The transcription factor c-Jun does not appear to be essential during SC development but multiple arguments demonstrate its central function in SC reprogramming. Firstly, the time course of c-Jun expression supports its role in SC plasticity. Indeed, c-Jun is down-regulated post-natally during SC differentiation and myelination and is highly up-regulated under pathological conditions such as peripheral nerve injuries, demyelinating diseases or other peripheral neuropathies (Stewart, 1995; Parkinson et al., 2008; Hutton et al., 2011; Hantke et al., 2014; Klein et al., 2014). Secondly, c-Jun is a negative regulator of myelination and a cross-antagonist of Krox-20, one of the main pro-myelinating transcription factors. Deletion of c-Jun in SCs in vitro strikingly facilitates Krox-20-induced expression of MBP or MPZ. Conversely, forced expression of c-Jun inhibits the expression of myelin proteins. Moreover, the expression of c-Jun and Krox-20 is mutually exclusive: while immature SCs express relatively high levels of c-Jun and low levels of Krox-20, myelinating cells switch to high levels of Krox-20 and low levels of c-Jun (Parkinson et al., 2004, 2008). Thirdly, strong evidence in vivo exhibits the involvement of c-Jun in major aspects of the SC response following nerve injury. Myelination during development occurs normally in the c-Jun conditional knockout mice (Parkinson et al., 2008). However, the absence of c-Jun following a nerve lesion results in a failure of generating the repair SCs, leading to impaired axonal regeneration, target reinnervation and functional recovery. The myelin sheath degradation and clearance are also strikingly slowed in the absence of c-Jun in vitro and in vivo. Moreover, mutant SCs fail to form the Büngner regenerative tracks and to express the adhesion molecules and trophic factors critical for nerve repair (Parkinson et al., 2008; Arthur-Farraj et al., 2012; Fontana et al., 2012).

The role of c-Jun in SCs after nerve injury is to activate a repair program to support regeneration. The high levels of c-Jun found in different peripheral neuropathies that do not involve nerve injuries like Charcot-Marie Tooth (CMT) raises the question on the precise function of c-Jun in these diseases. Does c-Jun counteract the pathology, attenuate its symptoms and promote repair or does its up-regulation cause the disease? In two models of demyelinating diseases, CMT1A and CMT1X, c-Jun is increased in SCs that retain myelin differentiation (Hantke et al., 2014; Klein et al., 2014). The myelin sheaths are also relatively normal in a mouse mutant showing a 5–8 fold c-Jun overexpression in SCs (levels that are lower than those seen after a nerve cut though) (Jessen and Mirsky, 2016). Moreover, c-Jun and repair-associated genes are up-regulated in a mouse model in which the serine/threonine-protein kinase (LKB1, also called STK11) is inactivated in SCs and in which axons are damaged without overt demyelination (Beirowski et al., 2014). This suggests that moderated c-Jun up-regulation is compatible with myelination, does not necessarily provoke demyelination and that it is sufficient to activate the transcription of injury-related genes (Jessen and Mirsky, 2016). In this context, it is likely that the role of c-Jun in those diseases is to promote neuroprotection and repair. However, if c-Jun is activated in an uncontrolled and exacerbated manner, c-Jun could also be detrimental. For example, the increase of c-Jun expression in some human schwannomas may directly or indirectly be associated with the abnormal proliferation and the absence of myelin markers (Shivane et al., 2013).

Neurogenic locus notch homolog protein (Notch) is a transmembrane receptor protein that, upon ligand binding, is cleaved and generates an intracellular domain (NICD), which acts as a transcriptional regulator. Notch controls SC proliferation and promotes the generation of immature SCs from Schwann cell precursors (SCPs) in vivo but also acts as a negative regulator of myelination. It is down-regulated as myelination proceeds and its inactivation or over-activation leads to premature or delayed myelin formation, respectively. Moreover, the inhibition of Notch signaling in adult mice decelerates myelin breakdown following a nerve lesion (Woodhoo et al., 2009). The addition of recombinant jagged1, an activator of Notch, in rat nerves after injury improves nerve regeneration and functional recovery (Wang et al., 2015). This indicates that stimulation of Notch signaling in SCs could represent an interesting therapeutic strategy to efficiently promote nerve repair. However, aberrant activation of Notch signaling in SCs may also participate to the development of malignant peripheral nerve sheath tumors (Li et al., 2004).

