Combinatorial expression of glial transcription factors induces Schwann cell‐specific gene expression in mouse embryonic fibroblasts

Abstract

Background

Schwann cells provide peripheral nerve trophic support, myelinate axons, and assist in repair. However, Schwann cell repair capacity is limited by chronic injury, disease, and aging. Schwann cell reprogramming is a cellular conversion strategy that could provide a renewable cell supply to repair injured nerves. Here, we developed a plasmid‐based approach to test the Schwann cell conversion potential of four glial transcription factors.

Results

We employed four transcription factors implicated in Schwann cell differentiation and repair: Sox10, Sox2, Jun, and Pax3. Expression vectors were generated for Sox10 alone and two triple transcription factor combinations: Jun‐Pax3‐Sox2 (triple 1, T1) and Sox10‐Jun‐Sox2 (triple 2, T2). Mouse embryonic fibroblasts (MEFs) were transfected with these vectors, transferred to glial inductive media, and Schwann cell‐marker expression was in assessed by immunostaining, flow cytometry, and qPCR. All expression vectors repressed fibroblast‐specific gene expression. However, T2 was most efficient at generating O4⁺ Schwann cell‐like cells, which had some capacity to myelinate denervated axons from explanted dorsal root ganglia. In comparison, T1 more efficiently induced repair Schwann cell‐marker expression in converted O4⁺ cells.

Conclusions

T1 and T2 convert MEFs to Schwann cells with different efficacies and gene expression profiles, and may provide cell‐based therapies for peripheral nerve repair.

Key Findings

• Sox10, T1 and T2 suppress fibroblast gene expression in MEFs.

• T2 is most efficient at converting MEFs to Schwann cell‐like cells, which have some myelinating capacity.

• T1 more efficiently induces repair Schwann cell‐specific gene expression in converted O4+ cells.

1. INTRODUCTION

It is a common misconception that since the peripheral nervous system (PNS) has the ability to regenerate, ¹ repair strategies are not required. However, repair processes are only transiently activated in response to peripheral nerve injury (PNI) and often do not persist during the entire phase of axonal regrowth, which is exceedingly slow, at a rate of about 1 mm/day in humans. ² Thus, with larger injuries, there is a reduced likelihood that the regenerating axon will successfully innervate the stump. Additionally, advanced age markedly reduces repair potential in older patients. ³ , ⁴ Interventions are thus required to facilitate repair to prevent chronic sensory and/or motor impairments and life‐long neuropathic pain. ⁵ The current standard of care is autologous grafting of a healthy nerve to fill the gap between the proximal and distal stumps of the injured peripheral nerve. ⁶ A complication of this procedure is potential donor‐site morbidity in what was otherwise a healthy nerve. Furthermore, success rates for nerve grafts are low, with only ~25% of patients recovering full motor function and ~3% regaining sensory function. ⁷ Treatment strategies are thus urgently needed.

Mature Schwann cells are either myelinating or non‐myelinating Remak cells. Upon PNI, peripheral axons distal to the site of damage undergo Wallerian degeneration, with the loss of axonal contact triggering Schwann cells to de‐differentiate and enter a repair state. ⁸ The repair state involves a transient downregulation of myelin genes to support axonal repair and regeneration, and the re‐expression of several transcription factors (TFs) involved in Schwann cell differentiation in the embryo. ⁹ , ¹⁰ , ¹¹ , ¹² We focused on four TFs implicated in the Schwann cell repair response in this study: Sox10, Sox2, Pax3, and Jun. Sox10 is expressed throughout Schwann cell lineage development and is crucial for the survival, proliferation, and repair phenotype of Schwann cells in the adult, controlling myelin production and supporting nerve regeneration. ¹³ Given its central importance, SOX10 mutations are also observed in peripheral demyelinating neuropathies. ¹⁴ Sox2 is expressed in Schwann cell precursors (SCPs) and immature Schwann cells during development. ¹⁵ The re‐initiation of Sox2 expression in repair Schwann cells inhibits myelination and promotes proliferation. ¹⁶ Similarly, Pax3 is induced in repair Schwann cells and is also expressed in neural crest cell (NCC) precursors and SCPs during embryonic development, supporting Schwann cell proliferation and repressing myelin gene expression. ¹⁰ , ¹⁷ , ¹⁸ , ¹⁹ Finally, Jun is expressed in late immature Schwann cells/pro‐myelinating Schwann cells and is rapidly upregulated in Schwann cells post‐injury, inhibiting myelination and initiating Schwann cell de‐differentiation. ⁴ , ²⁰ Notably, in Jun knockout mice, peripheral nerve regeneration is impaired. ⁹ , ¹⁰ , ²⁰ , ²¹ , ²²

Despite the acquisition of a repair Schwann cell phenotype, endogenous PNI repair is inefficient since the repair capacity of de‐differentiated Schwann cells declines with time post‐injury, increased age, in chronic injuries, and with pre‐existing ailments. ⁴ , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ Schwann cell transplants have been used to improve functional recovery in animal models. ²⁸ However, a renewable Schwann cell source is required for clinical applications. Several strategies have been employed to generate Schwann cells in vitro, such as the directed differentiation of pluripotent ⁸ , ²⁹ , ³⁰ , ³¹ , ³² or multipotent stem cell populations, including skin‐derived precursor cells (SKPs), ³³ , ³⁴ mesenchymal cells, ³⁵ , ³⁶ , ³⁷ bone marrow stromal cells, ³⁷ , ³⁸ bone marrow derived neural crest cells, ³⁹ adipose‐derived stem cells, ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ muscle derived stem cells, ⁴⁵ dental pulp stem cells, ⁴⁶ , ⁴⁷ and epidermal neural crest stem cells. ⁴⁸ Small molecules, including TGFβ and GSK3 inhibitors and neuregulin 1 (NRG1) can accelerate the differentiation of stem cells along the Schwann cell lineage. ³⁰ , ³¹ , ³² , ⁴⁹ , ⁵⁰ , ⁵¹ However, in general, Schwann cell differentiation protocols are lengthy, for instance, requiring 5–6 weeks when using SKPs. ³³ , ³⁴ This long timeline may limit clinical utility for generating autologous transplants. Furthermore, there is a risk that undifferentiated, proliferative stem cells with the potential to form tumors are not completely removed from cell products.

Direct cellular reprogramming involves the transition of one somatic cell type to another without passing through a pluripotent stage, mitigating the inherent stem cell risk. ⁵² , ⁵³ Induced Schwann cells have been generated from mouse embryonic fibroblasts (MEFs), human fibroblasts, ⁵⁴ and skin fibroblasts. ²⁸ , ⁵⁵ , ⁵⁶ , ⁵⁷ , ⁵⁸ Cellular conversion has been achieved using lineage‐specifying TFs, including Sox10. ¹¹ , ⁵⁹ , ⁶⁰ , ⁶¹ When Sox10 overexpression is combined with small molecules (GSK3 and HDAC inhibitors), fibroblasts are converted into NCCs that subsequently can be differentiated into Schwann cells in the presence of NRG1, cAMP, and CNTF. ⁴⁹ By expressing Sox10 in combination with Early growth response 2 (Egr2), a myelination TF, fibroblasts can be converted into induced Schwann cells that can accelerate myelination and regeneration of an injured peripheral nerve. ⁵⁴ , ⁶²

Our goal was to generate induced Schwann cells with an enhanced repair‐like phenotype using a non‐integrative, episomal approach. We reasoned that the non‐integrative, non‐replicating nature of the plasmid‐based system would make the cellular reprogramming method safer compared to the use of viral vectors. Herein, we compared the capacity of Sox10 alone versus two triple TF‐expression vectors (TTFEVs), Jun‐Pax3‐Sox2 (triple 1, or T1) and Sox10‐Jun‐Sox2 (triple 2, or T2) to initiate Schwann cell‐specific gene expression in MEFs. Through the analysis of select panels of Schwann cell markers by immunostaining, flow cytometry, and qPCR, we demonstrate that, in general, Jun‐Pax3‐Sox2 (T1) has an enhanced capacity to induce repair Schwann cell‐marker expression, while Sox10‐Jun‐Sox2 (T2) more efficiently converts MEFs. We thus suggest that T1‐ and T2‐derived Schwann cells should be further investigated as a potential cell source for PNI repair.

