An axon-intrinsic loop restricts nerve regeneration through axonal protein synthesis

Many lines of evidence indicate that ax onally synthesized proteins provide an ini tial response to injury and then l ater promote regenerati on of the injured ax ons through local acti ons of the newly synthesized proteins ( 17, 18). While most of the ax onally-synthesized proteins studied to date promote ax on growth or neuronal survival after injury , ax onal translation of Khsrp impedes ax on growth by 5 promoting decay of ARE-containing axonal mRNAs ( 4 ). Khsrp translation is increased by ax opl asmic Ca 2+ elevati on following injury , but its protein product stays elevated in axons w ell beyond the initi al ax otom y- induced Ca 2+ elevation subsides ( 4 ). Here, w e show that axonally-synthesized REG3A protein provides a means to i ntermi ttently activate Khsrp translati on in regenerating ax ons. Consequently , axonal transl ation of R e g3a slow s PNS nerve regeneration. The slow rate of PNS nerve regenerati on can result 10 in prolonged target tissue denervation and ultimately unsuccessful recovery from injury for lesions requiri ng ax on growth of more than a few centimeters to reach target tissues ( 3 ). Thus, in terventions that can accel erate ax on regeneration have great potenti al to increase recovery from PNS nerve injuries, and our data point to an ax on-i ntri nsic REG3A → Ca 2+ → KHSRP pathw ay as a target for accelerating nerve regenerati on (Figure 4I). 15 REG3A is a member of a family of lectin-lik e proteins that includes REG3B and REG3G. Murine REG3A is about 59% identical to REG3B and REG3G proteins. REG3 proteins have been functionally link ed to regenerati on of pancreatic islet cells and REG3A has bactericidal activity in the intestine ( 7 ). EXTL3 protein w as reported as a cell surface receptor for REG3 proteins ( 19). Several inflammatory cytokines have been reported to increase REG3A expression in epi thelial systems where REG3A binding 20 to EXTL3 activates the phosphatidyl-inosi tol-3 kinase (PI3K) to AKT signaling pathw ay ( 20). REG3G has also been reported to activate PI3K-AKT as w ell as the RAS/RAF-ERK signaling pathw ays through EXTL3 binding ( 21-23). Though EXTL3 has w ell-established roles as a glycosyl-transf erase in the Golgi apparatus (24), cell surface localizati on of EXTL3 has been reported for hippocampal neurons, N2a cells and (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 1 0 neuron-lik e PC12 cells ( 25). Further , REG1 α protein, which shares less homology to REG3A than the REG3 family members (42% sequence identi ty), w as reported to promote neuri te growth in PC12 cells by binding to EXTL3 ( 25). PI3K, AKT , Ras/Raf and ERK signaling are typically growth-promoting in neuronal contexts rather than the clear attenuation of growth that w e see with both recombinant and endogenous REG3A in the experiments here. REG3B w as also shown to activate JAK/ST A T3 signaling in 5 carcinoma cells ( 26), but ST AT3 activation after injury promotes nerve regeneration and neuronal survival (27). Thus, REG3A’ s attenuation of axon growth seems unlik ely to occur through previously defined signaling events downstream of EXTL3 binding. REG3A’ s function as a bactericide is link ed to its glycoprotein-binding activity , with hexamers of REG3A protein dimers permeabi lizing the bacterial membrane ( 28, 29). Cell mem brane permeabilization 10 by a naturally-occurring peptide fragment of REG3A (termed P AP) has also been reported to stimul ate pancreatic beta-cell prolif erati on ( 30). P ore-forming activities reported for REG3A could promote Ca 2+ entry after REG3A secretion or recREG3A exposure, but w e find that neither the ax on growth attenuation nor the ax onal KHSRP elevati on by REG3A w as aff ected by depletion of extracellular Ca 2+ . Activation of PERK and subsequent eIF2 α phosphoryl ation increasing axonal KHSRP synthesis is 15 consistent with release of ER Ca 2+ stores i n response to both recombinant and endogenous REG3A. Though Ca 2+ influx can trigger subsequent ER Ca 2+ release ( 31), extracellular Ca 2+ chelation did not prevent eff ects of recREG3A on DRG ax ons, so REG3A-dependent Ca 2+ en try cannot explain the eff ects of REG3A seen here. T aken together , our data argue that the eff ects of ax onally-synthesized REG3A are through a different receptor and signaling mechanism than previously published for REG3A. Further our 20 data suggest that this is an autocri ne, axon-intrinsic mechanism that slow s ax on growth. Several lines of evidence indicate that axotom y triggers an unfolded protein or integrated stress response locally that increases eIF2 α PS51 levels to impact transl ati on in ax ons ( 6, 1 1, 32-34 ). Consistent with this, w e see that recREG3A increases ax onal levels of cal reticulin protein, whose mRNA is translated (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 11 in axons by elevated Ca 2+ and PERK activation ( 6, 11, 12 ). Inflammation and injury have been reported to increase REG3A expression in rodent sensory neurons withi n 2 d following the inflammation or injury (35). Although the authors did not address function of the neuronal REG3A, they concluded that increased peripheral nerve REG3A levels derive from anterograde transport based on nerve ligati on studies ( 35). Ax onal localization of R eg3 a , combined with localized transl ati on of an ax onal reporter 5 mRNAs with Re g 3 a ’ s 5’ and 3’UTRs indicate that the increases in ax onal REG3A protein after ax otom y derive from anterograde mRNA transport and subsequent translation in ax ons. T ogether , these observations suggest that targeti ng the si gnaling mechanisms underlying this REG3A to KHSRP signaling can bring new strategies for accelerating nerve regeneration. 10 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 12

S 1. B. Grafstein, Role of slow axonal transport in nerve regeneration. Ac t a Neur opa th o l 5 , Suppl 5:144- 152 (1971). 2. P . N. Hoffman, R. J. Lasek, Ax onal transport of the cytosk eleton in regenerating motor neurons: 5 constancy and change. Brai n R es 202 , 317-333 (1980). 3. A. Hok e, Mechanisms of Disease: what factors limit the success of peripheral nerve regeneration in humans? Na t Clin Pract Neur ol 2 , 448-454 (2006). 4. P . P atel et al. , Intra-ax onal transl ati on of Khsrp mRNA slow s ax on regeneration by destabilizing localized mRNAs. N uclei c Aci ds R es 50, 5772-5792 (2022). 10 5. P . Briata et al. , Diverse roles of the nuclei c acid-binding protei n KHSRP in cell diff erenti ation and disease. Wiley I nt erdisci p R ev RNA 7 , 227-240 (2016). 6. P . Sahoo et al ., A translational switch drives ax onal stress granule disassembly through Casein Kinase 2α. Curr Biol 30, 4882-4895.E4886 (2020). 7. J. H. Shin, R. J. Seeley , Reg3 Protei ns as Gut Hormones? E n docri nol ogy 160 , 1506-1514 (2019). 15 8. A. L. K alinski et al ., mRNAs and Protein Synthetic Machi nery Localize into Regenerating Spinal Cord Ax ons When They Are Provided a Substrate That Supports Growth. J Ne urosci 35, 10357-10370 (2015). 9. Z. Zhang et al. , Automated preclinical detection of mechanical pain hypersensitivity and analgesia. Pa i n 163 , 2326-2336 (2022). 20 10. S. L. Olguin et al ., The RNA binding protei n KHSRP attenuates ax onal and dendritic growth, synaptic transmission, and memory consolidation via dysregulation of neuronal gene expression. Comm un Biol 5 , 672 (2022). 11. A. Pacheco, T . T . Merianda, J. L. T wiss, G. Gallo, Mechanism and rol e of the i ntra-ax onal Cal reticulin transl ation in response to ax onal injury . Exp Ne urol 323 , 113072 (2020). 25 12. D . V uppalanchi et al. , L ysophosphatidic acid diff erenti ally regul ates ax onal mRNA translation through 5'UTR elements. Mol Cell N eurosci 50, 136-146 (2012). 13. H. P . Harding, Y . Zhang, A. Bertolotti, H. Zeng, D . R on, P erk is essen tial for translational regulation and cell survival during the unfolded protein response. Mol Cell 5 , 897-904 (2000). 14. D . S. Smith, P . Sk ene, A transcription-dependent switch controls competence of adult neurons for 30 distinct modes of axon growth. J Ne urosc i 17, 646-658. (1997). 