Abstract
TFEB and TFE3 (TFEB/3), key regulators of lysosomal biogenesis and autophagy, play diverse roles depending on cell type. This study highlights a hitherto unrecognized role of TFEB/3 crucial for peripheral nerve repair. Specifically, they promote the generation of progenitor-like repair Schwann cells after axonal injury. In Schwann cell-specific TFEB/3 double knock-out mice of either sex, the TFEB/3 loss disrupts the transcriptomic reprogramming that is essential for the formation of repair Schwann cells. Consequently, mutant mice fail to populate the injured nerve with repair Schwann cells and exhibit defects in axon regrowth, target reinnervation, and functional recovery. TFEB/3 deficiency inhibits the expression of injury-responsive repair Schwann cell genes, despite the continued expression of c-jun, a previously identified regulator of repair Schwann cell function. TFEB/3 binding motifs are enriched in the enhancer regions of injury-responsive genes, suggesting their role in repair gene activation. Autophagy-dependent myelin breakdown is not impaired despite TFEB/3 deficiency. These findings underscore a unique role of TFEB/3 in adult Schwann cells that is required for proper peripheral nerve regeneration.
Significance Statement
Peripheral nerves have been recognized for their efficient regenerative capabilities compared with the central nervous system neurons. This is due to the remarkable ability of Schwann cells to undergo a reprogramming process, transforming into progenitor-like repair Schwann cells that actively contribute to axon regeneration and overall nerve repair. However, the specific transcriptional regulators responsible for initiating this transformation in the adult peripheral nerve have remained elusive. Our study elucidates a previously undescribed, injury-responsive function of transcription factors TFEB/3 in adult Schwann cells, showcasing their ability to promote tissue repair. Our findings hold important implications for enhancing nerve regeneration by bolstering the regenerative capacity of glial cells, thereby contributing to advancements in the field of neural tissue repair.
Introduction
Schwann cell plasticity in adults underlies the unique ability of peripheral nervous system neurons to regenerate after injury. Following nerve injury, Schwann cells distal to the site of injury undergo molecular reprogramming to become repair Schwann cells (reviewed in Jessen and Arthur-Farraj, 2019). This process entails downregulation of myelin-associated genes and the simultaneous upregulation of injury-responsive genes, allowing repair Schwann cells to acquire the phenotypes conducive to nerve regeneration and repair. Recent molecular profiling studies have unveiled that repair Schwann cell generation entails the activation of genes associated with epithelial-to-mesenchymal transition (EMT) and stemness (Arthur-Farraj et al., 2017; Clements et al., 2017). This molecular reprogramming equips Schwann cells with a new capacity for self-renewal, enhanced mobility, and adaptable morphology, which are vital attributes for tissue remodeling and nerve repair. Consequently, deciphering the molecular mechanisms governing the genesis and function of repair Schwann cells promises invaluable insights for designing therapeutic strategies aimed at fostering peripheral nerve repair.
Many signaling pathways, including ERK1/2, p38-MAPK, JNK, and mTORC1, have been identified as regulators of repair Schwann cell functions (Napoli et al., 2012; Yang et al., 2012; Shin et al., 2013; Norrmen et al., 2018). These pathways converge onto c-JUN, which is important for many repair Schwann cell functions. Interestingly, c-JUN deficiency has only a minimal impact on repair Schwann cell number in distal nerves. Furthermore, only a small subset of injury-induced genes in distal Schwann cells has been attributed to c-JUN activity (Arthur-Farraj et al., 2012; Hung et al., 2015; Ramesh et al., 2022). Therefore, it is likely that other transcription factors participate in the generation of repair Schwann cells.
Transcription factor EB (TFEB) and TFE3 belong to the MiTF-TFE family of basic helix-loop-helix leucine zipper transcription factors, which form homo- or heterodimers with each other (Fisher et al., 1991), and have been well characterized for their shared function in regulating lysosomal and autophagic processes (Martina et al., 2014; Pastore et al., 2017; Puertollano et al., 2018). Recent research has highlighted the role of TFEB or the combined function of TFEB and TFE3 in regulating the pluripotency transcription network, cellular stemness, and cell proliferation, in certain cases independently of its role in autophagy-lysosomal biogenesis (Palmieri et al., 2011; Doronzo et al., 2019; Tan et al., 2021; Yuizumi et al., 2021; Slade et al., 2022). In peripheral nerves, TFEB has been shown to translocate into the nuclei of distal Schwann cells soon after nerve injury; however, its function in the Schwann cell is unknown (Reed et al., 2020). Considering that conversion of distal Schwann cells to repair Schwann cells involves acquisition of stem cell-like features (Arthur-Farraj et al., 2017; Clements et al., 2017), we reasoned that TFEB alone or in combination with TFE3 may participate in the generation of repair Schwann cells.
To examine the combined role of TFEB and TFE3 (TFEB/3) in Schwann cells following nerve injury, we generated a mouse line in which floxed-Tfeb is deleted under a Schwann cell-specific promoter in a Tfe3-null background (Dhh-Cre+:Tfebflox/flox;Tfe3KO) (Tfeb/3 SC-dKO). Our study demonstrates that removal of TFEB/3 from Schwann cells impairs repair Schwann cell formation and proliferation in adult mice following peripheral nerve injury. RNA-seq analysis shows a significant impairment in the molecular reprogramming associated with repair Schwann cell generation. Additionally, Tfeb/3 SC-dKO mice exhibit hindered axon regrowth and target reinnervation and delayed recovery of sensorimotor function. Schwann cell-specific deletion of Tfeb alone or Tfe3 deficiency alone did not impact nerve regeneration. Overall, our results show that the combined function of TFEB/3 is important for repair Schwann cell generation and function during peripheral nerve regeneration and repair.
Materials and Methods
Animals
All animal procedures were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) at Rutgers University. Mice and rats were housed in microisolator cages on IVC racks in standard 12 h light/dark cycles with food and water provided ad libitum. Tfebflox/flox (C57BL/6N-Tfebtm1a(EUCOMM)Wtsi/Bay; RRID:IMSR_WTSI:490) mice (Settembre et al., 2012) were provided by Dr. Radek Dobrowolski (Reddy et al., 2016). Dhh-Cre [FVB(Cg)-Tg(Dhh-cre)1Mejr/J; RRID:IMSR_JAX:012929] mice (Jaegle et al., 2003) had been used in our previous study (Basak et al., 2015). Tfe3KO (B6.129S1-Tfe3tm1Est/Mmjax; RRID:MMRRC_042292-JAX) mice (Steingrimsson et al., 2002) were purchased from Jackson Laboratory. Mice were crossbred to generate Dhh-Cre-driven Schwann cell-specific Tfeb knock-out mice on a global Tfe3 knock-out background (Tfeb/3 SC-dKO) (Dhh-Cre+:Tfebflox/flox;Tfe3KO), control mice (Dhh-Cre−:Tfebflox/flox;Tfe3+), Tfeb SC-KO (Dhh-Cre+:Tfebflox/flox;Tfe3+), or Tfe3KO mice (Dhh-Cre−:Tfebflox/flox;Tfe3KO). Experiments were conducted using age-matched colony controls. Mice of either sex between 2 and 6 months of age were used for experiments. Sex differences were not examined in this study. Criteria for exclusion of animals during the experiment were animal illness or postsurgery illness. No animals or data were excluded during analysis. Surgeries for each cohort were performed on the same day, and collection of tissue was performed in the same order as surgery to minimize confounders. Timed pregnant SAS Sprague Dawley rats purchased from Charles River Laboratories were used for dorsal root ganglia neuron (DRG) culture at embryonic day 15 (E15) or for Schwann cell culture at postnatal day 2 (P2; detailed below). Rat embryos and pups are pooled regardless of sex.
Cell cultures
HEK293FT (Invitrogen, catalog #R70007; RRID:CVCL_6911) cells were used for lentiviral transduction (detailed below). Cells were grown in Dulbecco's modified Eagle's medium (DMEM; Corning, catalog #10-017-CV), 10% fetal bovine serum (FBS; Atlas Biologicals, catalog #F-0500-D), with 1 mM sodium pyruvate (Thermo Fisher Scientific, catalog #11360070), 1× MEM nonessential amino acids (Thermo Fisher Scientific, catalog #11140050), and Geneticin (G418 sulfate, catalog #11811031) at 37°C in a 10% CO2 incubator.
Primary rat Schwann cells were collected from P2 neonates as previously described (Monje, 2018). Briefly, sciatic nerves from P2 neonates were digested in 0.25% Trypsin (Thermo Fisher Scientific, catalog #15050057) with 0.1% collagenase (Worthington Biochemical, catalog #LS004194) for 30 min. Afterward, nerves were centrifuged at 50 × g for 5 min. The supernatant was removed, and the nerves were triturated with a glass pipette in DMEM with 10% FBS. The suspension was plated in media containing DMEM, 10% FBS, 1× Glutamax (Thermo Fisher Scientific, catalog #35050079), and 1× Penicillin/Streptomycin (Corning, catalog #30-002-CI) onto 60 mm plates coated with poly-ʟ-lysine (PLL; MilliporeSigma, catalog #P7890) at a concentration of 8 sciatic nerves/60 mm plate. Fibroblasts were removed by adding 10 µM cytosine-ß-arabinofuranoside hydrochloride (Ara-C) to media for 3 d, followed by incubation with 20 µg/ml mouse CD90 monoclonal antibody cloneT11D7e (Thy1.1; Bio-Rad, catalog #MCA04G) for 30 min followed by addition of 400 µl rabbit serum complement (MilliporeSigma, catalog #234400). Cells were then expanded and grown in Schwann cell growth media: DMEM, 10% FBS, 1× Glutamax, 1× Penicillin/Streptomycin, 10 ng/ml Recombinant Human NRG1-β 1/HRG1-β 1 EGF domain (Nrg1-EGF domain; VWR, catalog #10025-698), and 2 µM forskolin (MilliporeSigma, catalog #F3917) for experimentation.
For in vitro myelination assays, primary DRG were collected from E15 rat embryos as previously described (Monje, 2018). Briefly, E15 rats were dissected and their spinal cords exposed. DRGs were plucked from spinal cords and dissociated with 0.25% trypsin and plated at a density of 1.35 DRG/coverslip on coverslips precoated with growth factor reduced Matrigel (Corning, catalog #356231). DRGs were grown in DRG culture media: neurobasal media (Thermo Fisher Scientific, catalog #21103-049), with 0.08% d-(+)-glucose (MilliporeSigma, catalog #G27528), 1× Penicillin/Streptomycin, 1× B-27 supplement (Thermo Fisher Scientific, catalog #17504-044), 1× Glutamax, and 50 ng/ml 2.5S nerve growth factor (NGF; Thermo Fisher Scientific, catalog #NC0419584). To get a pure neuron culture, media was supplemented with 10 µM 5-fluoro-2′-deoxyuridine (FUDR) and 10 µM uridine (U) for 3 d. Afterward, DRG were maintained in DRG culture media with media changed every 2 d until Schwann cells were plated.
Surgical procedures
Sciatic nerve transection or crush injuries were performed on male or female mice aged 2–6 months. Under isoflurane, and in aseptic conditions, a skin incision was made in the thigh and the sciatic nerve was exposed. For transections, the sciatic nerve was cut distally adjacent to the sciatic notch using angled microdissecting spring scissors (Dumont, catalog #RS-5658). For crush injuries, the sciatic nerve was crushed using angled forceps (Dumont, catalog #RS-5058) for 30 s, then the forceps were rotated 90°, and the same site was crushed for an additional 30 s. The crush site was marked using charcoal-coated forceps. The wound was sutured, and an analgesic was administered. The sciatic nerve was collected at different time points for experimentation.
