1. Introduction
Responses to injury include a wound-healing structural remodeling of the lesion site and nearby regions culminating in the formation of a glial scar. This process has been observed in a wide range of CNS pathological conditions, including spinal cord injury (SCI), traumatic brain injury (TBI), ischemic stroke and in neurodegenerative diseases, such as multiple sclerosis (MS). The size and features of the glial scar varies in distinct CNS diseases and injuries, based on the different pathological dynamics, location, and severity of the insult. Yet, a fundamental arrangement of the glial scar can be recognized in all cases and includes three main compartments: a fibrotic lesion core encircled by a compact border of a heterogeneous population of glial cells, and, more externally, by an adjacent reactive neural parenchyma [
1,
2].
The lesion core or fibrotic scar is an area of severe tissue damage - often containing a central cystic cavity - populated by blood-borne cells (i.e., macrophages and lymphocytes), perivascular-derived macrophages and stromal cells. Early after acute injuries, blood-derived cells are recruited to the lesion site by resident astrocytes and microglia – which are the first responders to insults – secreting pro-inflammatory cytokines and chemokines. Then, overtime, the stromal component (i.e., fibroblasts and pericytes) progressively prevails over the blood-derived inflammatory cells and become the dominant cell population of the fibrotic scar, being embedded in a rich deposit of extracellular matrix (ECM) [
3,
4]. Newly generated astrocytes, oligodendrocyte progenitor cells (OPCs, or NG2 glia) and microglia are instead the main components of the surrounding ring, the proper glial scar [
2,
5], whereas the adjacent neural parenchyma comprises neurons actively engaged in neurite and synaptic remodelling [
6].
Formation of the glial scar has been traditionally interpreted as a protective process, instrumental to confine the inflammatory response and blood-derived cells to the lesion core, preserving the integrity of the healthy nervous tissue and allowing the repair of the blood-brain barrier. Such a protective role has been mainly ascribed to astrocytes and microglia, whose ablation/manipulation leads to increased influx of blood-derived macrophages and fibrotic cells and exacerbated neuronal cell death, axon loss and demyelination after injury [
2,
7,
8]. The role of the other glial cell types - and particularly of oligodendroglia - participating in the formation of the glial scar has been less investigated so far. Yet, an emerging corpus of data starts to be available. In this review, we will specifically focus on OPCs and oligodendrocytes, their behaviour and their molecular alterations in non-permissive inflammatory conditions, and potential therapeutic strategies to restore their functions and to promote lesion resolution.
2. Myelin as a Structural and Trophic “Organ” for Axons
In the CNS, oligodendrocytes are highly specialized cells responsible for the synthesis of myelin, a multilamellar structure characterized by up to 70% in lipid and myelin proteins that enable membrane compaction in a sheath that enwraps single axons thus allowing fast saltatory conduction [
9].
The composition of myelin is unique compared to other membranes, with an approximate 2:2:1 ratio for cholesterol:phospholipids:glycolipids [
10]. The extremely high cholesterol content in in myelin, that represent 80% of the cholesterol present in the brain, is ensured by both de novo biosynthesis and uptake of cholesterol precursors from other glial cells [
11]. Other lipids are synthesized using fatty acids as building blocks. Myelination is a dynamic process that requires continuous maintenance of membranes, thus making oligodendrocytes highly active in both protein and lipid metabolism.
Moreover, in the last decades, it has been clearly demonstrated that oligodendrocytes also provide axons with fuel to enable proper neuronal functions, thus further increasing their energy demand (for review, see Saab and Nave 2017 [
12]). Oligodendrocytes can directly transfer glucose, through the GLUT1 transporter; in addition, oligodendrocytes metabolize glucose to pyruvate through glycolysis, convert pyruvate to lactate, and then export lactate via monocarboxylate transporter 1 (MCT1) to the axons. Oligodendrocytes could also play a more indirect trophic role, fuelling axons with lactate produced by astrocytes and imported through gap junctions, rather than synthesising lactate by extensive glycolysis [
13]. Via gap junctions, oligodendrocytes also regulate the concentrations of ions, such as Na+, potassium K+, and chloride Cl-, in both intracellular and extracellular spaces in the neuron-oligodendrocyte synapse, thus contributing to the maintenance of the resting membrane potential [
14,
15,
16]. More recently, extracellular vesicles have been described as important components able to shuttle proteins, lipids, nutrients, and non-coding RNAs to axons, and maintain their homeostasis also in nutrient-deprived conditions [
17]. Axonal trophic support includes typical fatty acids, such as docosahexaenoic acid, which constitutes a major component of excitable membranes, produced through beta-oxidation by oligodendrocyte peroxisomes in the myelin sheath [
18].
