Paving the Axonal Highway: From Stem Cells to Myelin Repair
- First Online:
- Cite this article as:
- Peru, R.L., Mandrycky, N., Nait-Oumesmar, B. et al. Stem Cell Rev (2008) 4: 304. doi:10.1007/s12015-008-9043-z
- 249 Views
Multiple sclerosis (MS), a demyelinating disorder of the central nervous system (CNS), remains among the most prominent and devastating diseases in contemporary neurology. Despite remarkable advances in anti-inflammatory therapies, the inefficiency or failure of myelin-forming oligodendrocytes to remyelinate axons and preserve axonal integrity remains a major impediment for the repair of MS lesions. To this end, the enhancement of remyelination through endogenous and exogenous repair mechanisms and the prevention of axonal degeneration are critical objectives for myelin repair therapies. Thus, recent advances in uncovering myelinating cell sources and the intrinsic and extrinsic factors that govern neural progenitor differentiation and myelination may pave a way to novel strategies for myelin regeneration. The scope of this review is to discuss the potential sources of stem/progenitor cells for CNS remyelination and the molecular mechanisms underlying oligodendrocyte myelination.
KeywordsCentral nervous systemMyelin repairOligodendrocytesStem cellsRegenerationMultiple sclerosisTranscriptional regulationMyelination
For the past decade, neuroscience has been captivated by the revelation that, contrary to the prevailing dogma, adult neurogenesis does indeed occur. With the discovery that stem cells could generate specific cell-lineages, including neurons and oligodendrocytes, came the realization that nervous system repair was closer to becoming a reality. Understanding how we might tweak the molecular machinery controlling the fate of neural stem cells (NSC) is anticipated to yield new avenues for the treatment of currently incurable diseases such as neurodegenerative and demyelinating disorders.
Nervous system function is dependent upon efficient neural conduction and connectivity. At the heart of this adaptive prowess lies myelination, a process of axonal ensheathment that is necessary for maximizing the conduction velocity of neural impulses [1, 2]. The failure of myelin repair after injury aggravates many neurodegenerative diseases such as the devastating demyelinating disease, multiple sclerosis (MS).
In this review, we will discuss current issues in myelin formation and repair in the CNS by exploring the myelinating potential of embryonic and adult neural progenitor cells. We will also discuss current insights on the generation of myelin-forming cells from stem/progenitor cells based on the genetic and epigenetic mechanisms in the CNS.
CNS Myelin Repair By Oligodendrocyte Precursors
Effective axonal remyelination by OPCs in adults has been observed in many experimental situations where both structure and function are partially, if not completely, restored (reviewed in ). In the case of MS, this regenerative capacity progressively abates, leading to chronic demyelination and ultimately axonal loss. Depletion of OPCs is a possible reason for the remyelination failure in some MS lesions . However, some chronic MS lesions contain proliferating OPCs [6, 7], suggesting that remyelination failure in such an instance may not be accounted for by a scarcity of progenitor cells, but by their inability to differentiate into myelinating cells. This hypothesis was recently supported by the fact that a differentiation block of OPCs may be a major determinant of remyelination inefficiency in chronic MS lesions .
OPCs eventually cease to regenerate myelin sheaths after persistent chronic demyelination , suggesting that remyelination by OPCs may ultimately be halted by recurring insults in MS. It is worth mentioning that when remyelination episodes have sufficiently lapsed, the capability for remyelination is unimpaired . Recent studies indicate that the reduced remyelination efficiency of adult OPCs may be age-related, as aging was shown to affect both OPC recruitment and differentiation . Spontaneous remyelination in MS occurs in acute lesions and is often accompanied by inflammation. Macrophages activated in response to inflammation may remove myelin debris within active lesions, which presumably blocks OPC differentiation into myelinating oligodendrocytes . Moreover, OPC recruitment and remyelination can be induced by inflammatory signals associated with demyelination. Several lines of evidence indicate that the absence of macrophages impairs remyelination  and that upregulation of myelin-related proteins is achieved by the addition of macrophages into neuron-glia co-cultures . Although the precise cause for the inefficiency of remyelination in MS is not fully understood, recruiting OPCs and activating their differentiation capacity are crucial steps to target in the design of future therapies promoting endogenous myelin repair. In addition, efficiency of repair response depends upon the environmental context that is regulated by graft-to-host, host-to-host, and potentially graft-to-graft interactions [15–17].
Are Pluripotent Stem Cells Capable of Myelin Repair?
Interest in pluripotent embryonic stem cells (ESC) has gained momentum in recent years due to their experimental value to biomedical research and therapeutic potential in regenerative medicine. Many advances have been achieved in this field, such as the generation of cardiomyocytes from ES cells in the treatment of stroke patients [18, 19] or the selective induction of ES cells into functionally diverse neuronal phenotypes as a therapy for neurodegenerative diseases . The relative success of ESC research opens the possibility that ES cells could also be manipulated, either epigenetically or genetically, to generate myelinating cells, which could serve as future treatments in cell-based therapies of demyelinating diseases (reviewed in ) (Fig. 1).