2. RESULTS

2.1. Transcription factors expressed in Schwann cells after peripheral nerve injury

We set out to design a lineage conversion strategy in which fibroblasts could be reprogrammed into repair‐like Schwann cells for peripheral nerve repair (Figure 1A). For this purpose, we generated triple TTFEVs, combining TFs with reported roles in the Schwann cell repair process to determine whether we could more efficiently reprogram fibroblasts into a repair Schwann cell phenotype (Figure 1A). Since the molecular signature of repair Schwann cells resembles a mixed early/late embryonic phenotype, ¹¹ we focused on four TFs ‐ Sox10, Sox2, Pax3, and Jun, each implicated in Schwann cell specification, differentiation and in PNI repair. ¹⁰ , ¹¹ , ¹⁶ , ²⁰ By mining a bulk RNA‐seq dataset of purified Schwann cells collected at 3‐, 5‐, and 7‐days post‐PNI, Sox10, Sox2, Jun, and Pax3 transcripts were detected in the injured nerve, albeit with low levels of Pax3 ⁶³ (Figure 1B). To assess whether these genes were enriched in a particular population of Schwann cells, we mined a single‐cell (sc) RNA‐seq dataset collected from naive and injured (3‐, 12‐, and 60‐days post‐injury) peripheral nerves. ⁶⁴ Using graph‐based clustering and uniform manifold approximation and projection (UMAP) in the Seurat R package, a total of 18,411 cells were assigned a Schwann cell identity and subclustered into myelinating, Remak (non‐myelinating), repair, and dividing/repair cells (Figure 1C). Marker genes were used to assign subcluster identities and are visualized in a dotplot, showing the average expression levels and the percentage of cells expressing each gene within the subclusters (Figure 1D). Sox10, a canonical Schwann cell marker, was expressed in all Schwann cell subclusters, whereas Egr1 and Gfap were highly expressed in Remak cells, Mbp and Mag transcripts were elevated in myelinating Schwann cells, Cdk1, Top2a, and Mki67 were highly expressed in dividing/repair Schwann cells, and Ngfr and Slc22a3 transcripts were enriched in repair Schwann cells that were not proliferating (Figure 1D).

FIGURE 1.

Analysis of repair Schwann cell transcription factor expression post‐peripheral nerve injury. (A) Schematic of the lineage conversion strategy used and the potential application for peripheral nerve repair. (B) Bulk RNA‐sequencing in fragments per kilobases of transcript per 1 million mapped reads (FPKM) of Sox10, Sox2, Jun and Pax3 at 3, 5 and 7‐days post‐PNI. ⁶³ (Accession: GSE177037). (C–F) UMAP of scRNA‐seq data collected from naive and injured (3, 12, and 60 post‐injury) peripheral nerves. ⁶⁴ Subclusters include: Repair (blue), repair/dividing (green), myelinating (pink), Remak or non‐myelinating (brown) and unknown (gray) (C). Dotplot showing the expression levels of genes used to assign Schwann cell cluster identities (D). Dotplot showing the expression of the four genes contained in the TTFEV vectors: Sox10, Sox2, Pax3 and Jun, in the different Schwann cell subtypes (E). Feature plots showing Sox10, Sox2, Jun and Pax3 expression in Schwann cell subclusters (F). (Accession: GSE216665).

We next used the scRNA‐seq data to visualize the expression of Sox10, Sox2, Pax3, and Jun within the Schwann cell subclusters, which are the four genes used in our TTFEV vectors. Sox10 was highly expressed in all subclusters, albeit at lower levels in repair/dividing Schwann cells. Sox2 had the highest expression in the repair subtype, followed by Remak cells. Jun was similarly expressed at the highest levels in Remak cells, followed by repair and myelinating cell subtypes (Figure 1E). However, Pax3 transcripts were not detected in any subcluster, a reflection of the low expression levels of this gene and the lower sensitivity of scRNA‐seq (Figure 1E) compared to bulk RNA‐seq (Figure 1B). Accordingly, in feature plots, Sox10 was broadly expressed in all Schwann cell populations, as was Jun, albeit with a slight enrichment in the repair cluster (Figure 1F), consistent with previous reports. ⁶⁵ Taken together, these data confirm that Sox10, Sox2, Jun, and Pax3 are expressed in the injured nerve in the first few days post‐PNI, albeit with very low levels of Pax3 detected.

2.2. Establishing a plasmid‐based system for Schwann cell reprogramming

To develop a lineage conversion strategy for repairing Schwann cells, we used MEFs as a starting somatic cell source. MEFs are easily accessible and have been used in other reprogramming studies, including to make induced pluripotent stem cells ⁶⁶ and Schwann cells. ⁶² MEFs were isolated from embryonic day (E) 12.5–E13.5 mouse embryos, and after 4 days of expansion, nucleofected with CAG‐mCherry (hereafter, control) or CAG‐Sox10‐v2a‐mCherry (hereafter, Sox10) expression vectors (Figure 2A). After 1 day, transfected cells were transferred to glial media, ³³ containing forskolin and Neuregulin (NRG1), which supports the conversion of MEFs into NCC‐like cells and their subsequent differentiation into Schwann cell‐like cells. ⁴⁹ To confirm transfection was successful, mCherry epifluorescence was monitored after 1, 3, 5, and 7 days (Figure 2B). We observed a relatively high transfection efficiency (~60%–65%) for both control and Sox10 constructs at 1 day post‐transfection, with an approximate 14‐fold decline in the percentage of transfected cells by day 7 due to plasmid dilution during cell division (Figure 2B).

FIGURE 2.

Analysis of an episomal plasmid‐based expression system for Sox10 in MEFs. (A) Schematic of the MEF isolation procedure, plasmid‐based transfection of control and glial lineage TFs (day 0), addition of glial media (day 1) and qPCR analysis (day 14). (B) mCherry expression in transfected MEFs at 1‐, 2‐, 5‐ and 7‐days post‐transfection, showing merged phase/fluorescent images and fluorescent images alone. (Scale bar = 100 μm). (C–E) Relative Sox10 transcript levels in Sox10‐transfected cells at 1, 3, 5 and 7‐days post‐transfection, normalized to negative control, and assayed using qPCR (C). Relative Sox10, Egr2, Sox2, Egr1, Tfap2a and Pax3 transcript levels in negative control and Sox10‐transfected MEFs, assessed by qPCR at 14‐days post‐transfection (D). Relative Thy1 and Col1a1 transcript levels in negative control and Sox10‐transfected MEFs, assessed by qPCR at 14‐days post‐transfection (E). Graphs show means ± s.e.m. Statistical comparisons were made using an ANOVA with a post‐hoc Tukey's test.

