We recently reported that deletion of Phb1 specifically in SCs causes a severe and fast progressing demyelinating phenotype in mice (Della-Flora Nunes et al., 2021). Mice lacking Phb1 in SCs (Phb1-SCKO) show the first signs of myelin loss at postnatal day 20 (P20), with demyelination peaking around P40-P60 and causing partial or complete hindlimb paralysis around P90 (Della-Flora Nunes et al., 2021). Concomitant with the demyelination, Phb1-SCKO mice also display axonal degeneration, which follows a similar temporal progression (Della-Flora Nunes et al., 2021). To extend this data, we report here similar findings from density analysis in electron micrographs from the same animals (Figure 1A and B). This pathology is likely caused by the profuse changes in mitochondrial morphology and function that follow deletion of Phb1 in SCs (Della-Flora Nunes et al., 2021). We also demonstrated that SC apoptotic cell death is not greatly altered by deletion of Phb1 in vivo, and that SC numbers remain normal in nerves of Phb1-SCKO, suggesting that a simple loss of viability cannot explain their pathology (Della-Flora Nunes et al., 2021). Nevertheless, the molecular mechanism linking mitochondrial damage to myelin destruction in Phb1-SCKO mice remains unknown.

Although little is known about how myelin is maintained long-term, it seems that preservation of myelin with the correct structure and thickness is an active process, requiring the continuous activation of certain cellular machinery in SCs (Bremer et al., 2010; Bremer et al., 2011; Decker et al., 2006). We therefore hypothesized that the SC response to mitochondrial dysfunction may inadvertently interfere with pathways critical for myelin maintenance. Seeking to find the molecular link between compromised mitochondrial function and demyelination, we investigated the status of the following molecular pathways previously reported to play roles in preservation of myelin: mTORC1 (important for myelination Beirowski et al., 2017; Figlia et al., 2017) and remyelination (Norrmén et al., 2018), ERK 1/2 (whose activation is sufficient to trigger demyelination Napoli et al., 2012), AKT (important to regulate myelin sheath thickness Domènech-Estévez et al., 2016), c-Jun (the master transcription factor of nerve repair Arthur-Farraj et al., 2012), and eIf2α (the core protein in the integrated stress response, ISR, which is important for myelin maintenance in the context of perturbed protein homeostasis in SCs Scapin et al., 2020; D’Antonio et al., 2013).

We had previously reported that phosphorylation of eIf2α is actually protective in Phb1-SCKO mice. In addition, Phb1-SCKO mice only show minor changes in the ERK 1/2 pathway, which are probably insufficient to initiate demyelination (Della-Flora Nunes et al., 2021). The same is true for the AKT pathway (Figure 1—figure supplement 1). On the other hand, Phb1-SCKO animals show continuous upregulation of c-Jun and activation of the mTORC1 pathway, as measured by the protein and phosphorylation levels of the downstream targets 40S ribosomal protein S6 (S6) and 4E-BP1 (Figure 1C and D and Figure 1—figure supplement 2). Importantly, these changes are evident even before overt demyelination. Nonetheless, we found no differences in the phosphorylation levels of mTOR at Ser2448 (Figure 1—figure supplement 1), suggesting that the AKT/PI3K pathway is not involved in activating the mTORC1 pathway in Phb1-SCKO mice. Interestingly, the changes we detected involve upregulation of both total and phosphorylated levels of 4E-BP1 and S6, while the ratio of phosphorylated to total levels of these proteins remained mostly unaltered (Figure 1—figure supplement 2 and Figure 1—figure supplement 3). Although we cannot completely rule out the contribution of other nerve cells to the levels of c-Jun, 4E-BP1 and S6, infiltrating macrophages do not seem to activate c-Jun and mTORC1 in sciatic nerves of Phb1-SCKO mice (Figure 1—figure supplement 4), suggesting that these alterations may primarily occur in SCs.

Both c-Jun and mTORC1 have been previously implicated in the cellular response to mitochondrial dysfunction. The mitochondrial unfolded protein response (mtUPR) is believed to be partly regulated by binding of an AP-1 transcription factor (postulated to be c-Jun) to the promoter of CHOP and C/EBPβ (Horibe and Hoogenraad, 2007). On the other hand, mTORC1 was shown to be upstream of the integrated stress response (ISR) in muscle, regulating the progression of a mitochondrial myopathy (Khan et al., 2017).

