed by Elevated Plus Maze testing as previously described.39 Activity of mice was determined by Open Field testing for 10 minutes
in a 40 cm square arena as outlined before.39 All behavioral and motor testing was done blinded to the genotype of mice and videos
were analyzed after testing session using the AnyMaze software package (Stoelting). Phenotype development of rNLS mice was
monitored during weekly inspection from day 21 rating gait, tremor, limb clenching and paralysis. Full hindlimb paralysis was the
ethical euthanasia criterion.

Immunohistochemistry
Staining of paraffin tissue sections, including antigen retrieval has been described previously.73 Primary antibodies (source and di-
lutions provided in brackets) used for staining were against human TDP-43 (ProteinTech; Cat. No. 60019-2-IG; 1:400), pan-TDP-43
(ProteinTech; Cat. No. 10782-2-AP; 1:400), phosphorylated TDP-43 (CosmoBio; CAC-TIP-PTD-P02; 1:2,000), NeuN (Merck Milli-
pore; Cat. No. ABN91; 1:1,000), mCherry (Abcam; Cat. No. AB167453; 1:800), GFP (Abcam; Cat. No. AB6673; 1:500), V5 (Sigma;
Cat. No. V8137; 1:1,000), mTDP-43 (generous gift from Dr. Virginia Lee; 1:100), 14-3-3q (Abcam; Cat. No. AB10439; 1:400) and
PSMA1 (Invitrogen; Cat. No. MA5-32806; 1:500). Secondary antibodies used were Alexa-Fluor coupled 488, 555 and 647 (Life Tech-
nologies) for immunofluorescence staining and coupled to HRP for conventional immunohistochemistry with DAB.73 Proximity Liga-
tion Assay (PLA) was done with the Duolink PLA Brightfield kit according to the manufacturer’s instructions (Millipore Sigma) using

Neuron 112, 1249–1264.e1–e8, April 17, 2024 e7

 ll
Article

primary antibodies to human TDP-43 (ProteinTech; Cat. No. 10782-2-AP; 1:400) and 14-3-3q (Abcam; Cat. No. AB10439; 1:400).
Stained slides were scanned using a Zeiss AxioScan Z1 microscope and images processed with the Zeiss ZEN software.

QUANTIFICATION AND STATISTICAL ANALYSIS

Quantification
Western blot band intensities were quantified with Fiji (Image J, NIH) using non-adjusted images exported in TIFF format from the
ImageLab software (BioRad) of the ChemiDoc MP imaging system (BioRad). Immuno-staining of histological tissue sections were
quantified either by counting cells of interest on scanned sections using the ZEN software (Zeiss), or by measuring fluorescence in-
tensity of selected cells using the measure function of Fiji (Image J, NIH).

Statistical analysis
Statistical analysis was done with GraphPad Prism 10.0. Student’s t-tests were used for comparing two groups, and ANOVA for multi
group comparison. N-numbers are provided in figure legends or as individual values superimposed on bar graphs. Details on test
parameters and individual tests used for each experiment displayed in the main figures have been listed in the Table S3.

e8 Neuron 112, 1249–1264.e1–e8, April 17, 2024