Lentiviral transduction
Lentiviral transduction was performed as previously described (Heffernan and Maurel, 2018). Doxycycline-inducible, flag-tagged TFEBS211A in a pSLIK-Neo backbone (Addgene, catalog #25735; Shin et al., 2006) was a gift from Dr. Radek Dobrowolski. HEK293FT cells were passaged onto 100 mm cell culture plates coated with PLL; 5 × 106 cells/plate. The next morning, cell media was changed to Advanced DMEM (Thermo Fisher Scientific, catalog #12491051) with 2% FBS, 1× Glutamax, 1% chemically defined (CD) lipid concentrate, and 0.03 mM cholesterol (MilliporeSigma, catalog #C4951). Cells were transfected using a calcium phosphate transfection kit (Thermo Fisher Scientific, catalog #K278001) using 11 µg packaging plasmid psPAX2 (Addgene, catalog #12260) and 5.5 µg envelope plasmid PMD2.G (Addgene, catalog #12259) with 16.5 µg FLAG-TFEBS211A pSLIK-Neo and 0.25 M CaCl2. After 48 h, viral supernatant was collected and mixed in a 3:1 ratio with Lenti-X Concentrator (Takara, catalog #631232) for 24 h. Rat Schwann cells were plated at a density of 3.5 × 105 cells in PLL-coated 60 mm plates. The next day, viral supernatant was added to Schwann cell cultures in Schwann cell growth media. Cells were then passaged to PLL-coated 100 mm plates with Schwann cell growth selection media: DMEM, 10% FBS, 1× Glutamax, 10 ng/ml Nrg1-EGF domain, and 2 µM forskolin with 400 µg/ml Geneticin for selection, with media changes every 2 d for 1 week. After which, media was changed to Schwann cell growth media.
To induce FLAG-TFEBS211A expression, doxycycline (Thermo Fisher Scientific, catalog #NC0424034) was added to media at a concentration of 1 µg/ml for at least 24 h and with fresh media supplemented with doxycycline every 48–56 h.
Differentiation assay and differentiation maintenance
For Schwann cell differentiation, Schwann cells were plated (5 × 104 cells) onto 10 mm glass coverslips in 24-well plates for IF or onto 6-well plates for WB (5 × 105 cells). Cell media was then replaced to Schwann cell growth media without Nrg1-EGF or forskolin for 24 h. Then media was replaced with DMEM, 1% FBS, 1× Glutamax, and 1× Penicillin/Streptomycin for an additional 24 h. To induce differentiation, cells were given new media supplemented with 0.75 mM N6,2′-O-dibutyryladenosine 3′,5′-cyclic monophosphate (db-cAMP; MilliporeSigma, catalog #D0627). Media changes with fresh treatment of db-cAMP would recur every 2.5 d or shorter depending on time course. Doxycycline was supplemented in media at different time points as indicated in Figures 6B and 7A.
Myelination assay
For Schwann cell–DRG cocultures, Schwann cells were plated onto the DRG coverslips with 1 × 105 cells per coverslip. Cultures were maintained in coculture media: minimum essential media (MEM; Corning, catalog #15-010-CM), 10% FBS, 0.4% d-(+)-glucose, 1× Glutamax, 1× penicillin and streptomycin, and 50 ng/ml NGF, with media changed every 2 d. After 5 d, allowing Schwann cells to reach confluency, myelination was induced by using new coculture media supplemented with 50 µg/ml of ʟ-ascorbic acid (MilliporeSigma, A5960). Doxycycline addition time courses are indicated in Figures 6G and 7D.
Collection of nerves for RNA-seq analysis
Mice were killed with CO2, and intact and distal sciatic nerves were collected with distal nerves cut to remove ∼1 mm from injury site. Nerves were placed into ice-cold DEPC treated PBS with RNasin where the epineurium was removed from intact nerves and then nerves were snap frozen in liquid nitrogen. Nerves were then placed into a 2 ml bead mill tube with 1 ml Tri Reagent. Nerves were homogenized in a Fisherbrand Bead Mill 24 homogenizer for 30 s at 5 m/s for two cycles. Chloroform was added and mixed and the tubes were centrifuged. The aqueous phase was removed and mixed with 70% EtOH. The mixture was then centrifuged through the NucleoSpin RNA Mini Kit (Macherey-Nagel, catalog #740955.50) and RNA extraction was completed following the manufacturer instructions. RNA was dissolved in RNAse-free water. For RNA sequencing, RNA samples were sent to PrimBio Research Institute for mRNA enrichment, library preparation, sequencing, and alignment. Raw reads from PrimBio were analyzed with R version 4.2.1 in R Studio using DEseq2 to obtain differential expression and normalized counts data (Love et al., 2014).
GO term enrichment analysis and KEGG pathway analysis of DEGs were performed using Enrichr and graphed using GraphPad Prism 10.1.0 (Chen et al., 2013; Kuleshov et al., 2016; Xie et al., 2021). Gene set enrichment analysis was performed using GSEA software (Mootha et al., 2003; Subramanian et al., 2005). Heat maps were made using Superheat R package (Barter and Yu, 2018). Volcano plots were made with Enhanced Volcano R package (Blighe et al., 2024).
Transmission electron microscopy and morphometric analysis and toluidine blue staining
Nerves for electron microscopy and toluidine blue staining were fixed in 4% PFA (Electron Microscopy Sciences, catalog #15710) and 2.5% glutaraldehyde (Electron Microscopy Sciences, catalog #16365) in 0.1 M sodium cacodylate (Electron Microscopy Sciences, catalog #12300) in phosphate buffer (Electron Microscopy Sciences, catalog #19340-72) for 24 h at 4°C. Nerves were postfixed with 1% OsO4 (Electron Microscopy Sciences, catalog #19152) and 1.5% KFeCN (Electron Microscopy Sciences, catalog #25154) in phosphate buffer for 5 h then washed with distilled water. Nerves were then dehydrated with graded ethanol steps. Nerves were then infiltrated with propylene oxide (Electron Microscopy Sciences, catalog #20412) followed by gradual replacement with resin: Embed 812 (Electron Microscopy Sciences, catalog #14900), DDSA (Electron Microscopy Sciences, catalog #13710), NMA (Electron Microscopy Sciences, catalog #19000), and DMP-30 (Electron Microscopy Sciences, catalog #13600) in a 2:1.6:0.8:0.077 ml ratio. Nerves were embedded in resin at 65°C for 24 h. For EM, nerves embedded in resin were sent for sectioning and processing to the Rutgers Robert Wood Johnson Medical School Core Imaging Lab. Sections were imaged using a ThermoFisher-FEI Tecnai 12 BioTwin TEM. Analyses were performed using Fiji (Schindelin et al., 2012). g-ratios were calculated using MyelTracer (Kaiser et al., 2021). Fibers possessing myelin abnormalities were excluded from the quantification. At least 100 axon-myelin units were counted per individual animal.
For semithin toluidine sections, blocks were sectioned using a Leica EM UC7 ultramicrotome at a 300–500 nm thickness. Sections were collected onto glass coverslips and stained with a 1% toluidine blue (MilliporeSigma, catalog #T3260) and 1% sodium borate solution in distilled water for 5–10 min. Excess stain was washed away, and coverslips were mounted with Permount (Thermo Fisher Scientific, catalog #Sp15-100). Images were taken with a Hamamatsu Orca-ER camera using a 60× objective on a Nikon Eclipse TE2000-U Epi-fluorescence microscope. Whole nerve cross sections were imaged and counted for myelin abnormalities in intact nerves or stages of myelin degeneration in injured distal nerves.
Immunofluorescence
For nerve collection, mice were killed under CO2. Nerves were dissected out and fixed in 4% PFA in PBS solution for 1 h. After washing, nerves were cryopreserved in 30% sucrose in water for 24–48 h. Nerves were then embedded using tissue freezing media [General Data, catalog #TFM-(color)] and frozen at −20°C. Nerves were then sectioned using a Leica CM3050S cryostat at a 12 µm thickness for all sections.
For staining tissue, nerve sections were thawed at room temperature for 10 min and then rehydrated in PBS. Nerves were then postfixed and permeabilized for 2 min each in 50% acetone in distilled water, 100% acetone, and then 50% acetone in distilled water. Sections were blocked in 0.3% Triton X-100 with 5% normal donkey serum for 1 h. For p75-NGFR staining, sections were permeabilized with 0.5% Triton X-100 for 20 min and then blocked using 1% BSA with 5% normal donkey serum in PBS for 1 h. For IBA1 staining, sections were permeabilized and blocked with 0.3% Triton X-100 with 5% normal donkey serum for 1 h. Primary antibodies were diluted into 0.3% Triton X-100 with 5% normal donkey serum, and sections were placed at 4°C overnight. The next day, sections were washed with PBS and then secondary antibody diluted in 0.3% Triton X-100 with 5% normal donkey serum added for 1 h at room temperature. After washing, DAPI (4′,6-diamidino-2-phenylindole; Thermo Fisher Scientific, catalog #D1306, 5 µg/ml) was added for 10 min in PBS, and slides were mounted with Fluoromount G (SouthernBiotech, catalog #0100-01).
For NMJ staining, whole hind legs were flayed and fixed in 0.5% PFA in PBS. Soleus and EDL muscles were dissected and postfixed in 4% PFA in PBS for 20 min. Muscles were treated with glycine for 30 min then stained with RBTX (tetramethylrhodamine-bungarotoxin) for 15 min. Muscles were chilled in methanol for 5 min at −20°C. Then, muscles were treated with 0.2% Triton X-100 for 1 h. Primary antibody was diluted in diluent at 4°C overnight. After washing in 0.2% Triton X-100, secondary antibodies diluted in diluent were added for 1 h. Muscles were washed and then mounted.
For cell cultures, cells on coverslips were fixed with 4% PFA in PBS for 20 min and then washed with PBS. For MBP staining of Schwann cell–DRG cocultures, cultures were additionally fixed with 100% methanol at −20°C for 20 min and then washed with PBS. Cultures were then blocked and permeabilized with 0.3% Triton X-100 with 10% normal donkey serum for 1 h. Primary antibody was diluted into 0.3% Triton X-100 with 10% normal donkey serum and sections were placed at 4°C overnight. The next day, sections were washed with PBS and then secondary antibody diluted in 0.3% Triton X-100 with 10% normal donkey serum was added for 1 h. After washing, the coverslips were DAPI stained and mounted onto slides. Myelin index was calculated as previously described (Syed et al., 2010). To account for variation of myelin formation among individual coverslips, for each condition, 2–3 coverslips were counted. For control conditions, pairwise comparisons were made between all coverslip leading to 4–9 ratios per condition per experiment. For experimental conditions, pairwise comparisons were made between all experimental coverslips and control coverslips leading to 4–9 ratios per condition per experiment. Ratios were then averaged to yield the mean ± SEM.