For this reason, we sought to investigate if c-Jun and mTORC1 were activated downstream of perturbations to SC mitochondria. We treated primary rat SCs with compounds that affect different aspects of mitochondrial function: Carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP), an ionophore that dissipates the mitochondrial membrane potential; Oligomycin, an inhibitor of ATP synthase; or Antimycin A, an inhibitor of mitochondrial complex III. We then evaluated the expression levels of c-Jun and downstream targets of the mTORC1 pathway (4E-BP1 and S6), as well as components of the other pathways we previously demonstrated to be altered in Phb1-SCKO mice (Della-Flora Nunes et al., 2021): eIF2α (which is phosphorylated downstream of different cellular stressors to activate the ISR), BiP (a chaperone also known as HSPA that is upregulated in response to stress in the endoplasmic reticulum, ER), and Opa1 (a protein essential for mitochondrial fusion that, under situations of mitochondrial damage, is proteolytically cleaved, inhibiting mitochondrial fusion). Short-term treatment (24 h) with these compounds resulted only in minor changes in these pathways (Figure 2—figure supplement 1A and B). Cells treated with FCCP for 24 hr presented with elevated levels of p-eIF2α and reduced ratio between the long and short isoforms of Opa1 (Figure 2—figure supplement 1A and B). This is in line with previous reports showing that the ISR can quickly be activated directly downstream of compromised mitochondria (Viader et al., 2013; Mick et al., 2020). Oligomycin and Antimycin A applied to SCs for 24 hr were unable to elicit significant changes in the conditions tested. On the other hand, a 7-day treatment of SCs with Oligomycin or Antimycin A triggered robust stimulation of the ISR (Figure 2A and B and Figure 2—figure supplement 2). Furthermore, this lengthier treatment regimen also induced a potent activation of the mTORC1 pathway (Figure 2A and B and Figure 2—figure supplement 2A). Hence, it is possible that the mTORC1 pathway participates in the SC adaptation to long-term mitochondrial impairments.

Next, we interrogated whether c-Jun or mTORC1 activity were associated with presence of compromised mitochondrial network in vivo. For this experiment, we analyzed Phb1-SCKO animals in which SC mitochondria were genetically labeled with the PhAM reporter. The PhAM mouse line contains a floxed STOP construct coding for a mitochondrially targeted version of the Dendra2 fluorophore (Pham et al., 2012). We previously reported that, at P40, about 20% of SCs in Phb1-SCKO mice show disrupted mitochondrial network, especially away from the cell body, where PhAM is almost completely undetectable (Della-Flora Nunes et al., 2021). Staining of individual myelinated fibers with c-Jun indicated that c-Jun immunoreactivity was significantly associated with the presence of compromised mitochondria (Figure 3A and B). Moreover, 78.81% of all SCs with damaged mitochondria showed strong c-Jun nuclear expression, while only 23.15% of SCs with reasonably well-preserved mitochondria stained positive for c-Jun (Figure 3C). In contrast, we did not identify any association between mitochondrial loss and expression of p-S6 (Figure 3D–F). Similar results were also found in nerves of 90-day-old Phb1-SCKO mice (Figure 3—figure supplement 1).

Thus, our data supports the hypothesis that both c-Jun and mTORC1 can be activated downstream of perturbations to mitochondria under certain circumstances, raising the possibility that c-Jun and mTORC1 may participate in the response to mitochondrial stress in SCs. The discrepancy in our in vitro and in vivo data may be because of the already elevated c-Jun expression in SCs in vitro (SCs in culture show an immature phenotype Schmid et al., 2014; Morgan et al., 1991). In addition, mitochondrial loss is a late event in vivo, while mTORC1 activation occurs relatively quickly, after only 1 week in vitro. Thus, mTORC1 and c-Jun may be activated in different cells or at different stages of the SC response to mitochondrial damage.

