Matrix stiffness mechanosensing modulates the expression and distribution of transcription factors in Schwann cells

Abstract After peripheral nerve injury, mature Schwann cells (SCs) de‐differentiate and undergo cell reprogramming to convert into a specialized cell repair phenotype that promotes nerve regeneration. Reprogramming of SCs into the repair phenotype is tightly controlled at the genome level and includes downregulation of pro‐myelinating genes and activation of nerve repair‐associated genes. Nerve injuries induce not only biochemical but also mechanical changes in the tissue architecture which impact SCs. Recently, we showed that SCs mechanically sense the stiffness of the extracellular matrix and that SC mechanosensitivity modulates their morphology and migratory behavior. Here, we explore the expression levels of key transcription factors and myelin‐associated genes in SCs, and the outgrowth of primary dorsal root ganglion (DRG) neurites, in response to changes in the stiffness of generated matrices. The selected stiffness range matches the physiological conditions of both utilized cell types as determined in our previous investigations. We find that stiffer matrices induce upregulation of the expression of transcription factors Sox2, Oct6, and Krox20, and concomitantly reduce the expression of the repair‐associated transcription factor c‐Jun, suggesting a link between SC substrate mechanosensing and gene expression regulation. Likewise, DRG neurite outgrowth correlates with substrate stiffness. The remarkable intrinsic physiological plasticity of SCs, and the mechanosensitivity of SCs and neurites, may be exploited in the design of bioengineered scaffolds that promote nerve regeneration upon injury.