Primary antibodies used for immunofluorescence are as follows: TFEB (Bethyl Laboratories, catalog #A303-673A, RRID: AB_11204751, 1:1,000), p75-NGFR (p75-NTR, MilliporeSigma, catalog #07-476, RRID: AB_310649, 1:1,000), RUNX2 (Cell Signaling Technology, catalog #8486S, RRID: AB_10949892, 1:500), SOX2 (MilliporeSigma, catalog #AB5603, RRID: AB_823640, 1:1,000), Ki67 (Cell Signaling Technology, catalog #9129, RRID: AB_2687446, 1:400), c-JUN (Cell Signaling Technology, catalog #9165S, RRID: AB_2130165, 1:800), FLAG (rabbit, Cell Signaling Technology, catalog #14793, RRID: AB_2572291, 1:1,000; mouse, MilliporeSigma, catalog #F1804, RRID: AB_262044, 1:1,000), PRX (Novus Biologicals, catalog #NBP1-89598, RRID: AB_11004050, 1:1,000), EGR2 (KROX20, Abcam, catalog #AB245228, RRID: AB_2934181, 1:1,000), MBP (mouse, BioLegend, catalog #836504, RRID: AB_2616694, 1:300), IBA1 (Fujifilm Wako, catalog #019-19741, RRID: AB_839504, 1:400), SV2 (DHSB, RRID: AB_2315387, 1:10), SMI312 (Sternberger Monoclonals, BioLegend, catalog #837904, RRID: AB_2566782, 1:1,000), SOX10 (Thermo Fisher Scientific, catalog #AF2864, RRID: AB_442208, 1:150), and CD31 (R&D Systems, catalog #AF3628, RRID:AB_2161028, 1:400). Staining of neuromuscular junctions was performed with RBTX (Molecular Probes, catalog #T-1175, 1:200).
Secondary antibodies used for immunofluorescence are as follows: donkey anti-rabbit Alexa Fluor 647 (Jackson ImmunoResearch, catalog #711-605-152, RRID:AB_2492288), donkey anti-goat Alexa Fluor 488 (Jackson ImmunoResearch, catalog #705-545-003, RRID:AB_2340428), donkey anti-rabbit Rhodamine (TRITC ;Jackson ImmunoResearch, catalog #711-025-152, RRID:AB_2340588), donkey anti-mouse Alexa Fluor 488 (Jackson ImmunoResearch, catalog #715-545-150, RRID:AB_2340846), donkey anti-mouse Rhodamine (TRITC ;Jackson ImmunoResearch, catalog #715-025-150, RRID:AB_2340766), donkey anti-rabbit Alexa Fluor 488 (Jackson ImmunoResearch, catalog #711-545-152, RRID:AB_2313584). All secondary antibodies were used at a 1:500 dilution.
All images of cell cultures were taken with a Hamamatsu Orca-ER camera using a 10×, 20×, or 40× objective on a Nikon Eclipse TE2000-U Epi-fluorescence microscope. At least 10 images were taken per coverslip and at least 2–3 coverslips used for individual experiments, with at least three experiments. All images of nerve sections were taken on a Zeiss LSM 510 Laser Scanning Confocal Microscope using a 10×, 20×, or 40× objective. For nerve sections, 3–5 images were taken per nerve section. Image analysis was performed using Fiji.
Western blot analysis
During dissection, 1 mm of distal or proximal nerves was cut to avoid the area containing the wound. Sciatic nerves were frozen with liquid nitrogen. Nerves were homogenized in bead mill tubes with lysis buffer containing 150 mM Tris-HCl (pH6.8), 6% sodium dodecyl sulfate (SDS), 0.001% bromophenol blue, 10% β-mercaptoethanol, 50 mM NaF, 1 mM NaVO4, and cOmplete Mini protease inhibitor cocktail (MilliporeSigma, catalog #11836170001) in a Fisherbrand Bead Mill 24 homogenizer with two 30 s cycles at 5 m/s. Homogenized nerves were kept on ice for 15 min, heated to 95°C for 3 min, and centrifuged at 16,200 × g for 15 min at 4°C. Supernatant was removed and stored at −80°C.
Cell cultures were lysed using a lysis buffer containing 95 mM NaCl, 25 mM Tris-Cl, 1 mM EDTA, 1% SDS, 50 mM NaF, 1 mM NaVO4, 2 µg/ml aprotinin, 10 µM leupeptin, and 1 mM PMSF. Lysates were kept on ice for 10 min and then boiled at 100°C for 10 min. Lysates were centrifuged at 16,200 × g for 10 min. Supernatant was removed and stored at −80°C. Lysates were loaded onto SDS-PAGE gels and then transferred to PVDF membranes. Membranes were blocked using 2% blotting-grade blocker (Bio-Rad, catalog #1706404) in TBS for 1 h. Membranes were incubated with primary antibody in 5% BSA in TBST overnight at 4°C. Membranes were washed in TBST and then incubated with secondary antibodies diluted in 2% blotting-grade blocker in TBST for 1 h, followed by washing with TBST. Membranes were imaged using a LI-COR Odyssey.
Primary antibodies used for Western blotting are as follows: TFEB (Bethyl Laboratories, catalog #A303-673A, 1:2,000), TFE3 (MilliporeSigma, catalog #HPA023881, RRID: AB_1857931, 1:1,000), β-actin (MilliporeSigma, catalog #A5441, RRID: AB_476744, 1:5,000), Calnexin (Enzo Life Sciences, catalog #ADI-SPA-860-D, RRID: AB_2038898, 1:1,000), p75-NGFR (MilliporeSigma, catalog #07-476, 1:1,000), RUNX2 (Cell Signaling Technology, catalog #8486S, 1:1,000), SOX2 (MilliporeSigma, catalog #AB5603, 1:1,000), c-JUN (Cell Signaling Technology, catalog #9165S, 1:1,000), FLAG (rabbit, Cell Signaling Technology, catalog #14793, 1:1,000), EGR2 (KROX20, Abcam, catalog #AB245228, 1:1,000 or gift from Dr. Dies Meijer, 1:1,000), PRX (Novus Biologicals, catalog #NBP1-89598, 1:2,000), MPZ (MilliporeSigma, catalog #AB9352, RRID: AB_571090, 1:500), MBP (rabbit, EMD Millipore, catalog #AB980, RRID: AB_92396, 1:1,000), LC3I and LC3II (Cell Signaling Technology, catalog #4108, RRID:AB_2137703, 1:1,000), and GAPDH (MilliporeSigma, catalog #MAB374, AB_2107445, 1:1,000).
Secondary antibodies used for Western blotting are as follows: goat anti-rabbit Alexa Fluor 790 (Jackson ImmunoResearch, catalog #111-655-144, RRID:AB_2338086), goat anti-mouse Alexa Fluor 680 (Jackson ImmunoResearch, catalog #115-625-146, RRID:AB_2338935), goat anti-rabbit Alexa Fluor 680 (Jackson ImmunoResearch, catalog #111-625-144, RRID:AB_2338085), donkey anti-mouse Alexa Fluor 790 (Jackson ImmunoResearch, catalog #715-655-150, RRID:AB_2340870), and donkey anti-chicken Alexa Fluor 790 (Jackson ImmunoResearch, catalog #703-655-155, RRID:AB_2340382). All secondary antibodies were used at a 1:10,000 dilution.
Enhancer motif analysis
Motif matrices for TFEB and TFE3 were pulled from JASPAR (Castro-Mondragon et al., 2022). Find Individual Motif Occurrences (FIMO) from the MEME suite (Bailey et al., 2015) was used to identify enrichment of TFEB/3 motifs across ShamDB and InjuryDB enhancer peaks (Hung et al., 2015).
LC3II autophagic flux assay
Sciatic nerves were dissected out of mice and placed into DMEM, and the epineurium was removed using forceps. Nerves were then maintained in DMEM with 10% FBS, 1× Glutamax, and 1× Penicillin/Streptomycin for 5 d. Three hours prior to lysis, new media was added with either 20 mM ammonium chloride or vehicle after which nerves were homogenized and lysed for Western blotting.
Beam walk analysis
The beam walk test was performed as described with modifications (Luong et al., 2011). Briefly, mice were placed on one end of a 12-mm-wide and 1-m-long wooden beam and walked to the other side. After having each mouse walk the length of the beam at least three times for training, mice were tested on a 6-mm-wide and 1-m-long wooden beam. Each trial was recorded with a camera pointed at the mouse in a rear view. The number of missteps were counted for each trial and set as a ratio to the total number of steps taken by the two hind legs.
Toe pinch analysis
Animals were scruffed and the three most lateral toes of the hind limbs were examined with the most medial toe of each foot labeled as Toe 3, the most lateral toe as Toe 5, and the toe in between as Toe 4. Each toe was pinched three times and given a score of 1 if the mouse responds and 0 if it does not. For each toe, there is a maximal possible score of 3 if the mouse responds to every pinch and a minimal score of 0 if the mouse has no response. Mice were examined prior to injury and every 3 d subsequently.
Experimental design and statistical analysis
Experimenter blinding was performed for all the analyses. Sample sizes of 3–5 animals per experimental condition were used, similar to the range used in previously published studies in the field (Arthur-Farraj et al., 2012; Napoli et al., 2012; Beirowski et al., 2017; Roberts et al., 2017; Norrmen et al., 2018; Reed et al., 2020; Li et al., 2021; Daboussi et al., 2023). Statistical analysis was performed using GraphPad Prism software version 10.1.0. The following tests were used: unpaired two-tailed Student's t test, multiple unpaired two-tailed Student's t test, one-way ANOVA, or two-way ANOVA. One-way ANOVA was followed by Tukey's multiple-comparisons test. Two-way ANOVAs were followed by Šídák's multiple-comparisons test when only comparisons indicated by significance brackets were performed, Tukey's multiple-comparisons test when all possible comparisons were performed, or Dunnet's multiple-comparisons test when comparisons to a single control were performed. A hypergeometric test was used to compare genes that either fail to upregulate or fail to downregulate properly after injury with that of GEO GSE38693 (Arthur-Farraj et al., 2012). Statistic test used for individual experiments are listed within figure legends. For DEseq2 analysis, genes were considered differentially expressed when the log2FC ≥ |1| and padj ≤ 0.05. For GSEA analysis, a gene set permutation was used with an FDR q value ≤ 0.05 considered significant. Data are presented as mean ± SEM, where p ≤ 0.05 was considered significant. Significant p values are displayed in graphs and nonsignificant p values are represented by ns.
Data availability statement
RNA-seq data have been deposited at the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) and are publicly available as of the date of publication. GEO accession number: GSE249114.
Results
TFEB/3 deficiency impairs transcriptional reprogramming toward repair Schwann cell formation
Schwann cell-specific Tfeb/3 double knock-out mice (Dhh-Cre+:Tfebflox/flox;Tfe3KO, Tfeb/3 SC-dKO) developed normally without any gross myelin abnormalities in the sciatic nerves when compared with the control mice (Dhh-Cre−:Tfebflox/flox;Tfe3+), assessed at P14 and at 3–4 months of age (Fig. 1A,B). Ultrastructural analysis of the adult nerves showed normal axon integrity (Fig. 1C), myelin thickness (g-ratio; Fig. D–F), and Remak bundle organization (Fig. 1G,H). Furthermore, bulk RNA-seq analysis of sciatic nerves of age-matched controls and mutant mice showed similar gene expression profiles (Fig. 1I). Therefore, Schwann cell TFEB/3 are dispensable for proper organization of the peripheral nerve.