Mitigating the optimism for the ESC-based therapy is that there have been limited degrees of success when applied into a degenerated or lesion area in vivo, particularly, when it comes to the usage of human ES (hES) cells, which are less understood than mouse ES cells [22, 23]. Understanding the differentiation of myelinating oligodendrocytes from human ES cells still remains a challenge. Besides, there is an issue regarding the ethics of ESC use. In response to the heated public debate of the ethical use of ESC, there has been enormous political backlash that has hindered the widespread application of ESC in cell-based therapy .
Despite the above handicaps, recent data has provided compelling evidence that ES cell-derived progenitor cells can be systematically induced to differentiate into functional oligodendrocytes. These oligodendrocytes can then facilitate axonal ensheathment after transplantation into demyelination sites [25–27] either by stimulating the remyelinating potential of endogenous progenitors, or by the integration of exogenous progenitors into the host tissue and subsequently differentiating into remyelinating cells . Transplantation of ES cell-derived NSCs facilitates motor recovery in experimental demyelinating animal models. Although transplanted NSCs in spinal cord lesions were found to differentiate into astrocytes, neurons, and oligodendrocytes , McDonald et al. (1999) demonstrated that recipient mice could partially recover motor skills. When ESC-derived progenitors treated with fibroblast growth factor 2 (FGF2) and platelet-derived growth factor (PDGF) were injected into the spinal cord in a myelin-deficient rat model of Pelizaeus–Merzbacher disease, differentiated oligodendrocytes were observed to remyelinate axons of the spinal cord and restore motor function . Similarly, when ESC-derived oligospheres were injected into the spinal cord of myelin basic protein (MBP) deficient Shiverer mice and spinal cord lesions inflicted by chemical demyelination, they survived, migrated, and differentiated into myelinating oligodendrocytes . Furthermore, by transplanting hES cell derived OPCs into adult spinal cords after injury, Keirstead et al. (2005) found that both remyelination and functional recovery improved when cells were injected 7 months post-injury, but not after 10 months . These data suggest that exogenous myelin repair can take place during a specific window of time at which cells may be successfully recruited into a permissive environment for myelin formation, but whose myelinogenic potential can eventually be repressed by later development of an inhibitory environment.
The primary reasons for caution in the use of stem cells in therapeutic settings are the difference in cell harvesting and differentiation techniques, the failure to recapitulate the in vitro efficiency in vivo, and the inherent susceptibility of ES cells to become teratomas or other cell types upon transplantation . However, given their myelinating potential, ES-derived oligodendrocyte progenitors cells have relevant medical applications for myelin repair and neuroprotection once their proliferation and differentiation capacity in vivo is harnessed. In this regard, significant progress has been made towards the implementation of more refined experimental approaches using growth factors or modulating early transcription factors to instruct human ES cells to become specific neural cell types such as neurons or oligodendrocytes . However, further development and implementation of such protocols in clinical trials are needed to fully explore the medical application of ES cells.
Neural Stem/progenitor Cell-Based Therapy for Myelin Regeneration
Multipotent neural stem cells (NSCs) are neurogenic and gliogenic within the CNS throughout both development and adulthood and were extensively studied for their capacity to regenerate myelinating oligodendrocytes in vitro and in vivo after transplantation (Fig. 1). These cells can be expanded in vitro by exposure to FGF2 and epidermal growth factor (EGF), in adhering or free-floating conditions. In the later case, these cells give rise to neurospheres, which can be passed and cultured for an extensive period of time . Upon adhesion and withdrawal of growth factors, multipotent NSCs have the potential to differentiate into neurons, oligodendrocytes, and astrocytes. These cells have also been isolated from fetal and adult human brains, and can be amplified in vitro as neurospheres. Neurospheres have the ability to generate both neurons and glia. However, when compared to rodent cells, the differentiation of human NSCs into functional neurons or glia is more complex. Current technologies to direct the differentiation of human NSCs into specific cell types, and especially into oligodendrocytes, require further optimization.
Transplantation of neural stem/precursor cells has proven a more effective strategy for myelin repair in lieu of the slow rate of endogenous OPCs in myelin repair. Results from a number of transplantation experiments with neural progenitor cells and OPCs have generated interest regarding the myelinogenic and restorative properties of these cells in various experimental rodent models of dysmyelination and induced demyelination . Efficient NSC migration towards sites of demyelination upon transplantation and differentiation into myelinating oligodendrocytes can occur in the injured CNS. A recent study indicated that neonatal transplantation of human OPCs could effectively restore the myelination in congenital hypomyelinated shiverer mice. These data point to a clinical potential of cell-based myelin repair strategies in the treatment of disorders of congenital and perinatal hypomyelination . Despite these encouraging results in experimental models of induced-demyelination, transplantation strategies have been difficult to devise due to the multifocal nature of demyelination in MS.