More recently, the function of a transcriptional repressor, Zeb2, has been investigated in SC myelination and nerve repair by two different teams (Quintes et al., 2016; Wu L. M. N. et al., 2016). They both showed that mice lacking Zeb2 in SCs develop a severe peripheral neuropathy, caused by a failure of SC maturation. Zeb2 transcriptional repression of negative regulators of myelination such as Notch and SRY-related HMG-box gene 2 (Sox2) is necessary for SC lineage progression. Although Zeb2 seems dispensable for the establishment of the repair SCs following nerve injury, its absence disrupts nerve regeneration and remyelination (Quintes et al., 2016). Mechanistically, Zeb2 controls SC differentiation and remyelination by recruiting histone deacetylases 1 and 2 (HDAC1/2) and NuRD complexes and by inhibiting a Notch–Hey2 signaling axis (Wu L. M. N. et al., 2016).

The nuclear factor κB (NF-κB) transcription factor regulates a lot of physiological processes and mediates the physiopathological inflammatory response in many diseases. Several studies suggest that NF-κB activation is essential for the differentiation and myelination of SCs in vitro (Nickols et al., 2003; Yoon et al., 2008; Limpert and Carter, 2010). Furthermore, it has been shown that the deacetylation of NF-κB mediated by HDAC 1/2 is necessary to control myelination in the PNS (Chen et al., 2011). Paradoxically, by using transgenic mice in which NF-κB activation is inhibited in SCs, Morton and colleagues demonstrated that NF-κB is dispensable for developmental myelination in vivo (Morton et al., 2013). However, they showed that the inhibition of NF-κB signaling in SCs following sciatic nerve crush transiently delays axonal regeneration and remyelination (Morton et al., 2012). This highlights again the differences between immature SCs during development and transdifferentiated repair-SCs. Importantly, NF-κB activity is robustly enhanced after peripheral nerve damage and in inflammatory neuropathies (Andorfer et al., 2001; Laura et al., 2006; Smith et al., 2009; Fu et al., 2010). The different NF-κB-induced transcriptional targets responsible for SC beneficial effects after nerve injury still have to be determined. Of note, the placental growth factor (PlGF) which is up-regulated by NF-κB in SCs following injury, is critical for SC proliferation, myelin sheath degradation, macrophage invasion and axonal regeneration (Chaballe et al., 2011). Targeting NF-κB or its downstream targets in SCs may offer new therapeutic strategies for PNS regeneration. Nevertheless, it is essential to keep in mind that aberrant activation of NF-κB may also lead to schwannoma (Ammoun et al., 2014; Dilwali et al., 2015).

Other factors that regulate gene transcription have been studied in SC dedifferentiation. These are most of the time negative regulators of myelination that functionally counteract positive regulators of myelination including Krox-20 or Oct-6 and are likely to have major roles after nerve damage (Jessen and Mirsky, 2008). For instance, the transcription regulators Sox-2, paired box protein 3 (Pax-3), DNA-binding protein inhibitor 2 (Id2) and early growth response proteins 1 and 3 (Egr-1 and Egr-3) could represent good candidates for regulating SC dedifferentiation in damaged nerves. As for c-Jun or Notch, they are inactive or expressed at low levels in myelinating cells and re-expressed at high levels under conditions that lead to SC dedifferentiation such as nerve injury (Kioussi et al., 1995; Nikam et al., 1995; Stewart et al., 1997; Le et al., 2005; Parkinson et al., 2008; Doddrell et al., 2012). In vitro, Sox2, Pax3, and Id2 decrease or increase myelin gene expression when overexpressed or depleted, respectively (Kioussi et al., 1995; Mager et al., 2008; Parkinson et al., 2008; Doddrell et al., 2012). Egr1 and Egr3 are both necessary to up-regulate p75NTR in SC primary cultures, known to be re-expressed in SCs following a nerve injury (Nikam et al., 1995; Topilko et al., 1997; Gao et al., 2007). It remains to be unraveled by in vivo evidence whether those different factors are required for SC response after nerve injury.

