Broad brain biodistribution conferred by an AAV to restore TDP-43 function mitigates Frontotemporal Demenia-like deficits

32 TDP-43 dysfunction is an early pathogenic determinant of frontotemporal lobar 33 degeneration with TDP -43 pathology (FTLD-TDP), a devastating disorder currently 34 without effective therapy . Here, we exploit a blood-brain-barrier (BBB)-permeable AAV 35 (AAV-PHP.eB) that confers broad brain biodistribution to restore TDP-43 function in a 36 TDP-43 deficient model (CamKIIa-CreER;Tardbp mice) that mimics the early stage of 37 TDP-43 dysfunction occurring in FTLD-TDP. Intracerebroventricular delivery by AAV -38 PHP.eB of CTR, our previously characterized splicing repressor, revealed its 39 accumulation in ~40% of adult hippocampal neurons. Remarkably, treatment of adult 40 CamKIIa-CreER;Tardbpf/f mice with AAV-PHP.eB-CTR restored TDP -43 function, 41 attenuated neuronal aberrant activity and memory deficits, and rescued neuron loss. 42 Importantly, we showed that TDP-43's autoregulatory element restricts CTR expression 43 to a physiological range. No overt phenotype was observed after long-term exposure to 44 AAV-PHP.eB-CTR in aged mice , highlighting a favorable safety profile for this gene 45 therapy. These results validate that BBB-crossing AAVs can deliver CTR with a 46 biodistribution in the adult brain that is broad enough to rescue FTD-like phenotypes, 47 supporting clinical testing of this gene therapy for FTLD-TDP. 48

49 Emerging evidence1–3 supports the notion that loss of nuclear TAR DNA/RNA-binding 50 protein 43kDa ( TARDBP, TDP-43) and its splicing repression underlies frontotemporal 51 dementia (FTD) and amyotrophic lateral sclerosis (ALS), devastating adult -onset 52 neurodegenerative disorders4–12 currently without disease modifying therapy. TDP -43, a 53 highly conserved nuclear RNA binding protein (RBP), was first proposed to primarily 54 induce cytoplasmic neuronal inclusions and drive neuron loss through such aggregates 55 in this disease spectrum13. A major neuronal function of TDP-43 was subsequently shown 56 to be the regulation of cryptic exons splicing, the loss of which is implicated in ALS-FTD1. 57 This discovery led to the proposal that deficits in TDP-43 cryptic splicing can be an early 58 pathogenic event that drive s neuron loss. This view is supported by observation s that 59 cryptic exons encoded peptides, such as that found in cryptic HDGFL2, can be found in 60 biofluids of ALS -FTD patients during early and presymptomatic stages of disease, 61 indicative of early loss of TDP -43 function 2,3. Findings of TDP -43 splicing dysfunction 62 preceding cytoplasmic inclusion in the human aging brain by at least a decade 14 also 63 supports this view. This model is further strengthened by the observations that mutations 64 in TDP-43 linked to ALS 15–17 impact a TDP -43 cryptic exon of stathmin-2 (STMN2) in 65 human iPSC derived neurons independent of TDP -43 cytoplasmic aggregates 18–20. 66 UNC13A, another TDP-43 target, encodes a strong risk allele for ALS and FTLD -TDP 67 which influences the inclusion of its cryptic exon21,22. Observations of neuronal TDP -43 68 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint nuclear depletion in a presymptomatic C9orf72 patient 23 and some FTLD-TDP cases24, 69 along with cryptic exons in FTLD -TDP and AD -TDP cases25,26, all without cytoplasmic 70 aggregates, provide additional data. These data support a model that deficits associated 71 with TDP-43 cryptic splicing is triggered during the presymptomatic stage of disease to 72 drive neuron loss. Thus, therapeutic strategies designed to restore TDP -43 function 73 represent a potential mechanism-based therapy for FTLD-TDP. 74 Using our conditional knockout mouse model with TDP-43 depletion in forebrain 75 neurons ( CamKIIa-CreER;Tardbpf/f) that exhibits selective neuron loss mimicking the 76 prodromal phase of FTLD -TDP, we previously showed that loss of TDP -43 dependent 77 cryptic splicing leads to activation of caspase -327,28, forebrain circuit abnormalities and 78 memory deficits29. Importantly, we developed an AAV -based therapeutic strategy to 79 restore the loss of splicing repression by TDP-43 using a novel chimeric splicing repressor, 80 termed CTR1,30. To validate the therapeutic efficacy and potential toxicity associated with 81 such AAV -mediated delivery of CTR to forebrain neurons , we delivered CTR 82 intracerebroventricularly to adult CamKIIa-CreER;Tardbpf/f mice utilizing a BBB crossing 83 serotype, AAV-PHP.eB31,32 that facilitates efficient transduction of adult central neurons. 84 We report here the safety and efficacy of AAV-PHP.eB-CTR in restoring TDP-43 function 85 and attenuating forebrain neuron loss and rescuing neuronal circuit and cognitive deficits. 86 Importantly, the inclusion of TDP -43 autoregulatory element as a “safety switch” in the 87 payload ensured CTR levels remained within the normal range in hippocampal neurons. 88 Furthermore, long-term exposure of AAV-PHP.eB-CTR in aged mice showed no evidence 89 of untoward events. These outcomes strongly support the favorable safety profile of this 90 gene therapy. Together, this validation serves as a crucial step toward establishing the 91 clinical viability of AAV-mediated CTR delivery as a potential therapy for FTLD-TDP. 92

93 Intracerebroventricular delivery of AAV -PHP.eB-CTR broadly transduces adult 94 forebrain neurons without excessive accumulation of CTR 95 We previously demonstrated that AAV9 delivery of CTR to P0 mice lacking TDP -43 96 in spinal motor neurons can complements TDP-43 loss of function (LOF)30. Build on this, 97 the current study asked whether restoration of TDP -43’s splicing repression function 98 could mitigate downstream consequences of TDP -43 loss in the CamKIIa-99 CreER;Tardbpf/f mouse model , which recapitulates early features of FTLD-TDP, while 100 also avoiding the known risks associated with TDP -43 overexpression and aggregation 101 toxicity33. To achieve this, we designed a chimeric therapeutic construct , termed CTR1, 102 which retains the N-terminal RNA recognition domains (RRM1 and RRM2, amino acids 103 1–267, ~30 kDa) of TDP -43 to preserve its RNA -binding specificity, while replacing the 104 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint glycine-rich C-terminal domain, which is implicated in pathological aggregation, with an 105 unrelated splicing repressor domain from RAVER134–36 (amino acids 450–643, ~20 kDa), 106 an unrelated RBP(Supplementary figure 1C). 107 Physiological regulation of expression was ensured by incorporating the 3′ 108 untranslated region (3 ′UTR) of human TARDBP, which contains the polyadenylation-109 dependent autoregulatory element that enables TDP -43 to bind its own transcript and 110 limit its expression levels 37,38 (Supplementary figure 1 C). This negative feedback 111 mechanism is critical for maintaining normal nuclear TDP-43 homeostasis and preventing 112 toxic accumulation and essential for neuronal survival39. Our RT-PCR analysis (Fig. 2G) 113 confirmed that CTR expression is higher in forebrains of TDP -43–deficient ( CamKII-114 CreER;Tardbpf/f) animals as compared to that of Tardbpf/f control littermates, indicating 115 preserved autoregulation via this 3′UTR of TARDBP. 116 To validate such ability of CTR to rescue cell death of forebrain neurons lacking TDP-117 4327,29, we delivered intracerebr oventricularly (ICV) AAV9-CTR to P0 CamKIIa-118 CreER;Tardbpf/f pups. We found that AAV9 -CTR attenuates neuron loss occurring in 119 CA2/3 hippocampal neurons of mice lacking TDP -43 ( Supplementary figure 2A-C). 120 Corroborating these results, we show ed that TDP-43 cryptic exons are suppressed in 121 rescued mice (Supplementary figure 2D-F), confirming that the failure to repress TDP-122 43 cryptic exons in central neurons underlies neuron loss. However, for clinically relevant 123 context, it is critical to establish delivery of CTR using an AAV serotype that broadly 124 transduces central neurons to restore TDP-43 LOF in the adult brain. 