A New Acquaintance of Oligodendrocyte Precursor Cells in the Central Nervous System

Abstract

Oligodendrocyte precursor cells (OPCs) are a heterogeneous multipotent population in the central nervous system (CNS) that appear during embryogenesis and persist as resident cells in the adult brain parenchyma. OPCs could generate oligodendrocytes to participate in myelination. Recent advances have renewed our knowledge of OPC biology by discovering novel markers of oligodendroglial cells, the myelin-independent roles of OPCs, and the regulatory mechanism of OPC development. In this review, we will explore the updated knowledge on OPC identity, their multifaceted roles in the CNS in health and diseases, as well as the regulatory mechanisms that are involved in their developmental stages, which hopefully would contribute to a further understanding of OPCs and attract attention in the field of OPC biology.

Introduction

Oligodendrocyte precursor cells (OPCs, also known as NG2 glial cells) were first described over half a century ago, which were initially termed as either glioblasts, early oligodendrocytes, young oligodendrocytes, immature oligodendroblasts, or oligodendrocyte-type II astrocyte (O2A) progenitor cells, but were later identified and termed as oligodendrocyte progenitors or precursors [1–3]. OPCs are now considered the fourth major type of glial cells [4, 5], constituting 5–8% of glial cells in the adult central nervous system (CNS) [6].

OPCs have received increasing attention over the past years being recognized as a critical component in CNS health and diseases. OPCs can differentiate into oligodendrocytes (OLs) to myelinate axons during development and in adulthood [7] while generating Schwann cells during the repair of CNS demyelination and astrocytes in certain CNS regions [8–10]. OPC is also the only glial cell type in the CNS that forms synapses with neurons [11–13], which impacts OPC developmental events such as migration, proliferation, and differentiation [12, 14–18]. Recent advances also uncover the myelin-independent roles of OPCs, including the regulation of neuronal development and activity, astrocyte maturation, vascular formation, and function, as well as neuroinflammation [13, 19–24]. These discoveries expand our knowledge of the multifaced role of OPCs, cultivating novel research areas in OPC biology. Additionally, with the up-and-coming new technologies, our understanding of the OPC developmental mechanism is also renewed, including the extrinsic and intrinsic regulation of OPC migration, proliferation, and differentiation.

In this review, we will cover the progress in the primary issues surrounding OPCs, including: (1) What are OPCs? (2) The function of OPCs. (3) OPC heterogeneity. (4) The regulatory mechanisms of OPC development. We hope to provide an outline for new researchers interested in this field.

What are OPCs?

The definition of OPCs is still in development since their discovery. The existence of OPCs was first suggested by ultrastructural electron microscopy studies between the 1960s and 1970s [2] and was further identified with the discovery of their markers, NG2, PDGFRα, and A2B5, in primary cultures and in vivo [25–27]. Since then, a more holistic understanding of OPC identity, characteristics, and functions, has been reached through complex biochemical and genetic manipulation, as well as lineage tracing techniques. It is now considered that OPCs as neuroectoderm-origin multipotent progenitors that are NG2/PDGFRα positive, highly ramified, mainly generating OLs, but are also capable of giving rise to other CNS cell types including astrocytes.

Derivation of OPCs

During early embryonic development, oligodendroglial lineage cells are derived from radial glia, which are multipotent neural stem cells near the ependymal cell layer. They generate OPCs through asymmetric division [28], and OPC specification is regulated by several transcription factors, most notably Olig2 and Sox10, which together define the oligodendroglial lineage [29]. OPCs initially emerge in the CNS during middle gestation from Olig2+ neuroepithelial progenitors found in the ventricular zones of the spinal cord and brain [7]. During this specification progression, OPCs emerge in three originating waves from distinct domains in different regions of the forebrain, firstly from the medial ganglionic eminence at embryonic day (E)12.5, then the lateral ganglionic eminence at E15.5, and finally the dorsal cortex at birth [30, 31]. They then migrate extensively throughout the whole CNS along the brain vasculature as their guiding scaffolds [32]. Furthermore, they can also emerge indirectly through intermediate progenitors, such as glial restricted precursors (GRP), and potentially from other types of immature neural cells, including OPC pre-progenitors or pre-OPCs [33–35]. These pathways illustrate the complex and multifaceted origins of OPCs within the CNS, highlighting their crucial role in the development and maintenance of neural networks. The oligodendrogenesis process is not limited to prenatal stages but continues postnatally and into adulthood, with new OPCs being generated from neural stem cells in the forebrain's subventricular zone (SVZ) [36–39], or from Nestin+/CD13+ immature pericytes/mesenchymal cells in meninges and perivascular regions [40], upon demyelinating insults or ablation of OPCs.

Markers and Morphology

Various markers have been identified to label OPCs in different developmental states. PDGFRα and NG2 have been used to identify OPCs [27, 41–43], but recent evidence has challenged their veracity as exclusive OPC indicators. For example, NG2 is also a marker of pericytes and is expressed by reactive macrophages/microglia following injury, leading to potential misinterpretation [44–47]. Labeling with PDGFRα suffers from similar complications. PDGFRα, an essential surface receptor involved in OPC proliferation, and migration and as a regulator for timely differentiation [48–50], is also expressed by vascular and leptomeningeal cells (VLMCs) [51]. Nonetheless, NG2 and PDGFRα remain the most reliable and widely used markers for OPCs, which could reveal their ramified morphology, a characteristic distinguishing OPCs from NG2-positive pericytes or PDGFRα-positive VLMCs, which display a flat morphology and are lack of processes.