Besides its likely role in SC plasticity, it has been shown that Sox-2 is necessary in SCs for the formation of a nerve bridge after a complete transection. This process is mediated by an ephrin-B/EphB2 signaling between fibroblasts and SCs that initiates cell sorting, followed by directional collective cell migration of SCs out of the nerve stumps to guide axons across the injury wound (Parrinello et al., 2010).

Harrisingh and coworkers were the first ones to demonstrate the role of the Ras/Raf/Erk signaling in SC dedifferentiation (Harrisingh et al., 2004). They showed that the ectopic activation of Raf, an upstream activator of Erk, suppresses the differentiation of primary SCs induced by cyclic adenosine monophosphate (cAMP). Raf activation in cultured differentiated SCs also drives their dedifferentiation. Importantly, they demonstrated that sustained activation of the Ras/Raf/Erk pathway in myelinated dorsal root ganglion (DRG) neuron-SC co-cultures induces demyelination even in the presence of normal axon signaling. Consistent with this idea, another paper reported that the dual specificity phosphatase 15 (Dusp15) is necessary for full activation of Erk and that its activation represses the expression of several myelin genes including MBP (Rodríguez-Molina et al., 2017). While high and sustained activation of Ras/Raf/Erk signaling seems essential for SC plasticity, Ras over-activation might also be deleterious and lead to the development of SC-derived tumors in neurofibromatosis type 1 patients (Harrisingh et al., 2004). A strong activation of Erk has actually been reported in human primary schwannoma cells (Ammoun et al., 2008).

More recently, the strategy of Raf high activation has been used in vivo to determine whether this signaling was central in controlling SC plasticity (Napoli et al., 2012). Napoli et al. demonstrated that a tamoxifen-inducible Raf transgene in SCs is sufficient to induce the down-regulation of myelin proteins and the expression of dedifferentiation markers such as p75NTR, even in the absence of axonal damage. The duration of Raf activation, controlled by the period of tamoxifen treatment, determines the time frame of SC demyelination and proliferation. A prolonged tamoxifen treatment and activation of Raf in SCs also induces an inflammatory response, which includes the breakdown of the blood-nerve barrier and the infiltration of macrophages, which promote nerve regeneration.

Although the repression of Erk signaling with specific inhibitors blocks SC mitosis, cytokine expression and demyelination, it only prevents partially the induction of p75NTR or GDNF and the shutdown of myelin gene expression (Napoli et al., 2012; Shin et al., 2013). Thus, even if the sustained and pronounced activation of the Ras/Raf/Erk pathway is fundamental in regulating SC plasticity, this supports the idea that several independent mechanisms control multiple aspects of SC plasticity. Moreover, different studies showed that Erk activation is actually pro-myelinating and that its inhibition blocks SC differentiation and myelination in vivo (Grossmann et al., 2009; He et al., 2010; Newbern et al., 2011). Therefore, Erk signaling is necessary both for the differentiation of SCPs and for the dedifferentiation of adult SCs following a nerve injury. The reconciling explanation could be that distinct levels of Erk activity would define the state of SC differentiation. Low or basal activity would be required for SC differentiation while high Erk levels would drive dedifferentiation and proliferation (Napoli et al., 2012; Newbern and Snider, 2012).

Finally, the role of c-Jun as a downstream target of Erk signaling is an attractive idea but it is actually quite controversial and further in vivo studies are necessary to better understand the signaling mechanisms downstream of Erk in SC plasticity. Indeed, Syed et al. showed that Erk inhibition abolishes both the demyelination and the expression of c-Jun induced by Neuregulin 1 (Nrg1) treatment in SC-DRG neuron co-cultures (Syed et al., 2010). Additionally, the Lloyd group indicated that c-Jun is strongly up-regulated after Raf over-activation in SCs in vivo (Napoli et al., 2012). However, others reported that the induction of c-Jun and SC dedifferentiation triggered by cAMP removal in vitro depends on JNK rather than on Erk activity (Monje et al., 2010; Shin et al., 2013). Actually, other data revealed that Rac/JNK and Raf/Erk might be responsible for different complementary functions in SC plasticity. For example, microarray experiments showed that the expression of regeneration-associated genes, including p75NTR or GDNF is dependent on JNK but not on Erk signaling. In addition, it has been demonstrated that most Erk-dependent genes in dedifferentiated SCs do not require Rac activity (Shin et al., 2013).