125 For efficient CTR delivery in adult forebrain neurons of CamKIIa-CreER;Tardbpf/f mice, 126 we used a BBB crossing AAV-PHP.eB that facilitates efficient transduction of adult central 127 neurons31,32 via the ICV route of administration. To model the early stage of TDP-43 LOF 128 occurring in human disease 2, we selectively deleted TDP -43 from forebrain neurons of 129 adult CamKIIa-CreER;Tardbpf/f mice. Upon tamoxifen administration, Cre recombinase 130 was activated in Camk2a-positive excitatory neurons, resulting in the excision of exon 3 131 in Tardbp and subsequent loss of TDP -43 in targeted regions of the brain 132 (Supplementary figure 1A-B). Immunostaining confirmed robust nuclear depletion of 133 TDP-43 in the hippocampus one month after tamoxifen exposure in 5 month-old CamKIIa-134 CreER;Tardbpf/f mice, whereas control littermates ( Tardbpf/f mice) maintained normal 135 level of TDP-43 (Fig. 1A). 136 Delivery of AAV-PHP.eB-CTR to adult brain restores splicing repression in TDP-43 137 deficient hippocampal neurons 138 To evaluate the in vivo expression and function of the chimeric repressor CTR, we 139 first assessed its distribution following ICV delivery of AAV -PHP.eB–CTR in CamKIIa-140 CreER;Tardbpf/f or Tardbpf/f mice. Immunohistochemical analysis using an antibody 141 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint specific to the N-terminus of human TDP-43 revealed widespread expression of the CTR 142 fusion protein throughout the hippocampus, with robust signal detected in CA2 and CA2/3 143 subregions and detectable expression in CA1 ( Fig. 1B); quantification of transduction 144 efficiency showed a stable transduction rate of approximately 40% of neurons in 145 CamKIIa-CreER;Tardbpf/f mice (Fig. 1C). 146 We next asked whether CTR expression was sufficient to restore TDP -43’s splicing 147 repressor function in vivo . RT-PCR of hippocampal tissue from treated CamKIIa-148 CreER;Tardbpf/f mice revealed a marked reduction in a panel of TDP-43–regulated cryptic 149 exons, including those in Adnp2, Ap3b2, Bud23, Camk1g, Cerm, Ggct, Unc13a, Synj2bp, 150 Tbc1d1, Usp15, Tecpr1, and Washc4 (Fig. 1E,F), indicating broad repression of aberrant 151 cryptic splicing. These cryptic exons span a range of genes involved in neuronal function, 152 and are known to be directly regulated by TDP-43 in the mouse brain. Their coordinated 153 repression by CTR suggests that the chimeric construct can functionally reconstitute 154 TDP-43’s splicing repressor activity across multiple endogenous targets. This broad 155 repression pattern reflects the molecular reach of CTR in restoring splicing fidelity in TDP-156 43–deficient CNS neurons when delivered to the adult brain. 157 To further validate cryptic exon repression at the cell ular level, we performed in situ 158 hybridization using a BaseScope probe targeting the cryptic exon in mouse Unc13a. 159 UNC13A is a known ALS genetic risk factor and harbors a TDP -43–sensitive cryptic 160 exons that are not conserved between mouse (within intron 1) and human (within intron 161 22)22. Upon TDP-43 depletion, cryptic Unc13a RNA accumulated in multiple hippocampal 162 regions, with the highest signal in the dentate gyrus (DG), followed by CA1 and CA2/3 163 (Fig. 1D). This regional pattern may reflect differential vulnerability to TDP-43 depletion 164 within the hippocampus. Notably, CTR treatment of CamKIIa-CreER;Tardbpf/f significantly 165 reduced cryptic Unc13a expression in all subregions as compared to those of GFP-166 treated CamKIIa-CreER;Tardbpf/f mice (Fig. 1D), providing evidence that CTR effectively 167 represses disease-relevant cryptic exons in vivo. 168 To ensure CTR expression is subject to TDP -43–like autoregulation, we analyzed 169 CTR mRNA levels using RT-PCR with a forward primer targeting the TDP-43 N-terminus 170 and a reverse primer targeting the RAVER1 C-terminus. Despite equivalent AAV doses, 171 CTR transcript levels were significantly higher in CamKIIa-CreER;Tardbpf/f mice as 172 compared to that of Tardbpf/f control littermates (Fig. 1G), consistent with the increased 173 protein expression observed by immunohistochemi cal analysis (Fig. 1B). This result 174 suggests that CTR retains the autoregulatory features of endogenous TDP -43 via its 175 3′UTR, leading to transcript stabilization and translation in the context of TDP-43 loss. 176 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint AAV-PHP.eB-CTR ameliorates CA2/3 neuron loss and hippocampal atrophy 177 To assess the long-term neuroprotective effects of CTR in the context of TDP-43 loss, 178 we examined neuronal survival in the hippocampus of CamKIIa-CreER;Tardbpf/f mice 12 179 months post-treatment. Immunohistochemical analysis using antibody recognizing NeuN 180 revealed severe neuronal loss of ~60% in the CA2/3 region of CamKIIa-CreER;Tardbpf/f 181 mice treated with AAV-PHP.eB-GFP, with near-complete depletion of NeuN-positive cells 182 compared to control Tardbpf/f mice (Fig. 2A). In contrast, ~70% of neurons survive in 183 CamKIIa-CreER;Tardbpf/f mice treated with AAV-PHP.eB-CTR, (Fig. 2C). Indicating that, 184 on average, AAV-PHP.eB-CTR treated CamKIIa-CreER;Tardbpf/f mice protected ~30% 185 of the CA2/3 neurons from death , which closely matched the local transduction rate 186 observed for AAV -PHP.eB-CTR (Fig. 2C). The extent of neuronal rescue in CA2/3 is 187 consistent with the ~40% transduction rate in this area, supporting a direct impact 188 between CTR accumulation and cell survival. 189 Importantly, long-term AAV-PHP.eB-CTR exposure to Tardbpf/f mice—where TDP-43 190 expression remains normal—did not lead to any overt phenotypes, including reduction in 191 hippocampal neuron number or area as compared to those observed in untreated or AAV-192 PHP.eB-GFP treated Tardbpf/f control littermates. These findings support the idea that 193 AAV-PHP.eB-CTR is highly efficacious, well-tolerated and non -toxic, highlighting its 194 safety profile as a promising therapy to be evaluated for its clinical impact in FTLD -TDP 195 patients. 196 To evaluate the temporal dynamics and durability of CTR-mediated neuroprotection, 197 we also quantified CA2/3 neurons at different timepoints—6 months and 12 months post-198 treatment (Supplementary figure 3A-C). At 6 months, CA1 neuron numbers did not differ 199 among groups, whereas CA2/3 neurons were significantly reduced in CamKIIa-200 CreER;Tardbpf/f mice relative to Control . This CA2/3 loss at 6 months was blunted by 201 CTR (CamKIIa-CreER;Tardbpf/f + CTR > CamKIIa-CreER;Tardbpf/f + GFP; partial rescue 202 toward Control , P= 0.0536) At 12 months, CA 2/3 neuron numbers in CamKII-203 CreER;Tardbpf/ + CTR mice were significantly higher compare to in controls ( Fig. 2C), 204 indicating a clear rescue at this later time point. These findings suggest that 205 neurodegeneration progresses gradually in this model and that AAV-PHP.eB-CTR 206 provides sustained protection across the progression of disease. Notably, the 207 preservation of neuron s in AAV-PHP.eB-CTR treated CamKIIa-CreER;Tardbpf/f mice 208 persisted up to 12 months post -treatment which corresponds to the late adult phase of 209 the mouse, underscoring the long-term efficacy of this therapeutic strategy. 210 Outside the CA2/3 region, neuronal loss was not consistently observed in other 211 hippocampal subfields in our model (Supplementary figure 3A, C). However, we noted 212 a marked reduction in overall hippocampal size in CamKIIa-CreER;Tardbpf/f mice treated 213 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint with AAV -PHP.eB-GFP as compared to Tardbpf/f control littermates. Morphometric 214 analysis of CV-stained sections confirmed a significant rescue of hippocampal area in the 215 AAV-PHP.eB-CTR treated group (Fig. 2D; Supplementary figure 4), indicating that CTR 216 expression alleviates TDP-43–dependent neurodegeneration at both cellular and 217 anatomical levels. 