To this end, recent studies have challenged the regenerative capacity of NSCs by demonstrating their immunomodulatory properties in repair. With improved modes of systemic delivery, Pluchino et al. (2003) demonstrated the regenerative potential of adult NSCs in the experimental autoimmune encephalomyelitis (EAE) mouse model. Remarkably, injection of adult NSCs either intra-ventricularly or intra-cerebrovascularly in sites of experimentally induced multifocal lesions in EAE mice promotes remyelination of severed axons and functional recovery from EAE . Further studies have indicated that the transplanted NSCs appear to promote neuroprotection by maintaining undifferentiated features and exerting immune-like functions . Hence, the therapeutic effects appear to be primarily due to immunomodulation rather than cell replacement [35, 36]. Recent studies have pointed out that high levels of cell adhesion molecules (CAMs), integrins, and chemokine receptors in neural progenitors promote their migration across the brain blood barrier or to reactive-astrocyte rich regions .
Furthermore, a recent study by Ziv et al. (2006) demonstrated that transplantation of adult neural progenitors into the cerebrospinal fluid, concomitant with T cell-based vaccination of mice, synergistically promotes functional recovery after spinal cord injury. These results suggest that immunosuppression with myelin-specific peptides can disrupt the inhibitory local immune response and enables adult neural progenitors to migrate to the lesion site . Collectively, these studies provide evidence that adequate mode of delivery and proper immunomodulation enable neural progenitors to survive for long periods of time, migrate to injury sites, and differentiate into cell types required to promote repair.
Sources of Myelinating Cells
The vast majority of remyelinating cells in the CNS are believed to descend from the endogenous OPCs. This notion stems from various lines of evidence indicating that most remyelinating cells arise from proliferating OPCs at the site of demyelination and that the provision of isolated adult progenitors into demyelinated regions induces myelin repair [39–41]. These adult precursors can be identified with a series of precursor markers including NG2, platelet-derived growth factor receptor alpha (PDGFRα), Nkx2.2, and oligodendrocyte lineage transcription factors 1 and 2 (Olig1/2) [6, 42–46]. While adult OPCs are speculated to retain the myelin-forming properties, it remains unresolved whether all, or a subset of progenitors with the same antigenic identity are capable of generating remyelinating cells after injury [39, 47–49]. The identification of precursor sources of OPCs with effective myelinogenic capacity remains a prerequisite for understanding endogenous repair mechanisms in the adult CNS.
Sources of Oligodendrocyte Precursors in the Developing CNS
Although initial progenitors from the ventral region require Sonic hedgehog (Shh) activity, the emergence of later waves of oligodendrocyte progenitors is independent of Shh [56, 57]. At a later embryonic stage when bone morphogenetic protein (BMP) and/or Wnt signaling are diminishing, a Shh-independent OPC subpopulation emanates from the dorsal spinal cord and constitutes 10–15% of the entire OPC population . Fate-mapping of the progeny of dorsal progenitors expressing developing brain homeobox 2 (Dbx2) in the Dbx2-Cre line suggests that a subset of oligodendrocytes are derived from the dorsal neural epithelium during late embryogenesis . Thus, ventral and dorsal progenitors cells can sequentially generate OPCs to populate the entire spinal cord at different stages (Fig. 2c).
In a similar fashion to the spinal cord, initial OPCs in the forebrain are generated from the ventral telencephalon which includes the medial and lateral ganglionic eminences (MGE and LGE), beginning around e12 to e14 . OPC populations in the thalamus, hypothalamus, and cerebellum appear to be derived from the anterior entopeduncular area (AEP) of the ventral diencephalon [51, 55, 61] (Fig. 2d).
Forebrain progenitors derived from ventral telencephalic regions constitute a transient and migratory population of cells that initially competes for resources against dorsal progenitors. Eventually, ventral progenitors are obliterated and replaced by their dorsal counterparts, which will ultimately populate the grey and white matter regions of the brain [58, 62].
The respective contributions of regionally diverse OPCs to cortical myelination remain an issue of contention in regard to the repair of prominent myelin losses in the cortex after injury or disease. A recent study indicated that forebrain dorsal progenitors play a critical role in cortical myelination. It has also been shown that although ventral progenitors contribute to cortical oligodendrogenesis, they are insufficient to overcome myelin deficits due to the defects in dorsal progenitor cells of the cortex . Hence, progenitor sources proximal to demyelination sites could be more efficient for myelin regeneration than those distal from the lesion site.