TFEB/3 expression increases in response to nerve injury. A, Representative toluidine blue stained semithin images of control and Tfeb/3 SC-dKO intact sciatic nerves at P14 (left) and 3–4 months (P90–110; center). Scale bar, 5 µm. Representative transmission electron micrographs of intact sciatic nerves of control and Tfeb/3 SC-dKO of mice aged P90–110. Scale bar, 2 µm. B, Graph of percentages of abnormal myelin sheaths categorized as infolds, outfolds, focal thickenings, or degenerated myelin of mice aged P90–110. n = 3–4, tinfold(5) = 1.507, toutfold(5) = 0.22946, tfocal thickening(5) = 1.672, tdegenerate(5) = 0.2175, multiple unpaired t tests. C, Graph of percentage of large diameter axons that show signs of degeneration (shriveled or darkened appearance) of mice aged P90–110. n = 3, t(4) = 0.5905, unpaired t test. D, Graph of g-ratios of intact sciatic nerves of control (black open circles) and Tfeb/3 SC-dKO (blue open squares) of mice aged P90–110 (left). E, Average g-ratio of control and Tfeb/3 SC-dKO nerves (right). n = 3, t(4) = 0.3028, unpaired t test. F, Binned g-ratios of control and Tfeb/3 SC-dKO nerves (right). n = 3, t0–2(4) = 0.08022, t2–3(4) = 0.06084, t3–4(4) = 0.4832, t4–5(4) = 0.09103, t>5(4) = 1.041, Multiple unpaired t tests. G, Average number of axons per Remak bundle of mice aged P90–110. n = 3, t(4) = 0.2672, unpaired t test. H, Binned axon frequency per non-myelinating Schwann cell. n = 3, t1(4) = 0.718, t2–5(4) = 0.3394, t6–10(4) = 0.5162, t11–15(4) = 0.2051, t15–20(4) = 0.2480, t20–25(4) = 0.3517, t>25(4) = 1.082, multiple unpaired t tests. I, Volcano plot of DEGs in Tfeb/3 SC-dKO versus control intact nerves (top) of mice aged P90–110. DEGs with Log2FC ≥ |1| and padj ≤ 0.05 represented as red circles; all others represented as gray x's. Select genes marked with black label. Table of DEGs (bottom). J, Representative immunostaining images of control intact and 1DPI distal nerves for TFEB (magenta) and DAPI (blue). Scale bar, 20 µm. Graph of TFEB+ cell percentage. n = 3, t(4) = 5.126, unpaired t test. K, Western blot of TFEB in intact contralateral and 7DPI distal nerves of control and Tfeb/3 SC-dKO. n = 3, FInteraction(1,8) = 2.769, Finjury(1,8) = 32.35, Fgenotype(1,8) = 17.54, two-way ANOVA with Šídák's multiple-comparisons test. L, Western blot of TFE3 in intact contralateral and 7DPI distal nerves of control and Tfeb/3 SC-dKO. n = 3, FInteraction(1,8) = 4.832, Finjury(1,8) = 8.41, Fgenotype(1,8) = 102.6, two-way ANOVA with Šídák's multiple-comparisons test.
Following sciatic nerve transection injury, we observed increased TFEB nuclear localization in distal cells as early as 1 d postinjury (DPI), indicating its activation (Fig. 1J). We also conducted Western blot analysis to assess TFEB/3 expression levels in distal nerves. In control mice, while TFEB/3 protein levels in intact nerves were low, both increased in the distal nerves at 7DPI. This TFEB/3 expression was abolished in the Tfeb/3 SC-dKO mice (Fig. 1K,L).
Following nerve injury, distal Schwann cells undergo transcriptional reprogramming to become repair Schwann cells. To determine whether Schwann cell TFEB/3 deficiency alters the injury-responsive transcriptome, we generated bulk RNA-seq data from distal nerves of control and Tfeb/3 SC-dKO mice at 5DPI, which corresponds to the period of repair Schwann cell generation and proliferation. Among the injury-responsive downregulated genes in the control mice, we found 258 genes that either failed to downregulate (122) or were downregulated to a lesser degree (136 genes) in Tfeb/3 SC-dKO mice (Fig. 2A, left, B; Extended Data Table 2-1). These included myelin protein genes including Mbp, Mpz, Prx, and Pmp2 and the promyelinating transcription factor Egr2 (Krox20; Fig. 2C, center and right; Extended Data Table 2-1). Junctional proteins (Cdh1, Cldn19, Cadm4) normally expressed in myelinating Schwann cells also remained at a higher level in the mutant nerves (Fig. 2C, center and right; Extended Data Table 2-1).
TFEB/3 deficiency impairs transcriptional reprogramming toward repair Schwann cell formation. A, Venn diagram comparing genes that failed to be properly downregulated between Tfeb/3 SC-dKO versus control 5DPI distal nerves (blue) and downregulated genes between control distal versus control intact (red) and Tfeb/3 SC-dKO distal versus Tfeb/3 SC-dKO intact nerves (green; left). Venn diagram comparing genes that failed to be properly upregulated between Tfeb/3 SC-dKO versus control 5DPI distal nerves (blue) and upregulated genes between control distal versus control intact (red) and Tfeb/3 SC-dKO distal versus Tfeb/3 SC-dKO intact nerves (green; right). See Extended Data Table 2-1 for list of DEGs. B, Heat map of all 1,051 DEGs between 5DPI distal Tfeb/3 SC-dKO versus control nerves. C, Volcano plots of DEGs in Tfeb/3 SC-dKO versus control 5DPI distal nerves. DEGs with Log2FC > |1| and padj ≤ 0.05 represented as red circles; all others represented as gray x's. See Extended Data Table 2-1 for the list of DEGs. D, KEGG pathway analysis of 475 DEGs that failed to downregulate between Tfeb/3 SC-dKO versus control 5DPI distal nerves (left). KEGG pathway analysis of 576 DEGs that failed to upregulate between Tfeb/3 SC-dKO versus control 5DPI distal nerves (right). E, Table of GSEA enrichment scores of Tfeb/3 SC-dKO versus control 5DPI distal nerves of MSigDB gene sets for embryonic stem cell modules, cell cycle, and regulation by polycomb repressive complex. F, GSEA enrichment plot of Tfeb/3 SC-dKO versus control distal nerves for “Apical Junctions” from MSigDB Hallmarks. G, GSEA enrichment plot of Tfeb/3 SC-dKO versus control distal nerves for “Reactome S-phase” (top). GSEA enrichment plot of Tfeb/3 SC-dKO versus control distal nerves for “G2 M Checkpoint” from MSigDB Hallmarks (bottom). H, GO term biological processes for 576 downregulated DEGs between Tfeb/3 SC-dKO versus control distal nerves (left). GO term biological processes for 475 upregulated DEGs between Tfeb/3 SC-dKO versus control distal nerves (right).
Table 2-1
List of DEGs corresponding to Venn diagrams from Figure 2A Table of DEGs in 5DPI distally transected sciatic nerves of Tfeb/3 SC-dKO relative to control compared to upregulated or downregulated DEGs of 5DPI distally transected sciatic nerves relative to intact sciatic nerves for both control and Tfeb/3 SC-dKO. Download Table 2-1, XLSX file.
Among the injury-responsive upregulated genes in the control mice, we found 423 genes which failed to upregulate (137) or exhibited a diminished increase (286) in the mutant nerves (Fig. 2A, right, B; Extended Data Table 2-1). These included many genes associated with repair Schwann cells including p75-Ngfr, Gdnf, Runx2, Sox2, and Olig1 (Fig. 2C, left; Extended Data Table 2-1). Despite the hindered expression of many repair Schwann cell genes in the mutant mice, there was no significant alternation in c-jun expression (Fig. 2C, left). After injury, distal Schwann cells, as they convert to repair Schwann cells, become enriched for genes involved in EMT, cell cycle, and embryonic stem cells while the polycomb-related module is typically repressed (Clements et al., 2017). Tfeb/3 SC-dKO distal nerves showed a significant decrease in EMT, cell cycle, and embryonic stem cells gene signatures (Fig. 2D–H), while polycomb-regulated genes remained high in the mutant distal nerves (Fig. 2E; Ben-Porath et al., 2008; Liberzon et al., 2015). Inhibition of the polycomb complex has been shown to increase expression of repair genes after injury (Ma et al., 2018). Overall, results from transcriptomic analysis suggest that TFEB/3 deficiency negatively impacts injury-induced transcriptional reprogramming toward repair Schwann cell formation in the distal nerves.
Repair Schwann cell generation and proliferation are impaired in the absence of TFEB/3
Next, we investigated the impact of TFEB/3 deficiency on repair Schwann cell formation. We used distal nerves collected at 3DPI and 7DPI, which correspond to the period of repair Schwann cell formation and active proliferation, respectively. Proximal or contralateral intact nerves served as controls for the absence of repair Schwann cells.
During development, p75-NGFR is highly expressed in proliferative immature Schwann cells, but, as Schwann cells differentiate, the expression becomes restricted to non-myelinating Schwann cells in mature nerves (Zimmermann and Sutter, 1983; Scarpini et al., 1988). After nerve injury, p75-NGFR reappears in repair Schwann cells (Taniuchi et al., 1986). Therefore, we first used p75-NGFR to mark the presence of repair Schwann cells in control and Tfeb/3 SC-dKO distal nerves using immunohistochemistry (Fig. 3A) and Western blot analysis (Fig. 3B). The expression was detected at a low level in intact nerves likely coming from the non-myelinating Schwann cell population. After nerve injury, while p75-NGFR expression increased drastically from 3 to 7DPI in control mice, it was significantly inhibited in Tfeb/3 SC-dKO nerves at both time points (Fig. 3B), thus indicating a defect in the appearance of repair Schwann cells following injury.
Repair Schwann cell generation and proliferation are impaired in the absence of TFEB/3. A, Representative immunostaining images of control and Tfeb/3 SC-dKO intact, 3DPI, and 7DPI distal nerves for p75-NGFR. Intact contralateral nerves have been modified with a higher brightness for visualization. 3DPI and 7DPI distal nerves were imaged with the same settings. Scale bar, 100 µm. B, Western blot of p75-NGFR in proximal and distal 3DPI (top and left graph) or 7DPI nerves (bottom and right graph) of control and Tfeb/3 SC-dKO. n = 3, 3DPI: FInteraction(1,8) = 15.18, Finjury(1,8) = 1.742, Fgenotype(1,8) = 1.828, 7DPI: FInteraction(1,8) = 8.586, Finjury(1,8) = 30.27, Fgenotype(1,8) = 8.242, two-way ANOVA with Šídák's multiple-comparisons test. C, Western blot of RUNX2 in proximal and distal 3DPI (top and left graph) or 7DPI nerves (bottom and right graph) nerves of control and Tfeb/3 SC-dKO. n = 3, 3DPI: t(4) = 5.494, 7DPI: t(4) = 3.221, unpaired t test. D, Representative immunostaining images of control and Tfeb/3 SC-dKO 3DPI distal nerves for RUNX2 (gray) and DAPI (blue; top) and 7DPI distal nerves for RUNX2 (magenta) and SOX10 (green; bottom). Scale bar, 20 µm. Graphs of RUNX2+ cell percentage or RUNX2+;SOX10+ cells in the SOX10+ population. n = 3, 3DPI: t(4) = 3.38, 7DPI: tRUNX2/DAPI(4) = 6.054, tRUNX2:SOX10/SOX10(4) = 6.904, unpaired t test. E, Western blot of SOX2 in intact contralateral and 7DPI distal nerves of control and Tfeb/3 SC-dKO. n = 3, t(4) = 14.08, unpaired t test. F, Representative immunostaining images of control and Tfeb/3 SC-dKO 7DPI distal nerves for SOX2 (magenta) and SOX10 (green). Scale bar, 20 µm. Graph of SOX2+ cell percentage and SOX2+;SOX10+ cells of SOX10+ population. n = 3, 7DPI: tSOX2/DAPI(4) = 4.42, tSOX2:SOX10/SOX10(4) = 3.029, unpaired t test. G, Representative immunostaining images of control and Tfeb/3 SC-dKO 3DPI distal nerves for Ki67 (magenta) and SOX10 (green). Scale bar, 20 µm. Graph of percentage of Ki67+;SOX10+ cells of SOX10+ population (left). n = 3, t(4) = 4.579, unpaired t test. Graph of SOX10+ cells per 1mm2 (right) of distal 3DPI and 7DPI nerves. n = 3, FInteraction(1,8) = 8.407, Finjury(1,8) = 20.71, Fgenotype(1,8) = 8.213, two-way ANOVA with Šídák's multiple-comparisons test. Data represented as mean ± SEM, ns = not significant, significant p values displayed in graphs.