218 AAV-PHP.eB-CTR rescues social and cognitive deficits 219 Since we showed behavioral deficits in CamKIIa-CreER;Tardbpf/f mice29, we asked 220 whether AAV-PHP.eB-CTR treatment rescues behavioral impairments associated with 221 TDP-43 depletion. As before, we conducted a battery of behavioral tests approximately 222 4-5 months after AAV-PHP.eB-CTR treatment. During the first round of behavioral testing, 223 both male and female CamKIIa-CreER;Tardbpf/f mice and Tardbpf/f littermate controls 224 underwent assessments including the open field test, light –dark box test, novel object 225 recognition (NOR) test, and the social behavior test. No significant differences were 226 observed in locomotor activity or anxiety -related behavior between genotypes or 227 treatment groups in the open field and light–dark box tests(Supplementary figure 5A-E). 228 However, deficits emerged in the novel object and social novelty preference paradigms. 229 In the NOR test, CamKIIa-CreER;Tardbpf/f mice with AAV-PHP.eB-CTR or control AAV-230 PHP.eB-RFP treatment spent equal time exploring familiar and novel objects, indicating 231 impaired object recognition memory. In contrast, CamKIIa-CreER;Tardbpf/f mice treated 232 with AAV-PHP.eB-CTR displayed restored object novelty preference, like Tardbpf/f control 233 littermates (Fig. 3A, D). 234 A comparable rescue pattern was observed in the social behavior test. During the 235 sociability phase, all groups exhibited normal preference for a novel conspecific over an 236 empty enclosure. However, during the social novelty phase, CamKIIa-CreER;Tardbpf/f 237 mice treated with AAV -PHP.eB-RFP failed to show a preference for a second novel 238 mouse over the relatively familiar conspecific, consistent with impaired social recognition. 239 This behavioral deficit was rescued in AAV-PHP.eB-CTR treated CamKIIa-240 CreER;Tardbpf/f mice, who displayed normal social novelty preference, comparable to 241 Tardbpf/f control animals ( Fig. 3B, E). These results indicate that AAV-PHP.eB-CTR 242 treatment reverses specific cognitive and social deficits associated with TDP -43 loss in 243 forebrain neurons. 244 AAV-PHP.eB-CTR restores neuronal calcium activity impaired by TDP-43 loss 245 Following the initial behavioral assessments, mice underwent brain surgeries 246 including viral injection of AAV -CaMKII-GCaMP6f and GRIN lens implantation targeting 247 the prelimbic cortex to enable longitudinal in vivo calcium imaging. Perioperative 248 outcomes are summarized in Supplementary Table 1: 7 of 41 male mice (17%) and 5 of 249 26 female mice (19%) died during or shortly after surgery. Of the 67 mice subjected to 250 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint both procedures, 30 (45%) yielded high -quality calcium imaging data, with each 251 experimental group represented. 252 To quantify neuronal activity, 15-minute calcium imaging recordings were processed 253 using a custom Python and MATLAB pipeline. The firing frequency of active neurons was 254 extracted and pooled across sexes for analysis. TDP-43 LOF in CamKIIa-CreER;Tardbpf/f 255 +RFP mice significantly reduced the frequency of calcium transients compared to that of 256 Tardbpf/f control littermates (Fig. 3C), indicating functional impairment of excitatory 257 neuronal activity. Remarkably, AAV-PHP.eB-CTR treatment restored calcium firing 258 frequency to levels comparable to those observed in Tardbpf/f control mice ( Fig. 3C). 259 These results suggest that AAV-PHP.eB-CTR rescues not only behavioral deficits, but 260 also forebrain circuit abnormalities. 261 Together, these results demonstrate that AAV-PHP.eB-CTR not only corrects 262 molecular splicing defects but also preserve s forebrain neuronal circuit activity and 263 restores behavioral performance, supporting its therapeutic potential for testing in the 264 clinic for FTLD-TDP. 265

266 Our findings establish the success of a chimeric splicing repressor (CTR) designed 267 to restore TDP -43’s splicing repression activity in vivo while avoiding its aggregation. 268 Using a conditional TDP-43 knockout mouse model that mimics the early stage of FTLD-269 TDP, we demonstrate that ICV delivery of AAV-PHP.eB-CTR resulted in broad repression 270 of cryptic exon inclusion, long-term rescue of neuronal survival in vulnerable hippocampal 271 neurons, and restoration of cognitive behavior and neuronal calcium dynamics. These 272 findings validate CTR as a disease-modifying payload capable of restoring key functional 273 aspects of TDP-43 in the adult forebrain and provide preclinical proof -of-concept for an 274 AAV serotype amenable for broad biodistribution to the adult brain designed to restore 275 TDP-43 function for C9orf72-FTD or FTLD-TDP. 276 Current therapeutic strategies aimed at mitigating TDP -43 loss-of-function primarily 277 rely on antisense oligonucleotides (ASOs) designed to suppress individual cryptic exons, 278 such as those found in STMN2 and UNC13A19,40–42. While these approaches have 279 demonstrated transcript -level correction and partial phenotypic rescue in preclinical 280 models, they are inherently constrained in scope. Each ASO targets a single downstream 281 consequence of TDP-43 dysfunction, necessitating multiple independent interventions to 282 address complex disease phenotypes43,44. This piecemeal approach may be insufficient 283 to restore network-level RNA homeostasis in the affected neurons. 284 In contrast, CTR represents a unified strategy that intervenes at the level of the 285 splicing repression mechanism itself. Rather than correcting the results of TDP -43 286 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint dysfunction one transcript at a time, CTR restores the upstream regulatory function that 287 governs a broad range of cryptic exon inclusion, as well as other functions such as 288 alternative polyadenylation. This distinction not only provides therapeutic breadth but also 289 aligns more closely with the pathophysiology of the disease. 290 Moreover, whereas full -length TDP -43 re -expression poses risks due to its 291 aggregation-prone domains and overexpression toxicity, CTR is structurally designed to 292 mitigate these issues1,44. CTR avoids reintroduction of the aggregation-prone C-terminal 293 domain of TDP -43, which has been implicated in pathological phase separation and 294 neurotoxicity45. Moreover, by retaining the 3 ′UTR of human TARDBP, CTR preserves 295 autoregulatory feedback, minimizing the risk of overexpression —a critical safety feature 296 especially in the context of viral gene delivery39. By combining mechanistic precision with 297 reduced off-target liability, CTR offers a complementary and potentially more versatile 298 alternative to ASO -based interventions—particularly in disorders like FT LD-TDP or C9 -299 FTD where multiple gene networks are disrupted in parallel. 300 Finally, this demonstrated efficiency of AAV-PHP.eB in CNS-wide transduction of 301 adult mice highlights the potential of developing new BBB permeable AAV serotypes to 302 enable efficient delivery of payload to the human adult bra in46,47 to include other 303 neurodegenerative disorders exhibiting pathology of TDP -43 such as LATE 48, and AD -304 TDP49–51 Moreover, the therapeutic efficacy of CTR appears tightly linked to local rate of 305 transduction, emphasizing the need for improved delivery platforms for clinical 306 application—such as next -generation capsids with broader tropism and non -invasive 307 administration routes52,53. 308