To assess the onset of Sox10 expression, we performed qPCR at 1, 3, 5, and 7 days post‐transfection, normalizing Sox10 transfections to the negative plasmid transfection controls. Sox10 transcript levels reached their peak at 1‐day post‐transfection, followed by a decline in Sox10 mRNA levels due to dilution of the expression vectors in rapidly dividing MEFs (Figure 2C). Although Sox10 expression declined over time, we reasoned that it could induce a more permanent change in endogenous gene expression. Control‐ and Sox10‐transfected MEFs were thus cultured in glial media for 14 days before harvesting to assess the expression of a panel of Schwann cell markers by qPCR (Figure 2D). At 14 days, Sox10‐transfected cells expressed higher levels of several Schwann cell‐expressed genes, including Sox10, Egr2, Sox2, and Tfap2a, whereas Egr1 and Pax3 were not induced (Figure 2D). As another measure of the conversion of fibroblasts to a new cell type, we asked whether the fibroblast‐specific genes Thy1 and Col1a1 were downregulated. ⁶² We found that Thy1 and Col1a1 were expressed at 6.0‐ and 6.8‐fold lower levels in Sox10‐transfected MEFs compared to negative control transfections, respectively (Figure 2E). In summary, the overexpression of Sox10 in MEFs cultured in glial media increases the expression of a set of Schwann cell markers and represses fibroblast‐specific gene expression.

2.3. Design of triple transcription factor expression vectors (TTFEVs) for Schwann cell reprogramming

While Sox10 expression alone converts fibroblasts to NCCs that can subsequently be differentiated into Schwann cells when cultured in glial induction media, the combined expression of Sox10 and Egr2 is more efficient. ⁵⁴ , ⁶² We therefore set out to determine whether combinations of three TFs might further improve the conversion of NGFR‐negative MEFs to Schwann cells, focusing on Sox10, a Schwann cell determinant, and TFs implicated in Schwann cell repair: Sox2, Pax3, and Jun ¹⁰ , ¹¹ , ¹⁶ , ²⁰ (Figure 3A). We generated a single Sox10 expression vector, and two triple TTFEVs; Pax3‐Jun‐Sox2 (Triple 1, or T1) and Sox10‐Jun‐Sox2 (Triple 2, or T2) (Figure 3B). Sox2 and Jun were kept in common in both triple vectors, given their importance in defining a repair Schwann cell identity. ⁴ , ¹⁰ , ¹⁶ To drive expression, we used a CAG promoter/enhancer and separated each cistron with a viral 2A (v2A) self‐cleaving peptide to allow stoichiometric expression of each gene. ⁶⁷ We verified that T1 and T2 induced expression of the corresponding three TFs by Western blotting 24 h after the transfection of NIH‐3 T3 cells (Figure 3C). As expected, SOX2 and JUN were both detected in T1‐ and T2‐transfected cells, while PAX3 and SOX10 were detected only in T1‐ and T2‐transfected cells, respectively (Figure 3C). Notably, JUN was expressed in NIH‐3 T3 cells at relatively high levels, so we could not conclude definitively that it was also expressed from the triple TF vectors, prompting us to use qPCR to assess transcript levels in transfected MEFs.

FIGURE 3.

Design and testing of triple transcription expression vectors (TTFEVs) with transcription factors expressed after injury. (A) Schematic of the MEF isolation procedure, plasmid‐based transfection of control and glial lineage TFs (day 0), addition of glial media (day 1) and qPCR analysis (day 14). (B) Schematic of four expression vectors: mCherry control (‘negative’), Sox10‐mCherry (Sox10), Pax3‐Jun‐Sox2 (triple 1, or T1) and Sox10‐Jun‐Sox2 (triple 2, or T2). (C) Western blot analysis of JUN, SOX10, PAX3, SOX2 and GAPDH (loading control) in T1‐ and T2‐transfected NIH‐3T3 cells. (D) qPCR analysis of relative Sox10, Sox2, Pax3, and Jun transcript levels in negative, T1‐ and T2‐ transfected MEFs at 14‐days post‐transfection. Graphs show means ± s.e.m. (N = 3). Statistical comparisons were made using an ANOVA with a post‐hoc Tukey's test.

Compared to negative control transfections, Sox10 transcript levels were significantly elevated in MEFs transfected with Sox10 alone or with T2, a vector that contains the Sox10 cistron, at 14 days post‐transfection (Figure 3D). Sox2, contained within T1 and T2 expression vectors, was expressed in T1‐ and T2‐transfected MEFs after 14 days, whereas Pax3 expression was only elevated in T1 transfections, which contain the Pax3 cistron (Figure 3D). Finally, Jun was expressed above baseline in T2 transfections after 14 days, as expected given that Jun is part of the T2 cistron. However, Jun transcripts were not elevated above baseline in T1 transfections in this time frame, even though Jun was part of T1 as well (Figure 3D). Notably, Jun transcript levels were above baseline controls at 21 days post‐transfection (see Figure 5D). Taken together, these data support the efficacy of our TTFEV strategy to promote the ectopic expression of triple TF combinations in MEFs.

FIGURE 5.

Analysis of Schwann cell mRNA marker expression in NGFR⁺O4⁺ Schwann‐like cells isolated from NGFR‐negative MEFs transfected with negative, Sox10‐, T1‐ or T2‐expression vectors. (A) Schematic of the lineage conversion strategy, with a negative FACS sort to remove NGFR⁺ Schwann cell lineage cells contaminating MEFs prior to transfection. Cells were analyzed by qPCR at 21‐days post‐transfection. (B) FACS plot showing the NGFR‐negative cell population sorted out (red asterisk). (C) Schematic showing the isolation of nerve‐derived Schwann cells, used as a positive control for qPCR. (D) Relative Sox10, Sox2, Pax3 and Jun transcript levels in negative control and Sox10‐, T1‐ and T2‐transfected NGFR‐negative MEFs, assessed by qPCR at 21‐days post‐transfection. N = 3. Graphs show means ± s.e.m. Statistical comparisons were made using an ANOVA with a post‐hoc Tukey's test. (E) Relative Ngfr, Gap43, Tfap2a, Cdh19, Egr1, Egr2, S100b, Mpz, Olig1, and Col1a1 transcript levels in negative control and Sox10‐, T1‐ and T2‐transfected NGFR‐negative MEFs, assessed by qPCR at 21‐days post‐transfection. N = 3. Graphs show means ± s.e.m. Statistical comparisons were made using an ANOVA with a post‐hoc Tukey's test. (F) FACS plot, showing the fluorescence intensity of O4 and NGFR labeling of MEFs transfected with negative, Sox10, T1 and T2 expression vectors after 21 days. NGFR⁺O4⁺ cells were sorted out, including O4⁺ cells with or without NGFR expression.