| INTRODUCTION
Schwann cells (SCs) are of fundamental physiological importance for the peripheral nervous system (PNS). Their remarkable biological plasticity enables them to rapidly adapt to tissues beyond the PNS microenvironment, as they have the capacity to form myelin around axons after transplantation into the central nervous system (CNS) spinal cord. 1 These striking properties make SCs a powerful candidate for cell-based PNS and CNS regenerative therapies. 2 SC plasticity is reflected by the fact that they undergo significant morphological transformations during PNS development and repair among others. 3,4 Apart from their canonical role of myelin formation around peripheral axons, SCs clear debris from lesion sites after nerve damage and promote nerve regeneration via diverse interactions with their microenvironment. 5,6 It is well documented that nerve regeneration is associated with changes in the biochemical properties of the involved structures including SCs. Compelling evidence shows that the nerve lesion site also undergoes significant changes in the biomechanical properties of acellular and cellular constituents that are crucial for nerve regeneration, but the underlying mechanisms remain unclear.
Upon nerve injury, mature SCs dedifferentiate and reprogram into a repair SC phenotype. Repair SCs are characterized by downregulation of the expression of pro-myelinating genes, such as Krox20, and the upregulation of the transcription factor c-Jun, a key regenerative marker that promotes nerve regeneration. 7 Later on, SCs transform into a pro-myelinating phenotype to re-myelinate regenerated axons. During the differentiation and dedifferentiation phases, the expression profile of SC-specific markers changes. 5,8 We have recently shown that mechanosensing of the extracellular matrix (ECM) impacts physiological processes including embryonic outgrowth of neurites from dorsal root ganglions (DRGs), as well as the spreading area, migration velocity, and biomechanical properties of SCs. 9 It remained unexplored, however, whether ECM stiffness mechanosensing affects SC differentiation and the regulation of specific phenotypes. In the present work, we generated elastic substrates that cover the physiological range of nerve tissue stiffness to investigate the modulation of SC stage-specific markers, including c-Jun, Krox20, Oct6, and Sox2.
Here, we show that soft matrices upregulate the expression of the repair-associated transcription factor c-Jun, whereas stiff substrates upregulate the expression of transcription factors Sox2, Oct6, and Krox20, suggesting a link between substrate mechanosensing and regulation of gene expression in SCs. Beyond refining our neurophysiological knowledge of the PNS, our findings may be exploited to advance clinical nerve regeneration. They may also be of some relevance for the priming of SC phenotypes/functions for possible future application in the PNS and potentially in the CNS.
2 | RESULTS AND DISCUSSION 2.1 | YAP nucleo-cytoplasmic shuttling in SCs is regulated by ECM stiffness Our recent study demonstrates that SCs are highly mechanosensitive, and that ECM stiffness directs different physiological responses in SCs, which include morphological changes, cell-ECM adhesion, motility, and cell mechanics. 9 It remains an open question, however, whether the mechanosensitivity of SCs also modulates their biochemistry and differentiation stages. To address this question, we generated ECMcoated polyacrylamide (PAAm) substrates with constant biochemical composition but varied biomechanical properties (Young's modulus) (see Figure S1, supporting information). Substrate stiffness was appropriately selected to cover the physiological stiffness range of the native microenvironment of SCs and axons. 10,11 We coated the generated elastic substrates with the ECM protein laminin owing to its importance in the peripheral nerve microenvironment, and its essential role in mechanosensing and mechanotransduction. 12 Mechanosensing of the ECM stiffness is a key player in mechanotransduction pathways, including signaling via the serum response factor, NF-kB, zyxin/paxillin, integrin, E-cadherin, Wnt, and the Hippo-signaling pathway among others. [13][14][15] In the present work, we analyzed in nerve-derived isolated SCs the modulation of the transcriptional regulator Yap, F I G U R E 1 Nucleocytoplasmic YAP localization in SCs is modulated by substrate stiffness. Labeling of YAP (green) on SCs seeded on a compliant (a) and stiff (b) laminin-coated PAAm substrate. (c) Quantification of nuclear/ cytoplasmic ratio. (d) Projected SC area. Cells were stained with DAPI (white) and rhodaminephalloidin (red) for nucleus and Factin labels, respectively. *P <.05, Mann-Whitney test. n = 50 cells for each substrate. Abbreviation: SC, Schwann cell a mechano-transducer of the Hippo pathway, in response to changes in ECM stiffness. Figure 1 shows representative confocal images of SCs, seeded on laminin-coated substrates. YAP (green), nuclei (white), and cytoskeletal F-actin are visualized following staining with anti-Yap antibody, DAPI, and rhodamine-phalloidin, respectively. We found that in SCs exposed to compliant environments (n = 50) of 1.1 kPa, YAP is localized mainly in the cytoplasm, whereas it shuttles to the nucleus when cells are exposed to stiffer (27.7 kPa) substrates (n = 50) [ Figure 1(a-c)].
The accumulation of YAP in the nucleus is accompanied by a significant increase in SC spreading area [ Figure 1(d)]. These results support previous data showing that SCs sense and respond to changes in matrix rigidity by modulating the nucleocytoplasmic transport of YAP/TAZ. 16 In previous studies, we measured the stiffness changes in the nerve microenvironment during PNS development and disease, using atomic force microscopy. 30,31 We demonstrated that the basal lamina, a special type of ECM, conveys crucial biomechanical resilience to myelinated axons, 10 and that it is a major contributor to the overall stiffness of the nerve tissue microenvironment during development and maturation. 11 SCs interact not only physiologically but also biomechanically with the basal lamina. This interaction significantly impacts the behavior of SCs, which includes their interaction with axons, and eventually myelination. 17,18 Hence, ECM stiffness and the mechanosensitivity of SCs are of particular importance for the physiological functions of SCs. Next, we set out to investigate the influence of matrix stiffness on the regulation of protein expression profiles in SCs.

