White matter injury across neurodegenerative disease

Abstract

Oligodendrocytes (OLs), the myelin-generating cells of the central nervous system (CNS), are active players in shaping neuronal circuitry and function. It has become increasingly apparent that injury to cells within the OL lineage plays a central role in neurodegeneration. In this review, we focus primarily on three degenerative disorders in which white matter loss is well documented: Alzheimer’s disease (AD), Parkinson’s disease (PD), and amyotrophic lateral sclerosis (ALS). We discuss clinical data implicating white matter injury as a key feature of these disorders, as well as shared and divergent phenotypes between them. We examine the cellular and molecular mechanisms underlying the alterations to OLs, including chronic neuroinflammation, aggregation of proteins, lipid dysregulation, and organellar stress. Lastly, we highlight prospects for therapeutic intervention targeting the OL lineage to restore function.

Oligodendrocytes are key players during neurodegenerative disease

Hallmarks of neurodegenerative disease have been long described as neuronal loss, reactive gliosis of astrocytes and microglia, and accumulation of misfolded proteins in disease-specific regions of the CNS. These findings are associated with activation of both the innate and adaptive immune systems that in turn generates a neuroinflammatory milieu and further contributes to neuropathology. While much research has focused on the convergence of neuroinflammation on neurons, emerging evidence has implicated white matter and oligodendrocyte (OL) injury as a generalizable phenomenon across numerous neurodegenerative diseases, raising the question of whether alterations in white matter function may be primary driver of dysfunction, especially in instances where white matter pathology precedes neuronal injury. In fact, myelin deficits are associated with cognitive, behavioral, and motor dysfunction across a spectrum of neurodegenerative disorders. In this review, we highlight the evidence for white matter injury in chronic neurodegenerative and neuroinflammatory conditions, including Alzheimer’s disease (AD), Parkinson’s disease (PD), and amyotrophic lateral sclerosis (ALS), and focus on the underlying mechanisms triggering OL dysfunction. Lastly, we discuss the therapeutic potential of targeting the OL lineage to alleviate neuropathological and cognitive/behavioral dysfunction in chronic, neurodegenerative diseases.

White matter injury is a common feature of chronic neurodegeneration

In the central nervous system (CNS), OLs form myelin, an extension of their plasma membrane that concentrically wraps around axons to insulate and facilitate rapid and efficient action potential conductance. The OL lineage arises from embryonic neural progenitors that give rise to the major proliferating cell in the CNS (~5% of all cells), oligodendrocyte precursor cells (OPCs) [1]. OPCs remain proliferative throughout life and the proliferation rate is location- and age-dependent [2]. The generation of premyelinating (immature) and eventually mature, myelinating OLs from their precursors is a generally protracted process that in humans continues well into the sixth decade of life [3,4]. It was recently proposed that the temporal sequence of myelination can be divided into five different stages of human life: (1) early childhood; (2) childhood; (3) adolescence; (4) adulthood; and (5) age-related decline in myelin [5]. In three of these periods, separate and rapid waves of synchronized myelin changes occur: during early childhood, adolescence, and aging [5]. The age of onset for numerous psychiatric and neurodegenerative disorders overlaps with these periods of significant myelin alterations, suggesting OLs and white matter may be uniquely vulnerable during disease processes, including neuroinflammation. Accumulating evidence from human imaging and post-mortem studies lends support to this notion and indicates that white matter injury might be a core feature in chronic, neuroinflammatory diseases.

Clinical evidence for white matter injury in AD.

AD is the most common cause of dementia and is characterized by extracellular senile plaques comprised of β-amyloid (Aβ) and intracellular neurofibrillary tangles (NFTs) composed of hyperphosphorylated tau [6]. White matter hyperintensities (WMH), signal abnormalities in white matter visualized via magnetic resonance imaging (MRI), predict incident AD and the rate of cognitive decline; additionally, they are associated with genetic risk factors (e.g. ApoE4) for late onset AD [7-9]. This relationship between AD and WMH is even more apparent in individuals with full penetrant familial mutations for AD, where WMH volume is increased up to 20 years prior to symptom onset, contemporaneous with elevated CSF levels of Aβ and tau [10]. Intriguingly, in cognitively normal adults, greater amyloid beta (Aβ) burden as assessed via positron emission tomography (PET) imaging is associated with lower myelin content in several brain regions, further strengthening the view that myelin injury is one of the earliest abnormalities to occur prior to the onset of AD symptoms [11]. These alterations in overall myelin content are supported by single-cell RNA sequencing (scRNA-seq) studies from post-mortem human tissue that have identified highly cluster-specific transcriptomic changes in OLs based on disease state, including upregulation of gene ontology (GO) terms associated with axonogenesis, synapse organization, and cholesterol metabolism (cluster 1 and 2), and downregulation of GO terms associated with synapse transmission, synaptic vesicle, ion transmembrane transport, and metabolism (cluster 0) [12].