Given the possibility that c-Jun and mTORC1 could be activated downstream of mitochondrial dysfunction, and the importance of these pathways for formation of repair SCs, we asked if demyelination in Phb1-SCKO mice was preferentially happening in SCs with overactive c-Jun or mTORC1. With this goal, we immunostained individual myelinated fibers of Phb1-SCKO animals for myelin proteins (P0 and MBP) in conjunction with c-Jun or p-S6. We then identified SCs undergoing demyelination by the presence of myelin fragments inside SCs (myelin ovoids). In our analysis, all the evaluated fibers containing myelin ovoids showed intense nuclear c-Jun staining (Figure 4A–C), suggesting a strong association between c-Jun expression and demyelination. Similarly, we also identified a correlation between p-S6 and the presence of myelin ovoids (Figure 4D and E), and a greater proportion of the SCs containing myelin ovoids were also positive for p-S6, although this difference did not reach statistical significance (Figure 4F).

We then tested if Phb1-SCKO mice showed activation of other pathways known to participate in breakdown and degradation of myelin. First, we probed Phb1-SCKO animals for genes modulated by c-Jun after nerve injury. After nerve damage, c-Jun leads to upregulation of genes involved in regeneration and trophic support, such as glial-derived neurotrophic factor (Gdnf) and Sonic hedgehog (Shh), and to increased expression of the transcription factor (Olig1), all of which are specific to the repair SC and are not expressed during SC development (Arthur-Farraj et al., 2012). On the other hand, as myelin gets degraded, and due to the antagonistic actions of c-Jun on EGR2 (Parkinson et al., 2008), there is downregulation of several molecules. One is Cdh1, coding for Cadherin-1, also called E-cadherin (Arthur-Farraj et al., 2012), an adhesion molecule present in adherens junctions that stabilizes the non-compact regions of myelin (Young et al., 2002; Tricaud et al., 2005). Also downregulated are the myelin genes Myelin basic protein (Mbp) and Myelin protein zero (Mpz). We found that the mRNA expression of the majority of the genes evaluated is altered in Phb1-SCKO mice in the expected direction: upregulation of Gdnf and Shh; downregulation of Cdh1, Mbp and Mpz (Figure 4—figure supplement 1). We did not, however, identify any change in expression of Olig1. Among these genes, Gdnf is a direct c-Jun target (Fontana et al., 2012), Shh has a c-Jun binding site on its enhancer (Hung et al., 2015), and the Olig1 enhancer has a binding site for Runx2 (a transcription factor proposed to mediate activation of some injury-responsive genes downstream of c-Jun Hung et al., 2015). Finally, Mpz and Mbp are directly regulated by EGR2 (LeBlanc et al., 2006; Denarier et al., 2005), which is known to have a cross-antagonistic relationship with c-Jun (Parkinson et al., 2008). Although this provides further evidence for involvement of c-Jun in the nerve pathology of Phb1-SCKO mice, we cannot rule out that other pathways are also regulating the expression of the evaluated genes.

Phb1-SCKO mice also showed overexpression of Mixed lineage kinase domain-like (MLKL), a protein recently implicated in dismantling the myelin sheath after nerve injury (Ying et al., 2018; Figure 4—figure supplement 1B and D). Moreover, deletion of Phb1 in SCs also resulted in upregulation of the autophagic machinery, which is important for myelinophagy and is regulated by c-Jun after nerve injury (Gomez-Sanchez et al., 2015; Figure 4—figure supplement 1C and D).

Taken together, these results indicate that nerves of Phb1-SCKO mice show molecular similarities to nerves undergoing myelin breakdown after nerve injury, including upregulation of myelin degradative pathways such as MLKL and autophagy, increased expression of the Gdnf and Shh trophic factors, and reduced levels of the Mbp and Mpz myelin genes. After nerve injury, many of these effects depend on the activation of the mTORC1/c-Jun axis. Moreover, we found a strong association between demyelination and activation of both mTORC1 and c-Jun in Phb1-SCKO mice. Given these results, it is tempting to hypothesize that c-Jun and mTORC1 could be key pathways orchestrating the demyelination process in Phb1-SCKO mice.