RUNX2 and SOX2 are injury-responsive repair Schwann cell-specific transcription factors (Parkinson et al., 2008; Parrinello et al., 2010; Hung et al., 2015; Li et al., 2021). We further assessed the impact of Schwann cell TFEB/3 loss on repair Schwann cell generation by examining the appearance of RUNX2+ or SOX2+ Schwann cells in the distal nerves. Western blot analysis showed RUNX2 induction in the distal nerves of the control mice at 3 and 7DPI. The expression was significantly inhibited in the mutant nerves (Fig. 3C). Immunostaining also showed a significant reduction in the number of RUNX2+ cells in the Tfeb/3 SC-dKO, with 55 ± 3.9 and 40 ± 2.4%, respectively, in control and mutant at 3DPI, and 52 ± 1.4 and 33 ± 2.0% at 7DPI (Fig. 3D). Coimmunostaining for SOX10 showed a significant decrease in RUNX2+ cells among the Schwann cell population at 7DPI, with 93 ± 0.2 and 77% ± 2.3% in control and mutant mice, respectively (Fig. 3D). A similar result was observed for SOX2 both by Western blot and immunostaining (Fig. 3E,F). By comparison, single deletion of either Tfeb (Dhh-Cre+:Tfebflox/flox;Tfe3+, Tfeb SC-KO) or Tfe3 (Dhh-Cre−:Tfebflox/flox;Tfe3KO, Tfe3KO) did not affect the number of RUNX2+ or SOX2+ Schwann cells in the distal nerves (Fig. 4). Altogether, these results show that the combined function of TFEB/3 is important for repair Schwann cell generation after nerve injury.
Tfeb SC-KO and Tfe3KO nerves display unaltered repair Schwann cell protein expression. A, Representative immunostaining images of control and Tfeb SC-KO 3DPI distal nerves for RUNX2 (magenta), SOX10 (green), and DAPI (blue). Scale bar, 20 µm. Graphs of RUNX2+ cell percentage (left) or RUNX2+;SOX10+ cells in the SOX10+ population (right). n = 3, tRUNX2/DAPI(4) = 0.172, tRUNX2:SOX10/SOX10(4) = 1.261, unpaired t test. B, Representative immunostaining images of control and Tfe3KO 3DPI distal nerves for RUNX2 (magenta), SOX10 (green), and DAPI (blue). Scale bar, 20 µm. Graphs of RUNX2+ cell percentage (left) or RUNX2+;SOX10+ cells in the SOX10+ population (right). n = 3, tRUNX2/DAPI(4) = 1.589, tRUNX2:SOX10/SOX10(4) = 0.4664, unpaired t test. C, Representative immunostaining images of control and Tfeb SC-KO 3DPI distal nerves for SOX2 (magenta), SOX10 (green), and DAPI (blue). Scale bar, 20 µm. Graph of SOX2+;SOX10+ cells in the SOX10+ population. n = 3, t(4) = 0.8796, unpaired t test. D, Representative immunostaining images of control and Tfe3KO 3DPI distal nerves for SOX2 (magenta), SOX10 (green), and DAPI (blue). Scale bar, 20 µm. Graph of SOX2+;SOX10+ cells in the SOX10+ population. n = 3, t(4) = 0.1284, unpaired t test. Data represented as mean ± SEM, ns = not significant, significant p-values displayed in graphs.
RNA-seq analysis predicts a defect in repair Schwann cell proliferation in the absence of TFEB/3 (Fig. 2D,E,G,H). Repair Schwann cells enter the cell cycle at ∼3DPI (Abercrombie and Johnson, 1946; Stierli et al., 2018). There was a significant reduction in the percentage of Ki67+ Schwann cells (SOX10+) in Tfeb/3 SC-dKO nerves (7 ± 1.5%) compared with control mice (17 ± 1.4%; Fig. 3G, left graph). The failure in cell cycle entry at 3DPI suggests a defect in the subsequent expansion of the repair Schwann cell population. Control mice showed a 2.6-fold increase in the Schwann cell (SOX10+) density from 3DPI to 7DPI. The increase was not seen in the mutant nerves (Fig. 3G, right graph). Therefore, TFEB/3 are essential for injury-induced repair Schwann cell proliferation and expansion in the distal nerves.
Repair Schwann cell defects occur in the continuous presence of c-JUN
The transcription factor c-JUN is an important regulator of repair Schwann cell functions. Nerve injury induces c-JUN expression in distal Schwann cells and its absence impairs the regenerative function of repair Schwann cells (Arthur-Farraj et al., 2012). Our RNA-seq analysis showed that Schwann cell TFEB/3 deficiency did not impact injury-induced c-jun expression in the distal nerves (Fig. 2C, left). Western blot analysis at 3DPI showed that c-JUN was induced normally in Tfeb/3 SC-dKO nerves, comparable with control mice (Fig. 5A, left graph). By 7DPI, however, the level was lower than in the control (Fig. 5A, right graph). To determine whether this was due to failed c-JUN expression by the mutant Schwann cells or due to an overall decrease in the repair Schwann cell population (Fig. 3G), we coimmunostained the distal nerves for c-JUN and SOX10 (Fig. 5B). At 3DPI, the numbers of SOX10+ Schwann cells were not significantly different between control and Tfeb/3 SC-dKO distal nerves, and the percentages of c-JUN expressing Schwann cells (c-JUN+/SOX10+) were also comparable (Fig. 5B, top graphs). At 7DPI, while there were fewer SOX10+ Schwann cells in the mutant mouse nerves, the percentage of c-JUN+ cells within the Schwann cell population was not significantly different when compared with the control mice (Fig. 5B, bottom left and center graphs). We also compared c-JUN fluorescence intensity in individual Schwann cells. There was no significant difference in the average fluorescence intensity per Schwann cell between control and the mutant mice (Fig. 5B, bottom right graph). This result indicates that the decrease in c-JUN protein expression seen in Tfeb/3 SC-dKO mice at 7DPI was not due to a defect in c-JUN induction, but rather due to an overall reduction in the distal Schwann cell population. We conclude that TFEB/3 deficiency does not impair injury-induced c-JUN induction in the distal Schwann cells.
Deficiency in repair Schwann cell generation occurs despite continuous presence of c-Jun. A, Western blots of c-JUN proximal and distal 3DPI (top and left graph) and 7DPI (bottom and right graph) nerves of control and Tfeb/3 SC-dKO. n = 3, unpaired t test. B, Representative immunostaining images of control and Tfeb/3 SC-dKO 3DPI and 7DPI nerves for c-JUN (magenta) and SOX10 (green). Scale bar, 25 µm. Graphs of SOX10+ cell percentage and c-JUN+;SOX10+ cell percentage of SOX10+ population at 3DPI (top) and 7DPI (bottom left and center). n = 3, 3DPI: tSOX10/DAPI(4) = 0.2926, tc-JUN:SOX10/SOX10(4) = 1.605, 7DPI: tSOX10/DAPI(4) = 3.542, tc-JUN:SOX10/SOX10(4) = 0.6085, unpaired t test. Graph of c-JUN fluorescent intensity normalized to control mean. Black closed circles represent mean percentage per biological replicate. Open gray circles represent individual nuclei. n = 3, t(4) = 0.4632, unpaired t test. C, Venn diagrams of Tfeb/3 SC-dKO 5DPI (red) and c-Jun SC-KO 7DPI (blue) comparing genes that failed to properly upregulate (top) and failed to properly downregulate (bottom). p values displayed below Venn diagrams. Hypergeometric test. See Extended Data Table 5-1. D, Injury-induced enhancer sequences revealed in a previous study of nerve injury in rat sciatic nerves (Hung et al., 2015) examined for enrichment of TFEB and TFE3 motifs and the indicated percentage of injury-induced enhancers (InjuryDB) had TFE motifs. A comparison of enhancers that are diminished after injury (ShamDB) shows a lower percentage of TFE motifs. E, Profiles of injury-induced enhancers through analysis of the active enhancer mark H3K27ac, in sham or injured nerve in the Runx2, Hmga2, p75-Ngfr, and Met genes. The pink shaded injury-induced enhancers have TFEB/TFE3 motifs. Data represented as mean ± SEM, ns = not significant, significant p values displayed in graphs.
Table 5-1
List of DEGs corresponding to Venn diagrams from Figure 5C Table of DEGs in 5DPI distally transected sciatic nerves of Tfeb/3 SC-dKO relative to control compared to 7DPI distally transected sciatic nerves of c-Jun SC-KO from previously published micro-array data (Arthur-Farraj et al., 2012). Download Table 5-1, XLSX file.
TFEB/3 binding motifs are enriched in injury-induced enhancers
Our data shows that the repair Schwann cell defects in Tfeb/3 SC-dKO persist despite the continuous presence of c-JUN. Since only a small subset of injury-induced genes has been attributed to c-JUN activity (Arthur-Farraj et al., 2012), we asked whether TFEB/3 and c-JUN modulate distinct sets of injury-induced target genes. A previous microarray analysis reported 106 genes that failed to upregulate in the distal nerves at 7DPI in mice with Schwann cell c-JUN deficiency (c-jun SC-KO; Arthur-Farraj et al., 2012). In comparing with the 576 genes from our RNA-seq data that failed to increase in Tfeb/3 SC-dKO mice at 5DPI, only 25 genes were commonly found in the two gene sets, including Gdnf, Olig1, and Runx2 (Fig. 5C, top; Extended Data Table 5-1). Among the genes that failed to downregulate, 66 and 475 genes, respectively, in c-jun SC-KO and Tfeb/3 SC-dKO nerves, 11 genes were shared, including Cdh1, Gjb1, Mbp, and Mpz (Fig. 5C, bottom; Extended Data Table 5-1). We cannot rule out the possibility that the difference in gene expression observed between the two mutant mouse lines may reflect the temporal change in gene expression from 5DPI to 7DPI. However, considering that many injury-responsive genes also maintain their expression profiles from 3 to 7DPI (Kalinski et al., 2020; Li et al., 2021; Brosius Lutz et al., 2022), our result suggests that while c-JUN and TFEB/3 may function cooperatively, each may also play a distinct role in injury-responsive gene activation.
