Experimental animals

All experiments were performed using Wfs1∆exon8 mice (Wfs1 KO mice) and wild type littermates (129S6/SvEvTac and C57BL/6 mixed background). This mouse model was originally created by Luuk et al. and has a disrupted exon 8 in the Wfs1 gene [44]. Mice were bred under standard laboratory conditions and had free access to food and water. Mice of both sexes were used, at the age of 3 months (± 1 week), 4.5 months (± 1 week), 6 months (± 2 weeks) or 7.5 months (± 2 weeks). Occasionally, animals had an opaque cornea. These were excluded from the study.

Optical coherence tomography

Upon general anesthesia (75 mg/kg ketamine and 1 mg/kg medetomidine) and pupil dilatation (0.5% tropicamide), optical coherence tomography scans of the retina (1000 A-scans, 100 B-scans, 1.4 × 1.4 mm) (Envisu R2200, Bioptigen) were acquired. Anesthesia was reversed with atipamezole (1 mg/kg). Retinal layer thickness was measured using InVivoVue Diver 3.0.8 software (Bioptigen), at 16 locations equally spaced around the optic nerve head and averaged per mouse, all as described in [78].

Optomotor response test

Visual performance was tested using the virtual reality optomotor test (Optomotry, Cerebral Mechanics), as previously described [72, 78]. Briefly, the unrestrained mouse was placed on a platform in the testing arena and vertical black-white sine-wave gratings were projected on the screens. A video camera was placed on top of the arena to provide real-time video data. Spatial frequency thresholds were measured for each eye under 100% contrast, via a simple staircase procedure. The highest spatial frequency that the mouse could track was identified as the visual acuity. For contrast sensitivity, a similar approach was used, but contrast was systematically reduced until the contrast threshold was identified. Contrast threshold was identified at three spatial frequencies (0.103, 0.192, 0.272 cyc/deg), and calculated as described in [61].

Electroretinograms and visual evoked potentials

Following overnight dark adaptation, electroretinograms (Celeris, Diagnosys) were recorded for anesthetized mice (75 mg/kg ketamine and 1 mg/kg medetomidine) as described in [11]. Lens electrodes with integrated stimulators were placed on the cornea after pupil dilation (0.5% tropicamide and 15% phenylephrine), and a needle scalp electrode was inserted above the midline near the visual cortex. The ground and reference electrode were placed in the tail base and cheek, respectively. Eyes were alternately stimulated, and full field visual evoked potential responses were recorded at a single flash intensity of 0.05 cd*s/m2, averaging 300 brief flashes with an inter-sweep delay of 690 ms. To measure the scotopic threshold response, the responses from 50 light flashes with a single-flash intensity of 0.0001 cd·s/m2 and an inter-sweep delay of 1 s were averaged. Anesthesia was reversed with atipamezole (1 mg/kg). Data were analyzed with Espion v6.65.1 software (Diagnosys) according to [11]. Visual evoked potential and scotopic threshold response amplitudes were defined as the amplitude from the baseline to the trough of the negative visual evoked potential response, and from the baseline to the peak of the positive scotopic threshold response, respectively.

Ex vivo compound action potentials

Acute optic nerve preparations for compound action potential (CAP) recordings were carried out as previously described [38]. In brief, optic nerves were isolated after isoflurane anesthesia and decapitation and placed in an interface perfusion chamber (Haas Top, Harvard Apparatus), perfused with artificial cerebrospinal fluid (ACSF containing in mM: 126 NaCl, 3 KCl, 2 CaCl2, 1.25 NaH2PO4, 26 NaHCO3, 2 MgSO4, 10 glucose, pH 7.4) at 37 °C using a TC-10 temperature control system (npi electronic), and continuously oxygenated with 95% O2 and 5% CO2. Nerve ends were inserted into custom-made suction electrodes filled with ACSF and stimulated evoking the CAP. The optic nerve was first stimulated at 0.4 Hz for 1 min to obtain baseline values, followed by a stepwise increase in stimulation frequency, with intervals of 1 min at frequencies of 1, 10, 25, and 50 Hz and a recovery period of 10 min with 0.4 Hz.