JNK is another MAPK that plays important functions in SCs. It has been demonstrated that c-Jun activation by JNK is essential for SC migration and proliferation in vitro (Parkinson et al., 2001, 2004; Yamauchi et al., 2003). Parkinson and colleagues showed that overexpressing the MAPK kinase 7 (MKK7), a protein upstream of JNK, in SC primary cultures, inhibits myelin gene expression by elevating c-Jun, suggesting a role for JNK signaling in SC plasticity (Parkinson et al., 2008). This was confirmed by Monje et al. who revealed that a reduction of cAMP in vitro is sufficient to induce SC dedifferentiation through a JNK-dependent mechanism (Monje et al., 2010). More recently, it has also been demonstrated that the active Rac1 GTPase (Rac), an upstream regulator of MKK7 and JNK, inhibits SC differentiation. In SC cultures, Rac up-regulates c-Jun and down-regulates Krox-20 through the MKK7/JNK pathway, but not through the Raf/Erk pathway. Furthermore, MKK7 activation and the induction of c-Jun observed in sciatic nerves after axotomy were blocked by Rac inhibition (Shin et al., 2013). In addition, Rac activation has been observed in the distal stump of injured nerves, and specific Rac inhibition with a dominant negative Rac1 and Rac1-RNAi decelerates myelin sheath fragmentation. Rac would actually promote myelin sheath fragmentation by controlling actin polymerization in the Schmidt-Lantermann incisures of SCs (Jung et al., 2011).

The cyclin-dependent kinase inhibitor 1 (Cdkn1c), also called p57Kip2, is a protein first known to block G₁/S transition during cell cycle that also regulates actin filament dynamics through the translocation of LIM domain kinase 1 (LIMK1), a downstream effector of Rac/Cdc42/Rho (Yokoo et al., 2003; Heinen et al., 2013). Knocking down Cdkn1c in cultured SCs leads to cell cycle exit, morphological changes and up-regulation of myelin genes. In SC-DRG neuron co-cultures, Cdkn1c depletion accelerates myelination. A global gene expression analysis also revealed a shift of the transcriptional expression program toward the pattern of differentiating SCs, strongly suggesting a role for Cdkn1c as an important negative regulator of myelination (Heinen et al., 2008). The same team also showed that Cdkn1c is a direct target of the enhancer of zeste homolog 2 (EZH2) and identify Hes5 as a new transcriptional repressor of myelin genes (Heinen et al., 2012). Indeed, EZH2 binds Cdkn1c promoter and its suppression results in a decrease of histone H3K27 trimethylation, an induction of Cdkn1c expression and a down-regulation of myelin genes in cultured primary SCs.

Yamauchi and co-workers reported that the atypical guanine-nucleotide exchange factor dedicator of cytokinesis protein 7 (Dock7), activates Rac/Cdc42/JNK signaling in SCs to positively regulate migration and negatively regulate differentiation and myelination (Yamauchi et al., 2008, 2011). Indeed, knocking down Dock7 in primary SCs results in an increase of cAMP-mediated differentiation and in a shorter duration of activation of Rac/Cdc42 and JNK while silencing Dock7 in mice increases myelin thickness.