Adult NSCs and Myelin Repair
Adult multipotent NSCs have recently been identified as potential sources of precursors for myelin repair in demyelinating diseases . ‘Germinal niches’ of adult NSCs are harbored in the sub-ventricular zone (SVZ) lining the lateral ventricle and in the sub granular zone (SGZ) of the dentate gyrus . Like other stem cells, adult NSCs not only display self-renewing features and untapped proliferative capacity, but they also generate more lineage-restricted neural cell types.
SVZ neural stem cells have been identified as a cortical source of astrocytes and oligodendrocytes during postnatal development  and may have an important role in myelin repair in the adult brain following demyelination [65–67]. Glial fibrillary acidic protein (GFAP) positive type B stem cells in the adult SVZ, which produce mainly new neurons in the olfactory bulb, are capable of giving rise to NG2+OPCs and mature oligodendrocytes in normal conditions , as well as after chemically- or inflammatory-induced demyelination . It is worth mentioning that OPCs are present in the corpus callosum and other white matter regions and undergo cell division during both normal and injured conditions. At present, the extent to which mobilization of adult SVZ stem cells or local endogenous OPCs in the gray and white matters contributes to myelin repair after injury remains unresolved. Aguirre et al. (2007) argued that over-expressing epidermal growth factor receptor (EGFR) in CNP-expressing SVZ progenitors and OPCs can increase NG2+, Mash1+ (mammalian achaete-schute Homolog 1) and Olig2+ progenitor pools, oligodendrocyte production, and myelination . This data strongly suggests that the activation of neural progenitors and endogenous OPCs could be important for the generation of oligodendrocytes for myelin repair.
The SVZ also persists in the adult human brain and harbors cells expressing the same markers as in rodents, such as Poly-Sialated Neural Cell Adhesion Molecule (PSA-NCAM), GFAP, nestin, and the epidermal growth factor receptor (EGFR). Adult human SVZ cells also retain the capacity to self-renew and generate neurons, astrocytes, and oligodendrocytes in vitro [70, 71]. Despite these similarities, the adult human SVZ differs significantly from rodents in its cellular organization. In the human brain, this structure may be formed by a ribbon of GFAP+ presumptive stem cells that are separated from the ependymal wall by a hypocellular gap. The presence of a rostral migratory stream like structure in the human brain  is still a matter of discussion. Recent data has demonstrated that the human SVZ is re-activated in MS and contains early glial progenitors expressing Olig2 and Sox10 transcription factors .
Enhancing the Differentiation Capacity of OPCs: A View of Extrinsic and Intrinsic Control
Extrinsic Regulation of OPC Differentiation
The functional roles of neurotrophins in CNS myelination remains to be established, although recent data indicates that members of distinct neurotrophin families have divergent roles in the PNS and CNS. For example, brain-derived neurotrophic factor (BDNF) and Neurotrophin-3 (NT-3) promote oligodendrocyte differentiation, while having inverse effects on Schwann cell myelination [83, 84]. In contrast, nerve growth factor (NGF) signaling appears to be inhibitory in CNS myelination and promoting in PNS myelination . Although the molecular mechanisms underlying the divergent effects of neurotrophins in CNS and PNS myelination are not clear, it is suspected that the expression of distinct receptors in neurons and oligodendrocytes may contribute to the difference in neurotrophic effects. In addition, the balance of different neurotrophic factors has a critical role in modulating myelination. These observations suggested that growth factors modulate receptivity to myelination depending on intrinsic axonal properties .
NT-3 appears to directly promote the proliferation and survival of oligodendrocytes. In conjunction with PDGF, NT-3 stimulates clonal expansion of oligodendrocyte precursor cells [85, 86]. Transplantation of genetically modified fibroblasts or Schwann cells that secrete NT-3 and/or BDNF into the injured rat or demyelinated mouse spinal cord was shown to increase the production of new oligodendrocytes and promote myelination . NT-3 and BDNF-induced myelinogenesis results, at least in part, from the expansion of endogenous oligodendrocyte progenitors. These findings may have significant implications for the development of treatment strategies for chronic demyelinating diseases or other CNS injuries.
Other neurotrophic factors of the ciliary neurotrophic factor (CNTF) family such as leukemia inhibitory factor (LIF), cardiotrophin-1, and oncostatin M, elicit a promyelinating effect . Even though they play a role in astrocyte differentiation, the CNTF family members also act via the gp130 kDa receptor on oligodendrocytes to promote their final maturation . In the PNS, the neuropoietic cytokine CNTF promotes Schwann cell myelination via Sox10 transcription activation, which is also required for terminal oligodendrocyte differentiation .