309 1. Ling, J. P., Pletnikova, O., Troncoso, J. C. & Wong, P. C. TDP -43 repression of 310 nonconserved cryptic exons is compromised in ALS -FTD. Science 349, 650 –655 311 (2015). 312 2. Irwin, K. E. et al. A fluid biomarker reveals loss of TDP -43 splicing repression in 313 presymptomatic ALS–FTD. Nat. Med. 30, 382–393 (2024). 314 3. Seddighi, S. et al. Mis-spliced transcripts generate de novo proteins in TDP-43–related 315 ALS/FTD. Sci. Transl. Med. 16, eadg7162 (2024). 316 4. Renton, A. E. et al. A hexanucleotide repeat expansion in C9ORF72 is the cause of 317 chromosome 9p21-linked ALS-FTD. Neuron 72, 257–268 (2011). 318 5. Gao, F.-B., Richter, J. D. & Cleveland, D. W. Rethinking Unconventional Translation in 319 Neurodegeneration. Cell 171, 994–1000 (2017). 320 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint 6. Elahi, F. M. & Miller, B. L. A clinicopathological approach to the diagnosis of dementia. 321 Nat. Rev. Neurol. 13, 457–476 (2017). 322 7. Grossman, M. et al. Frontotemporal lobar degeneration. Nat. Rev. Dis. Primer 9, 40 323 (2023). 324 8. DeJesus-Hernandez, M. et al. Expanded GGGGCC hexanucleotide repeat in 325 noncoding region of C9ORF72 causes chromosome 9p -linked FTD and ALS. Neuron 326 72, 245–256 (2011). 327 9. Mead, R. J., Shan, N., Reiser, H. J., Marshall, F. & Shaw, P. J. Amyotrophic lateral 328 sclerosis: a neurodegenerative disorder poised for successful therapeutic translation. 329 Nat. Rev. Drug Discov. 22, 185–212 (2023). 330 10. Benatar, M. et al. The Miami Framework for ALS and related neurodegenerative 331 disorders: an integrated view of phenotype and biology. Nat. Rev. Neurol. 20, 364–376 332 (2024). 333 11. Taylor, J. P., Brown, R. H. & Cleveland, D. W. Decoding ALS: from genes to 334 mechanism. Nature 539, 197–206 (2016). 335 12. Mehta, P. R., Brown, A. -L., Ward, M. E. & Fratta, P. The era of cryptic exons: 336 implications for ALS-FTD. Mol. Neurodegener. 18, 16 (2023). 337 13. Neumann, M. et al. Ubiquitinated TDP-43 in frontotemporal lobar degeneration and 338 amyotrophic lateral sclerosis. Science 314, 130–133 (2006). 339 14. Chang, K. et al. Loss of TDP -43 splicing repression occurs early in the aging 340 population and is associated with Alzheimer’s disease neuropathologic changes and 341 cognitive decline. Acta Neuropathol. (Berl.) 147, 4 (2023). 342 15. Sreedharan, J. et al. TDP-43 mutations in familial and sporadic amyotrophic lateral 343 sclerosis. Science 319, 1668–1672 (2008). 344 16. Kabashi, E. et al. TARDBP mutations in individuals with sporadic and familial 345 amyotrophic lateral sclerosis. Nat. Genet. 40, 572–574 (2008). 346 17. Gitcho, M. A. et al. TDP-43 A315T mutation in familial motor neuron disease. Ann. 347 Neurol. 63, 535–538 (2008). 348 18. Melamed, Z. et al. Premature polyadenylation -mediated loss of stathmin -2 is a 349 hallmark of TDP-43-dependent neurodegeneration. Nat. Neurosci. 22, 180–190 (2019). 350 19. Klim, J. R. et al. ALS-implicated protein TDP-43 sustains levels of STMN2, a mediator 351 of motor neuron growth and repair. Nat. Neurosci. 22, 167–179 (2019). 352 20. Prudencio, M. et al. Truncated stathmin -2 is a marker of TDP -43 pathology in 353 frontotemporal dementia. J. Clin. Invest. 130, 6080–6092 (2020). 354 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint 21. Brown, A.-L. et al. TDP-43 loss and ALS-risk SNPs drive mis-splicing and depletion 355 of UNC13A. Nature 603, 131–137 (2022). 356 22. Ma, X. R. et al. TDP-43 represses cryptic exon inclusion in the FTD –ALS gene 357 UNC13A. Nature 603, 124–130 (2022). 358 23. Vatsavayai, S. C. et al. Timing and significance of pathological features in C9orf72 359 expansion-associated frontotemporal dementia. Brain J. Neurol. 139, 3202 –3216 360 (2016). 361 24. Nana, A. L. et al. Neurons selectively targeted in frontotemporal dementia reveal early 362 stage TDP-43 pathobiology. Acta Neuropathol. (Berl.) 137, 27–46 (2019). 363 25. Sun, M. et al. Cryptic exon incorporation occurs in Alzheimer’s brain lacking TDP-43 364 inclusion but exhibiting nuclear clearance of TDP -43. Acta Neuropathol. (Berl.) 133, 365 923–931 (2017). 366 26. Calliari, A. et al. HDGFL2 cryptic proteins report presence of TDP -43 pathology in 367 neurodegenerative diseases. Mol. Neurodegener. 19, 29 (2024). 368 27. LaClair, K. D. et al. Depletion of TDP-43 decreases fibril and plaque β -amyloid and 369 exacerbates neurodegeneration in an Alzheimer’s mouse model. Acta Neuropathol. 370 (Berl.) 132, 859–873 (2016). 371 28. Baghel, M. S. et al. Depletion of TDP -43 exacerbates tauopathy -dependent brain 372 atrophy by sensitizing vulnerable neurons to caspase 3 -mediated endoproteolysis of 373 tau in a mouse model of Multiple Etiology Dementia. 2024.06.26.600814 Preprint at 374 https://doi.org/10.1101/2024.06.26.600814 (2024). 375 29. Liang, B. et al. Aberrant neural activity in prefrontal pyramidal neurons lacking TDP -376 43 precedes neuron loss. Prog. Neurobiol. 215, 102297 (2022). 377 30. Donde, A. et al. Splicing repression is a major function of TDP -43 in motor neurons. 378 Acta Neuropathol. (Berl.) 138, 813–826 (2019). 379 31. Chan, K. Y. et al. Engineered AAVs for efficient noninvasive gene delivery to the 380 central and peripheral nervous systems. Nat. Neurosci. 20, 1172–1179 (2017). 381 32. Goertsen, D. et al. AAV capsid variants with brain -wide transgene expression and 382 decreased liver targeting after intravenous delivery in mouse and marmoset. Nat. 383 Neurosci. 25, 106–115 (2022). 384 33. Wils, H. et al. TDP-43 transgenic mice develop spastic paralysis and neuronal 385 inclusions characteristic of ALS and frontotemporal lobar degeneration. Proc. Natl. 386 Acad. Sci. U. S. A. 107, 3858–3863 (2010). 387 34. Rideau, A. P. et al. A peptide motif in Raver1 mediates splicing repression by 388 interaction with the PTB RRM2 domain. Nat. Struct. Mol. Biol. 13, 839–848 (2006). 389 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint 35. Gromak, N. et al. The PTB interacting protein raver1 regulates alpha -tropomyosin 390 alternative splicing. EMBO J. 22, 6356–6364 (2003). 391 36. Pongrácová, E., Buratti, E. & Romano, M. Prion-like Spreading of Disease in TDP-43 392 Proteinopathies. Brain Sci. 14, 1132 (2024). 393 37. Budini, M. & Buratti, E. TDP -43 Autoregulation: Implications for Disease. J. Mol. 394 Neurosci. 45, 473–479 (2011). 395 38. Sun, Y., Arslan, P. E., Won, A., Yip, C. M. & Chakrabartty, A. Binding of TDP -43 to 396 the 3′UTR of Its Cognate mRNA Enhances Its Solubility. Biochemistry 53, 5885–5894 397 (2014). 398 39. Ayala, Y. M. et al. TDP-43 regulates its mRNA levels through a negative feedback 399 loop. EMBO J. 30, 277–288 (2011). 400 40. Baughn, M. W. et al. Mechanism of STMN2 cryptic splice -polyadenylation and its 401 correction for TDP-43 proteinopathies. Science 379, 1140–1149 (2023). 402 41. Keuss, M. J. et al. Loss of TDP-43 induces synaptic dysfunction that is rescued by 403 UNC13A splice-switching ASOs. BioRxiv Prepr. Serv. Biol. 2024.06.20.599684 (2024) 404 doi:10.1101/2024.06.20.599684. 405 42. Faller, K. M. E., Chaytow, H. & Gillingwater, T. H. Targeting common disease 406 pathomechanisms to treat amyotrophic lateral sclerosis. Nat. Rev. Neurol. 21, 86–102 407 (2025). 408 43. Ou, K. et al. Application of antisense oligonucleotide drugs in amyotrophic lateral 409 sclerosis and Huntington’s disease. Transl. Neurodegener. 14, 4 (2025). 410 44. Hallegger, M. et al. TDP-43 condensation properties specify its RNA -binding and 411 regulatory repertoire. Cell 184, 4680-4696.e22 (2021). 412 45. Scialò, C. et al. Seeded aggregation of TDP -43 induces its loss of function and 413 reveals early pathological signatures. Neuron S0896-6273(25)00180–1 (2025) 414 doi:10.1016/j.neuron.2025.03.008. 415 46. Huang, Q. et al. An AAV capsid reprogrammed to bind human transferrin receptor 416 mediates brain-wide gene delivery. Science 384, 1220–1227 (2024). 417 47. Reich, M. et al. Peripheral expression of brain -penetrant progranulin rescues 418 pathologies in mouse models of frontotemporal lobar degeneration. Sci. Transl. Med. 419 16, eadj7308 (2024). 420 48. Nelson, P. T. et al. Limbic-predominant age-related TDP-43 encephalopathy (LATE): 421 consensus working group report. Brain J. Neurol. 142, 1503–1527 (2019). 422 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint 49. Josephs, K. A. et al. Updated TDP-43 in Alzheimer’s disease staging scheme. Acta 423 Neuropathol. (Berl.) 131, 571–585 (2016). 424 50. Amador-Ortiz, C. et al. TDP-43 immunoreactivity in hippocampal sclerosis and 425 Alzheimer’s disease. Ann. Neurol. 61, 435–445 (2007). 426 51. Uryu, K. et al. Concomitant TAR -DNA-binding protein 43 pathology is present in 427 Alzheimer disease and corticobasal degeneration but not in other tauopathies. J. 428 Neuropathol. Exp. Neurol. 67, 555–564 (2008). 429 52. Gasca-Salas, C. et al. Blood-brain barrier opening with focused ultrasound in 430 Parkinson’s disease dementia. Nat. Commun. 