Magnetic resonance imaging

Mice were anaesthetized with isofluorane (2.5% induction, 1-1.5% for maintenance). Body temperature was kept constant at 37 ± 1 °C, respiration rate was monitored during scanning and kept above 80 min− 1 by adjusting isoflurane concentrations. MR images were acquired using a 9.4 Tesla preclinical MR scanner with a horizontal bore of 20 cm (Biospec 94/20, Bruker Biospin) and equipped with an actively shielded gradient set of up to 600 mT m− 1. Brain images were acquired using a linearly polarized resonator (transmit) in combination with an actively decoupled surface coil for the mouse brain (Bruker Biospin) as previously described [71]. For the acquisition of manganese enhance MRI, mice received a unilateral intravitreal injection of 2 µl of 0.5 M MnCl2 20 to 24 h before image acquisition, similar to previous reports [36].

MRI was performed at the age of 3, 4.5, 6 and 7.5 months. After initial localizer images, T2-weighted two-dimensional (2D) MRI scans of the brain were acquired in axial orientation using a spin-echo sequence (rapid acquisition with relaxation enhancement, RARE). Acquisition parameters: repetition time (TR) = 2500ms, echo time (TE) = 33ms, RARE factor (RF) = 8, 20 slices, 0.7 mm slice thickness, 70 μm in plane resolution and number of averages (NA) = 1. For the analysis of Mn2+ distribution T1-weighted MRI were acquired using a 3D gradient echo sequence (FLASH) with TR = 30 ms, TE = 5 ms, 60º pulse, NA = 2, field of view (FOV) = 2.0 × 1.5 × 1.5 cm and an isotropic spatial resolution of 156 μm. T1 maps were determined using a spin echo multiple TR sequence (RAREVTR) with TE = 20 ms, RF = 8, NA = 2, TR = 500, 750, 1500, 2500, 6000 ms and geometric parameters as for the 2D with an in-plane resolution of 125 μm. Finally, diffusion MRI was acquired for the calculation of apparent diffusion coefficients (ADC), mean diffusivity (MD) and fractional anisotropy (FA). Parameters for diffusion MRI were as following: TR = 750 ms, TE = 20 ms, NA = 1, spin echo readout, 6 directions, b-values = 0, 100, 600, 1200 s/mm3, and geometric parameters as for the 2D with an in-plane resolution of 125 μm. The Bruker software ParaVision was used for image acquisition, processing and calculation of parametric maps. Examples for the respective images are shown in supplementary figure S1.

Human induced pluripotent stem cell culture

Two WS patient derived induced pluripotent stem cell (iPSC) lines (kindly provided by Prof. F. Urano, Washington University School of Medicine), and their genetically corrected isogenic controls made by adenosine base editing [51], have been used in this study: patient 1, with a homozygous point mutation at c.2002 C > T [p.Q668X], and patient 2, with compound heterozygous mutations at c.376G > A [p.A126T] and c.1838G > A [p.W613X]. Phenotypic characteristics of these patients have previously been described in [45].

Both patient and isogenic iPSC lines were cultured as described in [85], on human Matrigel (VWR)-coated 6-well plates (Corning) in E8 basal medium (Gibco) complemented with E8 supplement Flex (Gibco) and 5 U/ml Penicillin-Streptomycin (Gibco), and passaged two times a week using 0.5 mM EDTA (Gibco) (in PBS). Routine checkups for mycoplasma contamination were done using MycoAlert Mycoplasma Detection Kit (Lonza) to ensure data reliability.

Differentiation of iPSCs to OPCs and pre-myelinating oligodendrocytes

All iPSC lines were differentiated to OPCs and pre-myelinating oligodendrocytes (OPCs/pmOLs), using overexpression of the transcription factor SOX10, as described in Garcia-Leon et al. [17]. SOX10 overexpression iPSCs were dissociated with accutase (Sigma) and plated at 25,000 cells/cm2 in a 6-well plate. Oligodendrocyte induction was done by adding neurobasal medium supplemented with 0.1 µM RA (Sigma), 10µM SB431542 (Tocris), and 1µM LDN-193,189 (Milteny) for the next 6 days. From day 7 on, medium was replaced with neurobasal medium containing 0.1 µM RA, and 1µM SAG (Milllipore) for 3 days. The cells were then dissociated with accutase (Sigma) and replated at 50,000–75,000 cells/cm2 in a poly-L-ornithin-laminin coated 6-well plate with neurobasal medium supplemented with 10 ng/ml PDGFaa (Peprotech), 10 ng/ml IGF1 (Peprotech), 5ng/ml HGF (Peprotech), 10 ng/ml NT3 (Peprotech), 60 ng/ml T3 (Sigma), 100 ng/ml Biotin (Sigma), 1 µM cAMP (Sigma), and 5 µg/ml doxycycline (Sigma) (from here on referred to as oligodendrocyte maturation medium (OMM)) to induce oligodendrocyte fate, until day 23. On day 24, the cells were dissociated with accutase, characterized for pmOL expression markers (O4 and MBP) by either fluorescence-activated cell sorting (data not shown) and/or immunofluorescence, and used for experiments or cryopreserved in liquid nitrogen for future use.