15. T . T . Merianda et al. , A functional equival ent of endoplasmic reticulum and Golgi i n axons for secretion of locally syn thesized proteins. Mol Cell Neur osci 40, 128-142 (2009). 16. R. D . Cummings, E. Chif foleau, Y . van K ooyk, R. P . McEver , in Essen tials of Gl y c obi ol og y , A. V arki et al. , Eds. (Cold Spring Harbor (NY), 2022), pp. 455-474. 35 17. H. Naw abi, S. Belin, T ranslational machinery and translation regulation in axon regeneration. Ne ural R egen R es 20, 1392-1394 (2025). 18. T . P . Smith, P . K. Sahoo, A. N. Kar , J. L. T wi ss, Intra-ax onal mechanisms driving axon regeneration. Brain R es 1740 , 146864 (2020). (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 13 19. S. Kobayashi et al. , Identificati on of a receptor for reg (regenerati ng gene) protei n, a pancreatic beta- cell regenerati on factor . J Bio l Chem 275 , 10723-10726 (2000). 20. Y . Lai et al. , The antimicrobial protei n REG3A regulates k eratinocyte proliferation and diff erenti ation after skin injury . I mmu nit y 37, 74-84 (2012). 21. T . Ito e t al., IL-22 induces Reg3gamma and inhibits allergic inflammation in house dust mite-induced 5 asthma models. J Ex p Med 214 , 3037-3050 (2017). 22. E. Sugisaw a et al. , Nociceptor-derived Reg3gamma prevents endotoxic death by targeting kynurenine pathw ay in microglia. Cell R e p 38, 110462 (2022). 23. J. H. Shin et al ., The gut peptide Reg3g links the small intestine microbiome to the regulation of energy balance, glucose levels, and gut function. Cell Me t a b 34, 1765-1778 e1766 (2022). 10 24. S. Y amada, Specific functions of Ex ostosi n-lik e 3 (EXTL3) gene products. Cell Mol B iol Let t 25, 39 (2020). 25. I. Acquatella-T ran V an Ba et a l. , Regenerating islet-derived 1alpha (Reg-1alpha) protei n is new neuronal secreted factor that stimul ates neuri te outgrowth via ex ostosin T umor-li k e 3 (EXTL3) recep tor . J Biol C hem 287 , 4726-4739 (2012). 15 26. C. Loncle et al. , IL17 Functi ons through the Novel REG3beta-JAK2-ST A T3 Inflammatory Pathw ay to Promote the T ransi tion from Chronic Pancreatitis to Pancreatic Cancer . Ca ncer R es 75, 4852-4862 (2015). 27. K. Ben-Y aak ov et al. , Ax onal transcription factors signal retrogradely in lesioned peripheral nerve. EMBO J 31, 1350-1363 (2012). 20 28. H. L. Cash, C. V . Whitham, C. L. Behrendt, L. V . Hooper , Symbiotic bacteria direct expression of an in testinal bactericidal lectin. Science 313 , 1126-1130 (2006). 29. S. Mukherjee et al. , Antibacterial membrane attack by a pore-forming intesti nal C-type lectin. Na ture 505 , 103-107 (2014). 30. C. S. Levetan et al ., Discovery of a human peptide sequence signaling islet neogenesis. E n docr Prac t 25 14, 1075-1083 (2008). 31. M. J. Berridge, Neuronal calcium signaling. Neuro n 21, 13-26 (1998). 32. M. Onate et al ., Activati on of the unfolded protein response promotes ax onal regeneration after peripheral nerve injury . Sci R ep 6 , 21709 (2016). 33. Z. Ying, V . Misra, V . M. V erge, Sensing ner ve injury at the axonal ER: activated Luman/CREB3 serves 30 as a novel ax onal ly synthesized retrograde regeneration signal. Proc N atl Acad Sci U S A 111 , 16142- 16147 (2014). 34. Z. Ying et al. , The Unfolded Protein Response and Cholesterol Biosynthesis Link Luman/CREB3 to Regenerative Axon Growth in Sensory Neurons. J Neuros ci 35, 14557-14570 (2015). 35. S. Q. He et a l. , Inflammation and nerve injury induce expression of pancreati tis-associated protein-II 35 in primary sensory neurons. Mol P ain 6 , 23 (2010). 36. W . J. Lin et al. , Posttranscriptional control of type I interferon genes by KSRP in the innate immune response against viral inf ection. M ol Cell Biol 31, 3196-3207 (2011). 37. G. J. Broussard e t al. , In vivo measurement of aff erent activity with axon-specific calcium imaging. Nat N euros ci 21, 1272-1280 (2018). 40 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 14 38. D . Y udin et al. , Localized regul ation of axonal RanGTPase controls retrograde injury signaling in peripheral nerve. Ne uron 59, 241-252 (2008). 39. C. A. Schneider , W . S. Rasband, K. W . Eliceiri, NIH Image to ImageJ: 25 years of image analysis. Na t Metho ds 9 , 671-675 (2012). 40. I. Rishal et a l. , WIS-NeuroMath enabl es versatile high throughput analyses of neuronal processes. 5 Dev Neurobi ol 73, 247-256 (2013). A CKN O WLEDGE MENTS – JL T is the incipient Universi ty of South Carolina SmartState Chair in Childhood Neurotherapeutics. 10 FUNDING – This w ork w as funded by grants from the NIH (R01-NS069833 and R01-NS117821 to JL T) and the Dr . Miri am and Sheldon G. Adelson Medical Research Foundati on (to JL T). A UTHOR CO NTR IBUTI ONS – CNB, JYL, LSV , SM, AK, and JL T designed experiments. CNB, JYL, LSV , SM, MCH, MC, LFT , AM, IDC, and MLDL performed experiments. 15 CNB, JYL, LSV , SM, MCH, MC, LFT , and AM provided data analyses. CNB, JYL, SM, MCH, MC, LFT , AM, MLDL, and ET performed animal husbandry and mouse colony maintenance. LSV and JL T supervised the w ork. CNB, JYL, LSV , AK, NPB, and JL T wrote and edited the manuscript. 20 NPB and JL T obtained grants funding the w ork. COMPETI NG INTER E S TS – CNB, JYL, LSV , AK, and JL T have a pending US Patent for REG3A modul ati on as a neural repair strategy . JL T is a co-founder of Rinnerva Therapeutics. D A T A AND MA TER IALS A V AILABILITY – Data included here will be submi tted to Zenudo online publi cally accessible database once accepted for publication. Materials will be freely shared subject to Univ SC 25 policy for Materials T ransf er Agreement. SUPPLEMENT A R Y MA TER IALS – Materi als and Methods Supplemental T able 1 Supplemental Figures S1-S5 + Figure legends 30 Supplemental Videos S1-S3 + Legends (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 15 FIGURE LE GENDS Figur e 1 : RE G3A slows ax on growth and regener ation . A, Quanti tation of smFISH images for naï ve vs. 7 d crush injured sciatic nerve (ScN) show s a significant increase in ax onal R e g3a mRNA. One-w ay ANO V A with T uk ey post-hoc analysis, mean ± SEM (n ≥ 6 mice per group). 5 B, Representative immunoblot of 0-28 d post-injury ScN lysates show s increase in REG3A protein signals in regenerating nerve with relatively equi valent levels of ERK1 protein. C-D , Representative images for SMI312 (NF) + T uJ1 immunofluorescence in AA V-shCntl vs -shReg3a transduced adult DRG cultures ( C). Quantitati on show s increased ax on length with shReg3a vs. shCntl (D ). See Suppl. Figure S1A-C for shRNA validati ons and ax on branching data. Student ’ s t-test, mean ± 10 SEM (n ≥ 75 neurons); scale bar = 100 µm. E-F , Representative images of exposure matched SCG10 immunostained ScN sections at 14 d post-nerve crush from mice transduced with AA V-shReg3a vs. -shCnt; injury site is mark ed by dashed line ( E ; le ft is pro ximal and righ t is distal). R egeneration indices for mice as in E ( F ). See Suppl. Figure S1D-J for in vivo shRNA validations and regeneration indices on 7, 21 and 28 d post ScN crush. T w o-w ay ANO V A with 15 Sidak post-hoc analysis, mean ± SEM, ** p ≤ 0.01 and *** p ≤ 0.005 (n ≥ 5 per group); scale bar = 500 µm. G-H , Representative images of extensor digitorum longus (EDL) neuromuscular junctions (NMJ) from AA V0-shCntl vs. -shReg3a transduced mice at 0 and 21 d post ScN crush injury ( I). Quantitati on of NMJ occupancy is shown in J . T w o-w ay ANO V A with T uk ey post-hoc analysis (n ≥ 84 NMJs across 6 animals; 20 mean ± SEM; scal e bar = 20 µm. I-J , Representative images of right (Rt) hind paw ( G ) from BlackBox One analyses of AA V-shRNA transduced mice as in E. AA V-shReg3a transduced mice show increased toe spreading beginning at 10 d post-injury . Quanti tation of toe spread is shown in J. See Suppl . Figure S2A-D for data on KHSRP -/- vs. KHSRP +/+ mice and Suppl. Figure S2E-H for additional behavioral assessments of AA V-shRNA transduced 25 mice wild type mice. T w o-w ay ANO V A with T uk ey post-hoc analysis, mean ± SEM, * p ≤ 0.05 and ** p ≤ 0.005 (n = 10 shCntl and 13 shReg3A). K-L , Representative SMI312 + T uJ1 immunofluorescence images for control vs. recREG3A treated adult DRG cultures (K ). Quantitation of ax on morphologies show s decreased length and increased branching after recREG3A exposure ( L ). Student ’ s t-test, mean ± SEM (n ≥ 163 neurons); scal e bar = 100 µm. 30 Figur e 2 : RE G3A increases ax onal tr ansl ation of K hsr p mRNA . A-B , Ax on l ength quanti tation ( A ) and axonal KHSRP ( B ) for DRG neurons treated ± recREG3A and vehicle (DMSO) vs. cycloheximide (CHX). See Suppl. Figure S3A-C for representative images and ax on branching data. One-w ay ANO V A with T uk ey post-hoc analysis, mean ± SEM (n ≥ 63); scale bar = 100 µm. C-D , Fluorescence recovery after photobleaching (FRAP) for terminal ax ons of GFP MYR 5’/3’Khsrp 35 transfected DRGs treated with recREG3A ± CHX shown as representative exposure-matched image sequences ( C) and quanti tation ( D ). T w o-w ay ANO V A with T uk ey post-hoc analysi s, mean ± SEM, * p ≤ 0.05, ** p ≤ 0.005, and *** p ≤ 0.0005, with colors matched to dataset vs. control (n ≥ 8 ax ons across at least three culture preparations); scal e bar = 10 µm. E-F , Representative exposure-matched i mage sequences ( E ) and quanti tation ( F ) for ax onal FRAP in 40 mCh MYR 5’/3’Reg3a transfected DRGs ± CHX. Repeated measures ANO V A with T uk ey post-hoc analysis, * p ≤ 0.05 and ** p ≤ 0.01, mean ± SEM (n ≥ 10 ax ons across at least 3 culture preparations); scale bar = 10 µm. G-J , Representative exposure-matched i mage sequences ( G,I) and quanti tations ( H, J ) for axonal FRAP in mCh MYR 5’/3’Reg3a + GFP MYR 5’/3’Khsrp transf ected DRGs treated with BAPT A-AM to chelate i ntracellular 45 Ca 2+ (G-H ) or Thapsigargin to increase i ntracellular Ca 2+ (I-J ). Mean ± SEM is shown. Repeated measures (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 1 6 ANO V A with T uk ey post-hoc analysis, * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.005, and **** p ≤ 0.00; scale bar = 10 µm. Figur e 3 : Ax onal RE G3A signaling requi r es release of i n tr acellular calciu m and a n ER s tress-li k e response . A-B , Ax on growth ( A ) and ax onal KHSRP levels ( B ) for DRG neurons ± recREG3A in presence of BAPT A-AM 5 to chel ate cytoplasmic Ca 2+ , BAPT A to chelate extracellul ar Ca 2+ , or vehicle control (DMSO). Suppl. Figure S3F show s axon branching for these cultures. One-w ay ANO V A with T uk ey post-hoc analysis; mean ± SEM (n ≥ 46 in A and ≥ 99 in B). C-D , Ax onal FRAP analyses are shown for GFP MYR 5’/3’Khsrp reporter in DRG neurons treated as in A. Representative exposure-matched images sequences shown in C and quanti tation in D . BAPT A-AM 10 attenuates recREG3A-dependent fluorescent recovery of the GFP reporter but BAPT A does not. T w o-w ay ANO V A with Sidak post-hoc analysis; mean ± SEM; * p ≤ 0.05 and ** p ≤ 0.005 vs. control and & p ≤ 0.05 and &&&& p ≤ 0.001 vs. RecREG3A (n ≥ 10 axons across at least three culture preparations); scale bar = 10 µm. E-F , Ax on growth ( E ) and ax onal KHSRP levels ( F ) for DRG neurons ± recREG3A in presence of dantrolene 15 to inhibi t the Ryanodine receptor or U73122 to attenuate production of IP3. Suppl. Figure S3G show s ax on branching for these cultures. One-w ay ANO V A with T uk ey post-hoc analysis, mean ± SEM, (n ≥ 36 neurons in E and ≥ 28 in F across ≥ 3 repli cate cultures). G-H, Ax on growth ( G ) and ax onal KHSRP levels ( H ) for DRG neurons ± recREG3A in presence of PERK inhibitor (PERK inh ) vs. vehicle. Suppl. Figure S4A-B show s representative images and axon branching for 20 these cul tures. One-w ay ANO V A with T uk ey post-hoc analysis, mean ± SEM (n ≥ 60 in A and ≥ 34 in B). I, Ax onal FRAP analysis for ax onal GFP MYR 5’/3’Khsrp transfected adul t DRG cultures treated as in G. T w o- w ay ANO V A with T uk ey post-hoc analysis, mean ± SEM; * p ≤ 0.05, ** p ≤ 0.005, *** p ≤ 0.0005, **** p < 0.0001 (n ≥ 10 ax ons across at l east three culture preparations with col ors matched to condition vs. control). 25 Figur e 4 : Endogenous RE G3A at t enua t e s ax on gro wth by incre asing ax opl asmic calcium and activati ng KHS RP s y n thesis in ax o ns. A-B , Representative exposure matched i mages of intact axons and quanti tation for DRGs transduced with AA V-GCaMP6s show elevati on of axonal GCaMP6s signals after 1 hr recREG3A treatment ( A ). Representative exposure matched images of intact ax ons and quanti tati on for DRGs co-transduced with 30 AA V-shRNAs + AA V-GCaMPS6s show reduction of ax onal GCaMP6s signals with shReg3a ( B ). Suppl. Figure S4E-F show s similar responses using the Fura-4 calcium indicator . Imaging parameters for panels A-B and Suppl. Figure S4E-F w ere set for recREG3A or shReg3A condi tions with control parameters matched within each experiment to opti mize detecting elevation vs. depression of Ca 2+ signals. Student ’ s t-test, mean ± SEM (n ≥ 75 neurons over ≥ 3 culture preparations); scal e bar = 10 µm. 35 C, Schematic for in vitr o ax otom y in DRG spot cultures with representative images before and after ax on severing; scal e bar = 100 µm. D-E, GCaMP6s signals in distal ax ons of AA V-shCntl vs. -shReg3a transduced DRG spot cultures are shown in representative exposure matched image sequences ( D ) and quantitation of mean signal intensiti es ± SEM ( E ) based on average signals across each time poi nt (mean GCaMP6s signals over the entire imaging 40 sequence are approximately 35% higher for shCntl vs. shReg3a-transduced cultures; also see Suppl . Figure S4K). P values by cross-correlation analysis; n = 10 ax ons across 4 separate culture preparations; scale bar = 10 µm. F-G, Ax on growth events were assessed i n the neurons from panel D based on BFP signals from the AA V- shRNAs. Representative exposure matched image sequences ( F ) and quantitati on of ax on growth events 45 (G ) as mean change ± SEM (see Suppl. Fi gure S4L-M for the overall average events and growth rates (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 1 7 across the sequences and Suppl. Video 1). P values by cross-correlation analyses (n = 10 axons across 4 separate culture preparati ons); scale bar = 10 µm. H, Mean growth cone fluorescence ± SEM for DRGs expressing GFP MYR 5’/3’Khsrp and mCh MYR 5’/3’Reg3a is shown for proximal severed ax ons in spot cultures over 16-20.5 h post-ax otomy imaging sequences. P value by cross-correl ati on analyses (n = 10 axons across 4 separate culture preparations). 5 I, Schematic for ax on-i ntrinsic transl ational regul ation of Khsrp mRNA by axonally synthesized REG3A in regenerating axons. 1 ) Injury induces an increase in cell body R eg3a mRNA that is transported into regenerating axons over 2-3 d following ax otom y; Re g 3 a mRNA is translated under low Ca 2+ levels when eIF2 α is not phosphorylated. 2-3) Nascent REG3A secreted from the di stal ax on activates a transmembrane signal resul ting in rel ease of ER Ca 2+ to activate PERK resulting in phosphoryl ation of 10 eIF2 α on S51. 4) Increased eIF2 α PS51 acti vates transl ati on of Khsrp mRNA, which promotes decay of regenerati on-associated gene (Rag) mRNAs. 