12, 779 (2021). 431 53. Rezai, A. R. et al. Ultrasound Blood –Brain Barrier Opening and Aducanumab in 432 Alzheimer’s Disease. N. Engl. J. Med. 390, 55–62 (2024). 433 434 435 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint Figures 436 Figure 1. CTR is broadly expressed in neurons of adult brain and restores cryptic 437 exon repression 438 439 (A) Representative immunohistochemistry images showing TDP-43 expression in the 440 hippocampus 1 month after tamoxifen treatment. TDP -43 expression is retained in 441 Tardbpf/f mice, but selectively lost in the TDP-43 conditional knockout (cKO) hippocampus, 442 confirming efficient Cre recombination. Scale bar: 500 μm 443 (B), Representative immunohistochemistry showing widespread CTR protein 444 expression in the hippocampus of TDP-43 CKO and control mice following ICV of AAV-445 CTR. CTR expression was detected using an antibody targeting the human TDP -43 N-446 terminus. Strong expression was observed in the CA2 and CA3 subregions, with 447 additional signal in CA1. 448 (C), Quantification of CTR viral transduction rate in TDP-43 CKO mice at 1-, 3-, and 449 6-months post-injection(n = 6,8,16 respectively). Infection rates remained stable over time 450 (~40%). 451 (D), Representative images of BaseScope in situ hybridization for cryptic Unc13a 452 RNA reveals spatial distribution of splicing defects across hippocampal subregions. 453 Cryptic Unc13a transcripts were elevated in dentate gyrus (DG), CA1, and CA2/3 of TDP-454 43 cKO-GFP mice and markedly reduced in CTR -treated animals , indicative of CTR 455 restoring TDP-43 splicing regulation. 456 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint (E), Representative RT-PCR analysis of hippocampal RNA showing repression of 457 multiple cryptic exon –containing transcripts in CTR -treated TDP-43 CKO mice. 458 Representative targets include Adnp2, Ap3b2, Bud23, Camk1g, Cerm, Ggct, Unc13a, 459 Synj2bp, Tbc1d1, Usp15, Tecpr1, and Washc4. 460 (F) Quantification of cryptic exon inclusion relative to wild -type transcript levels, 461 expressed as the ratio of cryptic exon band intensity to corresponding wild -type band 462 intensity from RT -PCR analyses (as in panel E). TDP -43 cKO mice treated with GFP 463 exhibited a significant increase in cryptic exon usage compared to CTR-treated cKO and 464 control groups, whereas CTR treated cKO mice were indistinguishable from control 465 littermates. 466 (G), RT-PCR quantification of CTR mRNA levels using primers spanning the TDP-43 467 N-terminal and RAVER1 C -terminal sequences. Despite equal AAV dosing, CTR 468 transcript levels were significantly higher in TDP-43 cKO mice than in controls, consistent 469 with autoregulatory stabilization of expression via the Tardbp 3′UTR. 470 Scale bars: a : upper panel 500 μm, lower panel 100 μm; d: 100 μm. Error bars 471 represent mean ± s.e.m. P -values were calculated using two -tailed unpaired t -tests or 472 Mann–Whitney tests as specified in figure panels; n values are defined in the 473 corresponding Methods. 474 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint Figure 2. AAV-PHP.eB-CTR mitigates CA2/3 neuron loss and hippocampal atrophy 475 476 (A), Representative NeuN immunohistochemistry images of the CA2/3 hippocampal 477 subregion in control and TDP-43 CKO mice 12 months after AAV -CTR or AAV -GFP 478 injection. TDP-43 cKO-GFP mice showed severe neuronal loss in this region, whereas 479 TDP-43 cKO-CTR mice retained more NeuN-positive cells. 480 (B), NeuN immunostaining overview of the entire hippocampus showing global 481 anatomical differences across groups. CTR -treated TDP-43 CKO mice exhibited visibly 482 preserved hippocampal structure. 483 (C), Quantification of NeuN+ neurons in CA2/3 at 12 months post -injection. CTR 484 treatment significantly increased neuron counts in TDP-43 CKO mice compared to GFP 485 controls, though not to the level of control animals. Neuron number in TDP-43 CKO-CTR 486 mice approximated the regional CTR infectivity rate (~30%) ( n = 19 (Control), 13 (TDP -487 43 cKO GFP), and 13 (TDP-43 cKO CTR) mice. One-way ANOVA, F(2,42) = 48.26, P < 488 0.0001; Tukey’s post hoc test: Control vs. TDP -43 cKO GFP, P < 0.0001; Control vs. 489 TDP-43 cKO CTR, P < 0.0001; TDP-43 cKO GFP vs. TDP-43 cKO CTR, P = 0.0014.). 490 (D), Quantification of total hippocampal area based on NeuN -stained sections. 491 Control GFP (n = 18) and Control CTR (n = 18) groups had significantly larger 492 hippocampal areas than TDP-43 cKO GFP mice (n = 12) (P < 0.0001). CTR treatment (n 493 = 15) significantly increased hippocampal area in cKO mice compared with GFP -treated 494 cKO animals (P < 0.0001), approaching control group values. 495 All images correspond to 12-month timepoints. 496 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint Scale bars: a: 200 μm, c: 500 μm. Error bars represent mean ± s.e.m. P-values were 497 calculated using one-way ANOVA followed by Tukey’s multiple comparisons test unless 498 otherwise noted. n values and statistical details are provided in Methods. 499 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint Figure 3. AAV-PHP.eB-CTR restores neuronal calcium activity and attenuates 500 memory deficits 501 502 (A), Schematic of the novel object recognition (NOR) test protocol. Mice were 503 exposed to two identical objects during a familiarization phase, followed by a testing 504 phase in which one object was replaced with a novel item. Time investigating each object 505 was measured. 506 (B), Schematic of the social behavior test. Mice underwent three 10 -minute phases: 507 habituation (two empty cages), sociability (one unfamiliar mouse introduced), and social 508 novelty (second unfamiliar mouse added). 509 (C), Analysis of in vivo neuronal activity using calcium imaging in the prelimbic cortex. 510 Compared to Control-RFP controls, TDP-43 cKO-RFP mice exhibited reduced neuronal 511 calcium spike frequency, consistent with impaired excitatory activity. CTR treatment 512 normalized activity levels. Imaging was performed through GRIN lenses following AAV -513 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint GCaMP6f injection. Group sizes: Control -RFP (n = 936 ROIs), Control-CTR (n = 1379 514 ROIs), TDP -43 cKO -RFP (n = 2 285 ROIs), TDP -43 cKO -CTR (n = 2 334 ROIs). **** 515 indicates P < 0.0001, ns = not significant. 516 (D), Quantification of exploration time during the novel object recognition (NOR) test. 517 TDP-43 cKO-RFP mice (n = 17) failed to show a preference for the novel object, indicating 518 impaired recognition memory. CTR treatment (n = 20) restored novel object preference 519 to levels comparable to Control-RFP (n = 16) and Control-CTR (n = 13) groups. 520 (E), Quantification of time spent investigating social targets. While all groups 521 displayed normal sociability, TDP-43 cKO-RFP mice failed to show a preference for social 522 novelty. CTR -treated TDP -43 cKO mice showed restored preference for the novel 523 conspecific, indicating rescue of social memory. 524 Data are mean ± s.e.m. P -values were calculated using Kruskal –Wallis ANOVA 525 followed by Dunn’s post hoc tests or Mann –Whitney tests, as noted in Methods. 526 ****P<0.0001, **P<0.01, *P<0.05, ns = not significant. 527 528 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint

529 Animal Models and Genotyping 530 All animal procedures were approved by the Institutional Animal Care and Use 531 Committee (IACUC) at Johns Hopkins University or University of Wyoming and conducted 532 in accordance with NIH guidelines. To model TDP -43 loss of function in a cell -type–533 specific manner, we utilized a previously established conditional Tardbp knockout mouse 534 line (Tardbpf/f), in which exon 3 of the Tardbp gene is flanked by LoxP sites54(The Jackson 535 Laboratory, stock #017591) .We crossed these mice with the CamkIIa -CreERT2 driver 536 line to generate a tamoxifen -inducible, excitatory neuron –specific Tardbp knockout 537 model55,56 (hereafter referred to as TDP-43 cKO). 538 Experimental cohorts included Tardbpf/f; CamkIIa-CreERT2 mice (referred to as TDP-539 43 CKO) and Tardbpf/f littermates lacking Cre (referred to as Control). Both sexes were 540 used, and groups were balanced by age and sex. Recombination was induced by feeding 541 mice tamoxifen-containing chow (Envigo, TD.130859, 400 mg/kg) for 4 weeks , followed 542 by a 2-week washout period. 543 Mice were housed in a 12 -hour light/dark cycle with ad libitum access to food and 544 water, and group -housed when possible. Genotyping was performed using DNA 545 extracted from ear biopsies. PCR was conducted using primers specific to the loxP -546 flanked Tardbp allele and the Camk2a-CreERT2 transgene: PCR products were analyzed 547 by agarose gel electrophoresis. Genotyping Primers are as follows: 548 CamKIIa-Cre F： GACAGGCAGGCCTTCTCTGAA 549 CamKIIa-Cre R: CTTCTCCACACCAGCTGTGGA 550 Tardbp FF F: AACTTCAAGATCTGACACCCTCCCC 551 Tardbp FF R: GGCCCTGGCTCATCAAGAACTG; 552 The expected product of the CamkIIa-Cre primers is 536bp. The expected products 553 of the Tardbp FF primers are 376 bp for the floxxed product and 230 bp for the wild-type 554 product. 555 AAV Vector Design and Delivery 556 The therapeutic construct CTR (Chimeric TDP-43 Repressor) was designed by fusing 557 the N-terminal RNA recognition domains (RRM1 and RRM2, residues 1 –267) of human 558 TDP-43 with the splicing repressor domain of RAVER1 (residues 450–643), replacing the 559 aggregation-prone C -terminal glycine -rich region 1. The construct includes the 560 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint endogenous TARDBP 3′ untranslated region (3′UTR) to preserve TDP-43 autoregulatory 561 feedback. The CTR transgene was driven by the CBh (chicken β-actin hybrid) promoter. 562 Control vectors included AAV9-CMV-GFP (Cat# AAV-0090, Virovek), with a reported 563 titer of 2E13 vg/mL, and AAV9-CMV-mCherry, both obtained from the same supplier. 564 For adult intracerebroventricular (ICV) delivery , mice were anesthetized with 565 isoflurane and placed in a stereotaxic frame. A total of 5 μL (1 E11 vg) was injected into 566 the right lateral ventricles at the following coordinates from bregma: AP –0.5 mm, ML 1.0 567 mm, DV –2.5 mm57. Injection was performed at a rate of 1 μL/min using a 5 μL Hamilton 568 syringe. Mice were allowed to recover on a heating pad and monitored until fully 569 ambulatory. 570 For neonatal injections, pups were injected within 8 h of birth. Cryo -anesthesia was 571 induced by brief exposure to wet ice (<2 min) with a protective barrier to prevent skin 572 injury. A pulled glass capillary needle (41 mm, ~0.5 mm tip) was inserted ~3 mm into the 573 lateral ventricle, and 3 μL of AAV9 (1×10^13 vg/mL) encoding either CTR or GFP was 574 slowly delivered over ~15 s. Trypan blue (0.05%) was included in the injection solution to 575 aid visualization. Following injection, pups were rewarmed under a heat source, placed 576 in bedding from the home cage to restore maternal scent, and returned to the dam once 577 active. Survival was ~80%, with early losses primarily due to incomplete recovery from 578 anesthesia (P0–P1) and occasional hydrocephalus developing at later stages (P14–P21). 579 Pups showing signs of distress were euthanized according to IACUC guidelines. 580 Tissue Collection and Immunohistochemistry 581 Mice were euthanized at designated timepoints (1-, 3-, 6-, or 12-months post-injection) 582 by isoflurane overdose, followed by transcardial perfusion with phosphate-buffered saline 583 (PBS) and then 4% paraformaldehyde (PFA) in PBS. Brains were carefully dissected and 584 post-fixed in 4% PFA at 4°C overnight, then dehydrated, embedded in paraffin, and 585 sectioned sagittally at 10 μm thickness using a rotary microtome (Leica HistoCore 586 MULTICUT). 587 For immunohistochemistry, slides were first oven -baked at 60°C for 30 mins, then 588 deparaffinized in xylene and rehydrated through a graded ethanol series (100%, 95%) to 589 water. Antigen retrieval was performed in 10 mM sodium citrate buffer (pH 6.0) by heating 590 slides in a microwave for 4 minutes. After cooling, sections were encircled with a 591 hydrophobic barrier pen and incubated in blocking solution containing 1.5% normal goat 592 serum and 0.1% Triton X-100 in PBS for 1 hour at room temperature. Primary antibodies 593 diluted in blocking buffer were applied overnight at 4°C in a humidified chamber. CTR 594 expression in mouse was expressed using N-terminal targeting human TDP-43 antibody. 595 Primary antibodies used: 596 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint human-TDP-43 (Ref#WH0023435M1, 1:1000, Millipore Sigma), 597 mouse-specific TDP-43 (C-Terminus, Ref#12892-1-AP, 1:1000, Proteintech), 598 Mouse-NeuN (Cat#MAB377,1:2000, MERCK). 599 Endogenous peroxidase activity was quenched by incubating slides in 0.3% hydrogen 600 peroxide (H₂O₂) in methanol for 30 minutes, followed by three 10-minute washes in PBST 601 (PBS + 0.1% Tween -20). Slides were incubated with biotinylated secondary antibodies 602 (Vector Laboratories, BP-9100-50, BP-9200-50) for 1 hour at room temperature. Signal 603 was amplified using VECTASTAIN Elite ABC reagent (Vector Laboratories, PK-7100) for 604 1 hour and visualized by DAB substrate reaction (Vector DAB Peroxidase Substrate Kit , 605 SK4100), monitoring under a microscope for optimal development. Slides were then 606 counterstained with hematoxylin, dehydrated through graded alcohols, cleared in xylene, 607 and mounted with mounting media. 608 Brightfield images were captured using a ZEISS microscope equipped with a high -609 resolution digital camera. Quantification was performed using ImageJ with uniform region-610 of-interest (ROI) definitions applied across experimental groups. 611 Reverse Transcription PCR (RT-PCR) 612 Total RNA was extracted from dissected hippocampal tissue using the RNeasy Mini 613 Kit (Qiagen, #74106) according to the manufacturer’s protocol. RNA concentration and 614 purity were assessed using a Nanodrop spectrophotometer (Thermo Fisher, 13-400-519), 615 and RNA integrity was confirmed by agarose gel electrophoresis. 616 First-strand cDNA was synthesized using the ProtoScript® First Strand cDNA 617 Synthesis Kit (New England Biolabs, E6300) following the manufacturer’s protocol. Each 618 reaction used 500 ng to 1 μg of total RNA and was primed with a mixture of enzyme mix 619 and random primers to ensure coverage of all transcripts. 620 To assess cryptic exon inclusion, we designed primers flanking known TDP -43–621 regulated cryptic exons (e.g., Adnp2, Ap3b2, Bud23, Camk1g, Cr em, Ggct, Unc13a, 622 Synj2bp, Tbc1d1, Usp15, Tecpr1, and Washc4)30. Products were separated on 1.5% 623 agarose gels and visualized using GelRed staining under UV illumination. Band 624 intensities were quantified using ImageJ and normalized to total transcript signal (i.e. 625 inclusion + exclusion bands). 626 To evaluate CTR autoregulation, a primer pair was designed targeting the N-terminal 627 human TDP -43 sequence (RRM1) and the C -terminal RAVER1 fusion domain. 628 Amplification of this junction-specific product confirmed the presence of CTR transcript. 629 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint All PCR reactions were performed in technical duplicates or triplicates, and at least 630 three biological replicates per condition were included. T ouch down t hermal cycling 631 conditions were optimized per primer set and are provided in Supplementary Table 2-4. 632 BaseScope In Situ Hybridization 633 BaseScope™ in situ hybridization was performed using the BaseScope ™ v2 Assay 634 (Advanced Cell Diagnostics, ACD) according to the manufacturer’s protocol. Paraffin -635 embedded sagittal brain sections (10 μm thickness) were processed using standard 636 deparaffinization, target retrieval, and protease digestion steps as described in the 637 BaseScope manual. 638 Probe sets were custom -designed by ACD to target exon -exon junctions spanning 639 cryptic exon inclusion events. Specifically, we used a probe targeting the 640 Unc13a(CAT#1182491-C1), Synj2bp(CAT#712191), and Ift81(CAT#712201) cryptic 641 exons along with one targeting CTR mRNA(CAT#1573691-C1). Signal amplification and 642 chromogenic detection were carried out using the BaseScope™ Red Detection Kit. Slides 643 were counterstained with 50% hematoxylin, air-dried, and mounted using AquaMount. 