RT-qPCR

Total RNA was isolated and purified using the Quick-RNA Microprep Kit (Zymo research), and cDNA preparation was done with the SuperScript™ III First-Strand Synthesis System kit (Thermofisher), on 500ng-1ug RNA, according to the manufacturer’s instructions. RT-qPCR for oligodendrocyte lineage markers was performed using Platinum SYBR Green qPCR Supermix-UDG (Thermofisher). The sequence of the primers is listed in supplementary table S1.

X-box-binding protein 1 (XBP1) splicing assay

RNA isolation was performed as described above, with the Quick-RNA Microprep Kit (Zymo research) and SuperScript™ III First-Strand Synthesis System kit (Thermofisher). Next, 140 ng cDNA was used to set-up a PCR reaction using Platinum® Taq mix (Invitrogen™) (Supplementary Table 1). The PCR product was run on 2.5% agarose/1x TAE gel with SYBR Safe. In addition to the spliced and unspliced amplicon, a hybrid amplicon was also seen in the gel. Quantification was done as the ratio of spliced amplicon (spliced amplicon + 50% of the hybrid amplicon) and total XBP1 amplicon (spliced + unspliced + hybrid amplicon) using ImageJ [65].

Real time cell metabolic analysis

OPCs/pmOLs were plated on a Seahorse XF24 (Agilent) Extracellular Flux Analyzer culture plate in OMM supplemented with 5 µg/ml doxycycline (Sigma) and RevitaCell supplement (Gibco). The next day, medium change with Mito Stress assay medium (Agilent) containing 10 mM glucose, 2 mM glutamine and 1 mM sodium pyruvate was done one hour before the Mito Stress assay. Addition of 1 µM oligomycin (Sigma-Aldrich), 2.0 µM FCCP (Sanbio) and 0.5 µM antimycin A (Sigma-Aldrich), allowed to detect the mitochondrial respiration activity. After the assay, cells were detached using accutase and counted with an automated cell counter (Nucleocounter 900-002, Chemometec) to enable normalization.

Intracellular Ca2+ measurements

OPCs/pmOLs were plated on a 4-chambered glass bottom dish (Cellvis) in OMM supplemented with 5 µg/ml doxycycline (Sigma) and RevitaCell (Gibco). After 2 days, Ca2+ recording was performed by loading 1.25 µM Cal-520AM (Abcam, ab171868) in OMM for one hour. PBS washing was performed twice, followed by addition of modified Krebs-Ringer solution (135 mM NaCl, 6.2 mM KCl, 1.2 mM MgCl2, 12 mM HEPES, pH 7.3, 11.5 mM glucose and 2 mM CaCl2) for imaging. For imaging (eclipse Ti2, Nikon), a baseline was recorded for 30 s followed by addition of different stimulants (10 µM ATP or 10 µM acetylcholine). Cal-520 was excited at 480 nm, after which fluorescent intensity alterations were measured at 510 nm. Area under the curve and peak intensity were quantified using ImageJ software. Fluorescent signals were plotted as F/F0, with F0 the mean Cal-520AM fluorescent intensity from the first 10 s of the baseline measurement [53].

Proximity ligation assay

The proximity ligation assay was done as described in [85] and based on the manufacturer’s protocol, using the Duolink® Proximity Ligation Assay kit (Sigma-Aldrich). Briefly, OPCs/pmOLs were plated on a 16 well CultureWell™ Chambered Coverglass (Invitrogen) in OMM supplemented with 5 µg/ml doxycycline (Sigma) and RevitaCell (Gibco). After 2 days, the cells were fixed with 4% paraformaldehyde (PFA) (Thermo Fisher) for 10 min, permeabilized with 0.1% Triton X100 in PBS and incubated overnight with the primary antibodies diluted in blocking serum (anti-protein tyrosine phosphatase interacting protein 51 (PTPIP51) (Abcam, ab224081) and anti-vesicle-associated membrane protein B (VAPB) (RnD Systems, MAB58551)) at 4 °C. Secondary antibodies were added to the cells in blocking serum for 1 h at 37 °C. Further ligation and amplification were performed by adding ligation buffer for 90 min, and amplification buffer for 100 min, at 37 °C. Lastly, slides were mounted with Duolink In Situ Mounting Medium containing 4’,6-diamidino-2-phenylindole (DAPI). Primary antibodies were omitted as a negative control. Quantification of VAPB-PTPIP51 interactions was done on confocal microscopy images (C2, Nikon), using ImageJ software, as described [40].