To complicate matters, a recent study has revealed that astrocytes are capable of mediating electrically induced myelination. Ishibashi et al. (2006) suggest that ATP released by axons during electrical stimulation acts on astrocytes and in turn stimulates the release of LIF, which promotes oligodendrocyte myelination [89, 91]. This finding suggests that astrocytes may have a novel role in axon-oligodendrocyte interactions and demonstrates that myelination by postmitotic oligodendrocytes is activity-dependent, consistent with the notion that cellular microenvironments play a crucial role in normal myelination. At present, the precise roles of astrocytes in modulating oligodendrocyte myelination remain to be fully understood. Thus, the integration of signaling pathways mediated by different neurotrophic factors may fine-tune the complex process of myelination in the CNS.
An axon-derived molecule, Neuregulin-1 (Nrg1) belongs to a family of membrane associated and secreted factors that have an EGF-like domain at their N-termini. It has three isoforms (type I, II and III) derived from alternative splicing. Neuregulins activate members of the EGF receptor subfamily of protein tyrosine kinases (PTKs) known as ErbBs, including ErbB2, ErbB3, and ErbB4 [92–94].
Although Nrg1 signaling determines the onset and extent of myelination by Schwann cells in the PNS, it is currently unclear whether or not and to what extent Nrg1 regulates oligodendrocyte myelination in the CNS. Overexpression of Nrg1 type III in neurons of transgenic mice induces hypermyelination around the axons by both Schwann cells and oligodendrocytes . However, its role in oligodendrocyte differentiation varies considerably between in vitro and in vivo studies. Studies on cultured OPCs have shown that Nrg1 has a trophic and mitogenic effect on OPCs, but it inhibits oligodendrocyte differentiation . In contrast, using spinal cord explants from ErbB2 knockout mice, Park et al. (2001) demonstrated that ErbB2-mediated signaling is necessary for the terminal differentiation of OPCs . At present, the extent to which myelination is dependant on Nrg1/ErbB signaling in the CNS is not clear. Further studies on the spatiotemporally specific mutagenesis of Nrg signaling are required to determine the role of Nrg/ErbB signaling in CNS myelination.
Intrinsic Regulation of Oligodendrocyte Differentiation
Oligodendrocyte specification and differentiation are complex processes that are regulated not only by various extracellular signals, but also by a network of intrinsic factors. The specificity and timing control of differentiation is achieved by the interplay between genetic and epigenetic factors. During development, glial promoting factors emerge near the end of neurogenesis, concomitant with the downregulation of proneural genes such as Ngn1/2 . These factors restrict the fate of neural progenitors in order to generate oligodendrocyte and astrocyte precursors in a spatiotemporally specific manner . The ability of OPCs to differentiate and myelinate axons is regulated by a network of DNA-binding transcription factors [99–101], which are far from being fully understood at present. In this section, we will emphasize on several transcription factor families and epigenetic factors that regulate oligodendrocyte specification and myelination.
Transcription Factors Modulating OPC Differentiation from Neural Stem Cells
bHLH factor family
The discovery of Olig1 and Olig2 points to a critical role of class II bHLH factors in oligodendrocyte fate specification analogous to neuronal cell type specification. Olig1 and Olig2 are expressed in the earliest OPCs of the ventral neural tube pMN domain [42, 43, 54]. During early neurogenesis, before E12.5 in rodents, Olig2 cooperates with proneural bHLH factors such as Ngn2 to promote motor neuron (MN) differentiation [102, 103]. At later embryonic stages when proneural genes become downregulated, Olig2 appears to promote oligodendrocyte specification in collaboration with other neural patterning genes such as Nkx2.2 . Olig1 and Olig2 compound mutations result in a complete absence of oligodendrocytes in the CNS. Disruption of Olig2 alone leads to the loss of oligodendrocyte formation in the developing spinal cord [105–107], while a localized population of OPCs arises within the ventral forebrain and the hindbrain. These studies suggest that the functions of Olig1/2 are partially redundant in oligodendrocyte development in the brain [105, 107].
The proneural bHLH factor Mash1 has recently been implicated in OPC specification in the CNS [108, 109], while it is also essential for neurogenesis [110, 111]. Mash1 is co-expressed with Olig1/2 and PDGFRα in an OPC population derived from the ventral ventricular zone. The formation of this OPC subpopulation is significantly reduced during early brain development (between E11.5 and E13.5). OPCs derived from Mash-negative progenitors are likely to compensate for early deficits in the developing CNS [108, 109]. These studies suggest that oligodendrocytes generated from different sources may depend on distinct transcription regulators, although functional differences among OPC subpopulations remain obscure in the CNS development.
Inhibitory HLH factors
Helix–loop–helix (HLH) factors (i.e. Id2 and Id4) lack a DNA-binding basic domain and negatively regulate oligodendrocyte differentiation. HLH factors inhibitor of DNA binding 2 (Id2) and inhibitor of DNA binding 4 (Id4) inhibit OPC specification by dimerizing with class II bHLH factors such as Olig2 . Over-expression of Id2/4 prevents OPC differentiation [113, 114], while a downregulation of Id2/4 and their extracellular nuclear translocation coincide with the onset of oligodendrocyte differentiation [113, 115].