2.4. Analysis of Schwann cell protein marker expression in NGFR‐negative MEFs transfected with negative, Sox10‐, T1‐, or T2‐expression vectors

Approximately 0.25% of dissociated cells in ‘pure’ MEF cultures are NGFR‐expressing NCCs and SCPs that could more readily differentiate into Schwann cells in the presence of glial supporting cues, such as forskolin and NRG1, present in our glial media. ³³ , ⁶² Thus, to ensure that we were truly monitoring the conversion of MEFs to induced Schwann cells, after 4 days of MEF expansion, a negative sort was performed to remove any contaminating NGFR⁺ Schwann cells using fluorescence‐activated cell sorting (FACS) (Figure 4A). To assess the impact of the Schwann cell TFs on lineage conversion, NGFR‐(negative) MEFs were transfected with control, Sox10, T1, or T2 expression vectors (Figure 4A). One day later, glial induction medium was added, followed by immunostaining at 21 days post‐transfection (Figure 4A). We first examined the expression of two proteins expressed in developing, non‐myelin forming, and repair Schwann cells, including growth‐associated protein 43 (GAP43), ⁶⁸ , ⁶⁹ and the low‐affinity nerve growth factor receptor (NGFR) for neurotrophins. ⁷⁰ GAP43 and NGFR were expressed in a few scattered cells transfected with the negative control vector, likely due to the glial induction media, but the expressions of both NGFR and GAP43 were comparatively elevated in Sox10, T1, and T2‐transfected MEFs (Figure 4B).

FIGURE 4.

Schwann cell protein marker expression in NGFR‐negative MEFs transfected with negative, Sox10‐, T1‐ or T2‐expression vectors. (A) Schematic of the lineage conversion strategy, adding a negative FACS sort to remove NGFR⁺ Schwann cell lineage cells prior to transfection. Cells were analyzed by NGFR and GAP43 immunostaining and O4⁺ flow cytometry at 21‐days post‐transfection. (B) Immunostaining of GAP43 and NGFR on NGFR‐negative MEFs transfected with Sox10, T1, and T2 after 21‐days in vitro (DIV). DAPI counterstain in blue. (Scale bar = 100 μm). (C) Flow cytometry to assess O4 expression in unstained MEFs (to set gate) and NGFR‐negative MEFs transfected with Sox10, T1, or T2 after 21‐days. The percentages of cells in the O4⁺ gate are shown.

To examine the induction of Schwann cell‐marker expression in a more quantitative way, we used flow cytometry to quantify the expression of O4, an immature and mature Schwann cell marker that is not detected in SCPs. ⁷¹ When NGFR‐negative MEFs were transfected with the negative control vector, O4 was expressed in 1.99% of cells, resembling the 2.76% of Sox10‐transfected, and 1.82% of T1‐transfected O4⁺ cells (Figure 4C). In contrast, 13.2% of T2‐transfected MEFs were O4⁺, suggesting that T2 can initiate O4 expression more effectively (Figure 4C). In summary, our analysis of protein expression suggests that in the presence of glial stimulatory cues, Schwann cell TFs can induce the ectopic expression of Schwann cell lineage proteins.

2.5. Establishing essential controls for the assessment of Schwann cell conversion

To determine whether the selected reprogramming TFs were able to induce Schwann cell‐specific gene expression in heterologous cells, we used MEFs as our starting cell type. We set up three important controls to ensure that our gene expression analyses were valid. First, to ensure that the cells used for transfection were pure MEFs and did not contain contaminating NCCs or SCPs, we used a negative FACS approach, employing a FITC‐labeled NGFR antibody (Figure 5A,B). In a FACS histogram of isolated MEFs labeled with NGFR‐FITC, NGFR‐negative and a smaller population of NGFR‐positive cells were detected (Figure 5B). NGFR‐negative MEFs were captured for further analyses, effectively removing the subset of cells that express NGFR at high levels from our starting MEF pool (Figure 5B).

Second, we employed a positive control for qPCR, harvesting mature Schwann cells from the sciatic nerves of 4‐month‐old mice (Figure 5C). Dissected nerves were cultured as explants for 2 weeks prior to dissociation to enrich for Schwann cells, allowing contaminating fibroblasts to migrate out of the nerve, while also inducing nerve demyelination, a trigger for Schwann cell expansion (Figure 5C). We acknowledge the caveat that the longer Schwann cells are cultured in vitro, the less representative they are of endogenous cells. Nevertheless, for the purpose of confirming the specificity of our panel of Schwann cell‐specific qPCR primers, we reasoned that our methodology would be sufficient. Typically, nerve‐derived Schwann cells isolated from the sciatic nerve are a mix of myelinating and non‐myelinating Schwann cells. To characterize gene expression in the nerve‐derived Schwann cells, we used qPCR, first asking whether the four TTFEV‐TFs were expressed in these cells. We observed that Sox10 was expressed at the highest levels in nerve‐derived Schwann cells, followed by low levels of Sox2 and Jun, whereas Pax3 transcripts were not detected (Figure 5D). Next, we examined the expression of a panel of Schwann cell markers not present in the expression vectors. Ngfr, Tfap2a, Cdh19, S100b, Mpz, Egr1, and Egr2 were all expressed in nerve‐derived Schwann cells, whereas unexpectedly, Gap43 and Olig1 were not (Figure 5E). Mpz and S100b transcripts were expressed at the highest levels, both of which are expressed in myelinating Schwann cells, suggesting that the nerve‐derived Schwann cells included a large subset of myelinating glia (Figure 5E). Finally, as anticipated, a fibroblast‐specific marker, Col1a1, was expressed in MEFs, repressed by all reprogramming TF‐expression vectors, and not expressed in nerve‐derived Schwann cells (Figure 5E).

Third, anticipating low reprogramming efficiency, we used FACS at 21 days post‐transfection to enrich for cells that had begun expressing O4, which included NGFR⁺ cells (hereafter, NGFR⁺O4⁺ cells, Figure 5F), before qPCR analyses. NGFR is expressed in NCCs and developing Schwann cells, and declines in expression with the onset of myelination. ⁹ , ⁷² O4 is expressed in immature and mature Schwann cells, but is absent from SCPs. ⁷¹ Together, NGFR and O4 capture Schwann cell lineage cells at different developmental stages. ⁶² The negative control vector induced the production of very few NGFR⁺O4⁺ cells, whereas the reprogramming TFs generated 3–4 times more (Figure 5F).

2.6. Analysis of Schwann cell mRNA marker expression in NGFR‐negative MEFs transfected with Sox10, T1, and T2

To evaluate the ability of the reprogramming TFs to induce Schwann cell‐specific gene expression in NGFR‐negative MEFs, we employed a qPCR approach. We first asked whether transcripts for the four TFs present in the expression vectors were detected in the sorted NGFR⁺O4⁺ cells by isolating RNA for qPCR analysis. Compared to MEFs that were transfected with a negative control construct, Sox10 transcripts were elevated in cells transfected with Sox10, T1, or T2 (Figure 5D). Sox10 transcript levels in the transfected cells were largely equivalent (Sox10, T2) or exceeded (T1) the levels of Sox10 detected in endogenous Schwann cells (Figure 5D). Strikingly, T1 induced the highest levels of Sox10 expression, even though Sox10 was not within this expression construct (Figure 5D). Similarly, Sox2 transcript levels were elevated above negative control and nerve‐derived Schwann cell levels in NGFR⁺O4⁺ sorted cells derived from MEFs transfected with Sox10, T1 or T2, although the Sox2 cistron was only in T1 and T2 (Figure 5D). Pax3 transcript levels were also elevated above negative control and isolated Schwann cell levels in NGFR⁺O4⁺ sorted cells transfected with Sox10, T1, or T2, although the Pax3 cistron was only in T1, suggesting that Sox10 and T2 can induce endogenous Pax3 expression (Figure 5D). In contrast, Jun transcript levels were only elevated above negative control levels and nerve‐derived Schwann cell levels in NGFR⁺O4⁺ enriched cells transfected with T1 and not T2, despite both vectors containing the Jun cistron (Figure 5D). In summary, at 21 days post‐transfection, the critical repair genes included in the Sox10, T1, and T2 expression vectors were induced in MEFs, in some instances, even when not part of the individual constructs. One exception is Jun, which was not expressed above baseline levels in T2‐transfected cells, at least by qPCR.