Lipidomics

Lipids were extracted from a pellet of 500,000 cells, homogenized in 700 µl water with a handheld sonicator, and mixed with 800 µl 1 N HCl: CH3OH 1:8 (v/v), 900 µl CHCl3 and 200 µg/ml 2,6-di-tert-butyl-4-methylphenol (BHT; Sigma-Aldrich), 3 µl of SPLASH® LIPIDOMIX® Mass Spec Standard (Avanti Polar Lipids, 330707), and 3 µl of Ceramides and 3 µl of Hexosylceramides Internal Standards (cat. no. 5040167 and 5040398, AB SCIEX). The organic phase was rated and evaporated by Savant Speedvac spd111v (Thermo Fisher Scientific) at RT for 1–2 h. The remaining lipid pellets were stored in -80 °C for further use.

Lipid pellets were reconstituted in 100% ethanol and ran on liquid chromatography electrospray ionization tandem mass spectrometry to identify several lipid classes. Sphingomyelin, cholesterol esters, ceramides, hexose-ceramides, and lactose-ceramides were measured in positive ion mode with a precursor scan of 184.1, 369.4, 264.4, 266.4, 264.4, and 264.4, respectively. Triglycerides, diglycerides and monoglycerides were measured in positive ion mode with a neutral loss scan for one of the fatty acyl moieties. Phosphatidylcholine, alkylphosphatidylcholine, alkenylphosphatidylcholine, lysophosphatidylcholine, phosphatidylethanolamine, alkylphosphatidylethanolamine, alkenylphosphatidylethanolamine, lyso phosphatidylethanolamine, phosphatidylglycerol, phosphatidylinositol, and phosphatidylserine were measured in negative ion mode by fatty acyl fragment ions. Lipid quantification was performed by scheduled multiple reactions monitoring, the transitions being based on the neutral losses or the typical product ions as described above.

Peak integration was performed with the MultiQuantTM software version 3.0.3. Lipid species signals were corrected for isotopic contributions (calculated with Python Molmass 2019.1.1). The total lipid amount and concentration per lipid class were determined with absolute values in nmol/mg DNA. The lipidomic dataset was analysed with the web-based application MetaboAnalyst 5.0 [55]. Missing values were addressed through imputation. Lipid species with > 50% missing values were removed, and the remaining missing values were substituted by LoDs (1/5 of the smallest positive value of each variable). To prepare data for univariate and multivariate analyses, data were log10 transformed and Pareto scaling was applied (mean-centred and divided by the square root of the standard deviation of each variable). Groups were compared using principal component analysis and hierarchical clustering heatmaps (distance measure: Euclidean, clustering algorithm: ward, standardization: auto-scale features). This multivariate-based approach identified a technical outliner with very high lipid concentrations compared to all the other samples, which was removed from the analysis.

Transcriptomics

Total RNA was extracted from OPCs/pmOLs using the Quick-RNA Microprep Kit (Zymo research). Libraries were prepared from 500 ng using the QuantSeq 3’ mRNA-seq library prep kit (Lexogen), and pooled equimolar to 2 nM for single-read sequencing on the HiSeq4000 (Illumina) with settings 51-8-8. Quality control of the generated raw fastq sequence files was performed with FastQC v0.11.7 [2], and adapters were filtered with ea-utils fastq-mcf v1.05 [4]. Splice-aware alignment was performed with HiSat2 [23] against the human reference genome hg38 using default parameters. Reads mapping to multiple loci in the reference genome were discarded. Resulting BAM alignment files were handled with Samtools v1.5. [34]. Reads per gene were calculated with HT-seq Count v2.7.14 and count-based differential expression analysis was done with R-based Bioconductor package DESeq2 [41]. Principle component analysis was performed with regularized log transformation for clustering, and limma package was used to remove batch effects. Cut-off values of adjusted p-value < 0.05 and log2 fold change > 1 were used for the enhanced volcano plot. Gene set enrichment analysis was performed using ClusterProfiler [80] and STRING v12.0 [73].