In addition, transcription factors NKX2.1 and DLX2 can label early OPCs in the human telencephalon [52], while O4 labels the non-migratory, late OPCs [53]. The arrival of the single-cell era also provides novel markers for identifying OPCs as well as different stages of oligodendroglial lineage cells. In the developing human brain, single-cell RNA seq identified a transitional cell type lying between multi-potent neural progenitor cells and lineage-committed early OPCs, defined as “pre-OPC”, which possesses expression of both neural progenitor cell markers including GFAP, VIM, NES, HES1, NOTCH2, and EGFR, and oligodendroglial lineage markers OLIG1, OLIG2, and PDGFRα [54]. The same study also identified a potential novel OPC marker PCDH15, which regulates daughter cell separation after division [54]. In addition, differentiation-committed OPCs were also identified, with a unique expression pattern of Tcf7l2, Itpr2, Tmem2, Gpr17, and Pdgfa [51]. Among them, TCF7L2, ITPR2, and GPR17 are adopted as a marker for differentiation-committed OPCs or newly-formed OLs [51, 55]. In addition, activated OPCs in CNS injury highly express Rnf43, which is then downregulated during OL differentiation; RNF43 is thus considered a marker of reactive OPCs [56].

OPCs display variable morphologies, as reflected by the immunofluorescent staining of PDGFRα (Fig. 1). In early development (embryonic stage till birth), OPCs exhibit an elongated shape with bipolar processes during migration along brain vasculature [32]. Later (at around postnatal day 7 (P7)), they change to a branched morphology upon arrival at their destination, interacting with surrounding neural axons and other glial cells [5]. Live imaging using two-photon microscopy showed that adult OPCs possess elegantly ramified but slightly less branched processes (compared to OPCs at P7) in the mature CNS, where they constantly survey the environment within their territories to respond to local changes in tissue homeostasis [57]. Upon CNS injuries, OPCs either convert to a hypertrophic morphology with increased cell processes and upregulated NG2 expression level, or return to a bipolar migratory shape to facilitate their recruitment into lesion areas [22, 58, 59].

Fig. 1.

Morphology changes of OPCs. A PDGFRα immunofluorescent staining and skeletonized image (right panel) visualize the altered OPC morphologies at different developmental stages. At postnatal day 0 (P0), most OPCs had a migratory morphology with some short and sparse protrusions (A1). At P7, the morphology of OPC processes was more complex, with more fine main protrusions and longer branches (A2). At P28, OPC protrusions and branches decreased (A3). Stars mark OPC soma. Immunostaining images, for illustration purposes, are courtesy of the laboratory of J.N. B Both migratory morphology OPCs with simple protrusions (B1) and reactivated OPCs with complex protrusions (B2) are visualized after lysolecithin-induced demyelinating injury at 3-day post lesioning (3 dpl). Stars mark OPC soma. Immunostaining images, for illustration purposes, are courtesy of the laboratory of J.N. C Schematics of OPC morphological changes.

The Function of OPCs

OPCs are a unique type of glial cell in the CNS, which could replenish the OL population for myelination or remyelination, transdifferentiation into astrocyte and Schwan cells, and could form synaptic connections with neurons. Recent advances also reveal the myelin-independent roles of OPCs. These characters further distinguish them from other glial cells, neurons, and neural progenitor cells [4, 6, 60].

Dynamic Architects in CNS Regeneration and Plasticity

OPCs are multipotent progenitors. While they mainly differentiate into OLs, they are also capable of generating astrocytes or transforming into a Schwann cell-like state in development and/or in CNS injury (Fig. 2A–D).

Fig. 2.

Origin, function, and heterogeneity of OPCs. A Developmentally, OPCs differentiate into OLs and form myelin that enwrap axons. B and C Some OPCs maintain their undifferentiated state into adulthood, become adult OPCs, and populate the entire adult CNS. Adult OPCs continue to generate OLs and perform remyelination under physiological and pathological demyelination conditions. D Adult OPCs also have the potential to differentiate into astrocytes in diseases. E OPCs are the only glial cell types that can form synaptic connections with neurons, including the neuron-OPC synapse as well as the OPC-neuron synapse. F OPCs display regional, sex, and age heterogeneity in the CNS.

(Re)myelination

During development, OPCs proliferate and differentiate into OLs to perform developmental myelination [7]. However, around 20% of OPCs remain in an undifferentiated state throughout the developmental process and convert to adult OPCs in the mature CNS, where they act as multifaceted regulators of CNS health and disease [61–63]. Many questions remained open about adult OPCs, including the difference between developmental OPCs and adult OPCs, and how adult OPCs maintain a quiescent, multipotent state. On the latter issue, it was proposed that contact with vasculature may participate in the maintenance of adult OPCs, as vascular endothelial cells are inhibitory for OPC differentiation [64]. Adult OPCs continue to generate OLs throughout life, albeit the rate decreases with aging [7]. While the innate OPC differentiation and myelination could take place independent of neuronal activity, accumulating evidence showed that neuronal activity could remodel myelin formation, a process called adaptive myelination, which involves changes in the thickness, length, and/or number of myelin sheaths on the axons [65, 66]. OPCs are actively involved in adaptive myelination via de novo OL generation [67]. This dynamic change of myelination in turn regulates signal transmission in neural circuits and implicates emotions and cognitive functions [68, 69].

OPCs also play a pivotal role in remyelination in demyelination injury. OPCs can sense myelin damage-related signaling and migrate into demyelinated lesions [70]. Genetic fate-mapping showed that upon demyelination by local injection of lysolecithins in the corpus callosum, parenchymal OPCs were rapidly recruited into demyelinating areas within 7 days and produced mature OLs within 14 days [71]. Interestingly, adult neural stem cells in the SVZ also contribute to remyelination in the adjacent demyelinated regions. Adult neural stem cells transit into OPCs and rapidly differentiate into OLs to aid remyelination [37–39]. However, compared to the parenchymal OPCs, neural stem cell-derived OPCs respond to the demyelination injury more slowly and have less contribution to the de novo OL generation [71, 72].