In addition, Notch signaling activation, which induces Mash1 expression, could instruct neural stem cells to adopt glial fate including OPC specification while suppressing neurogenesis. However, persistent activation of Notch signaling blocks further differentiation of OPCs by presumably upregulating the inhibitory bHLH factor Hes5  and the transcription factor NFIA . Sustained expression of Hes5 and NFIA inhibits OPC differentiation while promoting astrocyte development [117, 118]. Consistent with its repressive activity, Hes5 inhibits the activity of the MBP promoter by binding to Sox10 and Mash1 [100, 119].
A recent study indicated that the cooperation of Olig2 and Mash1 may counteract the proneural activity of neural patterning factors distal-less homeobox 1 (Dlx1) and distal-less homeobox 2 (Dlx2) . Expression of Dlx1/2 can promote neuronal over oligodendroglial fate acquisition in Olig2+ cells in the developing forebrain (Fig. 2d) . Collectively, a balance among oligodendrocyte promoting bHLH factors (i.e. Olig1/2 and Mash1), negative HLH factors (i.e. Id2/4 and Hes5), and pro-neuronal factors (i.e. Ngn1/2 and Dlx1/2) plays an important role in restricting oligodendrocyte specification and regulating the timing of OPC differentiation from neural progenitors.
Sox protein family
Among group E HMG domain-containing Sox proteins (SoxE), Sox9 and Sox10 play a central role in oligodendrocyte lineage development [121–123]. Expression of Sox9 precedes that of Sox8 and Sox10 during oligodendrocyte lineage development, but it is downregulated in differentiated oligodendrocytes. Recent studies have indicated that Sox9 is required for the formation of oligodendrocyte progenitors and may act as a molecular switch between neurogenesis and gliogenesis . How Sox9 controls this neuron-glia switch still remains to be established. Sox9 can partially compensate for the absence of Sox10 in terms of OPC formation [124, 125], however, disruption of Sox10 results in a failure of terminal differentiation of OPCs in the spinal cord . Moreover, overexpression of Sox10 promotes the expression of myelin genes such as MBP in vitro and in vivo [126, 127]. In addition, an oligodendrocyte-specific double Sox9/10 deletion, but not single deletion of either gene, leads to an aberrant migration pattern and OPC death due to the downregulation of PDGFRα, suggesting a collaborative role of Sox9 and Sox10 in OPC survival and migration .
On the other hand, SoxD proteins (e.g. Sox5 and Sox6) were found to compete against SoxE proteins for DNA binding sites and to negatively regulate oligodendrocyte differentiation. SoxD double mutants display an earlier transient increase in OPC production . Hence, SoxD may modulate SoxE protein functions for appropriate timing and progression of OPC differentiation at different stages of oligodendrocyte development .
Recently, Sox17, a member of the SoxF family, was identified to associate with oligodendrocyte lineage cells . Sox17 overexpression promotes myelin gene expression in OPCs and directly stimulates MBP gene promoter activity, while inhibition of Sox17 expression by siRNA maintains OPCs in a precursor stage. Thus, Sox17 may promote maturation via promoting OPC cell cycle exit and differentiation into myelinating oligodendrocytes. At present, it is not clear whether or how Sox17 coordinates with SoxE and SoxD proteins to regulate oligodendrocyte differentiation.
Homeodomain protein family
Neural patterning factors mainly encoded by homeodomain (HD) transcription factors including Iroquois homeobox 3 (Irx3), Nkx2.2, and Nkx6.1/6.2 in the ventral neural tube, play an important role in modulating neural subtype diversification. Combinatorial expression and cross-repressive interactions among Irx3, Nkx2.2, and Nkx6.1 appear to act as a HD transcription factor code to refine and maintain the domains of specific neuronal and glial progenitors [131, 132]. Expression of homeodomain proteins Irx3 and Nkx2.2, dorsal and ventral to the pMN domain, respectively, restricts Olig2 expression in the pMN domain to oligodendrocyte specification [105, 107] (Fig. 2b).
Although initial expression of Nkx2.2 is restricted to the p3 domain of the ventral ventricular zone adjacent to the floor plate, it appears to subsequently expand into Olig2+ populations . In cooperation with Olig2, Nkx2.2 can promote oligodendrocyte differentiation . Although Nkx2.2 is critical for oligodendrocyte differentiation, it is dispensable for OPC specification .
Nkx6.1 and Nkx6.2 appear to act upstream and regulate Olig2 expression in a dose-dependent manner . In Nkx6.1/Nkx6.2 double null mutants, Olig2 expression is diminished in the developing spinal cord [56, 57]. However, in the dorsal region of the spinal cord, oligodendrocyte formation is largely unaffected in the double mutants. This suggests that oligodendrocyte specification from dorsal and ventral precursors is mediated by independent mechanisms. At present, it is not clear whether oligodendrogenesis defects observed in Nkx6 mutants are due to a patterning defect in the ventral spinal cord or to the specific role of Nkx6 in oligodendrocyte lineage development.