5) Subsequent dephosphoryl ation of eIF2 α p romo tes transl ation of R e g3a to generating an oscillating translation of Khsrp and R e g3a as shown in panel F . (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 1 8 Supplementary Materials for An ax on-in t rinsic loo p restricts ra t e of n er ve regenera tion th roug h ax onal p ro t ei n s yn thesis Courtney N. Buchanan †, Jin young Lee †, Samaneh Matoo, Marie-Claire Honoree, Molly Conw ay , Lillian F . Thompson, Ansley McKay , Irene Dalla Costa, Moira Lopez De Leon, Elizabeth Thames, Nora Perrone- Bizzozero, Ashley L. Kalinski, Lauren S. V aughn and Jeff ery L. T wiss 1 5 † These auth ors c o ntrib uted eq uall y t o thi s work Correspondi ng Author: twiss@mailbox. sc.edu 10 This PDF file includes: Materi als and Methods T able S1 Figures S1-5 + Figure Legends Legends for Movies S1-3 15 Other Supple men t ary Ma t erials f or this manuscript i nclud e the f ollowing: Movies S1-3 MA TERIALS AN D MET HODS 20 K e y R eage n ts an d R eso urces – Suppl ementary T able S1 summarizes k ey reagents utilized in these studies and the sources of those reagents where indicated. Animal Use – The Institutional Animal Care and Use Committee of the Universi ty of South Carolina approved all animal procedures. Adul t (8-16 wks old) male and f emale Khsrp knock out ( Khsrp -/- ) (36 ) and 25 wildtype (C57Bl/6) mice w ere used for all experiments. Isoflurane inhal ation w as used for anesthesia in all survival surgery experiments (see below) and animals w ere euthanized by CO 2 asphyxiation as indicated for results. For peri pheral nerve injury , 8-12 w eek ol d, anesthetized mice w ere subjected to sciatic nerve crush at mid-thigh level as previous described ( 4 ). Briefly , the sciatic nerve w as exposed crushed using 30 #2 fine jew eler ’ s forceps twice for 15 sec each, ~1.5-2 cm from its origin, proximal to the trifurcation. Ax otom y w as moni tored by the initi al contraction of the hind limb upon applicati on of pressure to the sciatic nerve and then the lack of hind paw extension during and upon recovery from anesthesia. For animals undergoing unil ateral nerve crush, the contral ateral nerve w as exposed but not crushed ( i.e. , ‘sham’ control). 35 W e have previously shown that nerve injected AA V is retrogradely transported to the neuronal cell bodies whose axons transverse the nerve resul ting in a neuronal specific transduction ( 4 ). For shRNA mediated knock down experiments 4 µl of 5 x 10 11 particles of AA V9-T agBFP2-U6-shReg3a or T agBFP2- U6-scramble-shRNA (V ector Builder , Chicago, IL) w ere diluted in 600 mM NaCl (total volume = 5 µl) and injected i nto the proximal sci atic nerve at ~1.2-1.5 cm from origin. 14 d after viral transduction, animals 40 w ere subjected to a bilateral sciatic nerve crush at ~0.5 cm distal to injection si te on both the l eft and right sciatic nerve as above. For consistency betw een animals, a single experimenter performed the viral inf ections and crush injuri es within each series of animals. Cell Culture – Adult mouse dorsal root ganglia (DRG) w ere harvested and pl aced in ice-cold Hibernate-A medium (BrainBits). For experiments with naïve DRG neurons, all DRGs including l umbar , thoracic and 45 low er cervical w ere collected. For in vivo injury conditioning, only lumber segment 4-6 DRGs w ere (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 19 collected. DRGs w ere ri nsed five times wi th DMEM/F12 + 10U/mL Pen/Strep (CAT#15-140-122, Fisher , W altham, MA) and then ganglia w ere treated with 0.125U/mL collagenase type II CA T#17-101-015, Fisher) in DMEM/F12 for 30 min at 37°C, 5 % CO 2 . DRGs w ere tri turated using a fire polished Pasteur pipette and pelleted by centrifuged at 100 xg for 10 min. DRGs w ere w ashed twice by resuspension with DMEM/F-12 + Pen/Strep and centrifugati on. 5 Following w ashes, cells w ere resuspended in DMEM/F-12 media supplemented wi th 10 % f etal bovine serum (Gibco , #10437028) 10 µM cytosine arabinoside (Sigma, St Louis, MO; # C6645), 1x L- glutamine (Fisher; # A2916801), 1x N1 medium supplement (0.5 mg /mL Insulin (Sigma; # I9278), 0.5 mg /mL Sodium Selenite (Sigma; # S5261), 0.5 mg /mL T ransf errin (Sigma; #T8158), 1.6 mg /mL Putrescine (Spectrum; # P1834), 0.73 ug /mL Progesterone (Spectrum; # P1834), in Earle’ s Buff ered Salt solution, no 10 phenol (Thermo Scientific, AA J67559AE). Dissociated DRGs w ere pl ated immedi ately or transf ected and then pl ated (see below) on poly-L-lysine/laminin-coated glass substrates. For coati ng substrates, coverslips w ere incubated in 50 µg /ml poly-L-lysine (Fisher; #A005C) for 2 h at 37°C and 5 µg /ml laminin (Sigma) at 4°C overnight. 12 mm glass coverslips, 35 mm black w all glass bottom dishes (WillCo W ells from T ed Paella, Redding, CA; # HBSB-3522), or chambered coverslips (Ibidi, Gräfelfing, Germany; # 15 80297) w ere used. For DRG ‘spot cultures’ , isol ated and ganglia w ere treated with collagenase as above for 30 min at 37 °C, 5% CO 2 . DRGs were then triturated 10-15 times with a 1000 µl pipette tip follow ed by a second incubati on at 37°C, 5% CO 2 for 10 min. Ganglia w ere dissociated into single cell suspension using a 1000 µl pipette tip in 4 ml of media containi ng 1x penicillin/streptom ycin solution. Dissociated ganglia w ere 20 centrifuged for 10 min at 700 xg. Cell pell et w as resuspended in fresh DRG culture medium and plated at a density of approximately 2.8 DRGs per 7 µl (up to 45 DRGs harvested per mouse). For this, dissoci ated ganglia w ere gently tri turated 10 times with 20 µl pipette tip and 7 µl placed onto poly-D-lysine/laminin- coated chambered coverslips at up to 2 w ell-separated ‘spots’ per w ell. Spotted DRGs w ere incubated at 37°C. 5% CO 2 for 7 min and then DRG cul ture medium w as added al ong the w all of the culture vessel 25 adequate to fully cover the cells and not be at risk for evaporati on. These ‘spot cul tures’ w ere incubated at 37°C, 5% CO 2 for 7-8 d. For medium changes, half volume of the medium in each w ell w as removed and repl enished with fresh DRG cul ture medium at d 1 and 5 post-dissociation. T o ax otomize the DRG spot cultures for in vitr o regeneration assay , ax ons were cut on one side of the spot under a stereomicroscope using a sterilized 0.6 mm thick x 2.75 mm wide flat carbon steel surgical blade (Fine 30 Science T ools, 10035-10). Ax ons w ere cut at distance approximately equal to the radius of the spot ( i.e ., approximately 1 mm from the closest soma). Ax on regrowth from the injury si te w as evaluated at 16-30 h post-ax otom y as indicated in the results section. For transducing spot cultures with AA V9, dissociated ganglia resuspended in DRG media were pelleted by centrifugati on at 700 xg for 10 min. The pellet w as resuspended in fresh 35 DRG culture media containi ng 5 µl of 5 x 10 11 particles AA V9shReg3a-T agBFP2 or AA V9-shControl- T agBFP2 virus. For cDNA transf ections, dissoci ated gangl ia w ere pelleted after w ashing in DMEM/F-12 at 100 x g for 5 min and then resuspended in 100 µl ‘nucleof ector solution’ for Rat N euro n N ucleof e ctor kit (Lonza, Alpharetta, GA; # VPG-1003) or 20 µl solution for Small Cell Nu mber-SCN kit (Lonza; # VSPI-1003). 4-6 µg 40 plasmid w as electroporated for the Rat Neuron Kit and 1 µg for the Small Cell Number Kit using AMAXA Nucleofector apparatus (program G-013 for Rat Neuron kit and SCN-8 for SCN kit). T ransf ected DRGs w ere then pl ated as above. For stimul ati on and inhibi tion in DRG cultures the following agents were used: recombinant human REG3A-His (277.8 nM; SinoBiological, W ayne, P A; # 11235-H08H), cycloheximide (150 µg /ml, 45 Sigma), BAPT A-AM (3 µM; Sigma, St Louis, MO), BAPT A (3 µM; Cayman Chemical, Ann Arbor , MI), Thapsigargi n (1 µM; Sigma), GSK260614 (90 µM; Bio-T echne Corp /T ocris, Minneapolis, MN), Dantrolene (10µM, Sigma, # D9175 ), and U73122 ( 1 µM; Sigma, # 1268 ) 36 h after initial plating. Equal (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 2 0 concentration of vehicl e w as used for inhibitors and chel ators, and equivalent concentrati on of BSA w as used for recREG3A. For chel ators and inhibitors, cultures w ere incubated for 15 min prior to other manipulations. Cultures were then treated with recREG3A or control for 24 h follow ed by rinsing in 1 x PBS and fixation in 4% paraformaldehyde in 1x PBS. Plasmid Con s tru cts and vir us c o ns t ructs – All Fluorescent reporter constructs for analysis of RNA 5 transl ation were based on eGFP with m yristoylation el ement (GFP myr ; originally provided by Dr . Erin Schuman, Max Planck Insti tute) or mCherry plasmid with m yristoyl ation element (mCh myr ). Reporter construct containing 5’ and 3’UTR of Khsr p mRNA has been published (GFP MYR 5’/3’Khsrp) ( 4 ). For R eg3a transl ation reporter (mCh MYR 5’/3’Reg3a), the 5’ and 3’UTRs of mouse R eg3a were PCR-amplified and subcloned into mCh myr reporter . pAA V-hSynapsin1-GCaMP6s-P2A-mRuby3 Ca 2+ sensor plasmid w as 10 originally generated by Lin Tian and w as purchased from Addgene ( 37). AA V5 and AA V9 preparations w ere purchased