Another controversial issue surrounding remyelination is whether adult OLs also participate in the remyelinating process [73–75]. In an attempt to solve this, a study investigated the source of remyelinating OLs in human patients with MS using a ¹⁴C incorporation-based technique but showed that no new OLs were detected in remyelinating lesions [76], which is consistent with a previous non-human primate study [77]. While these studies suggest that residual OLs in the lesions may participate in remyelination, a recent report showed that compared to new OLs, the survived OLs have defective myelination regeneration, as they displayed limited myelination capacity [78] and could misguidedly myelinate neuronal cell body [79]. Meanwhile, several other studies failed to observe the contribution to remyelination from the pre-existing OLs in the galactocerebroside-antibody plus complement-induced demyelination model [80] or the lysolecithin-induced demyelination model [81]. Overall, these studies support the importance of OPCs in the remyelination process. However, one of the critical issues in human demyelination diseases is the failure of OPC migration into the demyelinated site [22, 82]. Therefore, stimulating OPC recruitment should benefit remyelination in demyelinated lesions.

Transformation and Conversion

It has long been proposed that OPCs have the potential to transdifferentiate into astrocytes in primary culture [26], hence OPC was initially named oligodendrocyte-type II astrocyte (O2A) progenitor cells [1]. However, observation in vivo using genetic fate mapping revealed that the transdifferentiation potential of OPCs is limited, and can only take place during early development or in CNS injury. Embryonic OPC transdifferentiation is more prominent in the ventral brain, while it is absent in the dorsal cortex [8]. This plasticity of OPC is regulated by transcription factor Olig2: knockout of Olig2 in NG2+ OPCs resulted in their transdifferentiation into ALDH1L1+ astrocytes [8]. Such Olig2-regulated OPC plasticity declines as mice reach adulthood [83]. In addition, engrafted human fetal OPCs, or a murine cerebrum OPC cell line, in the postnatal mouse brain could also differentiate into both OL and GFAP+ astrocytes [84, 85].

In CNS injury, OPCs can be regulated by signal molecules and transformed into astrocytes [1]. In a chronic ischemia model that yields severe myelin loss, BMP4-derived from pericytes promoted OPC transformation into astrocytes, which was proposed to hinder remyelination [86]. In the traumatic brain injury model, OPCs gave rise to astrocytes within 24 h post-injury but mainly differentiated into OLs to contribute to remyelination after 7 days [87]. The role of OPC-astrocyte conversion in CNS injuries is under debate. On the one hand, transdifferentiation of OPCs was proposed to hinder the process of remyelination, and targeting OPC transdifferentiation could be an effective strategy for the treatment of demyelination-related diseases and injuries [86]. On the other hand, OPC transdifferentiation might aid the formation of a glial barrier surrounding the injured niche that limits the spread of inflammation signal, which would facilitate post-injury tissue repairment [88, 89].

In spinal injury, OPC could also transdifferentiate into a Schwann cell-like state, expressing Schwann cell markers such as Periaxin (P0), and participating in myelin regeneration in the CNS [9, 90]. However, the identity of these Schwann cell-like states, as well as the regulatory mechanism, await further studies.

While it is recognized that OPC transdifferentiation into astrocyte and Schwann cells implicates disease progression, it remains unclear how OPC possesses the ability to transdifferentiate to astrocyte or Schwann cells. A combination of extracellular signals and intracellular regulation, e.g., epigenetic mechanism, might be at play to maintain OPC in a multipotent state, which would switch to the astrocyte- or Schwann cell fate upon pathological insult.

Neuron-OPC Synaptic Connection

OPC is the only resident glial cell type that forms direct synaptic connections with neurons in the CNS (Fig. 2E). Immuno-electron microscopic analysis revealed that silver-intensified gold reaction-labeled OPC processes approached neuronal dendritic spines and formed thinner postsynaptic membranes, while electrophysiological evidence confirmed a synaptic connection between glutamatergic neurons and OPCs [11]. The synaptic input from excitatory neurons to OPCs promotes myelination [14]. Meanwhile, OPCs also form inhibitory synapses with GABAergic interneurons [12]. GABAergic input to OPCs transiently suppresses the AMPA receptor-mediated glutamatergic current [12]. Additionally, functional inhibitory synapses between GABAergic interneurons and OPCs regulate various features of OPC development, including survival, proliferation, differentiation, myelination, and the length of CNS internodes [12, 15–18]. Recently, a study utilizing a novel monosynaptic retrograde tracing technique further revealed that local OPCs receive inputs from multiple brain regions, which may promote myelination to support brain-wide circuits [14, 91]. The post-synaptic elements in OPCs gradually reduce when they differentiate into OLs [92]. It remains a puzzle why both glutamatergic and GABAergic input promote OPC differentiation and myelination, and if other neuronal cell types, e.g., cholinergic or dopaminergic neurons, could form a synapse with OPC, and thus regulate OPC differentiation.

More importantly, why OPC is necessarily the only glial cell type that forms synapses with neurons, instead of just reacting to spillover neurotransmitters like other glial cells, such as astrocytes [93, 94]? From a parsimonious viewpoint, neuron-OPC synapse should bear physiological significance. It may be required for the precisive myelin formation, which is hinted by a recent study in the zebrafish model, showing that GABAergic post-synaptic hotspots in OPCs predict a subset of future myelin sheath formation sites [92].