We recently showed that the ISR is likely a beneficial response in Phb1-SCKO mice (Della-Flora Nunes et al., 2021). Activation of the ISR frequently leads to alterations in the mTORC1 pathway (Ryoo and Vasudevan, 2017; Zhang et al., 2019). Thus, we asked whether the ISR is upstream of the pathways analyzed in this study. Inhibition of the ISR using a daily injection of 2.5 mg/kg ISRIB as previously reported (Della-Flora Nunes et al., 2021) did not significantly alter p-S6, c-Jun or Opa1 expression in Phb1-SCKO mice (Figure 5). Nonetheless, this treatment was sufficient to result in a small reduction in levels of p-4E-BP1 and 4E-BP1 (Figure 5B and C and Figure 5—figure supplement 1A).

In order to test the hypothesis that c-Jun coordinates demyelination in Phb1-SCKO mice, we crossed those animals to Jun floxed mice. At P40, presence of one (Phb1-SCKO; JUN Het) or two Jun floxed alleles (Phb1; JUN SCKO) resulted in a significant and dose-dependent reduction in c-Jun protein levels in sciatic nerves (Figure 6A). As expected, this reduction in c-Jun also led to a dose-dependent reduction in the number of demyelinated fibers and myelinophagy events (Figure 6B and C). Nevertheless, deletion of Jun in Phb1-SCKO mice also caused a dose-dependent reduction in the number of myelinated axons in tibial nerves (Figure 6B and C). In addition, nerves of Phb1; JUN SCKO contained bundles with large non-myelinated axons and a visible division of the nerve in smaller fascicles (hyper-fasciculation) (Figure 6B and D). These phenotypes suggest that ablation of Jun in SCs already lacking Phb1 may exacerbate the mild developmental defects that we previously observed in Phb1-SCKO mice (Della-Flora Nunes et al., 2021). As a consequence, Jun deletion is not sufficient to rescue the motor deficits of Phb1-SCKO mice in the rotarod (Figure 6E).

To investigate if c-Jun was important to modulate other pathways of interest, we probed the protein levels of eIF2α, p-eIF2α, 4E-BP1, p-4E-BP1, S6, p-S6, Opa1, and BiP. Deletion of Jun in Phb1-SCKO mice had a dose-dependent effect on levels of p-4E-BP1 and also reduced p-S6 when both Jun alleles were deleted, suggesting that c-Jun may partially modulate mTORC1 (Figure 6—figure supplement 1A and B). A similar effect was observed on BiP levels (Figure 6—figure supplement 1A and B). c-Jun, however, does not seem to be important to regulate the ISR, since p-eIF2α levels are unaltered in Phb1-SCKO mice upon Jun deletion (Figure 6—figure supplement 1A and B). In agreement with this finding, ablation of Jun is also unable to alter the mRNA levels of ATF4 targets (Asns, Chac1, Pck2, and Ddit3, also known as Chop), which are upregulated downstream of p-eIF2α during ISR in Phb1-SCKO mice (Figure 6—figure supplement 1C).

Mice containing two floxed Jun alleles and one floxed Phb1 allele (JUN SCKO; Phb1 Het) are statistically indistinguishable from control mice in all the aforementioned analyses (Figure 6—figure supplement 2), indicating that deletion of Jun alone is unable to elicit any of the changes observed in Phb1-SCKO; JUN Het or Phb1; JUN SCKO mice.

Next, we sought to explore the role of mTORC1 in the pathology observed in Phb1-SCKO mice. Considering the important developmental role of mTORC1 (Beirowski et al., 2017; Figlia et al., 2017; Sherman et al., 2012), it would be necessary to inhibit mTORC1 after myelination is completed. Therefore, we opted for a pharmacological treatment instead of a genetic approach. We administered Phb1-SCKO mice and controls with daily injections of the well-established mTORC1 inhibitor rapamycin from P20 to P40 (Figure 7A). Rapamycin binds to FK506-binding protein (FKBP12), which becomes an allosteric inhibitor of mTORC1 (Li et al., 2014; Figure 7A). Rapamycin treatment was efficient, and significantly reduced the levels of p-4E-BP1 and p-S6 in sciatic nerves of P40 Phb1-SCKO mice (Figure 7B and C and Figure 7—figure supplement 1A). Interestingly, suppression of mTORC1 activity resulted in almost complete rescue of nerve morphology in Phb1-SCKO animals at P40. Mutant mice treated with rapamycin had the same number of myelinated axons, demyelinated axons, degenerating axons and myelinophagy as their littermate controls (Figure 7D and E). We also confirmed the effectiveness of rapamycin to reduce demyelination by assessing electron micrographs (Figure 7—figure supplement 2A and B). Interestingly, we found a few aberrantly myelinated Remak bundles in Phb1-SCKO mice treated with rapamycin, a finding that was not present in control animals treated with the same drug (Figure 7—figure supplement 2A). Deletion of Phb1 in Schwann cells did not alter myelin thickness on the myelinated axons that remained, and this was unaffected by rapamycin treatment (Figure 7F). Consistent with the results of our morphological analyses, rapamycin was also able to partially ameliorate the nerve conduction velocity in Phb1-SCKO mice (Figure 7G). Nonetheless, rapamycin could not improve the amplitude of the compound muscle action potential (CMAP) of mutant mice (Figure 7G) nor the motor deficits of Phb1-SCKO mice measured by the rotarod (Figure 7H). In summary, these results suggest that mTORC1 overactivation is key to induce demyelination in Phb1-SCKO mice, and that inhibition of the mTORC1 pathway can provide meaningful benefit to morphological parameters and nerve conduction velocity. It is, however, worth noting that, since we opted for a systemic treatment, effects of rapamycin in cells other than SCs could also be contributing to the observed outcome.