di rectly from Addgene (112005-AA V5 and 112005-AA V9). AA V9-shCntl-BFP and -Reg3a- BFP w ere purchased from V ectorBuil der . RNA isolatio n and an alyses – RNA w as isolated from cultured DRG neurons and sciatic nerves using the RNeas y Mi croisol atio n kit (Qiagen, Hilden, Germany). Dissociated DRG cultures were grown for 3 d, 15 w ashed bri efly with PBS then RNA w as isolated following manufacturer ’ s protocol . Sciatic nerves w ere cut into small pieces and di gested with collagenase at 37 °C for 30 min with intermittent tri turation. RNA w as isolated from the collagenase-treated nerves using the RN Aeas y Mi croisol ati on kit . RNA yield w as quantifi ed using fluorimetry with Ri bogr een reagent (ThermoFisher , W altham, MA) and 50 ng of RNA w as reverse transcribed using the Sensifa s t cDNA s y n t hesis kit (Bioline, London, UK). Droplet digital (dd) 20 PCR products w ere detected using E vagr een reagent on a QX200 droplet reader (Biorad, Hercules, CA). Mitochondrial 12S RNA (Mtrnr1) mRNA l evels w ere used for normalizi ng yields across diff erent isolates as indicated in resul ts. The following primers w ere used for ddPCR (all from Integrated DNA T echnologi es [ID T], Coralville, Iow a; all listed as 5’ to 3’): Mtrnr1, sense – GGCT ACACCTT GACCT AACG and antisense – CCTT ACCCCTTCTCGCT AA TTC; R eg3a , sense – TCT ACAAGAGAGACAAGA T GCTG and antisense –25 AGCT GGT ACGTGGAGAGG. Single Mole cule fluo rescence in-situ h yb ridiz a tion (smFI SH) – smFISH plus immunofluorescence (IF) w as used to detect Khsrp and R e g3a mRNAs i n dissociated DRG cul tures and sciatic nerve sections. Custom designed Cy3- and Cy5-labelled Stellaris probes (LGC Biosearch T ech, Middlesex, UK) for mouse Khsrp and R eg3a with Cy3- and Cy5-labelled scramble probes for control w ere used. Pri mary anti bodies for 30 smFISH/IF consisted of axon mark er anti body containing mouse anti-b3-T ubulin (TU J1, Biolegend; # 801201) and anti-Neurofilament mark er (SMI312; Biolegend; # 837904) cocktail (1:200 for tissues and 1:500 for dissociated cultures) and chick en anti-GFP (1:200 for tissues and 1:500 for dissociated cultures; Abcam, Boston, MA; # ab13970). FITC-conjugated donk ey anti-mouse and Alexa405-conjugated anti- chick en ( 1:500 each; Jackson ImmunoRes., W est Grove, P A ) w ere used as secondary antibodi es. 35 For cultured neurons, smFISH/IF w as performed as previously described with minor modificati ons ( 8 ). All steps w ere carri ed out at room temperature and all solutions w ere RNase free and made using DEPC-treated w ater unless otherwise noted. Coverslips w ere briefly ri nsed in 1 x PBS and then fixed in 2% PF A in 1 x phosphate-buff ered saline (1x PBS) for 15 min. Coverslips w ere rinsed 2 times in 1x PBS, then permeabilized in 0.3% T ri ton X-100 in 1x PBS for 10 min. Samples w ere prehybridized for 40 1 h in hybridization buff er (50% dextran sulphate, 10 µg /ml E. c oli tRNA, 10 mM ri bonucleoside vanadyl complex, 80 µg BSA, and 10% formamide in 2x SSC), and then incubated with 12.5 µM each Stellaris probe, mouse anti-b3-T ubulin (TU J1) and anti-Neurofilament mark er (SMI312) coc ktai l (1:500 each), and chick en anti-GFP (1:500) in hybridization buff er for 16 h at 37 °C. Coverslips w ere then w ashed in PBS + 0.3% T riton X-100 3 times, follow ed by incubation with FITC-conjugated donk ey anti-mouse and 45 Alexa405-conjugated donk ey anti-chick en for 1 h. After rinse in PBS, cells w ere post fix ed in 2% PF A in 1x PBS for 15 min and w ashed 3 times with 1x PBS then DEPC-treated water . Coversli ps w ere inverted and mounted on glass slides using Prolong Gold Antifade (ThermoFisher). (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 21 For sciatic nerve tissue, smFISH/IF w as performed as previously described with minor modificati ons ( 8 ). All steps w ere carri ed out at room temperature and all solutions w ere RNase free and made using DEPC-treated w ater unless otherwise noted. Sci atic nerve segments w ere fi x ed for overnight at 4 °C in 4% PF A, w ashed with PBS then cryoprotected overnight in 30% buff ered sucrose at 4 °C. Nerves w ere cryosecti oned at 20 µm thickness and stored at -20 °C until used). Sections were brought to room 5 temperature, w ashed three times in 1 x PBS for 5 min each, and then incubated wi th 20 mM glycine follow ed by 0.25 M NaBH 4 in PBS (3 times, 10 min each for both) to quench autofluorescence. Sections w ere quickly rinsed in 0.1 M T riethylamine (TEA) and then incubated in 0.1 M TEA + 0.25% acetic anhydride for 10 min. After w ashing in 2x SSC, sections w ere permeabilized with 0.3% T riton X-100 in PBS for 10 min. Sections w ere rinsed with 1x PBS and incubated with 2x SSC + 10% formamide for 10 10 min. Sections were prehybridized for 1 h in hybridizati on buff er (50% dextran sulphate, 10 µg /ml E. c oli tRNA, salmon sperm DNA, 10 mM ribonucleoside vanadyl compl ex, 80 µg BSA, and 10% formamide in 2x SSC), and then incubated with 12.5 µM each Stell aris probe in addi tion to mouse anti-b3-T ubulin (TU J1) and anti-Neurofilament mark er (SMI312) cocktail (1:200 each), and chick en anti-GFP (1:200) in hybridizati on buff er for 16 h at 37 °C. The following da y , sections w ere w ashed in 2x SSC + 10% 15 formamide at 37 °C twice for 30 min each, follow ed by a 10 min incubation in 0.5x SSC at 37 °C. Sections w ere bri efly rinsed in 1x PBS + 1% T riton-X100 and incubated with FITC-conjugated donk ey anti-mouse and Alexa405-conjugated donk ey anti-chi ck en for 1 h. Sections w ere w ashed in three times in 1x PBS and DEPC-treated w ater , then mounted under glass coverslips with Prolong Gold. smFISH and IF signals in tissue sections were imaged using Leica SP8X or Stellaris confocal 20 microscope. 63x/NA 1.4 oil immersion objective and pulsed whi te light laser w as used for imaging RNA in both culture and tissue samples. Scrambl e probe w as used to set the maximum image acquisiti on parameters (l aser power , HyD or PMT detec tor ene rgy , & offset) that gave mi ni mu m si gnal s wi th the control probes. XYZ image stacks w ere acquired across at l east three separate locations in each section scanned nerve sections. 25 Immuno fluore scence (IF) – Standard I F was performed as previously described with all steps at room temperatures unless specified otherwise. Coverslips w ere fixed with 4% paraformaldehyde in 1 x PBS for 15 min at room temperature and w ashed 3 times in 1 x PBS. PBS w ashed neurons w ere permeabilized with 0.3% T riton X-100 in PBS for 15 min and then pl aced in block buff er (1X PBS + 10% Normal Donk ey Serum [Jackson ImmunoRes]) for 1 h. Neurons w ere then incubated with primary antibodies diluted in 30 block buff er overnight in a humidified chamber at 4 °C. Primary antibodies consi sted of chick en anti- NFH/-NFM/-NFL cocktail (1: 500; Aves Lab, Tigard, OR; NFH # A B2313552, NFM # AB2313554, and NFL # AB2313553), mouse anti-b3-T ubulin (TU J1, Bio Legends, 1:500) , anti-Neurofilament mark er (SMI312, Bio Legends, 1:500) , rabbi t anti-KHSRP (1:500; Novus Biologicals, Centennial , CO; #NBP1-18910), rabbit anti- REG3A (1:200; ThermoFisher; P A5-76091), chick en anti-GFP (1:500; Abcam), mouse anti-eIF2 α (1:200; 35 Cell Signaling T ech., Danvers, MA; 2103S), and rabbit anti-eIF2 α PS51 (1:200; Cell Signaling T ech.; 9721S). After w ashes in 1x PBS, coverslips w ere incubated with a secondary antibody cocktail contai ning FITC- conjugated donk ey anti-chick en, Cy3-conjugated donk ey anti-mouse, or Cy5 conjugated donk ey anti- rabbit diluted in block buffer (all at 1:500; Jackson ImmunoRes., W est Grove, P A) for 1 h at room temperature. Coverslips w ere then w ashed 3 times in 1x PBS, rinsed with distill ed H 2 O , and mounted 40 with Prolong Gold. For regeneration studies on mouse sci ati c nerve and quantifying ax onal content of KHSRP and REG3A in viv o, sciatic nerve segments were fixed, cryoprotected, cryosectioned, and stored as above. Tissue sections w ere brought to room temperature, equilibrated in 1x PBS and incubated with 20 mM Glycine for 30 min follow ed by 0.25 NaBH 4 for 30 min to quench