Apart from the neuron-OPC synapse, recent research also suggested that OPCs can form functional GABAergic synapses onto neurons in the hippocampus of adult mice, and a single OPC is capable of forming approximately 146 presynaptic structures per cell for signal transduction to neurons in the hippocampal CA1 region [13], and receive inputs from 141 different axons in white matter [95]. Bidirectional communication between OPCs and interneurons has since been also discovered, where GABAergic interneurons regulate OPC myelination, and OPCs can also shape the inhibitory neuronal network through GABA receptors and the cytokine tumor necrosis factor-like weak inducer of apoptosis (TWEAK) signaling pathways [13, 19].

Myelin-Independent Functions

The last few years have witnessed a surge of research uncovering the myelin-independent roles of OPCs in health and diseases, which have been well-summarized in several reviews recently [96–99].

OPCs have been shown to modulate neuronal migration, axon arborization, and neural activity, and participate in synapse engulfment and synaptogenesis. Developmental OPC ventral-dorsal perivascular migration acts to steer the migration of their predecessor, the ventral-originated interneurons, away from the blood vessel to facilitate their dispersion to the dorsal brain [100]. OPC also regulates the arborization of axons and fine-tuning the neural circuit formation [101]. The aforementioned OPC-neuron synapse promotes the tonic GABA release to promote inhibitory synapse transmission [13]. The neuron-OPC synapse, on the other hand, may have a reciprocal effect on neurons, as conditional knockout of GABA receptor subunit GABA_BR in OPC using the NG2-CreERT line resulted in reduced inhibitory neuron apoptosis, leading to an increased inhibitory neuron density in the prefrontal cortex [19]. Electron microscopic evidence shows that OPC fine processes are frequently in touch with axons and contain numerous phagosomes, which suggests an axon-trimming role of OPCs [102]. In vivo, phagocytosis assay confirms that OPC could engulf synapses to participate in circuit remodeling [103]. In a mouse model of schizophrenia, pathological OPCs suppressed hippocampal neuron synaptogenesis via secretion of WIF1 [104].

OPCs also regulate other brain cell types and implicate neuroinflammation in disease conditions. During development, OPC-derived Wnt ligands Wnt7a/7b promote angiogenesis [20] as well as the maturation of the astrocyte network [21]. In demyelinating diseases, OPCs could directly interact with the vasculature and interrupt the blood-brain barrier function, aggravating the neuroinflammatory condition [22]. In addition, OPC is not only a “receiver” of inflammatory signals, but also an “instigator”: while OPCs respond to inflammatory cytokines or chemokines, which may disrupt the remyelination process, reactive OPCs could also release immune factors such as CCL2 and IL1β, and present antigens, which may participate in the recruitment of peripheral immune cells and promote the inflammatory response [23, 24].

The myelin-independent role of OPC and the underlying mechanism remain largely unexplored and are becoming an exciting field for future studies.

OPC Heterogeneity

OPCs are a heterogeneous glial cell population with various functions both in development and in the adult CNS [105–108] (Fig. 2F).

Regional Heterogeneity

OPCs exhibit heterogeneity across different brain regions. One important evidence comes from the radiation-mediated OPC-depletion experiments, where OPCs in different brain regions show diverse radiation resistance and post-radiation proliferative capability. In the cortex, corpus callosum, and hippocampus, OPCs were almost completely depleted 7 days post-radiation, followed by a progressive repopulation from the non-irradiated area [109]. In contrast, OPC loss occurred slower in the central regions of the brain (midbrain and hypothalamus), with 26% of OPCs still present 4 weeks after radiation [109]. Transplanted and endogenous OPCs quickly established in the depleted cortex, but not in the partially depleted central regions, which suggests a degree of regional heterogeneity [109]. A similar heterogeneous response to stress was also observed in a DNA damage model by ablation of citron-kinase or cisplatin treatment [110]. OPCs in the dorsal brain are more susceptible to DNA damage and rapidly undergo apoptosis, while the ventral ones only display senescence phenotype [110]. In demyelination models, dorsal OPCs outperformed their ventral counterparts in proliferation, recruitment, and differentiation [111, 112]. A recent lineage tracing study in mice made a surprising discovery that the dorsal-originated OPC population continuously expanded throughout life, while the ventral-originated one gradually diminished; at 6 months of age, dorsal-originated OPCs can account for almost all OPCs in the brain [61].

There is still controversy over the origin of such regional OPC heterogeneity. Some argue that it stems from the difference in the developmental origin of dorsal and ventral OPCs [61, 111, 112]. On the contrary, it is also proposed that the local environmental cues are the main contributors [110], which was well supported by the aforementioned OPC transplantation experiment [109]. In addition, while the behavioral differences between OPC in dorsal and ventral regions are well documented, their molecular bases await further studies.

OPC heterogeneity was also observed between grey matter and white matter, displaying different morphology and responsiveness to PDGF [113]. A study in the zebrafish model further illustrates how neuronal-rich area and axon-rich area shape OPC heterogeneity, characterizing the difference in transcriptome, morphology, and differentiation kinetics [108]. Morphology-wise, OPCs with somata residing in the neuron-rich area form a more elaborate process network than those within the axon-rich area [108]. Moreover, live-imaging showed that OPCs within the axon-rich area had higher process dynamics, as they completely remodeled their entire process network in 1 h, whereas their neuron-rich area counterparts only remodeled the process tips [108]. Differentiation-wise, OPCs in the axon-rich area tend to differentiate rapidly into myelinating OLs, while OPCs in the neuron-rich area proliferate in response to neural activity, and migrate to plenish the axon-rich area OPC pool [108]. These observations in zebrafish reveal how the microenvironments in the neuron-rich grey matter and axon-rich white matter may shape OPC heterogeneity.