Similar to the studies with c-Jun, we also evaluated if other pathways of interest could be modulated by mTORC1. Treatment with rapamycin results in a trend toward reduction of c-Jun expression in Phb1-SCKO mice (Figure 7—figure supplement 3A and B). It also significantly reduces the protein levels of BiP and p-eIF2α (Figure 7—figure supplement 3A and B), and the mRNA level of Ddit3 (Figure 7—figure supplement 3C). Therefore, mTORC1 may be a central pathway modulating c-Jun and the ISR in Phb1-SCKO mice.

All animal procedures have been approved by the Institutional Animal Care and Use Committee (IACUC) of the Roswell Park Cancer Institute (Buffalo-NY, USA), and followed the guidelines stablished by the NIH’s Guide for the Care and Use of Laboratory Animals and the regulations in place at the University at Buffalo (Buffalo-NY, USA). Animals were housed separated by gender in groups of at most five per cage and kept in a 12 hr light/dark cycle with water and food ad libitum. Mpz-Cre and Phb1 floxed animals were previously described (Feltri et al., 1999; He et al., 2011). Mice were also crossed to the PhAM reporter line (Jackson Laboratories Stock No: 018385) (Pham et al., 2012) and to Jun floxed mice (Behrens et al., 2002). Animals were kept in a C57BL/6 and 129 mixed genetic background and analyses were performed from littermates. Animals carrying one or two floxed Phb1 alleles but no Cre were used as controls, other than for the experiments with PhaM mice, where Control mice were Phb1^wt/wt; Mpz-Cre; PhAM, while Phb1-SCKO mice were Phb1^fl/fl; Mpz-Cre; PhAM. No animals were excluded from this study. Genotyping was performed from genomic DNA as previously described for Mpz-Cre (Feltri et al., 1999), Phb1 floxed animals (He et al., 2011), PhaM (Della-Flora Nunes et al., 2021), and Jun (Parkinson et al., 2008). Rapamycin (LC Laboratories Cat# R-5000) was prepared as described previously (Beirowski et al., 2017) and administered intraperitoneally at 10 mg/kg daily from P20 to P40. ISRIB (Cayman chemicals Cat# 16258) was prepared as described previously (Chou et al., 2017) and administered intraperitoneally at 2.5 mg/kg daily from P20 to P40. For the above treatments, animals were randomly allocated to Vehicle or treatment group.

Morphological analyses were performed as described previously (Della-Flora Nunes et al., 2021). Quantification of morphological parameters in electron micrographs in Figure 1 used data of ten randomly selected fields at ×2900 magnification, which resulted in the evaluation of ~95–150 axons per sample. Analysis reported on Figure 7—figure supplement 2 included about 20–30 randomly selected fields of view at ×2900 magnification, which resulted in the evaluation of ~250–450 axons per sample. Data were quantified using the cell counter plugin of ImageJ Fiji v1.52p (Rueden et al., 2017; Schindelin et al., 2012). Axons were considered to be degenerating when one of the following conditions was present: 1. visible transport defects that originated accumulation of organelles and vesicles inside the axon; 2. Axonal shrinkage resulting in blackened axon appearance under EM; and 3. Axonal swelling causing appearance of disperse distribution of microtubules and neurofilaments.