644 Brightfield images were acquired using a ZEISS microscope. For quantification, the 645 number of BaseScope puncta per nucleus was manually counted within defined 646 hippocampal regions of interest using ImageJ. At least three sections per mouse and 3 –647 5 mice per group were analyzed in a blinded fashion. 648 Behavioral Testing 649 Behavioral experiments were conducted at the University of Wyoming Animal 650 Behavior Core Facility under blinded conditions. All testing was performed during the light 651 phase (9:00 a.m. to 5:00 p.m.), and mice were habituated to the testing room for at least 652 30 minutes prior to each assay. 653 Open field testing was used to evaluate general locomotion and anxiety-like behavior. 654 Mice were placed in a 42 × 42 cm open-field arena for 15 minutes. Distance traveled and 655 time spent in the center zone were tracked using Noldus EthovisionXT 15 software. 656 Novel object recognition (NOR) was used to assess recognition memory. Mice were 657 exposed to a 42 × 42 cm open-field arena settled with two identical objects for 10 minutes. 658 After a 5-minutes delay, the arena was cleaned and one object was replaced with a novel 659 object of similar size, and mice were reintroduced to the arena for 10 minutes. Object 660 exploration was scored manually by trained observers blinded to group allocation. 661 Social behavior testing was performed and recorded in an adapted same-chambered 662 procedure as previously described 58. During habituation, mice were allowed to freely 663 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint explore a 42 x 42 cm arena settled with two empty wire cages at the opposite corners for 664 10 minutes. In the 10-minutes sociability test phase, an unfamiliar age- and sex-matched 665 conspecific (stranger 1) was enclosed in one of the empty wire cages. In the 10-minutes 666 social novelty test phase, a second unfamiliar age - and sex -matched conspecific 667 (stranger 2) was placed in the previously empty wire cage. Time spent sniffing or touching 668 each cage were scored manually frame by frame by trained observers blinded to group 669 allocation. 670 All behavioral assessments were performed by experimenters blinded to genotype 671 and treatment group. Data were analyzed as described in the statistical analysis section. 672 In vivo calcium imaging 673 In vivo calcium imaging was performed to monitor neuronal activity in the prelimbic 674 cortex of TDP-43 cKO and control mice. 500nl of AAV1-CamKII-GCaMP6f (Addgene, 2 675 E13 GC/mL) was injected stereotaxically into the prelimbic cortex ( coordinates: A /P 676 +1.9 mm, M /L 0. 5 mm, D /V 1. 75 mm) under isoflurane anesthesia as previously 677 described59. A 1 mm diameter GRIN lens (Grintech) was implanted into the injection site 678 to reach the depth of 1.8 mm and secured with dental cement as previously described59. 679 Mice were allowed to recover for at least 4 weeks prior to recording. 680 In vivo Ca2+ Imaging was performed using a previously established custom-built 681 miniature fluorescence microscope recording system60 during exploratory behavior in a 682 42 x 42 cm open-field arena. Three recording sessions, each lasting 5 minutes per mouse 683 were conducted. Calcium fluorescence signals were acquired at 10 Hz and preprocessed 684 using standard motion correction and ΔF/F0 normalization. Individual neurons were 685 identified and segmented using CNMF-E implemented in MATLAB. Calcium event rates 686 were quantified by thresholding deconvolved signals, and the average spike rate per 687 neuron was calculated for each animal as previously described61. 688 Mice were excluded from analysis if recordings lacked sufficient signal -to-noise or 689 spatially stable fields of view. Of 67 mice that underwent imaging procedures, 30 (45%) 690 produced data of sufficient quality for analysis (see Supplementary Table 1 for group 691 breakdown). Sample sizes per group ranged from 5 to 9 animals. All animals were 692 included in downstream group-level comparisons. 693 Data analysis 694 All statistical analyses were performed using GraphPad Prism (version 10 ) unless 695 otherwise specified. Data are presented as mean ± standard error of the mean (s.e.m.), 696 unless otherwise indicated. Group comparisons were evaluated using unpaired two-tailed 697 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint Student’s t-tests, Mann-Whitney U test, one-way ANOVA, or two-way ANOVA with Holm–698 Sidak post hoc correction, depending on experimental design and variance structure. 699

references 700 54. Chiang, P.-M. et al. Deletion of TDP -43 down-regulates Tbc1d1, a gene linked to 701 obesity, and alters body fat metabolism. Proc. Natl. Acad. Sci. U. S. A. 107, 16320–702 16324 (2010). 703 55. Madisen, L. et al. A robust and high -throughput Cre reporting and characterization 704 system for the whole mouse brain. Nat. Neurosci. 13, 133–140 (2010). 705 56. Feil, S., Valtcheva, N. & Feil, R. Inducible Cre mice. Methods Mol. Biol. Clifton NJ 706 530, 343–363 (2009). 707 57. Delivery of Therapeutic Agents Through Intracerebroventricular ICV and Intravenous 708 IV Injection in Mice. https://app.jove.com/v/2968/delivery -therapeutic-agents-through-709 intracerebroventricular-icv. 710 58. Liang, B. et al. Distinct and Dynamic ON and OFF Neural Ensembles in the Prefrontal 711 Cortex Code Social Exploration. Neuron 100, 700-714.e9 (2018). 712 59. Thapa, R., Liang, B., Liu, R. & Li, Y. Stereotaxic Viral Injection and Gradient -Index 713 Lens Implantation for Deep Brain In Vivo Calcium Imaging. J. Vis. Exp. JoVE (2021) 714 doi:10.3791/63049. 715 60. Barbera, G. et al. Spatially Compact Neural Clusters in the Dorsal Striatum Encode 716 Locomotion Relevant Information. Neuron 92, 202–213 (2016). 717 61. Liang, B. et al. Aberrant neural activity in prefrontal pyramidal neurons lacking TDP -718 43 precedes neuron loss. Prog. Neurobiol. 215, 102297 (2022). 719 720 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint

721 This work was supported in part by the National Institutes of Health grants R01 722 NS095969 (to P.C.W.), UG3/UH3 NS115608 (to P.C.W.), R33NS115161 (to P.C.W.), R01 723 NS129878 (to P.C.W. and Y.L.) , and the Intramural Research Program of the National 724 Institutes of Health (NIH). The contributions of the NIH author (D. -T.L.) are considered 725 Works of the United States Government. The findings and conclusions presented in this 726 paper are those of the author(s) and do not necessarily reflect the views of the NIH or the 727 U.S. Department of Health and Human Services. 728 Disclosure statement 729 J.P.L. and P.C.W. are inventors on patents that describes the use of CTR to restore 730 TDP-43 function for the treatment of ALS -FTD and other diseases that exhibit TDP -43 731 dysfunction. 732 Author contributions 733 T.C. and P.C.W. conceptualized, designed and interpreted the study. T.C., Y.L, J.P.L. 734 and P.C.W. wrote the manuscript. T.C., R.T., R.L., A.P.M., I.R.S, M.S.B, G.D.B., B.P. and 735 X.W. performed experiments. D.-T.L. provided the custom -build miniscope recording 736 system. All the authors reviewed and approved the final manuscript. 737 Competing interest declaration 738 The authors of this study have no conflicts of interest to report. 739 740 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint Supplementary data figure legend s 741 Supplementary figure 1. Conditional TDP-43 knockout model and AAV-CTR vector 742 therapeutic design 743 744 (A,B), Schematic of the inducible, excitatory neuron –specific Tardbp conditional 745 knockout (TDP-43 CKO) mouse model. LoxP sites flank exon 3 of Tardbp in Tardbpf/f 746 mice, and CreERT2 is driven by the Camk2a promoter. Upon tamoxifen administration, 747 exon 3 is excised in forebrain excitatory neurons, resulting in loss of functional TDP-43. 