Age Heterogeneity

Gene tracing studies indicate that the age-related decrease in the number of OPCs may result in reduced myelin renewal and memory impairment in mouse models [114]. Single-cell electrophysiological recordings showed that OPC heterogeneity becomes more prominent with age, even though they start as a relatively homogeneous population. In immunostaining studies, a subset of OPCs lacking OLIG2, ranging from 2 to 26%, has been identified in various brain regions including the cortex, corpus callosum, CA1, and dentate gyrus. These OLIG2-negative OPCs are enriched in the young brain and decrease with age and are rarely detected in the adult brain [115]. Single-cell transcriptome analysis also revealed that OPC molecular signatures change with age, which is reflected by that embryonic OPCs possess a higher expression of migrating signatures (expressing genes such as DCC and EPHB2) compared to OPCs from postnatal or older mice. OPCs within 2 weeks of life had stronger proliferative signatures (for example, PDGFRα and PTCH1) but reduced migratory molecular signatures. Molecular signatures of both differentiation and proliferation begin to decline from 2 weeks after birth to the following months and decline further in aged mice [116]. In another single-cell transcriptome dataset, four OPC cell clusters were identified that distinguish young and old mouse OPCs. Some subsets of OPC clusters are closely related to cell division and differentiation, but their numbers decrease in the aging brain, in-depth analysis of single-cell RNA sequencing data from all brain cells from young and old mice revealed upregulation of several aging-related genes in a subset of OPCs in the aging brain [117].

Sexual Heterogeneity

OPCs exhibit differences between genders in terms of proliferative and differentiative potential, metabolism, regional heterogeneity, and response to pathological insults. In primary culture, the proliferation and migration ability of OPCs from female animals is higher, whereas the differentiation ability of OPCs from male animals is stronger [118]. Similar differences in proliferation and differentiation capability can be seen in OPCs from different genders in vivo [118, 119], suggesting that the ability of OPCs to proliferate and differentiate is closely related to gender. OPCs exhibit gender-related differences in energy metabolism: the intracellular ATP level is significantly higher in female OPCs than in male OPCs [118]. A recent single-cell RNA sequencing study has shown that male OPCs are more abundant in the interventricular septum than females, indicating gender-specific regional differences in the adult brain [120].

Gender differences in OPCs may also play an important role in the susceptibility and response to neurological and psychiatric diseases. In Alzheimer's Disease (AD), OPCs and OLs show gender-specific differences in the transcriptomic response to AD pathology. Single-cell RNA sequencing from post-mortem human brain tissue shows that OLs from males upregulate transcriptional genes in pathological reactions, while OLs from female patients do not [121]. Additionally, OPCs from females tend to downregulate transcriptional genes in pathological reactions compared to their male counterparts [121]. It is suggested that this may be partially due to differences in OPC proliferation and differentiation caused by sex hormones or biological gender [1]. Similarly, OPCs have gender-specific differences in cell death in response to ischemic stress. Male OPCs are more susceptible to in vitro oxygen-glucose deprivation. After 7 h of oxygen-glucose deprivation (OGD), the cytotoxicity and activity of OPCs were evaluated in male OPCs, and measurements of the cytotoxicity ratio (cytotoxicity/activity) showed that the cytotoxicity of male OPCs is significantly higher than that of female OPCs after OGD [118]. As gender-dependent OPC responses to neurological conditions correlate with the gender-difference in disease-susceptibility, the related mechanistic investigation would yield valuable insight into disease pathogenesis.

The Regulatory Mechanisms of OPC Development

OPC development involves several critical steps, including migration, proliferation, differentiation, and finally myelination. These steps tightly connect, and are precisely regulated (Table 1).

Table 1.

Molecular regulations for OPC development

Development	Role	Type	Factor	Evidence
Migration	Attraction	Extracellular factor/signaling	PDGF [129–132]; FGF [132]; SEMA3F [137–139]; HMGB1 [140]; SHH [123]	In vitro
	Repulsion	Extracellular factor/signaling	BMP [122]	In vitro
			SEMA3A [137–139]	In vitro; In vivo
			CXCL1 [133]; Netrin1 [134, 135]	In vitro; Ex vivo; In vivo
Proliferation	Promotion	Extracellular factor/signaling	PCDH15 [54]	Ex vivo
			FGF2 [146, 149]; IGF1 [147]; GGF [150]	In vitro
			BDNF [148]; NT3 [148, 149]; PDGF [48, 144, 145, 149]	In vitro; In vivo
		Channel proteins	Cx29, Cx47 [143]	In vitro
		Transcription factors	ID2, ID4 [172, 196–198]; SOX2, SOX5, SOX6 [199–203]; SRF [153]	In vitro
		Epigenetic regulation	SIRT1 [214]; SIRT2 [215]	In vitro; In vivo
	Suppression	Extracellular factor/signaling	TGF-β [151]; BMP [152]	In vitro
		Transcription factors	ASCL1 [193]	In vivo
Differentiation	Promotion	Extracellular factor/signaling	IGF-1 [164, 165]; CNTF [166]; FGF [167]	In vitro
		Transcription factors	OLIG2 [8, 175, 176, 178, 179]; OLIG1 [182]; SOX10 [183, 184]; MYRF [187]; ZFP24 [190, 191]; PPARγ [206]; RXR [207]	In vitro; In vivo
			TCF7L2 [156, 192]; ASCL1 [193]; FOXG1 [204, 205]	In vivo
		Epigenetic regulation	SETDB1 [179]; CHD7, CHD8 [208–210]; DNMT1 [219]; TET1 [222]; DICER1 [223, 224]	In vivo
			PRMT5 [211–213]; SIRT2 [215]; miR-219, miR-338, miR-138 [223–225]; miR-146a [226]; miR-21-5p [227]	In vitro; In vivo
	Suppression	Extracellular factor/signaling	NOTCH [168, 169]; Wnt/β-catenin [156–159]; BMP [172–174]	In vivo
		Transcription factors	HES1, HES5 [194, 195]; ID2, ID4 [172, 196–198]; SOX2, SOX5, SOX6 [199–203]	In vitro
		Epigenetic regulation	SIRT1 [214]	In vitro; In vivo