3. More than Oligodendrocyte Progenitor Cells
Oligodendrocytes originate from OPCs, a population of highly proliferating glia that represents approximately 5–8% of the total cells in the CNS. OPCs derive from multiple niches in the developing brain and they migrate in both brain and spinal cord, where they will differentiate and contribute to myelination [
9]. OPCs can be identified based on their expression of the platelet-derived growth factor receptor α (PDGFRα), the proteoglycan NG2, and the transcription factors SOX10 and OLIG2. In this early stage, OPCs show small capacitance, currents mediated by K
V channels, small-medium sized Na
V channels, and glutamate receptors [
19,
20,
21]. As OPCs differentiate, they firstly form pre-myelinating oligodendrocytes, down-regulating PDGFRα, NG2, voltage-gated Na+ and Ca2+ channels, while they progressively express the sulfatide O4, the galactocerebroside GalC, and GPR17, a G protein-coupled receptor typical of this intermediate stage and acts as a transient inhibitor of myelination [
22]. During their maturation, pre-myelinating oligodendrocytes progressively downregulate GPR17, rewire their glucose and lipid metabolism, produce myelin proteins and lipids that allow the organization of the myelin sheath around axons. This three-stage classification is an oversimplification since OPCs and oligodendrocytes represent a continuum. Recent transcriptomic analysis in oligodendroglial cells isolated from different brain areas in young and adult mice have identified 13 distinct populations that represent different oligodendroglial functional states [
23]. This study highlighted that differentiation-committed oligodendrocyte precursors are distinct from OPCs and show lower levels of cell cycle markers while expressing genes involved in migration. One population of newly formed oligodendrocytes was shown to respond to motor learning, while another of vascular and leptomeningeal cells could migrate following vessels. This heterogeneity could be both intrinsic to their developmental origin and acquired by extrinsic cues at their final location. It was demonstrated that OPCs change their differentiation ability and show different responsiveness to growth factors when transplanted into other CNS areas [
24]. OPCs are the only glial cell type that receives neuronal synapses, which activity was shown to modulate their proliferation, differentiation, and migration and is important for myelination and remyelination (for review, see Moura et al. 2022 [
25] and Habermacher, Angulo and Benamer 2019 [
26]). Recently, OPCs have been shown to phagocytose synapses in the developing and mature mouse visual cortex thus contributing to remodelling of neural circuitries [
27]. OPCs are extremely plastic and adapt their program in response to a multitude of signals both in physiological and pathological conditions [
28].
4. Oligodendroglial Cells Responses to Injury
Myelin and oligodendrocyte loss starts immediately after acute injuries such as SCI or TBI (i.e., within 15 minutes and for 24-48h at the core of the injury), likely due to the combinatorial effects of bleeding, hypoxia, oxidative damage, ATP- and glutamate-mediated excitotoxicity and release of proinflammatory cytokines such as interleukin-1α (IL-1α), tumour necrosis factor-α (TNFα) [
6,
29] (
Figure 1). While neurons and astrocytes do not typically die beyond 24 h post-injury, oligodendrocyte apoptosis is instead protracted up to the subacute (i.e., 3-14 days after injury) and chronic phases, especially in distal degenerating axon tracts after SCI [
30]. By impairing axonal conduction and reducing the metabolic supply and ion buffering, such a protracted loss of oligodendrocytes and demyelination contribute to neuronal and circuit dysfunction, leading to post-injury functional impairment. Accordingly, the number of spared axons
per se does not always predict functional recovery and, in case of SCI, complete loss of motor and sensory functions frequently occurs also in presence of anatomically incomplete spinal lesions [
29].
While demyelination is still ongoing, OPCs are rapidly recruited at the lesion site and set up a robust proliferative response, peaking at around 5-7 dpi in the tissue directly bordering the lesion upon SCI or stroke [
29,
31,
32]. Such OPC expansion can be sustained by both parenchymal OPCs and periventricular neural stem cells [
33,
34,
35], partly as a consequence of the loss of cell-to-cell contact inhibition [
36,
37], and partly in response to developmental (e.g., fibroblast growth factor - FGF, platelet-derived growth factor-PDGF, WNTs) and proinflammatory factors present in the lesion milieu and in the glial scar [
2]. In case of SCI, OPC proliferative response is surprisingly long-lasting, being detected even at chronic stages [
29]. OPC proliferative burst is accompanied by a prominent reduction in neuron-to-OPC synaptic connectivity, as assessed by the transient disappearance of glutamatergic inputs onto virtually all proliferating OPCs responding to a demyelinating injury [
38].