If TFEB/3 play a role in gene activation after nerve injury, then it is likely to regulate gene regulatory elements through their preferred binding motifs. In a previous study, we had identified ∼4,000 enhancers marked by histone H3K27 acetylation (H3K27ac), which are induced in the distal nerves at 3DPI. Using this set of enhancers, we previously found enrichment of JUN-binding motifs (Hung et al., 2015), which have been recently validated (Ramesh et al., 2022), consistent with the important role of c-JUN in regulation of the nerve injury program. We determined if TFEB/3 binding sites are similarly enriched in injury-induced enhancers. Our motif analysis found that 19.7 and 46.8% of injury-induced enhancers had TFEB and TFE3 motifs, C-A-C-G-T-G-A-C and C-A/T-C-G-T/A-G, respectively (Fig. 5D). The TFE3 motif is somewhat more degenerate, giving rise to a higher prevalence of motifs. As a control, we also analyzed enhancers that decrease after injury (ShamDB), and the percentage of such enhancers with TFE motifs was significantly lower (Fig. 5D), confirming the specific enrichment of TFE motifs in injury-induced enhancers.
Since TFEB/3 binding motifs are enriched in injury-induced enhancers, we analyzed injury-induced regulatory elements in four known injury-induced genes that are also downregulated in Tfeb/3 SC-dKO mice: Runx2, p75-Ngfr, Hmga2, and Met. These genes have been identified to be induced in Schwann cells in a variety of gene profiling studies, including sorted cell and single cell RNA-seq studies (Clements et al., 2017; Kalinski et al., 2020; Brosius Lutz et al., 2022). In certain cases, loss-of-function studies have determined that these genes play a role in nerve injury (Song et al., 2006; Sachs et al., 2007; Tomita et al., 2007; Chen et al., 2016; Ko et al., 2018; Pellegatta et al., 2022). We found TFEB or TFE3 binding motifs on enhancer regions of these genes (Fig. 5E). Altogether, these results suggest that TFEB/3 function as transcriptional regulators of injury-responsive genes in Schwann cells.
TFEB activation triggers myelin loss
We conducted a series of gain-of-function experiments to assess TFEB function in Schwann cells. For the study, we generated Schwann cells expressing a constitutively active TFEB mutant (FLAG-TFEBS211A) under the control of a doxycycline-inducible promoter (Martina et al., 2012; Roczniak-Ferguson et al., 2012; Settembre et al., 2012). In this culture system, treatment with doxycycline increases TFEB nuclear translocation, thus activation within the FLAG+ Schwann cells as demonstrated by the colocalization of TFEB and FLAG (Fig. 6A). To investigate whether TFEB activation is sufficient to mimic the Schwann cell injury response, we first generated differentiated Schwann cells using the well-established cAMP treatment paradigm (db-cAMP), which induces Schwann cells to acquire a myelinating Schwann cell phenotype in the absence of axons (Jessen et al., 1991; Morgan et al., 1991; Fig. 6B). At Day 5 (d5) under the differentiating condition, many Schwann cells expressed periaxin, a differentiated Schwann cell marker. Cultures were then treated with doxycycline from d5 to d7 to induce TFEB activation. This led to a significant loss of periaxin+ cells: from 64 ± 5.7% (d5) to 34 ± 4.2% (+Dox, FLAG+; Fig. 6C). TFEB activation was also sufficient to induce RUNX2 expression in the Schwann cells (Fig. 6D). In Schwann cells, EGR2 (KROX20) and c-JUN exhibit an inverse relationship. In differentiated and myelinating Schwann cells, EGR2 (KROX20) is expressed while c-JUN is absent. When Schwann cells lose the myelinating phenotype, such as after nerve injury, EGR2 (KROX20) is downregulated and c-JUN increases (Parkinson et al., 2004; Parkinson et al., 2008; Monje et al., 2010). We observed that TFEB-induced periaxin loss was accompanied by a concomitant loss of EGR2 (KROX20) nucleoplasmic expression, from 75 ± 0.6% (d7, −Dox) to 14 ± 3.6% (d7 +Dox, +FLAG; Fig. 6E). However, c-JUN induction was not seen (Fig. 6F), indicating that TFEB functions independently of c-JUN expression. We next monitored the TFEB impact on myelin maintenance. In the absence of doxycycline, Schwann cells were cocultured with sensory neurons, and then ascorbic acid was added to initiate myelination (d0). After a sufficient number of myelin segments were formed (d21), cultures were treated with doxycycline for 4 d (Fig. 6G) and the myelin segments were visualized by immunostaining for MBP. TFEB activation (+Dox) resulted in a significant loss of MBP+ myelin segments indicating myelin degeneration (Fig. 6H). Altogether, our data show that TFEB disrupts the differentiated state and triggers myelin loss in Schwann cells.
TFEB activation triggers myelin loss. A, Representative immunostaining images for FLAG (yellow), TFEB (magenta), and DAPI (cyan) of undifferentiated FLAG-TFEBS211A Schwann cells after 48 h of doxycycline treatment. Arrowheads indicate FLAG+ cells. Scale bar, 20 µm. Graph of percentage of pixels of TFEB+ expression area of total nuclear area. Black closed circles represent mean percentage per biological replicate. Open gray circles represent individual nuclei. n = 3, F(2,6) 119.4, ordinary one-way ANOVA with Tukey's multiple-comparisons test. B, Time course of FLAG-TFEBS211A Schwann cell differentiation. Doxycycline addition indicated with arrow. Cells examined at d0, d5, d6, or d7. C, Representative immunostaining images for FLAG (yellow), PRX (magenta), and DAPI (cyan) of d5 and d7 differentiated Schwann cells. Arrowheads indicate FLAG+ cells. Scale bar, 50 µm. n = 3, F(3,8) = 15.55, ordinary one-way ANOVA with Tukey's multiple-comparisons test. D, Representative immunostaining images for FLAG (yellow), RUNX2 (magenta), and DAPI (cyan) at d5 and d7 in differentiated Schwann cells. Arrowheads indicate FLAG+ cells. Scale bar, 50 µm. n = 3, F(6,14) = 26.79, ordinary one-way ANOVA with Tukey's multiple-comparisons test. E, Representative immunostaining images for FLAG (yellow), EGR2 (KROX20; magenta), and DAPI (cyan) at d7 in differentiated Schwann cells. Merged image of EGR2 (KROX20) and DAPI. Scale bar, 20 µm. Graph of EGR2 (KROX20) nucleoplasmic expression at d5 and d7 ± Dox differentiated cultures. n = 3, F(3,8) = 96.35, ordinary one-way ANOVA with Tukey's multiple-comparisons test. F, Western blot of c-JUN at d0, d5, d6, and d7 in Schwann cell differentiation cultures. n = 3, F(5,12) = 97.71, ordinary one-way ANOVA with Tukey's multiple-comparisons test. G, Myelination time course of SC-DRG cocultures. Doxycycline addition indicated by arrow. Cells examined at d21 and d25. H, Representative immunostaining images for FLAG, MBP, and DAPI in d21 and d25 SC-DRG cocultures. Scale bar, 50 µm. n = 3, F(2,4) = 12.61, RM one-way ANOVA with Tukey's multiple-comparisons test. Data represented as mean ± SEM, ns = not significant, significant p values displayed in graphs.
TFEB activation inhibits Schwann cell differentiation and myelination
We next investigated the impact of TFEB activation during initiation of Schwann cell differentiation and myelination using the FLAG-TFEBS211A system (Fig. 7A). When FLAG-TFEBS211A was expressed prior to and during the course of db-AMP treatment, Schwann cells failed to differentiate, with only 23 ± 6.8% cells acquiring periaxin expression (d5, +Dox, FLAG+) compared with 70 ± 5.7% in control cultures (d5, −Dox; Fig. 7B). Western blot analysis also showed failed induction of EGR2 (KROX20) and myelin proteins (MPZ and PRX) in TFEB-active cultures (Fig. 7C). Interestingly, despite preventing myelin protein expression, TFEB did not restore c-JUN expression (Fig. 7C). This result indicates that TFEB activation suppresses Schwann cell differentiation without affecting c-JUN.
TFEB activation inhibits Schwann cell differentiation and myelination. A, Differentiation time course of FLAG-TFEBS211A Schwann cells. Doxycycline addition indicated by arrow. Cells examined at d0 and d5. B, Representative immunostaining images for FLAG (yellow), PRX (magenta), and DAPI (cyan) at d5 ± Dox of Schwann cell differentiation cultures. Arrowheads indicate FLAG+ cells. Scale bar, 50 µm. n = 3, F(2,6) = 19.37, ordinary one-way ANOVA with Tukey's multiple-comparisons test. C, Western blots of EGR2 (KROX20), PRX, MPZ, c-JUN, and FLAG at d0 and d5 in Schwann cell differentiation cultures. Dashed lines separate individual membranes from separate gel electrophoresis runs. n = 3, tEGR2(4) = 6.892, tPRX(4) = 12.31, tMPZ(4) = 3.29, unpaired t test for EGR2 (KROX20), PRX, and MPZ. n = 3, F(3,8) = 4.612, ordinary one-way ANOVA with Tukey's multiple-comparisons test for c-Jun. D, Time course of myelination of SC-DRG cultures. Doxycycline addition indicated with arrows. Cells examined at d10. E, Representative immunostaining images for DAPI, FLAG, and MBP of SC-DRG cocultures. Scale bar, 50 µm. n = 3, F(3,8) = 11.52, ordinary one-way ANOVA with Tukey's multiple-comparisons test. Data represented as mean ± SEM, ns = not significant, significant p values displayed in graphs.
In myelinating cocultures, TFEB activation (+Dox, d-5 → d10) in Schwann cells inhibited myelin formation, with ∼75% reduction in the number of myelin segments generated compared with control cultures (−Dox; Fig. 7D,E). A similar result was seen when FLAG-TFEBS211A expression was induced only during the myelination period (+Dox, d0 → d10; Fig. 7D,E). Limiting TFEBS211A expression to the premyelination period (+Dox, d-5 → d0) did not impact subsequent myelination (Fig. 7D,E). Altogether, these results show that ectopic TFEB activity is a negative regulator of the myelination program in Schwann cells, consistent with the elevated expression of myelin genes post-injury in the Tfeb/3 SC-dKO mice.
Schwann cell TFEB/3 deficiency impairs myelin clearance in regenerating nerves
TFEB/3 promote cellular degradation in part through the autophagic process (Palmieri et al., 2011; Settembre et al., 2011; Martina et al., 2014). Studies have shown that myelin breakdown after injury requires autophagy activation in Schwann cells (Gomez-Sanchez et al., 2015; Jang et al., 2016). To investigate the impact of TFEB/3 deficiency on myelin breakdown, we conducted morphological analysis of myelin from semithin section images of distal nerves of control and mutant mice collected at 3DPI and 7DPI. The myelin breakdown process was categorized into four stages: intact-looking myelin (P-fiber), myelin layer separation, myelin collapse (D-fiber), and multiple myelin ovoid-containing (Fig. 8A). Control and mutant mice showed similar progression in myelin breakdown, except that in the mutant, D-fiber formation (myelin collapse), was slightly attenuated at 7DPI (Fig. 8B). Myelin protein loss, assessed by Western blot analysis for MBP and MPZ, progressed normally in the mutant mice (Fig. 8C). This is in contrast to the RNA-seq data, which showed failed downregulation of myelin protein transcripts in the mutant nerves, indicating that the protein loss is triggered by a post-transcriptional mechanism. We then examined autophagic activation in distal sciatic nerves maintained ex vivo for 5 d. The ex vivo model allows assessing autophagic activation in the absence of invading macrophages. There was no significant difference in the amount of accumulated LC3II in the presence of lysosomal inhibitor (NH4Cl) in nerves from control and Tfeb/3 SC-dKO mice, indicating that the TFEB/3 loss did not impair autophagic activation (Fig. 8D). Furthermore, RNA-seq analysis indicates a lack of difference in autophagic gene expression (Fig. 8E). We concluded that Schwann cell TFEB/3 deficiency does not impact autophagy-induced myelin breakdown in distal nerves after injury.