Regulation of OPC Migration

OPCs migrate throughout the CNS and are guided by a variety of molecules, such as growth factors, morphogens [122, 123], neurotransmitters or neuromodulators [124–126], as well as extracellular matrix [127, 128]. Chemo-attractive and chemo-repulsive factors co-ordinate to regulate OPC migration and intricately determine their positions. For instance, PDGF [129–132], FGF [132], and SHH [123] represent a chemo-attractive force, while BMP [122], CXCL1 [133], and Netrin1 [134, 135] repel OPCs to direct ventral-dorsal migration, as well as to stop OPC migration when reaching their destination. Semaphorins (SEMA) are a family of membrane proteins that primarily regulate cell chemotaxis [136]. In development and demyelination diseases, it was shown that SEMA3A suppresses OPC migration while SEMA3F promotes OPC migration [137–139]. In demyelinating injury, High mobility group box-1 (HMGB1) acts as an autocrine chemo-attractant for OPC migration [140].

Recent advances highlighted the critical contribution of cell-cell interaction in the delicate control of OPC migration. In the brain or spinal cord, migrating OPCs use blood vessels as physical scaffolds to rapidly disperse throughout the neural tissue or toward demyelinated lesions [22, 32]. Later in development, the peri-vascular OPC migration is ended by the formation of an astrocyte endfeet that suppresses the interaction between OPC and vasculature [64]. However, further studies are required to address the distance and speed of OPC migration, as well as any differences between migration during development and regeneration.

Regulation of OPC Proliferation

Various factors influence the proliferation of OPCs during their development. OPC density is one of the critical regulators of OPC proliferation. Early evidence showed that in primary culture, high seeding density suppressed OPC proliferation [141]. OPCs in vitro display interesting properties that after cell division, the daughter cells rapidly repulse each other and migrate to the opposite direction, distinct from the LIFR+ or EGFR+ neural progenitors that tend to stay in contact with each other after division [54]. This self-repulsion of OPC is mediated by a membrane-located protocadherin protein PCDH15; failure of self-repulsion in PCDH15 KO OPC is accompanied by a dampened OPC proliferation [54]. These data suggest OPC density exerts a contact-inhibitory effect on OPC proliferation. Such balance between cell growth and self-repulsion persists till adulthood to maintain the OPC population [57]. However, our knowledge of the underlying mechanism remained rudimentary.

OPC proliferation is also sensitive to energy supply. OPC-expressed connexin proteins Cx29 and Cx47 form hemichannel and absorb energy substances such as glucose to maintain cell proliferation [142, 143]. During hypoglycemia, brain levels of high-energy phosphate molecules decrease, cell membranes release free fatty acids, intracellular calcium levels increase and membranes depolarize. We and other researchers have observed that the proliferation and migration of OPCs are inhibited at low glucose levels, suggesting that OPCs utilize glucose as an important energy source for growth and motility [142, 143].

OPC proliferation is dynamically regulated by growth factors or morphogens. Growth factors that promote OPC proliferation include PDGF [144, 145], FGF2 [146], IGF1 [147], brain-derived neurotrophic factor (BDNF) [148], neurotrophic factor 3 (NT3) [149], and glial cell-derived growth factor (GGF) [150]. Factors that inhibit OPC proliferation include transforming growth factor-beta (TGF-β) [151] and BMPs [152]. However, most of these conclusions are derived from in vitro data. The in vivo contribution of different growth factors in OPC proliferation during development and diseases required further investigation.

Transcription factors and epigenetic mechanisms also play a regulatory role in the proliferation of OPCs. Transcription factors such as SRF (serum response factor) [153] promote the proliferation of OPCs. The epigenetic mechanisms related to the proliferation of OPCs mainly include DNA methylation, histone modification, and non-coding RNA. These mechanisms can influence gene expression and thus regulate the proliferation of OPCs [154].

Regulation of OPC Differentiation

Extracellular Cues and Signaling Pathways

Extracellular factors play important roles in OPC differentiation via activation of intracellular signaling pathways. These include the Wnt/β-catenin signaling pathway, in which Wnt ligands act on the transmembrane Lipoprotein receptor-related protein (LRP)-Frizzled receptor complex, leading to the stabilization and nuclear translocation of β-catenin, which then binds with transcription factor TCF to activate gene transcription [155]. The Wnt/β-catenin signaling pathway plays a complex role in OPC differentiation. Hyper-activated Wnt/β-catenin signaling would perturb OPC differentiation [156–159]. However, the Wnt/β-catenin signaling pathway is required for the expression of TCF7L2 [160, 161], a transcription factor and an effector of Wnt/β-catenin signaling, which propel OPC differentiation and myelination. One key mechanism involves the induction of Axin2 by Wnt/β-catenin-TCF7L2 signaling, which suppresses Wnt signaling as a negative feedback loop, allowing subsequent OPC differentiation [162, 163].

Several extracellular factors have been shown to promote OPC differentiation. Insulin-like growth factor-1 (IGF-1) could suppress BMP signaling to instruct oligodendroglial lineage specification from neural progenitor [164], while in OPCs, IGF-1 and extracellular matrix protein Laminin cooperate to activate the MAPK pathway in the promotion of OPC differentiation [165]. In addition, ciliary neurotrophic factor (CNTF) is a neurotrophic factor produced by neurons and glial cells, which can promote OPC differentiation into mature OLs through the gp130-JAK signaling pathway [166]. FGF was proposed to be a “late stage” signal, acting on FGF receptor 2 of OLs to regulate myelin thickness through activation of the MAPK-mTOR pathway [167].