Newly born OPCs accumulate at the border of the lesion (e.g., at the stroke
penumbra) together with astrocytes, whereas reactive microglia enter the margins of the fibrotic scar [
39]. Such an OPC confinement depends at least in part by stromal cell-derived signals, as demonstrated by OPC infiltration in the central fibrotic core of the lesion in case of fibroblast experimental ablation [
40]. Acutely, OPCs produce fibronectin and laminin that protect axon growth cones from the neuroinflammatory milieu [
41]. However, in the long term, glial cells overexpress matrix metalloproteases (MMPs) resulting in the degradation of hyaluronan and other ECM components into low molecular weight fragments that can amplify inflammation [
42]. Glial cells, including OPCs, further modify the ECM composition by overexpressing chondroitin sulphate proteoglycans (CSPGs), among which brevican, neurocan, versican, and NG2 are highly represented, and known for contributing to hostile environment to axonal regrowth and neuronal migration [
43].
On one hand, most of the newly generated OPCs entrap axon growth-cones and hamper regeneration, on the other hand, they do not differentiate into myelinating oligodendrocytes and do not contribute to remyelination. An increased number of immature OPCs and immature oligodendrocytes was reported at the border of the lesion even at later times after stroke, SCI or cortical stab wound [
30,
31,
44], suggesting a blockade in their progression along the lineage or a local requirement for cells exerting roles other than differentiation into myelinating cells. The first hypothesis is corroborated by the presence of a plethora of environmental signals in the lesion milieu, that operate as inhibitors of OPC differentiation, including proinflammatory cytokines and CSPGs (
Figure 1). Accordingly, when plated on CSPGs-coated plates, OPCs show maturation delay, reduced process outgrowth, and impaired migration, suggesting a direct chemorepulsive effect that alters their potential to access to demyelinated sites and the consequent recovery process [
45]. Similar repellent effects of CSPGs in oligodendrocyte function were observed in scars of traumatic origin. After SCI in rats, treatments with chondroitinase increased the numbers of OPCs surrounding lesions in the first two weeks post-lesion, and the number of OPCs inside the lesion without altering OPC proliferation and cell death, likely as a consequence of improved OPC migration. Improving OPC migration into lesioned sites was shown to be necessary to enable the increased axonal sprouting and improved recovery observed at the second week after SCI [
46,
47]. Interestingly, CSPG-induced oligodendrocyte maturation impairment could not be reverted by previously reported pro-remyelinating compounds such as clemastine, benztropine, and miconazole, which suggests that remyelination strongly relies on the extracellular microenvironment [
48,
49,
50]. Furthermore, OPCs have functional receptors for many cytokines such as IL-10, IL-6, IL-4, IL-18, IFN-γ, and TNF-α [
51,
52,
53]. These factors, primarily secreted by astrocytes and microglia, were shown to have a negative impact on OPC maturation in vivo and in vitro [
54,
55].
OPC or oligodendrocyte transdifferentiation in other cell types may also contribute to such a failure in providing new myelinating elements. Specifically, fate mapping experiments revealed that a fraction (10-25%) of OPCs acquire features of astrocytes (e.g., expression of the glial cell fibrillary protein – GFAP, or Gja1, Aqp4, Kcnq4 mRNAs and ion currents pattern similar to that observed in cortical astrocytes) upon cerebral stroke, cortical stab wound, SCI and experimental autoimmune encephalomyelitis - EAE [
56,
57,
58]. These OPC-derived astrocytes are a transient cell population characterized by a high expression of proliferation and motility markers [
56]. As astrocytes have a limited capacity to migrate and amplify compared to OPCs (i.e., astrocyte numbers increase just by 10-20% after acute injuries; [
59,
60], it has been proposed that OPC-derived astrocytes might temporarily perform the functions of astrocytes (such as regulating extracellular ion and neurotransmitter homeostasis or providing energy supply) in proximity to the lesion [
56], and then die or re-enter the oligodendroglia lineage [
57]. Unexpectedly, a subpopulation of mature oligodendrocytes has been also shown to activate astrocytic genes and transgress via a transitional precursor phenotype toward the astroglial fate after acute brain injuries [
61]. Mechanistically, mature oligodendrocyte-to-astrocyte conversion has been shown to be microglia-dependent and attributed to the activation of the IL-6 signalling [
61]. The specific function of this additional population of astrocytes remains to be determined.