Schwann cell TFEB/3 deficiency impairs myelin clearance in regenerating nerves. A, Representative TEM representations of categories of myelin breakdown: intact-looking myelin (P-fiber; top), myelin layer separation (2nd from the top), myelin collapse (D-fiber; 2nd from the bottom), and multiple myelin ovoid-containing cells (bottom). B, Representative toluidine blue stained semithin sections of myelin breakdown at 3DPI and 7DPI control and Tfeb/3 SC-dKO distal nerves. Percentage of each category for all myelin sheaths. n = 3, 3DPI: tP-fiber(4) = 0.7145, tlayer-sep(4) = 0.8452 tD-fiber(4) = 1.951, tmultple-ovoid(4) = 0.1324, 7DPI: tP-fiber(4) = 0.9022, tlayer-sep(4) = 2.796, tD-fiber(4) = 5.481, tmultple-ovoid(4) = 0.2129, multiple unpaired t tests. C, Western blot of MBP and MPZ in 7DPI proximal and distal nerves of control and Tfeb/3 SC-dKO. n = 3, MBP: FInteraction(1,8) = 0.2711, Finjury(1,8) = 45.06, Fgenotype(1,8) = 1.904, MPZ: FInteraction(1,8) = 0.1190, Finjury(1,8) = 61.62, Fgenotype(1,8) = 0.1122, two-way ANOVA with Šídák's multiple-comparisons test. D, Western blot of LC3I and LC3II in 5DPI explant nerves ± NH4Cl of control and Tfeb/3 SC-dKO. n = 3, FInteraction(1,8) = 11.67, Finjury(1,8) = 76.84, Fgenotype(1,8) = 1.622, two-way ANOVA with Šídák's multiple-comparisons test. E, Heat map of autophagy related genes in intact contralateral and 5DPI distal Tfeb/3 SC-dKO vs control nerves. F, Representative immunostaining images of control and Tfeb/3 SC-dKO 3DPI nerves for IBA1. Scale bar, 20 µm. Graphs of percentage of IBA1+ cells per 1,000 µm2. n = 3, t(4) = 0.3145, unpaired t test. G, Representative 21DPI TEMs of distal control and Tfeb/3 SC-dKO sciatic nerves. Scale bar, 10 µm. Uncleared myelin indicated by asterisks (Schwann cells) and number signs (macrophages). Graph of myelin debris containing cells per 10,000µm2. n = 3, t(4) = 6.769, unpaired t test. H, g-ratios of sciatic nerves of 21 d post crush injury of control (black open circles) and Tfeb/3 SC-dKO (blue open squares; left). Average g-ratio of control and Tfeb/3 SC-dKO nerves (right). n = 3, t(4) = 5.127, unpaired t test. Data represented as mean ± SEM, ns = not significant, significant p values displayed in graphs.
Repair Schwann cells secrete cytokines to recruit macrophages which, along with repair Schwann cells, participate in myelin debris clearance in the distal nerves (reviewed in Yuan et al., 2022). Immunostaining of 3DPI distal nerves showed that the abundance of IBA1+ macrophages in Tfeb/3 SC-dKO mice was comparable with that in control mice (Fig. 8F). We then conducted EM analysis to examine long-term myelin clearance during nerve regeneration after nerve crush. At 21DPI, Tfeb/3 SC-dKO nerves contained significantly more cells harboring uncleared, degenerated myelin 20 ± 0.6 and 35 ± 2.1 per 10,000 µm2 in control and Tfeb/3 SC-dKO mice, respectively (Fig. 8G). Myelin was found in large intracellular vacuoles in macrophages as well as in Schwann cells, identified by the presence of basal lamina or axon-association. Therefore, Schwann cell TFEB/3 deficiency impairs myelin debris clearance during nerve regeneration.
Schwann cell TFEB/3 deficiency impairs axon regrowth, target reinnervation, and functional recovery
To determine the impact of Schwann cell TFEB/3 deficiency on axon regrowth and target reinnervation, we examined the neuromuscular junctions (NMJs) on the soleus and extensor digitorum longus (EDL) muscles of the hind limbs at 12DPI and 34DPI following nerve crush. The NMJs were visualized by labeling of the regenerated axons and presynaptic nerve terminal with antibodies to neurofilament and synaptic vesicle glycoprotein 2A (SV2) and the postsynaptic acetylcholine receptors with fluorescent-labeled bungarotoxin (Fig. 9A). At 12DPI in control mice, most of the previously denervated postsynaptic sites were reinnervated by the regenerating axons, with only 6 ± 2.5 and 22 ± 2.9% remaining denervated on soleus and EDL NMJ sites, respectively (Fig. 9A, left graphs). Axon outgrowths that extended past the postsynaptic site, i.e., escaped fibers (Kang et al., 2003), were frequently observed, which indicated that both muscles were recently innervated and escaped fibers were not yet retracted (Fig. 9A). In comparison, in Tfeb/3 SC-dKO mice, most postsynaptic sites remained denervated, with 97 ± 2.7 and 87 ± 0.8% on soleus and EDL, respectively (Fig. 9A, left graphs). Furthermore, at this time point, no regenerating axons were observed along intramuscular nerve or on muscle surface in the Tfeb/3 SC-dKO mice indicating a marked delay in axon regrowth (Fig. 9A). At 34DPI, all the postsynaptic sites in control mice were fully reinnervated by presynaptic axon terminals that rarely exhibited escaped fibers. Notably, Tfeb/3 SC-dKO mice at 34DPI resembled control mice at 12DPI: most postsynaptic sites were reinnervated by regenerating axons that frequently exhibited escaped fibers (Fig. 9A). These escaped fibers were often associated with small extrasynaptic clusters of acetylcholine receptors, indicating prolonged denervation of postsynaptic sites (Kang et al., 2014) in the mutant mice (Fig. 9A). Additionally, a small percentage of postsynaptic sites remained denervated in the mutant mice (Fig. 9A, right graphs). A recent study has shown that angiogenesis plays a critical role in axon regeneration in peripheral nerve repair (Bhat et al., 2024). To determine whether the regeneration defect in the mutant mice was due to the global absence of Tfe3 affecting regeneration of blood vessels, distal nerves were immunostained for CD31, an endothelial cell marker. There was no significant difference in the CD31+ expression area in the mutant mice. Altogether, these results demonstrate that TFEB/3 deficiency in Schwann cells impairs axon regeneration and target reinnervation following peripheral nerve injury. Additionally, we observed significant hypomyelination of the regenerated axons of the mutant mice at 21DPI, with the g-ratio of 0.74 ± 0.01% compared with 0.67 ± 0.01% in control mice (Fig. 8H).
Schwann TFEB/3 deficiency impairs axon regrowth and target reinnervation. A, Representative images of control and Tfeb/3 SC-dKO soleus muscle NMJs at 12DPI and 34DPI for SMI312 + SV2 (green) and RBTX (red). Scale bar, 100 µm (center and left) or 20 µm (right). Arrows indicate escaped fibers. Graphs of denervated NMJ percentage at 12DPI for control and Tfeb/3 SC-dKO or 34DPI for control, Tfeb/3 SC-dKO, and TFE3KO for soleus (top) and EDL (bottom). n = 3, tsoleus(4) = 24.24, tEDL(4) = 21.78, Unpaired t test (12DPI). n = 3, soleus: Fsoleus(2,6) = 1.000, FEDL(2,6) = 4.672, ordinary one-way ANOVA with Tukey's multiple-comparisons test (34DPI). B, Representative images of control and Tfeb/3 SC-dKO distal crushed nerves at 3DPI and 5DPI for CD31 (green). Scale bar, 100 µm. Graphs of percentage of pixels of CD31+ expression area of total intrafascicular area. n = 3, t3DPI(4) = 0.5539, t5DPI(4) = 1.287, unpaired t test.
To investigate the functional consequence of TFEB/3 deficiency, we subjected mice to a beam walk test, which assesses sensorimotor coordination (Luong et al., 2011). Prior to nerve injury, control and mutant mice performed equally on the beam (Fig. 10A, 0DPI). Sciatic nerve crush abolished the normal sensorimotor coordination, shown by an increase in the number of missteps made on the beam at 3DPI (Fig. 10A,D, top graphs). By 12DPI, control mice achieved full recovery and performed at levels equivalent to preinjury (Fig. 10A,D, top left graph). In comparison, Tfeb/3 SC-dKO mice showed markedly retarded recovery, failing to achieve full recovery until 30DPI (Fig. 10A,D, top right graph). The mutant mice also exhibited defects in sensory function recovery in response to toe pinching (Fig. 10E). Mice with a single ablation of either Tfeb or Tfe3 showed functional recovery comparable with that of control mice in both beam walk and toe pinching tests (Fig. 10B–D, bottom graphs; 10F,G). Altogether, these results show that TFEB and TFE3 function in Schwann cell is required for achieving proper functional recovery of injured peripheral nerves.
Schwann TFEB/3 deficiency but not TFE3 alone impairs functional recovery. A, Graph of errors/total hind steps ratio in beam walk test for control (solid line and circles) and Tfeb/3 SC-dKO (dashed line with squares). n = 5. FInteraction(20,132) = 1.781, Finjury(10,132) = 51.46, Fgenotype(2,132) = 41.68, two-way ANOVA with Tukey's multiple-comparisons test. B, Graph of ratio of errors/total hind steps in beam walk test for control (solid line and circles) and Tfe3KO (dotted line with triangles). Note controls are the same as in Figure 10A. n = 5. FInteraction(20,132) = 1.781, Finjury(10,132) = 51.46, Fgenotype(20,132) = 41.68 (note that ANOVA analysis was the same as A), two-way ANOVA with Tukey's multiple-comparisons test. C, Graph of ratio of errors/total hind steps in beam walk test for control (solid line and circles) and Tfeb SC-KO (dotted and dashed line with diamonds). n = 5, FInteraction(10,88) = 0.8209, Finjury(10,88) = 187.5, Fgenotype(1,88) = 3.42, two-way ANOVA with Tukey's multiple-comparisons test. D, Graphs of errors/total hind steps ratio in beam walk test for control (top left), Tfeb/3 SC-dKO (top right), Tfe3KO (bottom left), and Tfeb SC-KO (bottom right) relative to uninjured (0DPI). n = 5. For control, Tfeb/3 SC-dKO, and Tfe3KO F statistic is the same as 10A and 10B, for Tfeb SC-KO F statistic is the same as C. Two-way ANOVA with Dunnett's multiple-comparisons test. E, Graphs of score for toe pinching test ipsilateral to nerve crush. Toes 5, 4, and 3 from most lateral to medial. Control indicated with solid line and circles and Tfeb/3 SC-dKO with dashed line with squares. Note controls are the same as in F. n = 5. Toe 5: FInteraction(20,132) = 3.798, Finjury(10,132) = 28.06, Fgenotype(2,132) = 47.72, Toe 4: FInteraction(20,132) = 1.801, Finjury(10,132) = 20.28, Fgenotype(2,132) = 15.91, Toe 3: FInteraction(20,132) = 1.561, Finjury(10,132) = 32.83, Fgenotype(2,132) = 11.74, two-way ANOVA with Tukey's multiple-comparisons test. F, Graph of score for toe pinching test ipsilateral to nerve crush. Toes 5, 4, and 3 from most lateral to medial. Control indicated with solid line and circles and Tfe3KO with dotted line with triangles. Note controls are the same as in C. n = 5. Toe 5: FInteraction(20,132) = 3.798, Finjury(10,132) = 28.06, Fgenotype(2,132) = 47.72, Toe 4: FInteraction(20,132) = 1.801, Finjury(10,132) = 20.28, Fgenotype(2,132) = 15.91, Toe 3: FInteraction(20,132) = 1.561, Finjury(10,132) = 32.83, Fgenotype(2,132) = 11.74 (note that ANOVA analysis was the same as E), two-way ANOVA with Tukey's multiple-comparisons test. Data represented as mean ± SEM, ns = not significant, significant p values displayed in graphs. G, Graph of score for toe pinching test ipsilateral to nerve crush. Toes 5, 4, and 3 from most lateral to medial. Control indicated with solid line and circles and Tfeb SC-KO with dotted and dashed line with diamonds. n = 5. Toe 5: FInteraction(10,88) = 1.905, Finjury(10,88) = 28.47, Fgenotype(1,88) = 6.557, Toe 4: FInteraction(10,88) = 0.5122, Finjury(10,88) = 20.08, Fgenotype(1,88) = 4.402, Toe 3: FInteraction(10,88) = 1.244, Finjury(10,88) = 21.27, Fgenotype(1,88) = 7.432, two-way ANOVA with Tukey's multiple-comparisons test. Data represented as mean ± SEM, ns = not significant, significant p values displayed in graphs.