Western blot

Mice were euthanized by an intraperitoneal injection of 60 mg/kg sodium pentobarbital followed by transcardial perfusion with saline, after which the brain, retina and optic nerve were dissected and snap-frozen in liquid nitrogen. Samples were homogenized in lysis buffer (150 mM NaCl, 50 mM Tris-HCl, 10 mM CaCl2dihydrate, 1% Triton-X 100, 0.05% BRIJ-35), supplemented with EDTA-free protease inhibitor cocktail (Roche Applied Science). Western blot was performed as previously described [32]. Briefly, samples (30 µg for retina, 20 µg for optic nerve) were separated on a 4–12% Bis-Tris gel and transferred onto a PDVF membrane, followed by 2 h blocking with 15% milk powder (in TBST) and overnight incubation with rabbit anti-BIP (ref. 3183, Cell signaling, 1/1000 for retina, 1/500 for optic nerve) or mouse anti-CHOP (ref. 895, Cell signaling, 1/1000 for retina, 1/500 for optic nerve). Protein bands were visualized using horseradish peroxidase-labeled secondary antibodies and a luminol-based enhanced chemiluminescence kit. At least two Western blot experiments (i.e., two technical repeats) were performed for each of the lysates. Optical density measurements were normalized against a SWIFT Membrane total protein stain (G-Biosciences), and subsequently presented as a percentage relative to the naive condition.

OPCs/pmOLs were lysed by using M-PER buffer (Thermo Scientific™) containing PhosSTOP™ Phosphatase Inhibitor Cocktail (Roche Diagnostics) and complete™ Protease Inhibitor (Roche Diagnostics). Protein concentration was determined by using a Pierce™ BCA Protein Assay kit (Thermo Scientific™) and 20 µg protein was mixed with SDS-containing reducing sample buffer (Thermo Scientific™), followed by denaturation at 95 °C for 10 min and loading on 4–20% ExpressPlus™ PAGE gels (GenScript). The gel was allowed to run at 120 V for 2 h, followed by 0.2 μm nitrocellulose membrane transfer at 25 V, 2.5 A for 7 min using iBlot™ 2 dry Blotting Transfer system (Thermo Fisher). The membrane was blocked in Tris-buffered saline with Tween-20 (TBST) (Sigma-Aldrich) containing 5% skim milk (Sigma-Aldrich) for 1 h, incubated overnight with primary antibody (BiP, cat. 3183 S dilution 1:1000 or WFS-1 cat. PA5-76065 dilution 1:2000) in 3% bovine serum albumin-TBST at 4 °C. The next day, incubation with secondary antibody diluted in TBST was performed for 1 h. SuperSignal™ West Pico PLUS Chemiluminescent Substrate (Thermo Scientific™) was used for chemiluminescence imaging (ImageQuant LAS4000, GE Healthcare). Beta-actin (ref. 8457, Cell Signaling, 1/1000) or vinculin (ref. V9131, Sigma, 1/1000) were used as a loading control and for normalization of the measured optical densities.

Immunohistochemistry and morphometric analyses

Mice were euthanized by an intraperitoneal injection of 60 mg/kg sodium pentobarbital, followed by transcardial perfusion with saline and 4% PFA. For retinal flatmounts, the eyes were post-fixated in 4% PFA for 1 h at room temperature, retinas were flatmounted and post-fixated for another hour in 4% PFA at room temperature. For cryosections of the optic nerve, the nerves were immersed in 4% PFA for 30 min at room temperature.

For wholemount immunostaining, wholemounts were frozen for 15 min at -80 °C in PBS + 2% Triton-X and incubated overnight at room temperature with primary antibodies for RNA-binding protein with multiple splicing (RBPMS) (ref. 1830-RBPMS, PhosphoSolutions, 1/250) or TH (ref. AB152, Millipore, 1/1000), followed by incubation with corresponding fluorophore-conjugated secondary antibodies. Washes were performed with PBS + 2% Triton-X. Images were made with a DM6 fluorescence microscope (Leica), and analysed using ImageJ software [41]. RBPMS + retinal ganglion cells were counted using a validated automated counting method [26]. Tyrosine hydroxylase + amacrine cell numbers were counted manually [48].