These experiments were performed as reported before (Della-Flora Nunes et al., 2021).

Primary rat SCs were prepared as described previously (Brockes et al., 1979) and were not passaged more than four times. Cells were maintained in media containing high glucose DMEM (4.5 g/L glucose) supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, 2 ng/ml Nrg1 (human NRG1-β1 extracellular domain, R&D Systems 377-HB), and 2 μM forskolin. For the induction of mitochondrial stress, we prepared stock solutions of 10 mM FCCP (Sigma C2920) in ethanol, 40 mM Antimycin A (Sigma A8674) in ethanol and 5 mM Oligomycin (Millipore 495455) in DMSO and stored at –20°C until use. For the experiment, 215,000 primary rat SCs were plated in each well of a 12-well dish. Two days later, media was exchanged to SC media containing 5 μM FCCP, 2.5 μM oligomycin, 10 μM antimycin A or an equivalent dose of vehicle. For the seven-day treatment, SC media was exchanged every other day with media containing a fresh dilution of the drugs. At the end of the experiment, protein was extracted and western blot was carried out as described below.

Western blots were performed as described previously (Della-Flora Nunes et al., 2021). Briefly, sciatic nerves were collected, stripped from epineurium, snap frozen in liquid nitrogen and stored at –80°C until analysis. Nerves were pulverized and lysis was performed in buffer containing 50  mM Tris pH 7.4, 150  mM NaCl, 1 % IGEPAL CA-630, 1  mM EDTA, 1  mM EGTA, 0.1 % SDS, 0.5 % sodium deoxycholate, 1  mM sodium orthovanadate, 1  mM sodium fluoride, protease inhibitor cocktail (Sigma-Aldrich P8340), phosphatase inhibitor cocktail 2 (Sigma-Aldrich P5726) and phosphatase inhibitor cocktail 3 (Sigma-Aldrich P0044). This solution was then sonicated for 3 cycles of 20 s at 70% power and centrifuged at 13,200 ×  g for 15  min at 4 °C. Protein in the supernatant was quantified using a BCA protein assay kit and equal amounts of protein per sample were used in the SDS-PAGE. Protein was then transferred to activated PVDF membranes. After blocking with 5% BSA in TBS solution with 0.5% Tween (TBS-T), membranes were incubated overnight with antibodies of interest. Membranes were then rinsed in TBS-T and incubated for 1  hr with secondary antibodies. Blots were either imaged directly with Odyssey CLx infrared imaging system (Li-Cor) or developed using ECL Select (GE Healthcare) and imaged using a ChemiDoc XRS system. Quantifications were carried out in the Image lab 6.0 software (Biorad) for blots imaged with the ChemiDoc XRS or in the Image Studio Lite 5.2 (Odyssey) for blots imaged with the Odyssey CLx. The following primary antibodies were used: Opa1 1:500 (BD Biosciences Cat# 612606), β-tubulin 1:5000 (Novus Biologicals Cat# NB600-936), GAPDH 1:5000 (Sigma Cat# G9545), eIF2α 1:500 (Cell signaling Cat# 5324), p-eIF2α 1:500 (Cell signaling Cat# 3398), Bip 1:500 (Novus Biologicals Cat# NB300-520), p-4E-BP1 1:500 (Cell signaling Cat# 2855), 4E-BP1 1:500 (Cell signaling Cat# 9644), p-S6 1:500 (Cell signaling Cat# 4858), S6 1:500 (Cell signaling Cat# 2217), c-Jun 1:500 (Cell signaling Cat# 9165), p-AKT 1:500 (Cell signaling Cat# 9271); AKT 1:500 (Cell signaling Cat# 9272); MLKL 1:500 (Abgent Cat# AP14272b); LC3 1:500 (Cell signaling Cat# 12741); Atg7 1:500 (Cell signaling Cat# 8558); p62 1:500 (Enzo Life Sciences Cat# BML-PW9860); mTOR 1:1000 (Cell signaling Cat# 2983), p-mTOR 1:1000 (Cell signaling Cat# 5536). GAPDH or β-tubulin were used as loading controls to normalize protein expression. Raw files (on Image lab or Image Studio format), as well as annotated uncropped blots are presented as Source data files associated to each figure.