Discussion
Our transcriptome analysis shows that TFEB/3 function is necessary for injury-induced transcriptional reprogramming associated with repair Schwann cells in peripheral nerves. Loss of TFEB/3 hampers the induction of key repair Schwann cell genes (Sox2, p75-Ngfr, Runx2) and fails to shut off myelin gene expression (Mpz, Pmp22, Egr2). Furthermore, distal Schwann cells fail to acquire EMT gene signatures and the self-renewal capacity of repair Schwann cells. This is accompanied by defects in nerve regeneration. Consequently, we propose that injury-induced TFEB/3 activity in adult Schwann cells plays a pivotal role in initiating the transition toward the repair Schwann cell phenotype.
Repair Schwann cells exhibit robust proliferation in distal nerves (Abercrombie and Johnson, 1946; Stierli et al., 2018). Tfeb/3 SC-dKO mice exhibit a failure in postinjury repair Schwann cell proliferation. Recent studies have reported that TFEB deletion impairs expression of G1/S regulators, including CDKs and replisome component MCM2 (Palmieri et al., 2011; Doronzo et al., 2019, Slade et al., 2022). Consistent with these findings, our transcriptomic analysis reveals an inhibition of genes critical in G1/S-transition and S-phase (Ccna1/2, Cdc45, Cdt1, Pcna, and Mcm2/3/5/6/10) and G2/M-transition (Foxm1, Ccnb1/2, Cdk1, Aurka/b, and Plk1). Altogether, these results lead us to conclude that TFEB/3 regulate Schwann cell proliferation after injury. We have shown previously that lack of repair Schwann cell proliferation does not hamper nerve regeneration (Yang et al., 2008), indicating that proliferation is uncoupled from other repair functions. Therefore, it is unlikely that the regeneration defects in Tfeb/3 SC-dKO mice results from proliferation failure or insufficient repair Schwann cell numbers, but rather it results from loss of repair Schwann cell features that are conducive to nerve regeneration.
Tfeb/3 SC-dKO mice exhibit a significant delay in axon regrowth, target reinnervation, and sensorimotor function recovery. These impairments are not seen in Tfeb SC-KO or Tfe3KO mice indicating both a Schwann cell autonomous function of TFEB/3 in promoting axon regeneration and a potentially compensatory role between the two transcription factors. In regenerating nerves, repair Schwann cells promote axon regrowth through the secretion of factors, such as Gdnf (Trupp et al., 1995; Hoke et al., 2002; Hoke et al., 2006; Fontana et al., 2012), which shows impaired expression in Tfeb/3 SC-dKO distal nerves. Schwann cells also directly interact with and guide regrowing axons (Son and Thompson, 1995). It is noteworthy that adhesion molecules facilitating axo-glia interaction, such as Cdh2, Nrcam, and Cadm1 (Letourneau et al., 1990; Letourneau et al., 1991; Wanner and Wood, 2002; Maurel et al., 2007), are downregulated in the Tfeb/3 SC-dKO distal nerves. Further studies will be required to examine the roles of these molecules in the delayed axon regeneration seen in Tfeb/3 SC-dKO nerves.
MITF, another MiTF-TFE family member, has been shown to regulate expression of a subset of injury-responsive genes (Daboussi et al., 2023). Interestingly, Schwann cell-specific Mitf-KO mice do not exhibit defects in inducing gene signatures associated with stemness or self-renewal as seen in Tfeb/3 SC-dKO mice. Instead, Mitf deletion results in ectopic myelination and supernumerary wrapping of small diameter axons and aberrant Remak bundle organization in regenerated axons. The mice also show a later onset but more prolonged (28–60DPI) deficiency in nerve function recovery compared with Tfeb/3 SC-dKO mice. This suggests that members of MiTF-TFE transcription factors have distinct roles during different stages of repair Schwann cell function and work cooperatively to ensure proper nerve repair.
c-JUN, an important regulator of repair Schwann cell function, continues to be expressed in Tfeb/3 SC-dKO. This suggests that c-JUN alone is not sufficient to overcome the repair Schwann cell defects in Tfeb/3 SC-dKO mice. It also implies that TFEB/3 and c-JUN may hold distinct roles in regulating repair Schwann cell function. In c-jun SC-KO mice, SOX2+ repair Schwann cells are generated normally in distal nerves after injury, and c-JUN deficiency does not significantly impact Schwann cell numbers (Arthur-Farraj et al., 2012). However, in the absence of c-JUN, there is extensive sensory neuron death after nerve injury and consequently, a long-term deficit in axon regeneration. The gene signatures associated with DEGs in Tfeb/3 SC-dKO include stem cell, cell cycle, EMT, and polycomb-related modules, while c-jun SC-KO categories include neurogenesis, neuronal growth, and regeneration. These observations suggest that TFEB/3 play a distinct role by promoting stem cell-like features of repair Schwann cells, whereas c-JUN is important for the repair Schwann cell function that supports neuronal survival. It also raises a possibility that the trophic support function of repair Schwann cells may be independent of the stem cell-like property of the Schwann cells.
In our previous work, we used the histone mark H3K27ac to identify enhancers that become active in the distal nerve after injury (Hung et al., 2015). TFEB and TFE3 binding motifs align with a large proportion of these enhancers. Binding motifs for other transcription factors, such as c-JUN, RUNX, and ETS factors, also align with injury-responsive enhancers suggesting that multiple transcription factors modulate distal gene expression following nerve injury (Hung et al., 2015). Single ablation of TFEB or TFE3 lacked differences in repair Schwann cell generation or function (Figs. 4, 9, 10), suggesting redundant and cumulative roles of transcription factors activating the repair program. Our analysis shows that while TFEB/3 and c-JUN regulate distinct sets of genes, they also share several common target genes (Extended Data Table 5-1). It is possible that these genes serve as potential convergence points for TFEB/3 and c-JUN to regulate shared functions, such as promoting axonal guidance and regrowth (Gdnf, Nrcam), which are shown to be defective in both Tfeb/3 SC-dKO and c-jun SC-KO mice.
TFEB/3 are known for their roles in regulating lysosomal and autophagosomal genes (Sardiello et al., 2009; Palmieri et al., 2011; Settembre et al., 2011). Previous studies have demonstrated that autophagy inhibition causes the maintenance of intact-looking myelin sheaths and sustained myelin protein levels in injured nerves (Gomez-Sanchez et al., 2015; Jang et al., 2016). Surprisingly, in Tfeb/3 SC-dKO mice, morphological myelin breakdown at 3DPI and myelin protein loss through 7DPI proceeds normally. Additionally, we do not observe significant downregulation of autophagy-related genes, including those previously shown to be regulated by TFEB (Settembre et al., 2011). Therefore, we speculate that initial myelin breakdown does not require TFEB/3 function in Schwann cells. It may involve an alternative autophagy-activating or an autophagy-independent mechanism of myelin clearance as proposed in other studies (Arthur-Farraj et al., 2012; Gomez-Sanchez et al., 2015; Ying et al., 2018).
Despite normal early myelin breakdown, myelin debris shows an increased persistence in both Schwann cells and macrophages of regenerating nerves at 21DPI in Tfeb/3 SC-dKO mice. Importantly, the abundance of macrophages to the distal nerve appears normal, ruling out an insufficient number of macrophages as the cause of the myelin clearance defect. A similar impairment in myelin clearance has been reported in c-jun SC-KO mice following nerve injury (Arthur-Farraj et al., 2012). This raises the possibility that repair Schwann cells may play a role in activating macrophages for myelin clearance. Whether TFEB/3 or c-JUN is involved in macrophage activation is an intriguing question that warrants further investigation. Alternatively, in Tfeb/3 SC-dKO mice, which carry a global TFE3 deficiency, the TFE3 loss in macrophages may be sufficient to suppress the myelin clearance function (Pastore et al., 2016), although the Tfe3KO did not show deficits in behavioral recovery after injury. Overall, our study reveals a mechanistic distinction between early myelin breakdown and a later phase of myelin debris clearance.
It remains unclear how nerve injury activates TFEB/3 in distal Schwann cells. A recent study showed that calcineurin loss inhibits early TFEB nuclear localization in distal Schwann cells after nerve injury (Reed et al., 2020). Calcineurin is a Ca2+-dependent phosphatase that acts on both TFEB and TFE3 to activate them (Medina et al., 2015; Martina et al., 2016). Therefore, TFEB/3 activation in distal Schwann cells may involve a mechanism dependent on intracellular Ca2+ increase (Tricaud et al., 2022).
It is interesting to note that TFEB/3 loss in Schwann cells does not impact myelination or gene expression in intact peripheral nerve. In comparison, single ablation of TFEB in oligodendrocytes causes ectopic myelination (Meireles et al., 2018; Sun et al., 2018). These studies demonstrate the cell type-specific function of TFEB in regulating myelination in the nervous system.
Understanding the function of TFEB/3 in Schwann cells could offer insights into the pathogenic mechanisms associated with Charcot–Marie–Tooth disease type 1A (CMT1A). CMT1A arises from mutations in Pmp22, leading to aberrant protein accumulation in the ER (Tobler et al., 1999). Elevated TFEB activity is observed in CMT1A Schwann cells (Chittoor et al., 2013). While this heightened TFEB activity might be a consequence of increased cellular catabolism to targeting PMP22 for degradation (Notterpek et al., 1997), our in vitro findings suggest that increased TFEB activity could negatively impact myelin retention and myelin repair in diseased Schwann cells.
In conclusion, our study highlights the pivotal role of TFEB/3 in repair Schwann cell generation. These findings shed light on the intricate regulatory networks at play in nerve repair, ultimately contributing to our understanding of Schwann cell function and its potential implications for promoting nerve regeneration.
Footnotes
This work was supported by National Institutes of Health grants NS105796 to Y.J.S., NS130566 and by a core grant to the Waisman Center from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (P50 HD105353) to J.S., AG062475 to R.D. and NS118020 to H.A.K.
The authors declare no competing financial interests.
- Correspondence should be addressed to Haesun A. Kim at haekim{at}newark.rutgers.edu.