We next analyzed the expression of the remaining Schwann cell markers in NGFR⁺O4⁺ FACS‐enriched cells isolated 21 days after transfection with Sox10, T1, and T2. Both Ngfr and Gap43 transcript levels increased in Sox10, T1, or T2‐transfected cells, with T1 the most effective inducer (Figure 5E). These data are consistent with the induction of NGFR and GAP43 protein by all three expression vectors (Figure 4B). Notably, levels of Ngfr and Gap43 transcripts were higher in the transfected cells compared to nerve‐derived Schwann cells (Figure 5E). Additional Schwann cell genes induced by all three expression vectors included Tfap2a, Cdh19, and S100b, with T1 being the most efficient at inducing Tfap2a and Cdh19. Tfap2a was expressed at higher levels in reprogrammed Schwann cells compared to nerve‐derived Schwann cells, consistent with its upregulation during the repair process. ⁷³ Conversely, Cdh19 and S100b, which are also typically involved in repair, did not reach the levels seen in nerve‐derived Schwann cells (Figure 5E). In contrast, Mpz was only induced by Sox10 and T1 but not T2, Egr1 was only induced by Sox10, and Egr2 was not induced by any of the constructs. Notably, the levels of Mpz did not reach that of nerve‐derived Schwann cells. Some genes were instead repressed by the Schwann cell TFs, including Olig1, which was repressed by all three constructs even though Olig1 is expressed in non‐myelinating Schwann cells ⁷⁴ and upregulated post‐PNI ²⁰ (Figure 5E). Finally, as anticipated, a fibroblast‐specific marker, Col1a1, was repressed by Sox10, T1, and T2 to low baseline levels observed in purified nerve‐derived Schwann cells (Figure 5D). Taken together, these data suggest that all three expression vectors have the ability to induce Schwann cell‐specific gene expression while suppressing a fibroblast fate, but T1 is superior for the induction of several Schwann cell‐specific markers.

2.7. Reprogrammed Schwann cells express some repair genes and display some myelinating capacity

To better understand the phenotypes of MEFs transfected with Sox10, T1, and/or T2, we compared our gene expression data with published RNA‐seq and scRNA‐seq data sets. Among the genes turned on by all three constructs, S100b and Mpz were highly expressed in Schwann cells at 3‐, 5‐, and 7‐days post‐injury based on the bulk transcriptomic analysis. ⁶³ Surprisingly, Gap43, which was induced by Sox10, T1, and T2, and reportedly important in the Schwann cell repair process, ⁶⁸ , ⁶⁹ was not expressed in endogenous Schwann cells post‐injury (Figure 6A). Of the genes preferentially induced by T1, Cdh19 and Ngfr were also expressed in Schwann cells post‐injury, particularly on day 3, indicative of an early repair phenotype. Notably, Tfap2a was also selectively turned on by T1, and previously implicated in Schwann cell repair, ⁶⁸ yet Tfap2a transcripts were not detectable in endogenous Schwann cells at 3‐, 5‐, and 7‐days after injury (Figure 6A). Finally, Sox10 preferentially turned on Egr1, a gene that represses myelination and is expressed at low levels at 3, 5, and 7 days post‐injury (Figure 6A). Next, to better associate our findings with annotated Schwann cell subclusters, we generated a dotplot of the genes analyzed in our study. We found that Cdh19 and Ngfr were most enriched in repair Schwann cells, while Sox10 was most highly associated with mature myelinating and Remak cells (Figure 6B). In summary, T1 has an enhanced ability to induce the expression of many Schwann cell genes involved in early repair events, while Sox10 alone turns on Egr1, a myelin‐regulating gene (Figure 6A).

FIGURE 6.

Reprogrammed Schwann cells express some endogenous Schwann cell repair genes and show some myelinating capacity. (A) Expression of Col1a, Thy1, Olig1, S100b, Cdh19, Ngfr, Mbp, Mpz, Gap43, Pax3, Tfap2a, Egr1 and Egr2 at 3, 5 and 7‐days post‐PNI ⁶³ (from GSE177037). Arrows show genes that were induced in reprogrammed Schwann cells, either with all of T1, T2 and Sox10 (green arrows), or induced by higher levels by T1 (red arrows) or Sox10 (blue arrow). (B) Dotplot showing the expression levels of Col1a, Thy1, Olig1, S100b, Cdh19, Ngfr, Mbp, Mpz, Gap43, Pax3, Tfap2a, Egr1 and Egr2 in Schwann cell subtypes in naïve Schwann cells and at 3, 5 and 7‐days post‐injury ⁶³ (from GSE177037). Arrows show genes that were induced in reprogrammed Schwann cells, either with all of T1, T2 and Sox10 (green arrows), or induced by higher levels by T1 (red arrows) or Sox10 (blue arrow). (C, D) Schematic of DRG‐Schwann cell co‐culture assay. P2 wild‐type DRGs were co‐cultured with tdTomato‐labeled Schwann cells collected from P21 Sox10‐cre;Rosa‐tdTomato nerves (C). Immunostaining of DRG‐ Schwann cell co‐culture after 15 days with tdTomato (labeling Schwann cells in red) and TUJ1 (labeling DRG neurite outgrowth in green) and DAPI counterstain in blue (D). (Scale bars: 100 μm [top left images], the rest were digitally magnified at 75 and 50 μm).

Given that Sox10, T1, and T2 all induced Mpz expression, which encodes for a myelin‐associated protein, we assessed the functional myelinating capability of transfected cells using a DRG‐Schwann cell co‐culture assay. ³³ DRGs were obtained from 2‐day‐old mouse pups and NGF was added to promote neurite outgrowth, while cytosine arabinoside was added to remove Schwann cells and fibroblasts that contaminated the explant culture ⁷⁵ (Figure 6C). Next, nerve‐derived, tdTomato‐labeled Schwann cells were isolated from Sox10‐Cre; Rosa‐tdTomato mice and co‐cultured with the DRG explants. Over time, the Schwann cells migrated toward the DRG axons and began to wrap themselves around these axons (Figure 6C). Next, we used this assay to assess the myelinating capacity of T2‐transfected cells. DRG explants were co‐cultured with either negative‐ or T2‐transfected MEFs that were grown for 21 days in vitro. The DRG‐transfected cell co‐cultures were grown in myelin‐inducing media containing NRG1, forskolin, NGF, and ascorbic acid for 15 days (Figure 6C). MPZ immunolabeling was used to identify myelin, while DRG axons were immunolabeled with TUJ1. Strikingly, in DRGs co‐cultured with T2‐transfected MEFs, and not with control‐transfected MEFs, MPZ immunolabeling surrounded the DRG axons (Figure 6D). Thus, T2‐reprogrammed MEFs have some capacity to induce myelination of DRG axons.