Clinical evidence for white matter injury in PD.

PD, which is characterized by selective dopaminergic neuron degeneration and death, also has a significant white matter injury facet [13]. Widespread microstructural alterations in white matter of PD are associated with olfactory and motor deficits that occur early after diagnosis [14,15]. Furthermore, deep white matter hyperintensities and periventricular white matter hyperintensities are associated with bradykinesia and axial symptoms, regardless of the severity of dopamine depletion, suggesting that white matter injury may occur independently from dopaminergic neuron loss [16]. The progression of brain atrophy, an early neuropathological alteration in PD, over a 2-4 year period in PD patients was inversely related to the presence of OLs within the caudate, nucleus accumbens, hippocampus, and posterior cortical regions as assessed via MRI and virtual histology [17]. As in AD, diffusion tensor imaging (DTI) revealed changes in fractional anisotropy (FA) during prodromal PD; however, unlike AD, these changes appeared to be adaptive in nature as higher FA was seen in the motor cortex and corticospinal tract compared to controls [18,19]. Postmortem analyses of PD brains have revealed conflicting data, with one study showing increased expression of myelin-related genes in the frontal cortex, suggesting potential compensatory mechanisms to alleviate neuronal dysfunction; however, other work has shown higher expression of the PD risk gene LRRK2 in OPCs compared to other substantia nigra cells, enrichment of PD-GWAS-SNPs within OLs, and changes in OL subpopulations that display reduced myelination [20-23].

Clinical evidence for white matter injury in ALS.

ALS is a neurodegenerative disease characterized by the progressive degeneration of motor neurons in the spinal cord, brain stem, and motor cortex [24]. As observed in AD and PD, white matter alterations, including decreased fiber integrity assessed by DTI, in corticospinal tract fibers of ALS patients have been shown to occur prior to the appearance of clinical symptoms [25]. In line with clinical imaging data, post-mortem single-nuclei RNA sequencing (snRNA-seq) has shown that genomic risk for ALS mapped consistently to OLs, while another scRNA-seq analysis found that ALS common genetic risk was enriched within two specific modules, one of which was enriched in OL markers [26,27]. Another recent transcriptomic analysis from ALS patient spinal cord found downregulation of OL gene expression which the authors concluded might reflect genuine cell loss due to secondary demyelination; however, it is also possible that the changes observed were due to the relative shift in proportion compared to increased astrocyte and microglia numbers [28]. Immunohistochemical analysis has found a high burden of glial inclusions in the prefrontal cortex, precentral gyrus, and spinal cord of ALS patients, a majority of which were found in OLs [29]. A specific reduction in myelin basic protein (MBP), a major myelin protein, compared to the myelin component proteolipid protein (PLP) was also seen, suggesting that impaired mRNA transport may be responsible for this effect. Dysregulation of the subcellular trafficking of MBP mRNA has been seen in both sporadic and familial ALS patients; however, this did not result in differences in MBP protein levels or myelination [30]. The discrepancy between these two studies may be due to regional differences in OL vulnerability as they examined different CNS areas (e.g., spinal cord vs motor cortex). Despite these diverse neuropathological conditions, there are common cellular and molecular pathways that converge upon on the OL lineage to induce injury and dysfunction as outlined below.