autofluorescence. Secti ons w ere 45 permeabilized in 0.3% T ri ton X-100 in 1x PBS for 15 min and block ed for 1 h at room temperature in 1x PBS containing 20 mM Glycine, 0.1% T ri ton X-100, and 10% Normal Donk ey Serum. Primary antibodies (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 22 consisted of , rabbit anti-KHSRP (1:200; Novus Biologicals, Centenni al, CO; #NBP1-18910), rabbit anti- REG3A (1:200; ThermoFisher P A5-76091 ), rabbit an ti-Stathmin-2 /SCG10 (1:200; Novus Biologicals; #NBP1-49461), mouse anti-b3-T ubulin (TU J1, Bio Legends, 1:200) , anti-Neurofilament marker (SMI312, Bio Legends 1:200) (1:500 each ), and chi ck en anti-GFP (Abcam, 1:200). The GFP antibody cross-reacts with BFP , so it w as used to detect BFP signals from AA V-shRNA transducti on. Samples w ere incubated 5 overnight with primary antibodies in a humidified chamber at 4 °C. The following day , samples w ere w ashed twice with 1X PBS and incubated with secondary antibodies consisted of FITC-conjugated donk ey anti-chick en, Cy3-conjugated donk ey anti -mouse, and Cy5-conjugated donk ey anti -rabbit diluted in block buff er (all at 1:500; Jackson ImmunoRes.) for 1 h at room temperature. Secti ons were w ashed with three times with 1X PBS and once with H 2 O pri or to mounti ng with Prolong Gold Antifade. 10 For DRG spot cultures, chamber w ells were gently rinsed with PBS (37 °C) and then fix ed in 4% paraformaldehyde in PBS for 30 min at room temperature. After additional rinse in 1x PBS at room temperature, w ells were incubated with primary and secondary antibodies as outlined above. Following this, chambers were removed and coverslips w ere mounted on glass slides using Prolong Gold. All samples w ere mounted using Prolong Gold Antif a de and samples were visualized using 15 epifluorescent or confocal microscopy . Leica DMI6000 epifluorescent microscope with ORCA Flash ER CCD camera (Hamamatsu) w as used for epifluorescent imaging. Confocal imaging f or immunofluorescence was performed on a Leica SP8X microscope fitted with a gal vanometer Z stage and HyD detectors; HC PL Apo 63x/1.4 NA objective (oil immersion) w as used with acquisition parameters matched for individual experiments using LASX softw are. Z-stack images w ere post-processed by 20 Leica Lightni ng De c on v olu tion i ntegrated into LASX softw are. Deconvolved image stacks w ere projected into single plane images. Fluorescent re c overy af t er pho t obleac hi ng (FRAP) – FRAP w as performed using a Leica confocal microscope fitted with an environmental chamber to maintain living cells at 37 °C and 5% CO 2 . A 63x/NA 1.4 oil immersion objective w as used for imaging with the pinhole set to 3 Airy uni ts (AU) for pre-bleach, 25 bleach, and post-bleach sequences to ensure enti re thickness of axons w as exposed to l aser emission (38). Dissociated mouse DRG cultures transf ected with GFP MYR 5’/3’khsrp and/or mCherry MYR 5’/3’reg3a, as indicated, w ere equilibrated in complete culture media that excluded phenol red. 72 h following transfection, GFP and/or mCherry expressing neurons w ere chosen for FRAP . A region of interest (ROI) in the most di stal portion of the ax on (40 x 40 µm, ≥ 250 µm from the soma) w as photobl eached with 488 30 nm (GFP) and/or 555 nm (mCherry) argon laser set at 100% pow er for 80 fram es at 0.78 sec each. Pre- bleach and post-bleach signals w ere captured using 70% pow er for 488 and/or 555 nm laser line every 30 sec (2 images tak en pre-bleach, and 30 images for post-bleach). T ranslation dependence of recovery w as tested by pre-treating DRG cul tures with 150 µg /ml cycloheximide (Sigma) 15 min prior to photobleaching. For testing recREG3A- and Ca 2+ -dependent transl ation by FRAP , transf ected DRG 35 cultures were pretreated for 15 min with 5 µg /ml recREG3A, 3 µM BAPT A-AM, 3 µM BAPT A, or 90 µM GSK260614. For combined treatments with recREG3A, DRG cultures w ere pretreated with inhibitors/chelators 15 min prior to recREG3A treatment. Immunobl otting – DRG cultures w ere lysed, or sciatic nerves were minced and lysed in RIP A buffer (50 mM T ris-HCl [pH 8.0], 1% NP-40, 0.5% Sodium Deoxycholate, 0.1% SDS, 150 mM NaCl) plus protease 40 inhibitors (Thermo Scientific A32965). Lysates were centrifuged at 20,000 x g for 15;i18min at 4 °C. Protei n concentrations of supernatants were determined using Pierce BCA Protein Assay Kit (ThermoFisher; # 23227). After normalization for protein content, lysates were denatured in Laemmli sample buffer (62.5 mM Tris [pH 6.8], 2% SDS, 10% glycerol, 1% β-mercaptoethanol, 0.01% bromophenol blue) at 95 °C for 5;i18min followed by fractionation on standard SDS/PAGE. Fracti onated proteins were el ectrophoretically 45 transferred to PVDF membranes (GE Healthcare Life Sciences, Marlborough, MA). Membranes were blocked in 5% non-fat dried milk powder (BioRad, Hercul es, CA) in Tris-buffered saline (TBS) with 0.1% Tween-20 (TBST) for 1 h at room temperature. Blots were probes overnight at 4 °C with the following (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 23 antibodi es diluted in TBST plus 3% BSA: rabbit anti-REG3A (Invitrogen; # PA5-76091), rabbit anti-ERK1 (Abcam; # ab109282), or mouse anti-HIS (Abcam; # ab18184). Membranes w ere washed in TBST and then incubated with horseradish peroxi dase (HRP)-conjugated anti-rabbi t IgG (1:2000; Cell Signaling T ech.) diluted in blocking buff er for 1 h at room temperature. Bl ots w ere w ashed i n TBST and signals w ere detected by ECL Prime™ (GE Healthcare Life Sciences, Marlborough, MA). 5 Behavioral a nalyses of ax o nal rege nera t ion – Behavioral testi ng w as conducted using adult male and f emale C57BL/6J mice, including both KH SRP -/- and KHSR P +/+ genotypes that were age- and sex-matched. Mice underw ent a unil ateral sciatic nerve crush injury , with corresponding sham surgery to the contral ateral side on d 0, and behavioral recordings w ere acqui red at d 0 prior to sciatic nerve crush and every 3-4 d up to 28 d post-crush. A separate cohort of C57BL/6J mice received a bilateral sci atic nerve 10 injection of either shCtrl or shReg3a on d 0. On d 14 post-injection, these animals w ere subjected to a unilateral sciatic nerve crush, with a contralateral sham surgery . Behavioral recordings for this cohort w ere tak en at d 0, prior to injecti on, and then every 3-4 d up to 21 d post-crush (35 d post-injection). All mice w ere housed in standard clear pl astic cages under controlled conditions (temperature 20-26°C, humidity 30-70%, lights on 07:00-19:00) throughout the duration of experiments and all 15 recordings w ere done between 10:00 and 17:00 in the same room where the ani mals w ere housed using the Bla ckBo x O ne instrument (BlackBox Bio, Cambridge, MA US). The animal containment chamber consisted of an 18 (l) × 18 (w ) × 15 cm (h) black acrylic box that w as closed on 4 sides and top; bottom of the box consisted of 5 mm thick borosilicate float glass for video recordings with 850 nm near-infrared (NIR) LED strips aligned perpendicul ar to 2 opposing edges of the glass and 2 separate 850-nm NIR LED 20 strips positioned horizontally 10 cm below the glass floor . These NIR LEDs provide illumination of the animals for camera mounted beneath to recording paw placement/usage and ani mal movement as described (9 ). Animals w ere placed i nto individual chambers within the device and allow ed 10 min of habituation prior to each recording. After 5-10 min habituation period, animals were briefly removed to 25 clean the glass surface of urine, f eces, and any debris and then returned to individual chambers for a 20 min reco rding. Image analyses a nd pr ocessing – Im ag e J w as used to quantify protein and RNA levels in sciatic nerve tissues from optical pl anes of XYZ scans ( 39). Ax on only signals w ere extracted via Colocalizatio n pl ug-in that to project protein or RNA signals that overlap with ax onal mark ers in each plane to a separate 30 channel. These ‘ax on only ’ signals w ere then quantified in each XY plane of these ax on only channels. Ax on mark er signal area w as used to normalize signal intensiti