Finally, collagen triple helix repeat containing 1 (Cthrc1) is another protein that has been linked to Rac activation in SCs. Apra and colleagues investigated Cthrc1 function in SCs both in vitro and in vivo (Apra et al., 2012). They showed that its silencing inhibits SC proliferation and promotes their migration and myelination in vitro. By using a transgenic mouse line, they demonstrated that the overexpression of Cthrc1 in SCs leads to a delayed myelination with SCs maintaining a proliferative state. More recently, Zhou et al identified miR-9 as an important negative regulator of SC migration targeting Cthrc1 after nerve injury (Zhou et al., 2014). The up-regulation of miR-9 decreases SC migration in culture and following nerve injury. miR-9 has also been shown to directly target Cthrc1, which in turn inactivates downstream Rac. Finally, miR-9 is down-regulated at a speed comparable to the rate of SC migration during nerve regeneration, suggesting that miR-9 down-regulation could be linked to the phenotypic modulation of SCs after nerve injury.

All these data confirm that Rac/JNK signaling is essential for SC plasticity, proliferation and migration following nerve injury. However, stimulating this pathway for nerve repair therapy must be cautioned as its upregulation has also been observed in primary human schwannoma cells (Kaempchen et al., 2003).

Yang and co-workers provided the most relevant data that suggest a role for p38 MAPK in SC plasticity and their response to injury (Yang et al., 2012). They reported that inhibition of p38 MAPK activity in mice blocks SC demyelination and dedifferentiation following nerve injury, despite ongoing axon degeneration. In myelinating co-cultures, they also showed that p38 MAPK mediates myelin breakdown induced by Nrg1. Finally, while the ectopic activation of p38 MAPK is sufficient to provoke myelin breakdown and c-Jun induction, its inhibition in SC-DRG neuron co-cultures promotes myelination. Roberts and collaborators recently investigated the in vivo function of the major p38 isoform in mouse conditional knockouts. In agreement with previous findings, the lack of p38alpha in SCS accelerates myelination during development, delays myelin clearance and slightly increases remyelination following injury. However, axonal regeneration and functional repair were not affected in the absence of p38alpha (Roberts et al., 2016).

A recent study on bone morphogenetic protein 7 (BMP7) in SCs revealed a role for this protein as a negative regulator of myelination through the activation of p38 MAPK signaling (Liu et al., 2016b). Liu and collaborators showed that BMP7 expression is inversely correlated with myelin gene expression during normal peripheral myelination. In addition, BMP7 treatment decreases cAMP-induced myelin gene expression and up-regulates c-Jun in primary SC cultures. By using p38 MAPK inhibitor, gene silencing and rescue experiments, they reported that p38 MAPK is responsible for BMP7-mediated suppression of myelin gene expression. Finally, when injected in newborn rats, BMP7 retards peripheral myelination.

Phosphatidylinositol 3-kinase (PI3K) signaling has been implicated in the regulation of proliferation, survival and differentiation of different cell types (Rameh and Cantley, 1999). The pathway is strongly activated in primary cultured SCs following treatment with soluble Nrg1 or contact with neurite membranes. Moreover, PI3K signaling is required for SC proliferation and survival in vitro (Maurel and Salzer, 2000; Li et al., 2001). Its inhibition affects the initiation of myelination but is not necessary for myelin maintenance in SC-neuron co-cultures (Maurel and Salzer, 2000). In addition, the selective activation of Akt in SCs by gene transfer results in an increased myelination in SC-DRG neuron co-cultures and allogenic nerve graft experiments. This last paper also demonstrated opposing functions for PI3K/Akt and Erk pathways in controlling SC myelination (Ogata et al., 2004). The imbalanced activation of the PI3K/Akt and Erk pathways has also been revealed in transgenic rodent models of CMT1A in which SCs develop persistent differentiation defects during early postnatal development (Fledrich et al., 2014). The transgenic expression of constitutively active Akt in SCs, however does not seem to disrupt PNS myelination in vivo (Flores et al., 2008).