748 (C), Diagram of the CTR construct (Chimeric TDP -43 Repressor), composed of the 749 N-terminal RNA recognition motifs (RRM1 and RRM2) of TDP -43 (amino acids 1 –267), 750 fused to the splicing repression domain (RAVER1, amino acids 450 –643), and followed 751 by the endogenous human TARDBP 3′ untranslated region (3′UTR) to preserve 752 autoregulatory feedback. 753 (D), Timeline of the in vivo study design. Mice were fed tamoxifen at 5 months of age 754 for 1 month to induce recombination. After a 2 -week recovery, intracerebroventricular 755 injection of AAV -PHP.eB vectors expressing CTR or GFP was performed. Mice were 756 analyzed at 1-, 3-, 6-, and 12-months post-injection. 757 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint Supplementary figure 2. AAV9-ICV Delivery of CTR at P0 Attenuates Cryptic Exon 758 Inclusion of TDP -43 Target Genes and Preserves Neuronal Integrity in the 759 Hippocampus 760 761 (A), Representative cresyl violet staining of the hippocampus in Control + GFP, 762 Control + CTR, TDP-43 cKO + GFP, and TDP-43 cKO + CTR mice at 2.5 and 6 months 763 of age. 764 (B), Quantification of the ratio of neurons in the CA2/3 region, normalized to neuron 765 numbers in CA1 region, 2.5 m time point: Control + GFP (n = 4), TDP-43 cKO + GFP (n 766 = 4), and TDP-43 cKO + CTR (n = 4), 6 m time point: Control + GFP (n = 9), TDP-43 cKO 767 + GFP (n = 9), and TDP-43 cKO + CTR (n = 12) 768 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint (C), Quantification of neurons in the CA1 region in the same groups as panel B. 769 (D), Representative semi-quantitative RT-PCR analysis of three TDP-43 target cryptic 770 exon RNAs (Synj2bp, Ift81, and Sort) in hippocampus and cortex of Control + GFP (n = 771 3), TDP-43 cKO + GFP (n = 4), and TDP-43 cKO + CTR (n = 3) mice at 4 months of age. 772 (E), Quantification of cryptic Synj2bp, Ift81, and Sort RT -PCR cryptic products in 773 hippocampus from the same groups as in (d). 774 (F), Quantification of cryptic Synj2bp positive cells by In situ hybridization in 775 hippocampus from the same groups. 776 Band intensities were quantified using ImageJ. One-way ANOVA: **** P < 0.0001, ## 777 P < 0.01. Scale bars: 200um. 778 779 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint Supplementary figure 3. Neuronal survival after CTR Treatment 780 781 (A) Quantification of CA1 neuron numbers at 6 months post -treatment in 782 Control(n=29), TDP -43 cKO + GFP (n=9), and TDP -43 cKO + CTR (n=9) mice. No 783 significant differences were observed between groups. 784 (B) Quantification of CA2/3 neuron numbers at 6 months post-treatment in the same 785 groups as in (a). TDP -43 cKO + GFP mice showed a significant reduction compared to 786 Control, and CTR expression partially rescued neuron numbers. 787 (C) Quantification of CA1 neuron numbers at 12 months post -treatment in in 788 Control(n=19), TDP-43 cKO + GFP(n=13), and TDP-43 cKO + CTR(n=13) mice. 789 Data are presented as mean ± SD. Statistical analysis was performed using one-way 790 ANOVA followed by post hoc multiple comparisons. Statistical significance: p < 0.05 ( *), 791 p < 0.01 (**), p < 0.0001 (***), ns = not significant. 792 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint Supplementary figure 4. Morphometric analysis of cortical thickness, cerebellar 793 area, and hippocampal area after CTR treatment in Control and TDP-43 cKO mice 794 795 (A) Schematic illustration of the brain regions measured for cortical thickness and 796 area, cerebellar area, and hippocampal area. Red outlines indicate the regions of interest 797 (ROIs): cortex (1, 4–6), cerebellum (2), and hippocampus (3). 798 (G, I, K) Quantification of cortical thickness in the indicated cortical regions at 6 799 months after treatment across experimental groups : Control + GFP (n=10), Control + 800 CTR(n=15), TDP-43 cKO + GFP(n=16), and TDP-43 cKO + CTR(n=13). 801 (H, J, L) Quantification of cortical thickness in the indicated cortical regions at 12 802 months after treatment across experimental groups: Control + GFP (n=18), Control + 803 CTR(n=18), TDP-43 cKO + GFP(n=12), and TDP-43 cKO + CTR(n=15). 804 (B, D, F) Quantification of cerebellar area (F), cortical area (D), and hippocampal area 805 (B) at 6 months after treatment. 806 (C, E) Quantification of cerebellar area ( E) and cortical area ( C) at 12 months after 807 treatment. 808 Data are presented as mean ± SD. Statistical analysis was performed using one-way 809 ANOVA followed by post hoc multiple comparisons. Statistical significance: p < 0.05 (*), 810 p < 0.01 (**), p < 0.0001 (***), ns = not significant. 811 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint Supplementary figure 5. No significant differences were observed in locomotor 812 activity or anxiety-related behavior between genotypes or treatment groups in the 813 open field and light–dark box tests. 814 815 (A) Total distance traveled in the open field test. 816 (B) Time spent in the center of the open field arena. 817 (C) Time spent in the light compartment of the light–dark box test in male mice. 818 (D) Time spent in the center of the open field in female mice. 819 (E) Time spent in the light compartment of the light–dark box test in male and female 820 mice combined. 821 Groups: Control + RFP(N = 16), Control + CTR(N = 14), TDP-43 cKO + RFP(N = 18), 822 and TDP-43 cKO + CTR(N = 19). Data are presented as mean ± SEM; ns = not significant 823 by one-way ANOVA. 824 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint Supplementary Table s 825 Male(n = 41) Female(n = 26) Total(n = 67) Deaths during/after surgery, n (%) 7 (17%) 5 (19%) 12 (18%) Mice with high-quality calcium imaging data, n (%) 13 (32%) 17 (65%) 30 (45%) Breakdown of mice with high- quality data Control + CTR 3 4 7 Control + RFP 3 3 6 TDP-43 cKO + CTR 2 4 6 TDP-43 cKO + RFP 5 6 11 Supplementary Table 1. Perioperative mortality and calcium imaging data yield by 826 sex and experimental group 827 828 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint 829 Stage Temperature Time Cycles 1 95 1:00 1 2 95 0:30 10 64, descend 1 degree percycle 0:15 72 0:15 3 95 0:30 30 53 0:15 72 0:15 4 72 7:00 1 Final 4 ∞ Supplementary Table 2. RT-PCR protocol using touchdown PCR 830 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint name sequence Adnp2 WT DP F CTTGACAACATCAGGAAGGTGC Adnp2 WT DP R TTGTCCGGTATCTCGCTTTCT Adnp2 CE F6 TGGAAATGGACTGCTGAGTG Adnp2 WT R6 TCCAGAAGGTTCCCAAGGAGA Ap3b2 WT F (YHJ 2017) AGCCAGAATATGGCCACGAC Ap3b2 WT R (YHJ 2017) CACTATGATGGGCACACGGA Bud23 WT F CCAGAATGAAGCCCGGAAATAC Bud23 WT R CTGATGCAGCCATCAAAAGAGC Camk1g WT F (YHJ 2017) CTGGCCAAGATCACAGACTGG Camk1g WT R (YHJ 2017) CTGTGTAGACACCACGCTCT Crem WT F CAGTTCCTTTCTGCTTTGTCAAG Crem WT R TTGCTTCTTCTGCTAGTTGCTG Ggct WT DP F TGGATAAGTGGACATGGCGAG Ggct WT DP R TGGAAATTGCCGAAGTCGAG Ift81 WT F (Adonde 2019) AAGTGCGAGGACTTCGTGAG Ift81 CE R (Adonde 2019) CAGCGATCTGTCTGCTTTGC Synj2bp CE F CTCCAACGACAGTGGCATCT Synj2bp WT R GCTGCTTTCGGTATCTCACG Tbc1d1 CE F GGCATATGGAAGCCACGTCAC Tbc1d1 WT F CCTGGTGCAGATGGAGAAGAC Tbc1d1 WT R TGTCACTGAGAGGCGAGGAC Tecpr1 CE R CACCATCAGTCTATCCACACGTC Tecpr1 WT F AGTCAGACTGGTACGTGGATGAG Tecpr1 WT R GTGGCTGACATCCTCTCGG Unc13a CE F CATGTCTCTGCTGTGCGTGGGAG Unc13a WT F CAGGCGGTTGATCTCAAACATGA Unc13a WT R GGTGCCAGCCATCACTTTAAC .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint Usp15 CE R GCTCACAAAGGTCCCCAGAG Usp15 WT F GGTCCCTCTACTCCTAAGTCCC Usp15 WT R TGGCTGTTCATTGTTTCTTCCAG Washc4 WT DP F GGTGCCAGCCATCACTTTAAC Washc4 WT DP R GCATCTGTTGGTGAGGTCTT CTR N Term F ATGGGACCTAGACGGCTCTT CTR Raver1 R GCCACCTGGATTACCACCAA Supplementary Table 3. RT-PCR Primers. 831 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint Target WT band CE band Single Band Adnp2 187 339 128 Ap3b2 374 473 N/A Bud23 306 409 N/A Camk1g 423 509 N/A Crem 332 476 N/A Ggct 183 285 N/A Ift81 N/A N/A 187 Synj2bp N/A N/A 326 tbc1d1 128 238 197 Tecpr1 329 382 358 Unc13a 169 213 166 usp15 192 356 212 Wash4c 274 472 N/A CTR N/A N/A 345 Supplementary Table 4. Expected RT-PCR band sizes 832 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted September 2, 2025. ; https://doi.org/10.1101/2025.08.28.672900doi: bioRxiv preprint