3. DISCUSSION

In this study, we set out to develop a triple TF‐based lineage conversion strategy involving a non‐integrative, episomal approach to induce a repair Schwann cell‐like phenotype in MEFs. We found that misexpression of Sox10 alone, or two triple TF combinations; Jun‐Pax3‐Sox2 (T1) and Sox10‐Jun‐Sox2 (T2), could induce Schwann cell‐specific target gene expression above control levels in MEF‐derived NGFR⁺O4⁺ Schwann cell‐like cells. Comparatively, T2 generated the most NGFR⁺O4⁺ Schwann cell‐like cells, whereas T1 was generally more efficient at inducing repair genes, including Ngfr, Gap43, Tfap2a, and Cdh19. While we do not provide direct evidence that T1‐derived Schwann cell‐like cells have functional repair potential, we show that T2‐reprogrammed cells have some capacity to myelinate DRG axons in culture. Taken together, our data support the idea that reprogramming TFs act in a combinatorial fashion, and can have an additive effect to support Schwann cell lineage conversion. We suggest that the cooperative phenotypes are due to the ability of these TFs to co‐regulate gene regulatory networks. Indeed, there is evidence of cross‐talk between the TFs employed in this study. For instance, Pax3 can repress JUN protein levels, ¹⁹ while SOX10 and JUN can physically interact, and may influence each other's transcriptional activities. ⁷⁶ Future studies will be required to fully elucidate the regulatory interactions between the component TFs in T1 and T2.

Early Schwann cell conversion studies employed small molecules alone to direct the conversion of fibroblasts toward a transient neural precursor state, and their subsequent differentiation into myelinating Schwann cells. ⁷⁷ The next advance was to utilize glial lineage TFs to enhance lineage conversion efficiency. The first TF‐based Schwann cell reprogramming study combined Sox10 overexpression with a cocktail of small molecules (valproic acid, 5‐azacytidine, CHIR99021) to drive the conversion of human fibroblasts to an NCC identity. ⁴⁹ Induced NCCs were then differentiated into myelinating Schwann cells by supplementing appropriate glial supportive cues. ⁴⁹ The next innovation involved the combination of Sox10 with Egr2, which bypassed the intermediary NCC step and promoted direct conversion of fibroblasts to myelinating Schwann cells with repair potential. ⁵⁴ , ⁶² The induced Schwann cells generated in response to Sox10 and Egr2 misexpression had a mixed phenotype but tended toward an immature to mature Schwann cell state as identified by S100β and O4 expression. ⁵⁴ , ⁶² Notably, Egr2 is normally downregulated during peripheral nerve repair, resulting in a transient loss of the myelinating capacity of Schwann cells, which is required for the repair phenotype. ⁷⁸ We thus reasoned that other TFs might be better suited to the generation of this initial repair state of Schwann cells.

Building on these studies, we designed an efficient lineage conversion strategy that involved the misexpression of three lineage‐specifying TFs together, focusing on TFs that are upregulated during the Schwann cell injury response. Notably, a similar episomal vector approach was used to transform human fibroblasts into induced SCPs by misexpressing 5 genes (OCT4/POU5F1, SOX2, KLF4, MYCL1, LIN28) and silencing one other (p53/TRP53 shRNA), also supplementing cultures with NRG1 and small molecules (CT‐99021, RG108, 5′‐[N‐Ethylcarboxamido] adenosine). ⁵⁸ However, we reasoned that fewer TFs might be similarly efficient and reduce the inherent risks associated with the ectopic expression of individual genes. Based on our data, Jun‐Pax3‐Sox2 (T1) is the most efficient at inducing high levels of repair gene expression, including Ngfr, Gap43, Tfap2a, and Cdh19, whereas Sox10‐Jun‐Sox2 (T2) induced the production of more O4⁺ Schwann cells. In the future, sc and bulk RNA sequencing studies will greatly aid in assessing the repair phenotype of the T1‐ and T2‐Schwann cell‐like cells. Finally, while we do not provide direct evidence that T1‐derived Schwann cell‐like cells have functional repair potential, we show that T2‐reprogrammed MEFs have some capability to myelinate axons in a DRG co‐culture. Further in vivo investigations are thus warranted to compare the repair potential of the T1‐ and T2‐Schwann cell‐like cells in vivo.

In our study, we used episomal plasmids to express Schwann cell determinants in MEFs, while also exposing the transfected cells to glial induction cues (i.e., neuregulin, forskolin). Episomal plasmids have a safety advantage as they avoid the risk of mutations that can occur with lentiviral integration. With this strategy, we could induce ectopic Schwann cell‐marker expression within 21 days, and these markers could possibly turn on even earlier, which we did not assess. Using a small molecule approach, the conversion of fibroblasts to an immature Schwann cell fate with small molecules typically takes ~39 days, ⁷⁷ almost double the amount of time of our strategy. A more recent study has further reduced this timeline to 10 days, albeit with a higher rate of heterogeneity among the cell products. ⁵⁴ , ⁵⁵ Reprogramming fibroblasts using shRNA and small molecules also exhibits a long timeframe for conversion, as cells go through an SCP intermediate phase after 18 days and then are further differentiated into mature Schwann cells for an additional 7–17 days. ⁵⁸ Furthermore, lentiviral expression of Sox10 alone converts MEFs to an NCC state within 21 days, but a further 14‐day exposure to glial determinants is required to generate immature Schwann cells. Of note, we shortened this time frame by adding the glial determinants 1 day after TTFEV transfection. Finally, the combined expression of Sox10 and Egr2 using a viral approach can induce ectopic Schwann cell‐marker expression within 10 days, ⁵⁴ and after 21 days, ⁶² these reprogrammed cells can trigger repair in a model of peripheral nerve demyelination. A combinatorial TF approach, combined with small molecules, is an efficient way to generate a pool of myelinating Schwann cells for repair.

Unexpectedly, Jun transcripts were barely elevated when T1‐ or T2‐ were overexpressed in MEFs, even though the Jun coding sequence was included in both of these expression vectors. This finding was not expected since Jun is highly expressed during Schwann cell repair, as observed in the bulk transcriptomic and sc transcriptomic data we mined in this study. A likely reason is that Jun is already highly expressed in MEFs, and detecting an increase in gene expression against a high baseline is challenging with qPCR due to the technical limitations of this approach. Indeed, we used the delta–delta Ct method to calculate relative expression levels for each gene, and normalized to transcript levels in un‐transfected MEFs, which express high levels of Jun. Additionally, qPCR has a limited dynamic range, meaning it might not be able to accurately detect very high levels of gene expression, as the amplification reaction can saturate. The sensitivity of qPCR is thus lower than RNA‐seq, especially for genes with very high expression levels. More complex interactions between the TFs in T1 and T2 may influence Jun expression, which remains to be determined.

In summary, our triple TF‐based lineage conversion strategy offers a new molecular toolkit for the generation of Schwann cells with a repair‐like phenotype that may be clinically relevant. The potential to transplant Schwann cells with an enhanced repair capacity could pave the way for novel lines of treatment, such as a cell‐based therapy. Indeed, Schwann cell transplants at the site of injury have been shown to improve functional recovery in animal models. ²⁸ To facilitate repair, nerve‐like conduits containing Schwann cells are transplanted; ⁷⁹ however, this approach requires nerve biopsies to isolate Schwann cells. ⁸⁰ Reprogrammed Schwann cell‐like cells may fill this demand. Additional in vivo functional assays will provide insight into whether endogenous signals induced by PNI are sufficient to switch our reprogrammed repair Schwann cells into a mature Schwann cell when axon regeneration is complete. If the switch from a non‐myelinating repair state to a myelinating Schwann cell phenotype is not activated by endogenous signals, additional steps may be required to support both phases of repair.