Zinc finger protein family
Members of the zinc finger superfamily have been shown to play an important role in oligodendrocyte differentiation. Among them, myelin transcription factor 1 (Myt1) is cloned by virtue of its binding to cis-regulatory elements of the myelin proteolipid protein (Plp) gene, the most abundantly transcribed myelin genes in the CNS . Myt1 is a zinc-dependent and DNA-binding protein of the unusual Cys-Cys-His-Cys (C2HC) class. Myt1 appears to maintain oligodendrocytes in immature progenitor pools and modulates their transition from progenitors to terminally differentiated oligodendrocytes in vitro . However, the in vivo role of Myt1 in oligodendrocyte differentiation remains unclear.
Another Zn-finger protein Yin Yang 1 (YY1) appears to bind to and enhance the transcription of the Plp promoter . He et al. (2007) identified the consensus DNA binding sequence of YY1 as a common binding motif among the promoters of the genes regulated by histone deacetylases (HDACs) . By specifically ablating YY1 in oligodendrocyte lineage cells, the authors found YY1 to be essential for oligodendrocyte progenitor differentiation by regulating OPC cell cycle exit. In addition, YY1 appears to repress transcriptional inhibitors of oligodendrocyte differentiation including Id4 and transcriptional factor 4 (Tcf4). Tcf4 is a downstream effector that is mediated by the Wnt signaling pathway, whose activation has been shown to repress oligodendrocyte differentiation . In addition, YY1 appears to recruit histone deacetylase-1 (HDAC1) to Id4 and Tcf4 promoters to repress gene expression .
Transcription Factors Modulating Oligodendrocyte Maturation
Although a series of transcriptional regulators have been identified in oligodendrocyte precursors , few have been identified that are uniquely expressed in mature oligodendrocytes. Many transcription factors do not restrict to a specific phase of lineage development, and can be reutilized in different phases during the maturation process. Olig2 is not only required for oligodendrocyte fate specification from the pMN domain, but also for differentiation and maturation of oligodendrocytes [62, 105, 107, 141]. Conditional mutagenesis of Olig2 leads to severe defects in oligodendrocyte myelination in the postnatal developing brain. The mechanism of Olig2 in controlling oligodendrocyte development at different stages is still unclear. Olig2 may interact with different partners such as the ubiquitously expressed E2A bHLH proteins and the non-DNA binding HLH repressors such as Id2 and Id4 [113, 142], or factors that are restricted in the late phase of oligodendrocyte differentiation such as zinc finger protein 488 (Zfp488) . Olig2 may also interact with Nkx2.2 and Sox10 to form a cross-regulatory transcriptional network to promote myelin gene expression and terminal differentiation of oligodendrocytes [104, 125, 144]. Furthermore, Olig2 transcriptional activity can also be regulated by nucleus-cytoplasmic shuffling [112, 145]
In contrast to the loss of OPCs in Olig2 null mice, Olig1 mutation affects OPC maturation. An Olig1 mutant allele produced by the removal of the neomycin selection cassette from the targeting locus exhibits severe myelination deficits during postnatal CNS development . OPCs in the Olig1 mutant extend their processes to contact axons, but they fail to assemble myelin sheaths around axons. These data suggest that axonal recognition and myelination by oligodendrocytes are distinct processes . In addition, a recent study has indicated that remyelination does not occur in adult Olig1 mutant mice with experimentally induced focal demyelination , which highlights the importance of Olig1 during episodes of remyelination. Discrepancies in these studies may be due to different genetic backgrounds and aberrant transcription of neighboring genes of the Olig1 locus influenced by the PGK enhancer in the targeting cassette [146, 148]. In spite of this, all data supports an important role of Olig1 in oligodendrocyte myelination . Recently, coding variants of OLIG1 in demyelinating diseases such as MS have been identified. Currently, a correlation between OLIG1 and the susceptibility for the acquisition of myelin-related disorders remains to be elucidated .
Expression of zinc-finger transcription factor Krox20 is restricted to mature Schwann cells. Given the essential role of Krox20 in PNS myelination [150, 151], the identification of transcriptional regulators that are spatially and temporally limited to mature oligodendrocytes may contribute to unraveling the mechanisms that control the oligodendrocyte myelination process. In a screen for genes that are downregulated in Olig1 null mice, Zfp488, a previously uncharacterized nuclear zinc-finger transcriptional regulator, was recently identified . In the developing CNS, Zfp488 is largely restricted to differentiated oligodendrocytes. Its expression increases in parallel with that of major myelin genes such as Mbp and Plp. In the developing chick neural tube, Zfp488 promotes OPC maturation through its collaboration with Olig2 . Although the nature of Zfp488 interaction with Olig2 is not known, the interaction between zinc-finger proteins and proneural bHLH proteins has been a common theme in regulating terminal differentiation of neural subtypes [152–154].