es across the individual XY planes. The relative signal intensity w as then averaged for all tiles in each biological replicate. T o assess regeneration in viv o , til e scans w ere post-processed by Straig ht en plugi n for Imag eJ (http://imagej.nih.gov/ij/). SCG10 fluorescence intensi ty w as measured along the length of the nerve 35 using Image J . Regeneration index w as cal culated by measuring the average SCG10 intensity in 500 µm bins across the length of the section starting at the crush site. The crush si te w as defined by the posi tion along the nerve length with maximal SCG10 intensity (secondarily confirmed by D API signals and DIC images). Dissociated DRGs were immunostained with neurofilament antibodies as described above and ax onal morphology w as analyzed from epifluorescent til e scans using WIS-Ne uro mat h to provide length 40 and branch densi ty information for each neuron ( 40). For FRAP image sequences, raw images w ere analyzed for recovery in the bl eached ROI using Leica Confocal Softw are. Percent fluorescent recovery w as determined relative to pre-bleach and post- bleach signals, which w ere set at 100 and 0% to allow for comparisons betw een experiments and betw een neurons. For each treatment and construct tested, FRAP w as analyzed for at l east 7 neurons 45 across 3 separate transf ections. St a ti s tical Analyses – Gra phP ad Prism softw are package (La Jolla, CA) w as used for statistical analyses. All experiments w ere performed in at least tripli cate with statisti cal tests and post-hoc analyses as (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 24 indicated in the resul ts section. P ≤ 0.05 w as considered as statistically significant. SUPPLEMENT AL FIGU RE LE GENDS Figur e S1 : RE G3A slows ax onal grow th. 5 A-B , RT ddPCR and immunofluorescence data show depletion of R eg3a mRNA (A) and REG3A protein ( B ) in shReg3a transduced DRG cultures as in Figure 1C-D . Student ’ s t-test, mean ± SEM (n = 3 mice for A and 45 neurons across ≥ 3 replicate cul tures for B). C, Quanti tation of axon branching in AA V-shCntl vs. -shReg3a transduced DRG cul tures from Figure 1C-D shown as mean ± SEM. Student ’ s t-test (n ≥ 75 neurons across ≥ 3 repli cate cul tures). 10 D-E smFISH and IF data for sciatic nerve sections from mice transduced with AA V-shRNAs show decreased ax onal R eg3 a mRNA (D ; scramble smFISH probe w as used as negative control) and protein ( E ) with shReg3a vs. shCntl. Student ’ s t-test, mean ± SEM (n = 6 mice per group). F-G, RT ddPCR for R eg3a mRNA in sciatic nerve (ScN) ( F ) and L4-6 DRG lysates ( G ) f or naïve, non- transduced (ø) and shReg3a vs. shCntl transduced mice at 0-28 d post-ScN crush corresponding to Figure 15 1E-J and Suppl. Figure 1H-J. One-w ay ANO V A with T uk ey post-hoc anal ysis, mean ± SEM (n ≥ 3 per group). H-J , Regenerati on indices based on SCG10 immunofluorescence for AA V-shRNA transduced mice at indicated post-crush i ntervals as in Figure 1E-J. T w o-w ay ANO V A with Si dak post-hoc analysis; mean ± SEM; * p ≤ 0.05 ** p ≤ 0.005 (n ≥ 5 per group). 20 K, Quanti tati on of ax on branching in recREG3A vs. Cntl treated DRG cultures from Figure 1L shown as mean ± SEM. Student ’ s t-test, mean ± SEM (n ≥ 163 neurons across ≥ 3 replicate cultures). Figur e S2 : RE G3A depletion ac celerat es f unctional rec ove ry af t er ner ve c rush inju ry . A-D , BlackBox One analyses of KHSRP -/- compared to KHSRP +/+ mice show increased distance traveled overall ( A ) and, after right-sided ScN crush, increased utilizati on of the hind paw ipsilateral to ScN crush 25 injury compared to contral ateral paw ( B- C, Rt/Lt hind paw luminance and standing Rt/Lt hind paw luminance, respectively) and average front paw s ( D, Rt hind to front paw s luminance) . Tw o - w a y AN O VA with T uk ey post-hoc analysis; mean ± SEM; * p ≤ 0.05, ** p ≤ 0.005 (n = 18 mice per genotype). E-H, BlackBox One analyses for AA V-shCntl vs. -shReg3a transduced mice following unilateral (right) ScN crush injury . There is no diff erence in distance traveled ( E ), but R eg3a depleted mi ce show increased 30 ipsilateral to injury vs. contral ateral hind paw s luminance ( F-G , Rt/Lt hind paw and standing Rt/Lt hind paw luminance), ipsilateral to injury vs. average front paw s luminance ( H ) compared to shCntl- transduced mice as in Figure 1G. T w o-w ay ANO V A with T uk ey post-hoc analysis; mean ± SEM; * p ≤ 0.05, ** p ≤ 0.005, *** p ≤ 0.0005, **** p ≤ 0.0001 (n ≥ 10 mice per group). Figur e S3 : RE G3A regulat es ax on g ro wth and KH SR P le vels thr ough m odulati on o f ax oplasmic Ca 2 35 le vels. A-B , Representative images for NF + T ubJ1 immunofluorescence ( A ) and ax on branching quanti tation ( B ) for DRGs treated with recREG3A ± cycloheximide (CHX) as in Figure 2A. One-w ay ANO V A with T uk ey post- hoc analysis, mean ± SEM (n ≥ 63 across ≥ 3 replicate cul tures); scale bar = 100 µm. C, Representative exposure matched images of terminal ax ons for DRG cultures treated as in A showing 40 SMI312/T uJ1 and KHSRP immunofluorescence corresponding to Figure 2B; scale bar = 5 µm. D-E, Ax on growth ( D ) and axonal Calreticulin protein ( E ) is shown for DRG neurons cultured from KHSRP -/- mice ± recREG3A treatment. Neurons lacking KHSRP do not show decrease in ax on growth in response to recREG3A but still axonal Calreticulin levels increase in both KHSRP +/+ and KHSRP -/- DRG neurons. One- w ay ANO V A with T uk ey post-hoc analysis, mean ± SEM (n ≥ 94 neurons in D and n ≥ 143 in E across ≥ 3 45 replicate cutures). (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 25 F-G, Ax on branching in DRG cul tures ± recREG3A and BAPT A-AM vs. BAPT A ( F ), or dantrolene vs. U73122 (G ) corresponding to Figures 3A, and E. One-w ay ANO V A with T uk ey post-hoc anal ysis, mean ± SEM (n ≥ 46 neurons in F and ≥ 36 in G across ≥ 3 replicate cultures). Figur e S4 : RE G3A regulat es ax on g ro wth and KH SR P le vels thr ough m odulati on o f ax oplasmic Ca 2 le vels. 5 A, Ax on branching in DRG cul tures ± recREG3A PERK inhibitor corresponding to Figures 3G. One-w ay ANO V A with T uk ey post-hoc analysis, mean ± SEM (n ≥ 60 neurons across ≥ 3 repli cate cultures). B-D , Representative exposure-matched i mmunofluorescent images for eIF α and eIF α PS51 (B ) and eIF α and eIF α PS51 levels in growth cones ( C-D) of dissociated DRG cul tures treated ± recREG3a with PERK inhibitor (PERK INH ) vs. vehicle (DMSO) corresponding to Figure 3G (n ≥ 60 neurons across ≥ 3 repli cate 10 cultures); scale bar = 5 µm. E-F , Representative images and quantitations for FURA-4 Ca 2+ -indicator signals in ax ons of dissociated DRG cultures exposed to recREG3A ( E ) or transduced with AA V-shCntl vs. -shReg3a ( F ). V alues shown as mean ± SEM with Student ’ s t-test (n ≥ 24 neurons over ≥ 3 culture preparations); scale bar = 10 µm. G-H, Quanti tation of eIF2 α PS51 (G ) and KH SRP ( H ) immunofluorescence signals in distal ax ons of intact 15 DRG spot cultures ± recREG3A vs. control for 1 h; mean ± SEM with P values by Student ’ s t-test (n ≥ 75 neurons over ≥ 3 culture preparati ons). I-J , Quantitation of eIF2 α PS51 (I) and KHSRP ( J ) immunofluorescence signals in distal axons of AA V-shCntl vs. -shReg3a transduced DRG spot cultures at 16 h post transection; mean ± SEM with P values by Student ’ s t-test (n ≥ 75 neurons over ≥ 3 culture preparations). 