Conventional immunofluorescence experiments were performed as previously described (Della-Flora Nunes et al., 2021) using 10-μm-thick longitudinal sections of sciatic nerve obtained using a cryostat. For co-staining of F4/80 and c-Jun at least three fields per animal were imaged at ×40 magnification using a confocal microscope Leica SP5II running the LAS AF 2.7.9723.3 software (Leica). For co-staining of F4/80 and p-S6 at least three fields per animal were imaged at ×40 magnification using a Zeiss ApoTome microscope (Zeiss Observer.Z1 AX10) running the AxioVision 4.8.2.0 software (Zeiss). For teasing, sciatic or tibial nerves were dissected, fixed in 4% PFA for 30 min, washed with PBS and teased in slides coated with (3-Aminopropyl)triethoxysilane (TESPA; Sigma). Coating with TESPA was achieved by subsequently submerging glass slides in acetone for 1 min, 4% TESPA in acetone for 2 min and two times in acetone for 30 s each. The teasing procedure consisted of placing a small portion of the nerve in a PBS droplet over the TESPA-coated slide, followed by careful mechanical separation of individual fibers, first using insulin syringes (0.3 ml 31 G x 8 mm) and then using modified insulin syringes containing insect pins (Fine science tools Cat# 26002–10) attached to their needle. The immunofluorescence procedure consisted in the permeabilization of the teased fibers for 2 min using acetone, blocking for 1  hr at room temperature, incubation with primary antibodies overnight at 4 °C, incubation with secondary antibodies for 1  hr at room temperature, counterstaining with DAPI, and mounting of slides with Vectashield (Vector Laboratories). Blocking buffer consisted of 5% fish skin gelatin and 0.1% Triton X-100 in 1× PBS. The following primary antibodies were used: rabbit anti-c-Jun 1:200 (Cell signaling Cat# 9165), rabbit anti-p-S6 1:200 (Cell signaling Cat# 4858), chicken anti-P0 1:300 (Aves Cat# PZO), mouse anti-MBP 1:300 (Millipore Cat# MAB384) and rat anti-F4/80 1: (Biorad Cat# MCA497GA). For staining with MBP, blocking of endogenous immunoglobulins was achieved by incubation with a 1:10 dilution of unconjugated Fab Goat anti-mouse IgG (Jackson ImmunoResearch Laboratories Cat# 115-007-003) during the blocking step. Images from teased fibers were acquired at ×40 magnification and 1.5 X zoom using a confocal microscope Leica SP5II running the LAS AF 2.7.9723.3 software (Leica). Quantifications were performed using ImageJ Fiji v1.52p (Rueden et al., 2017; Schindelin et al., 2012). Four to five fields per animal were analyzed. Classification of the mitochondrial network morphology was made blind to the experimental condition and followed a qualitative assessment: the mitochondrial network was labeled as ‘damaged’ when there was a clear reduction in mitochondrial density along the SC length and, specially, in regions distant from the cell body and closer to the nodes of Ranvier; in all other cases, the mitochondrial network was classified as intact.

RNA was isolated and reverse-transcribed as published (Poitelon et al., 2016). qRT-PCR was performed as reported previously (Della-Flora Nunes et al., 2021) and used the Universal Probe Library (UPL, Roche) for the genes involved in the ISR and SYBR green for the genes activated by c-Jun. Primers used in this study are reported in the key resources table.

Experiments were not randomized, but data collection and analyses were performed blind to the conditions of the experiments and genotype of the mice. However, due to the severity of the phenotype, in some analyses it was not possible to completely prevent investigators from identifying if the animal was WT or mutant. No data were excluded from the analyses. For treatments with cells or animals, allocation to groups was made randomly. No power analysis was performed, but our sample sizes are similar to those generally used in the field. The statistical test used in each analysis is reported in the legend of each figure. Data are presented as mean ± s.e.m. p-values < 0.05 were considered to represent a significant difference, while 0.05< p < 0.1 was considered to represent a trend. Data were analyzed using GraphPad Prism 6.01. Raw data and output of statistical analyses are available as source data files associated to each figure.