| Matrix stiffness modulates the expression of pro-myelin transcription factors
Recent in vivo studies have shown that transcriptional regulators YAP/TAZ are involved in the upstream regulation of myelin genes in F I G U R E 2 Expression and localization of pro-myelinating transcription factors in SCs is modulated by substrate stiffness. Representative confocal images showing labeling of Krox20 and Oct6 on SCs seeded on compliant (a and e) and stiff (b and f) substrates. Quantification of fluorescence signal of nuclear Krox20 (c) and Oct6 (g). Western blot showing expression of levels of Krox20 (d) and Oct6 (h). Actin cytoskeleton and nucleus labeled in red and white, respectively. *P < .05 and ***P < .0001, Mann-Whitney test. Krox20: n = 349 and n = 235 cells for compliant and stiff substrates, respectively. Oct6: n = 156 and n = 183 cells for compliant and stiff substrates, respectively. Abbreviation: SC, Schwann cell mouse peripheral nerves. 19 The activation of YAP/TAZ signaling pathway in SCs is regulated via integrin-mediated signaling and G-protein mechanism. Once activated, YAP and TAZ bind to DNA binding partners such as TEAD1-4, which modulate the proliferation and expression of myelin genes along with ERG2 and Sox10. 16,20 It remains unclear, however, whether physical cues affect SCs differentiation.
How physical cues modulate different SC phenotypes has important implications for nerve regeneration strategies, based on bioengineered nerve scaffolds containing cellular components, which mostly include SCs. We studied in SCs the expression levels of the transcription factor Krox20 (also known as Egr2), a master regulator of the onset of myelination in the PNS. 21  The transcription factor Oct6 (also known as SCIP/Tst1) is an upstream regulator of Krox20 in SCs and is considered necessary for the transition from nonmyelinating to the myelinating stage in peripheral nerves. 22 We studied the expression levels of  (d) Quantification of neurite length. Actin cytoskeleton and nucleus labeled in red and blue, respectively. **P < .001 and ***P < .0001, Mann-Whitney test. n = 31 cells for compliant and stiff substrates, respectively. Abbreviation: DRG, dorsal root ganglion stem-cell fate. During PNS development Sox2 is expressed in progenitors and immature SCs, and it is also reexpressed in pro-regenerative SCs after nerve injury. 25 We found the levels of Sox2 in nuclei of SCs

| Impact of matrix stiffness on the morphology and outgrowth of adult DRG neurites
Another crucial aspect of biomedical nerve regeneration using bioengineered scaffolds is the enhancement of neuronal outgrowth.
Using 2D elastic substrates of PNS stiffness, we have recently shown in embryonic DRG organotypic explants that neurite elongation is promoted on stiffer substrates compared to compliant substrates. 9  properties may be appropriately adjusted to drive SCs to the desired phenotype, and eventually promote nervous system regeneration.

| MATERIALS AND METHODS
All chemicals were obtained from Sigma-Aldrich unless otherwise stated.

| Preparation of PAAm substrates
PAAm gels within the physiological nerve stiffness range were produced as previously described protocols. 9,27 Compliant (1.1 kPa) and stiff (27.7 kPa) gels with a thickness of 150 μm were incubated with poly-Dlysine overnight at 4 C followed by 2 hours incubation with 10 μg/mL laminin. Substrates were covered with cell culture medium and allowed to equilibrate at 37 C for 1 hour before seeding of cells. The laminin coating homogeneity on the PAA substrates and the consistency of coating between the two stiffnesses were tested by immunochemistry and confocal microscopy (see supporting information Figure S1).

| DRG culture
Adult (4-6 months old) C57BL/6J mice were killed by cervical dislocation and spinal cords were removed. DRGs were dissected and incubated in Neurobasal medium (NB, Invitrogen) containing 2.5 mg/mL collagenase and incubated for 1 hour at 37 C in 5% CO 2 . Then, the tissue was homogenized using fired-polished glass pipettes and DRG neurons were separated from axon stumps and myelin debris by generating a 14% bovine serum albumin layer and centrifuged for 8 minutes at 120 rpm. The pellet with DRG neurons was resuspended in NB containing 20 μL/mL B27 supplement 50Â (Gibco), 2 mM Glutamax, 10 μm/mL antibiotic-antimycotic, 0.01 μg/mL nerve growth factor (Invitrogen) and seeded on the PAAm substrates. DRG neurons were kept in an incubator at 37 C in 5% CO 2 .

| Immunofluorescence and image analysis
Either SCs or DRG neurons were seeded on laminin-coated PAAm substrates and maintained for 48 hours before fixation for 20 minutes with 4% PFA and processed for immunocytochemistry.  Table 2 in supporting information for analysis of protein bands.

| Statistical analysis
Data were exported to Origin Pro 9 software. The results are consid-

CONFLICT OF INTEREST
The authors declare no conflict of interest.

SUPPORTING INFORMATION
Additional supporting information may be found in the online version of the article at the publisher's website.