For cryosections of the optic nerve, tissues were cryoprotected in an ascending sucrose series and embedded in optimal cutting temperature medium (Tissue-Tek, Lab Tech) to make 12 μm thick longitudinal sections. Optic nerve cryosections were blocked with 20% pre-immune donkey serum for 45 min, followed by overnight incubation with the primary antibody for platelet-derived growth factor receptor alpha (PDGFRa) (ref. AF1062-SP, R&D systems, 1/200), OLIG2 (ref. 13999-1-AP, Proteintech, 1/400), or CC1 (anti-adenomatous polyposis coli clone CC1, ref. OP80, Calbiochem, 1/200) and a 2-hour incubation with the secondary fluorophore-conjugated antibody. Washes were performed with TBS with 0.1% Tween-20. Nuclei were counterstained with DAPI and sections were mounted with mowiol. Images were taken with an epifluorescence microscope (DM6 B, Leica Microsystems) and 20X objective. Two or three representative sections were chosen for each optic nerve, on which cells in 3 regions of interest (ROI) of 0.08 mm² were manually counted using the cell-counter plugin of ImageJ: one ROI close to the ONH, one ROI in the middle of the optic nerve and one ROI close to the optic chiasm.

For immunocytochemistry of OPCs/pmOLs, 30,000 cells per well were plated in a 96-well plate in OMM. The cells were maintained for 2 days, followed by fixation with 4% PFA for 15 min at room temperature. The cells were washed, blocked and permeabilized with 5% goat serum (Dako) and 0.1% Triton X-100 (Sigma) for 1 h. Overnight incubation with the primary antibody (MBP, ref. ab9348 dilution 1:75 and O4, ref. MAB1326 dilution 1:1000) diluted in 5% goat serum at 4 °C was followed by secondary antibody incubation, diluted in Dako REAL Antibody Diluent (Dako) for 1 h. Hoechst33342 (Sigma, dilution 1/2 000 in Dako REAL Antibody Diluent) was applied for nuclear counterstaining. Fluorescence imaging was performed using the Operetta High Content Imaging System (PerkinElmer) and analysis was done using automated segmenting and counting of objects using Columbus software (PerkinElmer). Briefly, the nuclear staining was segmented and used as a reference to have object count and cytoplasm segregation. The cell population was selected based on the mean fluorescence intensity in both cell regions (nuclear and cytoplasmic), using a threshold to select the positive population. The integrated fluorescence intensity was measured by dividing the mean fluorescence intensity by the cytoplasm area (in µm2).

Transmission electron microscopy and morphometric analyses

Mice were euthanized by an intraperitoneal injection of 60 mg/kg sodium pentobarbital (Dolethal, Vetoquinol) followed by transcardial perfusion with 2.4% glutaraldehyde and 4% PFA in 0.1 M Na-cacodylate buffer. Optic nerves were gently dissected from the brain, while being immersed in fixative, followed by overnight post-fixation at 4 °C. Next, ascending concentrations of acetone were used to dehydrate the samples, after which they were placed in a 1:1 mixture of araldite epoxy resin and acetone overnight, and embedded in araldite expoxy resin. Next, 70 nm cross-sections of the optic nerves were made, transferred to 0.7% formvar-coated copper grids and contrasted with 0.5% uranyl acetate and lead citrate. Images were made with an EM208 S electron microscope (Philips), Morada soft imaging system camera and iTEM FEI software (Olympus). ImageJ software was used for morphometric analysis of at least 100 axons (on 4 different sections) per optic nerve. Axon area and axon-plus-myelin area were measured by encircling these areas, from which inner and out axon diameter, respectively, were calculated, as well as the g-ratio. Myelin area was calculated by subtracting the axon area from the axon-plus-myelin area. White space or interaxonal space was calculated by subtracting the sum of the axon-plus-myelin areas within a given region of interest from the total area of that region. Both measures are shown relative to the total region of interest (in %). Axon density was defined by counting the number of axons (Cell Counter plugin) in a 20 × 16 μm region of interest.

Statistical analysis

The employed statistical analyses and number of mice (N) or number of experiments performed (n) are stipulated in the respective figure legends. For experiments with OPCs/pmOLs, a minimum of three independent repeats, based on at least three different differentiation batches, was performed. Statistical analyses were performed using Prism v.8.2.1 (GraphPad). Differences were considered statistically significant for two-sided p-values < 0.05 (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001).