4. EXPERIMENTAL PROCEDURES

4.1. Animals and genotyping

All animal studies were approved by Sunnybrook Research Institute (protocol #AUP‐16‐606) and are in compliance with the Guidelines of the Canadian Council of Animal Care. Animals used in this study were CD1 mice (Charles River Laboratories, Senneville, QC). We also used two transgenic mouse lines: Sox10‐Cre (B6;CBA‐Tg[Sox10‐cre]1Wdr/J; #025807) and Rosa‐tdTomato (B6.Cg‐Gt(ROSA)26Sortm14(CAG‐tdTomato)Hze/J; #007914) (Jackson Laboratory, ME, United States). Sox10‐Cre x Rosa‐tdTomato intercrosses were used to generate Sox10‐Cre; Rosa‐tdTomato progeny for experimental purposes. PCR genotyping was performed with the following primers: Sox10‐Cre; mutant forward primer: ACC AGG TTC GTT CAC TCA TGG; mutant reverse primer: AGG CTA AGT GCC TTC TCT ACA C. Rosa‐tdTomato; mutant forward primer: CTG TTC CTG TAC GGC ATG G; mutant reverse primer: GGC ATT AAA GCA GCG TAT CC. For embryonic staging, the morning of the vaginal plug was considered embryonic day (E) 0.5. Animal procedures were approved by the Sunnybrook Research Institute Animal Care Committee (protocol #AUP‐25‐769) in compliance with the Guidelines of the Canadian Council of Animal Care.

4.2. Vector construction

Triple TTFEVs, including triple 1 – T1, and triple 2‐ T2, were generated by cloning TF coding domains (Sox2, Pax3, Sox10, Jun) into a pCAGGS expression vector (Lablife) modified by the addition of an extended multiple cloning site fragment (3′ XbaI site at nucleotide 677 to 5′ XhoI site at nucleotide 740) from pBluescript KS+ (Addgene) to create pCAGGS2. In T1 and T2, TFs at the first, second, and third positions were Myc‐, HA‐, and FLAG‐tagged, respectively. A viral 2A (v2A) self‐cleaving peptide sequence was inserted between TF sequences to allow for stoichiometric expression of all three genes. ⁶⁷ Final products were pCAGGS2‐Jun‐Myc‐v2a‐Pax3‐HA‐v2a‐Sox2‐FLAG and pCAGGS2‐Sox10‐Myc‐v2a‐Jun‐HA‐v2a‐Sox2‐FLAG. pCIG2‐Sox10‐IRES‐GFP and pCIC:Sox10‐IRES‐mCherry were generated by cloning Sox10 coding domain into pCIG2 and pCIC expression vectors. ⁸¹

4.3. MEF isolation and transfection

MEFs were isolated from E12.5‐13.5 CD1 embryos by first decapitating and removing internal organs and the vertebral column. The remaining tissue was finely minced and dissociated in 2% Collagenase Type 4 (Worthington #LS004189) for 20 min. MEFs were cultured in DMEM (Wisent #319‐005‐CL) containing 10% fetal bovine serum (FBS; Wisent #080‐450) and penicillin/streptomycin (Wisent #450‐201‐EL) and maintained for a maximum of four cell passages. MEFs were transfected using the 4D‐Nucleofector™ (Lonza). In brief, 2 × 106 cells were mixed in 20 μL P3 reagent and transfected with 13 μg plasmid DNA using the CZ‐167 program on the nucleofector. Transfected cells were plated on PDL:Lam pre‐coated dishes and cultured in MEF media for 24 hr. before the addition of glial media in DMEM: Ham's F‐12 Nutrient Mix, supplemented with GlutaMAX (F12; Invitrogen #31765‐035) (3:1), FGF2 (10 ng/mL, Miltenyi Biotech GMBH #130‐093‐842), forskolin (4 μM, Sigma #F8668), NRG1 (100 ng/mL, R&D Biosystems #377‐HB), N2 supplement (1%, Wisent 305‐016‐IL), 1X L‐glutamine (Invitrogen), and penicillin/streptomycin which was changed every 3 days.

4.4. Isolation and culture of nerve‐derived Schwann cells

Mouse nerve‐derived Schwann cells were isolated from adult CD1 mice. Nerves were sliced into smaller fragments with the epineurium removed, and explants were cultured in DMEM (Wisent, 19‐005) with 10% FBS and 1% penicillin/streptomycin to enhance Schwann cell expansion and increase purity by removing excess fibroblasts. After 2 weeks, explants were dissociated with an enzyme cocktail consisting of DMEM, 10% FBS, 25% Dispase 5 U/mL (Stem Cell Technologies, 07913), 0.125% Collagenase Type 4, and 1% penicillin/streptomycin overnight (<16 h) at 37°C in 5% CO₂. Schwann cells were cultured on dishes coated with Poly‐D‐Lysine (PDL; 20 μg/mL; VWR‐BD #354210) and Laminin (Lam; 40 μg/mL; VWR‐BD #354232), with Schwann cell media consisting of DMEM, 10% FBS, 1% Glutamax (Gibco, 35050061), 1% penicillin/streptomycin, 50 ng/mL neuregulin (Peprotech, 100‐03‐50), and 5 μM forskolin (Sigma, f6886) replenished every 3 days. ⁸²

4.5. Fluorescence‐activated cell sorting (FACS)

To isolate NGFR‐negative MEFs, 4 days after the establishment of primary MEF cultures, dissociated cells were incubated with CD271‐VioBright FITC antibody (Miltenyi Biotech; ME20.4‐1.H4; 1:500) for 30mins. Stained cells were sorted on BD FACS ARIA III, and CD271⁻ cells were isolated and cultured in MEF media. To isolate O4⁺ reprogrammed cells, dissociated cells were incubated with O4‐APC (Miltenyi Biotech; 130‐119‐155; 1:500) antibodies. To isolate NGFR⁺ O4⁺ reprogrammed cells, dissociated cells were incubated with CD271‐VioBright FITC (Miltenyi Biotech; 130‐110‐115) and O4‐APC (Miltenyi Biotech, 130‐117‐357) antibodies. Stained cells were sorted on BD FACS ARIA III, and O4⁺ or CD271⁺ O4⁺ cells were collected and processed for RNA isolation.

4.6. Magnetic‐activated cell separation (MACS)

To obtain pure populations, cells were sorted using the cell surface marker p75^NGFR sorted using MACS as per manufacturer's protocol (Miltenyi Biotech). Schwann cells and MEFs were sorted for NGFR⁺ or against NGFR⁻, respectively. In short, cells were expanded from the initial plating before sorting. MEFs were passaged using 25 μg/mL trypsin (Sigma #T1005‐1G) for 5 min, and Schwann cells using TryplE (12605010, Gibco) for 2 min at 37°C, before being passed through a 40um filter. Resuspension was followed in cold MACS buffer comprised of 2 mM EDTA +0.5% bovine serum albumin in Hanks' Balanced Salt Solution (311‐512 CL, Wisent). Cells were incubated with CD271‐VioBright FITC (Miltenyi Biotech; 130‐110‐115) for 1 h on ice and then subsequently incubated with anti‐FITC microbeads (Miltenyi, 130‐048‐701) for 20 min at 4°C. After MS columns (Miltenyi Biotech, 130‐042‐201) were activated with cold MACS buffer, the cell suspensions were filtered through to collect either the negative NGFR MEFs or positive NGFR Schwann cells.