20 K, Average growth cone GCaMPs signal intensity regenerating ax ons of DRG spot cultures transduced with AA V-shCntl vs. -shReg3a shown as mean ± SEM; P value Student ’ s t-test (n ≥ 75 neurons over ≥ 3 culture preparations). L-M, Average of axon grow th eve nts ( L ) and growth rates ( M ) across live cell imagi ng experiments from Figure 4D-G for regenerating axons in AA V-shCntl vs. -shReg3a transduced DRG spot cultures are shown 25 as mean ± SEM. Student ’ s t-test, mean ± SEM (n ≥ 75 neurons over ≥ 3 culture preparations). Figur e S5 : Endogenous RE G3A mo dula t e s growt h c o ne calcium le vels and axon g rowt h e ven ts . A, Low magnification sti ll image for regions of interest (ROI) from regenerating axons in DRG spot cultures transduced with AA V-GCaMP6s and co-transf ected with GFP MYR 5’/3’Khsrp + mCh MYR 5’/3’Reg3a that w ere moni tored for distal ax on fluorescence in panel B; scal e bar = 25 µm. 30 B, Representative images for GFP MYR 5’/3’ Khsrp and mCh MYR 5’/3’Reg3a at indicated time points post- transection are shown for adjacent termi nal ax ons indicated as ROI 1 and 2 in panel A; scal e bar = 10 µm. C, Regeneration indices of BFP positive vs. BFP negative regenerating ax ons of shCntl- and shReg3a- transduced mice were identified by SCG10 immunofluorescence from sciatic nerves analyzed in Figure 1E-F . T w o-w ay ANO V A with Sidak post-hoc analysis, mean ± SEM; * p ≤ 0.05, ** p ≤ 0.005, *** p ≤ 35 0.0005, and **** p < 0.0001 (n ≥ 5 mice per group). D, Representative confocal XYZ maximum projection images for REG3A protein and NF from dissociated L4-6 DRG cultures harvested from naïve vs. 7 d injury-conditioned (‘Crush’) mice. These show increased REG3A protein in the injury-condi tioned cultures. REG3A is largely concentrated extracellularly adjacent to the ax onal membrane, as emphasized by the inset orthogonal XZ and YZ projections. Heparinase 40 treatment decreases the membrane adjacent REG3A signal, particul arly in the injury-conditioned DRG cultures. See Suppl. Video 2 for 3D rendering of the XYZ images shown here; scal e bars = 5 µm. E, Representative confocal XYZ maximum projection images for optically isol ated ax ons showing REG3A protein and NF in naïve vs. regenerating sciatic nerve (7 d post-crush) with insets showing orthogonal XZ and YZ projections. Note that REG3A is increased in regenerati ng nerve secti ons and appears mostly 45 outside of the ax on as seen in panel D . See Suppl. Video 3 for 3D rendering of the XYZ images shown here; scale bars = 10 µm. (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 2 6 SUPPLEMENT AL VI DE OS Video S1 : Depletion of R eg3a a c celera t e s ax on regene ratio n. Representative video sequences for BFP signals in regenerating ax ons of shCntl-BFP- vs. shReg3a-BFP- transduced DRG spot cul tures as indicated; scale bar = 10 µm. Video S2 : Increased RE G3A in injury c on ditioned DRG c ulture s is largely e xtracell ular . 5 Representative 3D renderings of REG3A protein and NF from dissociated L4-6 DRG cultures harvested from naïve vs. injury conditioned (‘Crush’) mice treated with heparinase as indi cated; scal e bar = 10 µm. Video S3 : Increased RE G3A in rege nerati ng perip heral ner ve is la rgely e xtracellul ar . Representative 3D renderings of REG3A protein and NF in naïve vs regenerating sciatic nerve (7 d post- crush); scale bar = 10 µm. 10 Ta b l e S 1 : Summa ry of k e y reage n ts an d sources. REAGENT or RESOURCE SOURCE IDENTIF IER Antibodies Rabbit polyclonal anti-KHSRP antibody Novus Biologicals NBP1-18910 Rabbit polyclonal anti-REG3A antibody Thermo Fisher PA5-76091 Rabbit polyclonal anti-REG3A antibody Thermo Scientific PA5102577 Chicken anti-neurofilament (light) Aves Cat# NFL; RRID: AB_2313553 Chicken anti-neurofilament (medium) Aves Cat# NFM; RRID: AB_2313554 Chicken anti-neurofilament (heavy) Aves Cat# NFH; RRID: AB_2313552 Chicken polyclonal anti-GFP Abcam Ab13970 Mouse Monocl onal anti-eIF2a Cell Signaling 2103S Rabbit Anti-phosphorylated eIF2a (Ser51) Cell Signaling 9721S Mouse monoclonal anti-Tubulin beta-3 chain (TUJ1) Biolegend Cat# 801201 Mouse monoclonal anti-Neurofilament marker (SMI-312) Biolegend Cat# 837904 Rabbit polyclonal anti-Stathmin-2/STMN2 (SCG10) Novus Biologicals Cat# NBP1-49461 Anti-rabbi t IgG, HRP-linked antibody Cell Signaling 7074 Anti-mouse IgG, HRP-linked antibody Cell Signaling 7076 Normal Donkey Serum Jackson ImmunoResearch 017-000-121 Cy5-conjugated donkey anti-rabbi t IgG (H+L) Jackson ImmunoResearch 711-175-152 Cy3-conjugated donkey anti-mouse IgG (H+L) Jackson ImmunoResearch 715-165-150 FITC-conjugated donkey anti-mouse IgG (H+L) Jackson ImmunoResearch 715-095-150 FITC-conjugated donkey anti-chicken IgG (H+L) Jackson ImmunoResearch 703-095-155 405-conjugated donkey anti-chicken IgG (H+L) Jackson ImmunoResearch 703-475-155 Bacteri al and virus s trai ns pAAV9-hsynapsin1-axon-GCaMP6s-P2A-mRuby3 Addgene RRID: 112005-AAV9 and 112005-AAV5 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 2 7 pAAV9-TagBFP2-U6-Ctrl shRNA VectorBuilder AAV9M(VB010000- 0023jze)-C pAAV9-TagBFP2-U6-Reg3a shRNA VectorBuilder AAV9M(VB221101- 1415tar)-K1 Biological samples DRGs and Sciatic Nerve isolated from C57Bl/6 and KHSRP -/- mice N/A N/A Chemicals, peptid es, and r eco mbinant p roteins Recombinant Human REG3A protein (His Tag) Sino Biological Cat# 11235-H08H Cycloheximide Sigma CAS# 66-81-9 BAPTA Cayman Chemical CAS# 85233-19-8 BAPTA-AM Sigma CAS# 126150-97-8 GSK2606414 (PERK inhibitor) Tocris CAS# 1337531-36-8 Dantrolene (RyR inhibitor) Sigma CAS# 14663-23-1 U73122 (PLC inhibitor) Tocris CAS# 112648-68-7 Thapsigargin Sigma CAS# 67526-95-8 DMSO Sigma 472301 SSC (20X), RNase-free Thermo Fisher AM9763 Dextran Sulfate Sigma S4030 E. Coli tRNA Sigma 10109541001 Ribonucleoside Vanadyl Complex NEB S1402S Bovine Serum Albumin Sigma 10711454001 Formamide Fisher BP227 Triethylamine Sigma T58300 Acetic Anhydride Fisher A10 Salmon Sperm DNA Sigma D7656 Heparinase I and III Bl end from Flavobacterium heparinum Sigma H3917 Critical comme rcial ass ays RNeasy Microisol ation kit Qiagen 74034 Ribogreen RNA quantificati on Fisher R11490 Sensifast cDNA synthesis kit Thomas Scientific C755H66 QX200 ddPCR EvaGreen Supermix BioRad 1864034 Automated droplet generation oil for EvaGreen BioRad 1864112 BCA Protein Assay Kit Thermo Fisher Cat = 23227 Cell Culture Reagen ts Poly-L-Lysine (50ug/mL) Fisher A005C Poly-D -Lysine hydrobromide Millipore-Sigma P7280 Laminin (5ug/m L) Millipore-Sigma CC095 DMEM and Ham’s F12 50/50, +L-Glutami ne, +Phenol Red Fisher MT10090CV DMEM and Ham’s F12 50/50, +L-Glutami ne, (- Phenol Red) Fisher MT16405CV (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 2 8 Fetal Bovine Serum BioAbChem 72-0400 Cytosine β-D-Arabinofuranoside hydrochloride Millipore-Sigma C6645 Pen-Strep (Gibco 10,000 U/mL) Fisher 15140122 L-glutamine Fisher A2916801 N21-Max Media Suppl ement R&D Systems AR008 Hibernate A BrainBi ts HA Collagenase Sigma C-9697 Earle’s Buffered Salt Sol ution, no phenol Thermo Fisher AAJ67559AE Insulin (0.5mg/mL) Sigma I9278 Sodium Seleni te (0.5 ug/mL) Sigma S5261 Transferrin (0.5 mg/mL) Sigma T8158 Putrescine (1.6 mg/mL) Sigma P5780 Progesterone (0.73 ug/mL) Spectrum P1834 Nucleofector kit for Primary Mammalian Neurons (SCN) Lonza VSPI-1003 Rat Neuron Nucleofector Kit Lonza VPG-1003 Prolong Gold Antifade with DAPI Fisher P36931 Prolong Gold Antifade without DAPI Fisher P36930 Experimental m odels: Orga nisms/ stra in s KHSRP knockout mice ( 36) N/A C57BL/6J mice Jackson Laboratori es RRID: IMSR_J AX:000664 Oligonucleoti des Stellaris probes agai nst Re g3a LGC Biosearch Tech Genbank ID # NM_011259.1 Stellaris probes agai nst Dapb - Quasar 570 LGC Biosearch Tech SMF-1063-5 Target sequence: TGATTCTTTAAAATATCCCGA TAGGCAGTCGTAAATTGA Stellaris probes agai nst Dapb - Quasar 670 LGC Biosearch Tech SMF-1065-5 Target sequence: TGATTCTTTAAAATATCCCGA TAGGCAGTCGTAAATTGA Mitochondrial 12S RNA ( Mtrnr1) PCR primer Integrated DNA Technologies sense – GGCTACACCTTGACCTAACG and antisense – CCTTACCCCTTCTCGCTAATTC Reg3a PCR primer Integrated DNA Technologies sense – TCTACAAGAGAGACAAGATG CTG and antisense – AGCTGGTACGTGGAGAGG Recombin ant DNA GFP MYR 5’/3’khsrp ( 4 ) mCh MYR 5’/3’reg3a This manuscript Soft ware and alg orith ms (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint Page 29 FIJI/ImageJ NIH http://fiji.sc/cgi- bin/gitweb.cgi NeuroMath Weizmann I nstitute http://www.weizmann.ac.il/ vet/IC/software/wis- neuromath Prism GraphPad Software, https://www.graphpad.com Stellaris Probe Designer BioSearch Technologies https://www.biosearchtech. com/support/tools/design- software/stell aris-probe- designer SnapGene SnapGene https://www.snapgene.com (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted October 29, 2025. ; https://doi.org/10.1101/2025.10.28.685020doi: bioRxiv preprint