In addition, after injury, NG2 cells can trans-differentiate into Schwann cells. Yet, the appearance of NG2 cell-derived Schwann cells is context-dependent (i.e., they are detected following SCI, but not after cortical stab wound injury) and their actual contribution to remyelination remains controversial [
62,
63].
In agreement with the idea that, upon injury, OPC roles go beyond the generation of new myelinating cells, recent studies have shown that OPCs
per se participate in tissue remodelling and healing. Although OPCs were shown to contribute to synapse and axon pruning during normal development, they were also shown to selectively internalize myelin debris through phagocytosis upon injury [
27,
54,
64,
65]. Moreover, in the acute phase after brain injury, OPCs release metalloproteinases (e.g., MMP9) which in turn promote blood-brain barrier (BBB) leakage and infiltration of blood-derived cells at lesion site [
66]. For instance, ablation of proliferating NG2 glia results in reduced astrocyte hypertrophy and disorganization of the glial scar, which eventually lead to delayed brain wound closure and worse functional outcome [
39,
67]. Similarly, upon an inflammatory challenge, OPC depletion leads to a profound downregulation of the expression of microglia-specific genes and an exacerbated inflammatory response [
68]. Consistently, manipulations of OPCs not resulting in their depletion (e.g., OPC-specific deletion of β-catenin) reduce astrocyte hypertrophy and GFAP expression, and promote microglia/macrophage acquisition of an anti-inflammatory/pro-regenerative phenotype [
69]. Such effects likely depend on OPC ability to release cytokines and chemokines and to present antigen when exposed to inflammatory conditions [
70]. Both these capabilities are shared by more advanced stages along the oligodendroglia lineage. For instance, immature oligodendrocytes can upregulate the expression of toll-like receptor Tlr3, IL-1β, IFN-β, Ccl2, and Cxcl10 when stimulated with IL-1β [
71]. Similarly, the chronic activation of NF-kB, the transcription factor downstream to TNF-α, IL-1 and toll-like receptors [
72], leads to exacerbated neuroinflammatory conditions in mature oligodendrocytes [
73]. Moreover, recent transcriptomic data allowed the identification of populations of disease-associated mature oligodendrocytes enriched in immune-related and antigen presentation genes (e.g., major histocompatibility complex -MHC- class II) in the brain of mouse models and Alzheimer’s Disease and Multiple Sclerosis patients [
54,
74,
75]. Together, this evidence suggests that upon injury oligodendroglia can actively participate in fuelling inflammation and might have a role in disease development and antigen presentation.
6. Conclusions
Due to the cellular heterogeneity and to the complex dynamics of the lesions in time and space, therapeutic intervention to treat glial scar or scar-associated neuroinflammation still remains a challenge. It is clear that glial scar formation is part of a regenerative attempt and should not be completely blocked. Instead, the inhibitory factors that progressively establish a hostile environment should be dissected to identify more focused therapies. The roles of oligodendrocytes and OPCs have been overlooked in glial scar. Early after injury, oligodendrocyte apoptosis leads to demyelination, whereas OPCs react and migrate toward the damage to replenish and repair. Acutely, they form synapse-like connections with the tips of transected axons and produce ECM components that protect axon growth cones from the neuroinflammatory milieu. However, in the long term, they overexpress matrix metalloproteinases (MMPs) and CSPGs, contributing to the production of aberrant unfavourable ECM components and to the generation of a hostile environment, which in turn impairs long-term axonal regeneration and inhibit OPC differentiation and remyelination. The production of high amounts of CSPGs within the lesions is associated with chemorepulsion and pathological behaviour of OPCs. NG2 is one of the major CSPGs expressed on the surface of early OPCs, whose accumulation at the lesioned areas could take part to a vicious cycle, impeding their own differentiation. Accordingly, depletion of CSPGs was shown to successfully restore OPC differentiation and remyelination in several models of disease.
Fostering remyelination is an unmet need in the context of demyelinating diseases, such as MS. However, since oligodendrocyte dysregulation and demyelination are pathological features of several neurodegenerative diseases or acute conditions, and often they take place in the early phases of neurodegeneration, it is possible to speculate that remyelination would be a potential strategy also in other disorders, trauma, and brain ischemia. It is worth observing that the generation of a favourable microenvironment is a key step to pave the way to remyelination. Accordingly, cell transplantation approaches have shown promising results when coupled to other interventions aimed at reestablishing a favourable microenvironment, targeting CSPG deposition and degradation. These results may explain why previously reported pro-remyelinating compounds have shown poor efficacy in the presence of an unfavourable CSPG-rich environment, suggesting that also these therapeutic approaches should be complemented with other compounds targeting the hostile microenvironment.