4.7. Western blotting

NIH/3 T3 cells (ATCC® CRL‐1658™) were cultured in Dulbecco's modified Eagle medium (DMEM; Wisent #319‐005‐CL) containing 10% FBS (Wisent #080‐450) and penicillin/streptomycin (Wisent #450‐201‐EL). NIH/3 T3 cells were transfected using P3 Primary Cell 4D‐Nucleofector X Kit (Lonza V4XP‐3032) as per the manufacturer's instructions. Whole cell pellets were collected post‐transfection and lysed in RIPA buffer (150 mM NaCl, 0.1% SDS, 1% NP‐40, 50 mM Tris–HCl, pH 8.0, 0.5% sodium deoxycholate) with protease (1X protease inhibitor complete, 1 mM phenylmethylsulfonylfluoride and phosphatase (50 mM NaF, 1 mM NaOV)) inhibitors. 10 μg of lysate was run on 10% SDS‐PAGE gels for Western blot analysis. Primary antibodies included: JUN (Abcam; #ab31419, 1:1000), PAX3 (DSHB; #p23760, 1:1000), SOX2 (Cell Signaling; #C70B1, 1:1000), SOX10 (Santa Cruz Biotech; #17343, 1:1000), and GAPDH (Cell Signaling #2118, 1:10000). Western blots were developed with an ECL Blotting Substrate (Bio‐Rad, Canada) according to the manufacturer's protocol.

4.8. Quantitative reverse transcription polymerase chain reaction (RTqPCR)

RNA extraction was performed using the RNeasy Mini and Micro kit (Qiagen) as per the manufacturer's instructions. 1 μg RNA was reverse transcribed using the RT2 First Strand synthesis kit (Qiagen). For MEFs vs. nerve‐derived Schwann cell, cDNA was made with SuperScript™ IV Reverse Transcriptase (Invitrogen, 18090010) using 16 ng of RNA. Reverse‐transcribed cDNA was amplified using the RT2 primer assay kit and Sybr Green (Qiagen). Qiagen primers used for gene expression analysis were: Sox10 (PPM05134D), Sox2 (PPM04762E), Pax3 (PPM24658A), Jun (PPM03037A), Ngfr (PPM04327F), Gap43 (PPM03303A), Tfap2a (PPM30132A), Cdh19 (PPM40369A), S100β (PPM04492A), Mpz (PPM41824A), Egr1 (PPM02938C), Egr2 (PPM04478F), Col1a1 (PPM03845F), Olig1 (PPM04503A) and Thyl (PPM31015A). Gapdh (PPM02946E), B2m (PPM03562A), and Hrpt (PPM03559F) were used as reference genes. The delta–delta Ct method was used to calculate relative expression levels for each gene, using Gapdh, B2m and Hrpt to normalize qPCR data from transfected MEFs (Figures 2 and 3) and Gapdh and B2m to normalize qPCR data from nerve‐derived Schwann cells vs. MEFs (Figure 5). Gene expression levels post‐transfection were normalized to levels in non‐transfected MEFs. A minimum of 3 experimental and 3 technical replicates was performed for each experiment. Statistical comparisons were made using an ANOVA with a post‐hoc Tukey's test using Prism software (GraphPad v6.0). Data are represented as mean ± standard error of the mean (s.e.m.).

4.9. Immunostaining

Transfected cells were plated on 6‐well chamber slides, and DRG‐Schwann cell culture explants were washed with phosphate‐buffered saline (PBS) and fixed in 4% paraformaldehyde (PFA)/PBS. Samples were washed and permeabilized in PBS /0.1% TritonX‐100 (PBT) followed by blocking with 10% normal horse serum/PBT (blocking solution). Primary antibodies were diluted in blocking solution and incubated on sections and kept overnight at 4°C. Primary antibodies included: NGF Receptor p75 (NGFR, Millipore #AB1554; 1:500) and GAP43 (Novus Biologicals; NB300‐143, 1:300), tdTomato (Abcam; #ab62341; 1:500), TUJ1 (BioLegend #801202; 1:200) and MPZ (Abcam #AB_881430, 1:200). Cells were washed with PBT and secondary antibody diluted in PBT was applied. Cells were counterstained with 4′,6‐diamidino‐2‐phenylindole (DAPI; Santa Cruz Biotechnology) diluted in PBT and mounted with coverslips using AquaPolymount (Polysciences). Species‐specific secondary antibodies were conjugated to Alexa 568 or Alexa 488 (Invitrogen).

4.10. Dorsal root ganglia‐Schwann cell co‐culture assay

The in vitro myelination assay was performed as previously described. ³³ Briefly, DRGs were isolated from P2 CD1 pups and grown as explants on 8‐well chamber slides coated with Matrigel (1:20; BD Biosciences), PDL, Lam, and Collagen IV (1:20; BD Biosciences). Explants were grown in DRG media: DMEM:F12 (3:1) with Nrg1 (50 ng/mL; R&D Biosystems), forskolin (5 μM; Sigma Aldridge), and Nerve Growth Factor (NGF; 50 ng/mL; R&D Biosystems). After 2 days in DRG media, cytosine arabinoside (CyaA; Sigma Aldrich) was added for 4 days (~2–3 mitotic cycles) to kill surviving Schwann cells and fibroblasts (i.e., dividing cells). Fresh DRG media was then added and refreshed every 2 days. One week post‐dissection, DRG explants were co‐cultured with 20,000–25,000 Schwann cells in myelination media: DMEM:F12 with Nrg1 (5 ng/mL), Forskolin (0.5 μM), NGF (5 ng/mL) and Ascorbic Acid (50 μg/mL; R&D Biosystems). Schwann cells were either obtained from wild‐type CD1 pups or from Sox10‐Cre; Rosa‐tdTomato mice.

4.11. Imaging and figure preparation

Images were captured with a Leica DMIL LED or DMRXA2 optical microscope or a Leica DMi8 inverted microscope (Leica Microsystems). Figures were created using Adobe Photoshop 2021. A license from BioRender was used to prepare schematics.

FUNDING INFORMATION

Canadian Institutes of Health Research (CIHR) Project Grant (PJT‐ 156366) Alberta Innovates – Health Solutions (AI‐HS) Collaborative Research and Innovation Opportunities (CRIO) Project grant (#201200859).

CONFLICT OF INTEREST STATEMENT

The authors declare no competing interests.

Belfiore L, Balakrishnan A, Touahri Y, et al. Combinatorial expression of glial transcription factors induces Schwann cell‐specific gene expression in mouse embryonic fibroblasts. Developmental Dynamics. 2026;255(6):548‐566. doi: 10.1002/dvdy.70054

DATA AVAILABILITY STATEMENT

Bulk RNA‐sequencing data from Figure 1B was mined from ⁶³ (Accession: GSE177037). sc RNA‐sequencing data in Figure 1C,D was mined from ⁶⁴ (Accession: GSE216665). Western blot data is deposited in Mendeley Data, V1, DOI:10.17632/wyk539cj8z.1 (Combinatorial expression of glial TFs induces Schwann cell‐specific gene expression in MEFs ‐ Mendeley Data).