Gene targeting analysis indicated that Sox10 and Nkx2.2 are required for terminal differentiation of oligodendrocytes in the developing spinal cord. Expression of Nkx2.2 or Sox10 is able to induce myelin gene expression such as Mbp and Plp [124, 134]. Studies of compound heterozygous mutant mice (Sox10+/−; Nkx2.2+/−) suggest that Sox10 and Nkx2.2 cooperate to control oligodendrocyte differentiation . Furthermore, in vitro experiments suggest that combinatorial interaction of different transcription factors such as Olig1/2, Mash1, Sox10, and Nkx2.2 is required for a synergistic activation of myelin gene expression and correct timing of oligodendrocyte myelination [114, 144]. Hence, a high level of myelin gene expression may support the production of myelin sheathes that spirally wrap around axons.
Although many transcription factors influence the myelination process, only a few of them are required for proper myelin ultrastructure and maintenance. For instance, HD protein Nkx6.2 is expressed uniquely in differentiated oligodendrocytes despite its expression in neural progenitors during early neural development. Loss-of-function studies have indicated that Nkx6.2 regulates axon-glial interactions and is necessary for myelin paranode formation through regulating paranodal proteins .
Epigenetic Regulation of Oligodendrocyte Myelination
It has become apparent that factors that induce modifications of chromatin, nucleosomal histones, and non-coding RNAs including microRNAs, termed ‘epigenetic regulation,’ can control cell type-specific gene expression. Recently, several epigenetic factors have been found to regulate neural development and stem cell pluripotency, including histone deacetylation, DNA methylation, and small non-coding RNAs such as microRNAs [156, 157]. It remains to be ascertained as to how these factors contribute to oligodendrocyte differentiation and myelin formation in the CNS.
Myelin specific gene expression is controlled by multiple levels of epigenetic modulation of chromatin components. Several reports indicate that histone deacetylases (HDACs), particularly, class I HDACs including HDAC1, 2, 3, and 8, may be required for oligodendrocyte maturation and expression of myelin-related genes in vitro [158, 159]. Inhibitors of HDACs (i.e. valproic acid and trichostatin A) can repress oligodendrocyte myelination in early postnatal development, although no effect was produced after the onset of myelination . Thus, further research is required to determine which HDAC members modulate oligodendrocyte differentiation and myelination. Conversely, the activity of histone acetyltransferases (HATs) on OPCs inhibits their lineage progression and in turn reverts them to a neural stem cell-like state that is more receptive to neuronal or astroglial differentiation inducing signals . Hence, HDACs and HATs act in concert to regulate, at least partially, OPC proliferation and differentiation. It remains to be determined whether oligodendrocyte development regulated by HDACs or HATs is stage-dependent.
In addition to acetylation and deacetylation of nucleosomal histones, DNA methylation may potentially regulate oligodendrocyte differentiation and myelination, particularly, by modulating the promoter activity of oligodendrocyte-specific transcription factors such as Sox10 . However, the precise mechanisms by which specific epigenetic regulators control the development of oligodendrocyte progenitors into functionally myelinating cells remains to be established. Although few studies have been conducted on epigenetic regulation of Schwann cell development and PNS myelination, uncovering differences in the epigenetic control of myelinating cells will be important for identifying new avenues in the study of myelinogenesis and repair.
The plasticity and self-renewable properties of stem cells holds enormous promise for new therapies for neurological disorders ranging from the devastating demyelinating disease MS or spinal cord injury, to later onset diseases including Parkinson’s and Alzheimer’s. Although stem/progenitor cell based therapies for myelin regeneration in experimental animal models are promising, many challenges remain regarding the identity of cell types and the factors that are activated during the repair process. In this review, we discussed the potential sources of stem/progenitor cells that may contribute to myelin repair, as well as intrinsic and extrinsic mechanisms that modulate multipotent stem/progenitor cell capacity for myelin regeneration. Although curative treatments for demyelinating diseases currently eludes our grasp, such knowledge will increase our understanding of the etiology of myelin related-diseases and will eventually provide the framework for the identification of potential cell types and drug targets for the treatment of demyelinating diseases such as MS.
We thank Ed Hurlock for discussions and Kelly Zhang for illustration. This study was funded by grants from the National Multiple Sclerosis Society, and the National Institutes of Health (QRL) as well as the French MS Society, the European Leukodystrophy Associations, and INSERM (BNO). QRL is a Harry Weaver Neuroscience Scholar and a Basil O’Connor Scholar.