On the other hand, the Notch signaling pathway suppresses OPC differentiation and maintains the OPC state. At the early stage of OPC development, JAGG1(or Jag-1) is highly expressed on axonal membranes and can activate the NOTCH receptor on OPCs, leading to the cleavage of the NOTCH receptor and the nuclear-translocation of its intracellular domain, which promotes the expression of transcription factor HES5, and inhibit OPC differentiation and myelin formation [168, 169].

BMP signaling is another important pathway in regulating OPC differentiation. BMPs signal through serine/threonine kinase receptors on the membrane, leading to the phosphorylation and nuclear-translocation of SMAD proteins and the subsequent transcription regulation [170]. Activation of BMP signaling in OPCs by BMP2/4 causes transdifferentiation to astrocytes via an upregulation of transcription factor ID2/4 and a downregulation of OLIG1/2 [171–173].

Transcription Factor-Mediated Regulation

The transcription factor regulatory network that controls OPC differentiation has been intensively studied, revealing several key regulatory factors, with OLIG2, SOX10, Nkx2.2, ZBP24, and MYRF being the major determinants [174]. OLIG2, a basic helix-loop-helix transcription factor, is a master regulator of oligodendrocyte lineage cells. OLIG2 ablation in OPCs using the NG2-CreERT line during development causes loss of OPC marker expression and the transdifferentiation to astrocyte, as well as the subsequent hypomyelination [8, 175], while overexpression of OLIG2 enhanced OPC differentiation [176]. Mechanistically, OLIG2 interacts with Nkx2.2 [177] and engages with SWI/SNF chromatin remodeler SMARCA4 (also known as BRG1) to promote gene expression that involves the specification of oligodendroglial lineage as well as OPC differentiation [178]. A recent report also reveals that OLIG2 could also recruit SETDB1 to achieve gene silencing, which is also essential for OPC differentiation and myelinogenesis [179]. However, Plp-CreER-mediated conditional knockout of OLIG2 in OLs facilitates myelination [180], possibly via the compensating upregulation of OLIG1 expression. OLIG1 is a transcription factors that show structural similarity with OLIG2; OLIG1 and OLIG2 display coordinated expression in OPC during development, and their functions are partially redundant [181]. Interestingly, as OPCs differentiate, phosphorylation of OLIG1 would cause its translocation from the nucleus to the cytoplasm, which is crucial for OL morphological maturation and myelination [182].

OLIG2 directly promotes the expression of SRY-box 10 (SOX10), a transcription factor crucial for OL maturation [183, 184], which in turn forms a positive feedback loop to maintain OLIG2 expression [185]. SOX10 is required for expression of myelination-related genes, including transcription factors MYRF and ZFP24; a complex positive feedback network among OLIG2, SOX10, MYRF, and ZFP24 is crucial for normal OPC differentiation [178, 186–191].

Transcription factor TCF7L2 is a recently identified newly-formed OL marker. It is an effector of Wnt/β-catenin signaling and interacts with transcription co-repressor ZBTB33 to block β-catenin signaling, a negative feedback regulation that is required for differentiation and myelination [156]. TCF7L2 also cooperates with SOX10 to promote the expression of MYRF and ZFP24 [192].

Apart from this OPC-specific regulation, transcription factors that have been at play since the neural progenitor stages also implicate OPC differentiation. This includes ASCL1 which balances OPC proliferation and differentiation: Pdgfrα-CreERT2-mediated conditional knockout of ASCL1 promoted OPC proliferation and impeded differentiation [193]. Transcription factors HES1 and HES5 are effectors of the Notch signaling pathway, which negatively regulate ASCL1 expression and suppress OPC differentiation [194, 195]. ID2 and ID4 transcription factors act downstream of BMP and GPR17 signaling to promote OPC proliferation and suppress OPC differentiation, possibly via the sequestration of OLIG1/2 [172, 196–198]. In addition, SOX2/5/6 inhibits OPC differentiation and promotes proliferation [199–203]. Recent research showed that forkhead box transcription factor FOXG1 in OPC promotes cell cycle exit and differentiation in development and demyelinating injury [204, 205].

OPC differentiation is also regulated by nuclear receptors that are ligand-activated transcription factors, such as PPARγ and RXR, which are required for differentiation and myelination [206, 207].

Epigenetic Regulation

Transcription factors team up with epigenetic modifications to regulate OPC differentiation, which include chromatin remodeling, histone modifications, DNA methylation, and non-coding RNA. Transcription factor Olig2 recruits chromatin modifiers such as the BRG1 complex to alter the chromatin-accessibility landscape, which promotes or silences target gene transcription [178]. Olig2 also recruits an H3K9 tri-methyltransferase SETDB1 to achieve gene silencing and promote OPC differentiation [179]. Several other chromatin remodelers such as CHD7 and CHD8 have been shown to associate with transcription factors such as SOX10 to promote OPC differentiation [208–210].