Studies investigating the role of specific proteins indirectly indicated a role for PI3K/Akt signaling in the regulation of SC proliferation and differentiation after nerve injury. For example, Src-associated in mitosis of 68 kD (Sam68) has been suggested to promote SC proliferation by enhancing the PI3K/Akt pathway and to act on regeneration following sciatic nerve crush (Wu et al., 2016b). DIX domain containing-1 (Dixdc1), which is also linked to PI3K/Akt activation, promotes SC proliferation after sciatic nerve crush. Dixdc1 is up-regulated following nerve damage and its depletion leads to reduced PI3K/Akt activation and SC proliferation (Wu et al., 2016a). Consistently, 17β-estradiol has been shown to promote SC proliferation and early remyelination in the bridge after nerve transection through Akt activation (Chen et al., 2016a). Src-suppressed protein kinase C substrate (SSeCKS) has been indicated as a suppressor of SC differentiation and myelination by inhibiting Akt activation. SSeCKS is effectively down-regulated during SC differentiation and its depletion promotes Akt phosphorylation and myelin gene expression in cAMP-treated primary cultures or myelination when co-cultured with DRG neurons (Ji et al., 2010). SSeCKS is also essential for SC adhesion, spreading and migration (Yan et al., 2009). In addition, Chen and collaborators showed that cell adhesion molecule 3 (Cadm3) might act as a negative modulator of PNS myelination. Cadm3 interferes with PI3K/Akt activation and inhibits Schwann cell myelination in vitro (Chen M. S. et al., 2016).

The mammalian disks large homolog 1 (Dlg1) inhibits PI3K/Akt signaling via its interaction with the phosphatase and tensin homolog deleted on chromosome 10 (PTEN). This mechanism is necessary to transiently inhibit the axonal stimulation of myelination and to limit myelin sheath thickness. Myelin outfoldings and overmyelination followed later by pathologic demyelination are actually observed after Dlg1 silencing via RNA interference in mouse sciatic nerves (Cotter et al., 2010). Dlg1/PTEN functions and consequently Akt/mTOR signaling are also disturbed in different mouse models of demyelinating neuropathies and peripheral nerve sheath tumors (Bolino et al., 2004; Cotter et al., 2010; Goebbels et al., 2010, 2012; Keng et al., 2012). More specifically, Goebbels et al showed that mice lacking PTEN develop a tomaculous neuropathy characterized by focal hypermyelination and myelin outfoldings. Moreover, treating these mice with rapamycin, a mTOR inhibitor, significantly ameliorates the pathology (Goebbels et al., 2010, 2012). Another study on PTEN signaling showed that the hypermyelination observed in mice lacking Dlg1 is actually transitory. This study also identified DNA damage-inducible transcript 4 protein (DDIT4) as a sustained negative regulator of myelination. High DDIT4 expression in SCs precedes the peak of Dlg1 and Akt activity in peripheral nerves. Loss of DDIT4 expression both in vitro and in vivo results in sustained hypermyelination and enhanced PI3K/Akt/mTOR activation (Noseda et al., 2013). Therefore, it is likely that PI3K/Akt signaling has to be precisely regulated and balanced to induce SC differentiation or dedifferentiation.

mTOR is a signaling pathway that integrates a lot of signals (growth factors, nutrients, energy and oxygen) necessary for growth and proliferation (Norrmén and Suter, 2013). Sherman and collaborators demonstrated that the inactivation of mTOR in SCs provokes an arrest of myelination after axonal sorting and reduces axon growth (Sherman et al., 2012). More recently, it has been reported that the conditional ablation of raptor, an essential component of the mTORC1 complex, in SCs leads to hypomyelination and abnormal lipid biosynthesis (Norrmén et al., 2014). The pathways that converge to mTOR are numerous and include Nrg1/ErbB, PI3K/Akt, MAPKS or AMP-activated kinase (AMPK) signaling (Norrmén and Suter, 2013). AMPK is a heterotrimeric serine/threonine protein kinase acting as an inhibitor of mTOR and described as a crucial energy sensor important to maintain energy homeostasis (Hardie et al., 2012). Recently, Liu and colleagues revealed that AMPK is a critical negative regulator of SC myelination in the PNS (Liu et al., 2016a). They first showed that AMPK is gradually decreased as myelination proceeds. The high activation of AMPK decreases myelin gene expression and stimulates c-Jun expression in SCs whereas its inhibition or depletion induces myelin gene expression. The injection of an inhibitor of AMPK in newborn rats downregulates c-Jun expression, enhances lipid and protein synthesis and increases myelin gene expression and myelin sheath thickness.