Experiments were also performed to examine the role of histone modification enzymes in the regulation of OPC differentiation. PRMT5, which is a methyltransferase of histone 4 arginine 3 (H4R3) and histone 3 arginine 8 (H3R8) and achieves gene silencing, was proposed to promote OPC differentiation and myelination via suppression of ID2/4 expression in primary culture [211]. PRMT5 was further confirmed to be necessary for OPC survival and differentiation in vivo in Olig1-Cre or Olig2-Cre-mediated conditional knockout models [212, 213]. It was further shown that reduced histone arginine methylation by Prmt5 knockout was followed by increased histone lysine acetylation, and that defective myelination in Prmt5 knockout OPC could be rescued by histone acetyltransferase (HAT) inhibitor [212]. Histone acetylation is established by HATs, while acetyl removal is maintained by histone deacetylases (HDACs). SIRT1, the nicotinamide adenine dinucleotide (NAD)-dependent class III HDAC, promotes OPC proliferation while suppressing differentiation as suggested by the knockdown experiment [214]. In contrast, SIRT2 could promote OPC differentiation [215]. It is interesting that while both are histone deacetylases, SIRT1 and SIRT2 have opposite roles in OPC differentiation, which might be due to their alternative interactors [215]. In the adult brain, deactivation of SIRT1 in neural stem cells promotes the expansion of oligodendroglial lineage cells but does not alter OPC differentiation or myelination [216]. Interestingly, HDAC can also regulate OL development in a histone-independent manner, such as by the deacetylation of transcription factors [217, 218].

DNA methylation usually correlates with gene silencing and is dynamically regulated during OPC differentiation. Increased DNA methylation level was found in OLs compared to OPCs, accompanied by a decreased expression of DNA methyltransferase DNMT1; however, knockout of Dnmt1 in oligodendroglial cells by Olig1-Cre line causes a defective OPC differentiation and the subsequent hypomyelination [219]. It is proposed that the non-selected DNA hypomethylation by ablation of DNMT1 causes defective alternative splicing and the subsequent ER stress, thus leading to impeded OPC differentiation [219]. Thus, the role and mechanism of dynamic DNA methylation in OPC differentiation remain to be studied. It was also shown that differentiation inhibiting factors ID2/4 are one of the subjects under DNA methylation control, which reduces their expression to facilitate OPC differentiation [220]. Age-related myelin loss might be also regulated by DNA methylation, as the level of genomic DNA methylation and the activity of methyltransferases gradually decrease in spinal cord OPCs during aging [221]. DNA hydroxymethylation, another form of DNA modification, is an intermediate product generated during demethylation catalyzed by the ten-eleven translocation (TET) proteins. Interestingly, Tet1 ablation in oligodendroglial lineage cells also causes defective proliferation and differentiation, leading to hypomyelination in adolescent mice, which was then normalized in adulthood [222]. This, again, highlights the role of DNA methylation dynamics in OPC differentiation. The investigation of the interaction between transcription factors, chromatin remodelers, and DNA methyltransferases might yield beneficial insight into how DNA methylation dynamics are involved in the regulation of OPC differentiation.

miRNA is another level of epigenetic regulation of OPC differentiation. miRNA is a type of small noncoding RNA molecule with an average length of 21–25 nucleotides that can complementarily bind to the target mRNAs, thereby silencing their expression. The necessity of miRNA-mediated regulation on OPC differentiation was first examined by two pioneer studies focusing on the miRNA processing enzyme DICER1. Dicer1 conditional knockout in oligodendroglial cells by Olig1-Cre, Olig2-Cre, or Cnp-Cre lines causes severe OPC differentiation defect and hypomyelination while promoting OPC proliferation [223, 224]. Several DISC1-processed miRNAs, which were upregulated during OL differentiation and reduced by Dicer1 knockout, were then confirmed to promote OPC differentiation by gain-of-function assays, including miR-219 [223, 224], miR-338 [224], and miR138 [224]. The target transcripts of these miRNAs were further identified. miR-219 targets genes essential for maintaining OPC proliferation (such as SOX6, FoxJ3, ZFP238, and PDGFRα), its increased expression thus stimulates OPC exit from the proliferative cycle and entry into differentiation [223–225]. miR-338 share several common targets with miR-219 including SOX6, and also target gene transcripts such as Mytl1 [223]. A follow-up study showed that miR-219 cooperates with miR-338 to promote myelination, as miR-338 knockout aggravates the hypomyelination of Olig1-Cre-mediated miR-219 conditional knockout mice, and further presents the therapeutic potential of miR-219 in models of multiple sclerosis [225]. Additionally, miR-146a and miR-21-5p were also shown to promote OPC differentiation in vitro [226, 227]. While these data substantiate the critical role of DICER1 and miRNAs in OPC differentiation and myelination, it remained a puzzle how the production of these miRNAs was dynamically regulated in OPCs.

Conclusion

OPCs play critical roles in CNS function and homeostasis. This review covers how OPCs participate in myelination and remyelination processes, transdifferentiate into astrocytes and Schwann cells, as well as the recently identified myelin-independent roles. We also discuss the heterogeneity of OPCs across different regions, ages, and sex. Finally, we review the molecular mechanism regulating OPC development. With the development of biochemical and genetic manipulation techniques, the overall understanding of OPCs has been improved. Nonetheless, our understanding of OPCs is still very rudimentary, with numerous open questions.

One key aspect is the physiological role of OPCs, including: How does adult OPC maintain multipotent and quiescent? Are the neuron-OPC synapses associated with precisive myelin formation? How are the OPC differentiation and myelination required for learning? What are the biological roles of OPC in CNS homeostatic support, beyond myelination?

The second aspect revolves around the pathological role of OPCs. What is the spectrum of OPC pathophysiological changes and how these are modified in different diseases or disease stages? What is the role of OPC transdifferentiation in demyelinated injuries? What prevents OPC differentiation and remyelination in demyelinated lesions and how can we overcome the obstacles? How do these pathological changes in OPC integrate into the nervous tissue in response to various pathological insults?

Only by answering all these questions will we be able to achieve a comprehensive understanding of OPC biology and pathophysiology, and hopefully will contribute to the development of OPC-oriented therapeutic strategies.

Conflict of interest

The authors declare that they have no conflict of interest with the contents of this article.