Cellular and molecular mechanisms underlying oligodendrocyte dysfunction in neurodegenerative diseases

Immune activation and inflammatory mediators

A hallmark of the above-mentioned disorders is the inability to resolve an acute inflammatory response, which transitions into chronic neuroinflammation involving resident CNS cells (e.g., microglia and astrocytes) and the infiltration of peripheral immune cells (Figure 1) [31]. The removal of cellular debris and toxic protein aggregates is mediated through a combination of transcellular transport mechanisms across the blood-brain barrier (BBB), phagocytosis by resident microglia and recruited monocytes/macrophages, and a recently (re)discovered glymphatic drainage system within the CNS [32-34]. In aged 5xFAD mice, a model of familial AD that expresses human APP and PSEN1 with a total of five human AD mutations, impaired meningeal lymphatic vessel function was observed and this correlated with a significant increase in Aβ deposition in the meninges [33,35]. This failure to remove toxic products and potentially increase their deposition results in cell-type specific alterations in microglia and astrocytes.

Figure 1. Neuroinflammation by the innate and adaptive immune systems induces oligodendrocyte injury.

The infiltration of activated peripheral immune cells, in particular T cells, has been observed in chronic neurodegenerative conditions, including AD, PD, and ALS. These cells can recognize specific antigens within the CNS, as well as produce inflammatory cytokines that are known to disrupt OPC differentiation. The resident cells of the CNS, microglia and astrocytes, also contribute to the inflammatory milieu through the increased secretion of proinflammatory cytokines and chemokines that directly inhibit OPC maturation in mature, myelinating oligodendrocytes. Additionally, microglia can adopt a specific phenotype to phagocytose damaged myelin; however, this can impair their ability to clear amyloid plaques and perpetuate damage via aggregated proteins as outlined in Figure 2. Abbreviations: Aβ, amyloid beta; OPC, oligodendrocyte precursor cell. Figure created with BioRender.com.

As the resident immune cells of the CNS, microglia play essential roles in development, homeostasis, and injury [36]. A recent study has shown that during the initial response to cortical amyloid deposition in humans, there is a loss of the major homeostatic microglia response and an expansion of three Aβ-reactive microglia subpopulations characterized by transmembrane glycoprotein NMB (GPNMB+) expression [37]. These microglia upregulate genes involved in autophagy, antigen processing and presentation, interleukin-15, and response to cytokines, demonstrating that there are neuroinflammatory alterations even at the earliest stages of amyloid plaque deposition. Interestingly, the presence of GPNMB+ microglia has also been reported in scRNA-seq data from both PD and ALS human post-mortem tissue, demonstrating the presence of common disease-associated microglia among different neurodegenerative conditions that are, in part, responsible for maintaining a neuroinflammatory environment [28,38]. Another unique sub-population of microglia, known as white-matte associated microglia (WAM), are associated with aging but emerge much earlier in 5xFAD mice [39]. These microglia contain MBP+ intracellular particles in the corpus callosum, a major white matter tract, that is not associated with amyloid plaques, demonstrating that there are microglia-specific subpopulations that respond to myelin debris [39]. Damage-associated microglia (DAMs) have been identified across neurodegenerative conditions; however, a recent study demonstrated that despite successful induction of DAMs in different mouse models of AD, defective myelin inhibits DAMs from phagocytosing Aβ plaques and promotes further plaque deposition [40]. Thus, there appears to be a bidirectional relationship between myelin and a successful immune response.

While astrocytes are not classically considered a part of the immune system, they are able to respond to injury, become reactive, and phagocytose myelin debris [41,42]. As with microglia, astrocyte subpopulations have been identified that appear to be specific to neurodegeneration [43]. In a study in mice, a subpopulation of astrocytes (cluster 8 in the publication) was identified as absent in the healthy mouse brain and was rarely observed in LPS-treated animals; however, the cluster was robustly identified in 5xFAD mice [43]. Astrocytes in this subpopulation express many interferon stimulated genes (ISGs), suggesting a gain of antigen-presenting function, as well as a high level of the chemokine, CXCL10 [43]. Elevated levels of CXCL10 and interferon response from human induced pluripotent stem cell (iPSCs) astrocytes containing the clusterin (CLU) risk allele for late onset AD inhibits OPC proliferation and OL myelination in vitro; additionally, decreased MBP expression was confirmed in post-mortem human tissue of CLU risk allele carriers [44]. Thus, it seems possible that this subpopulation of astrocytes may contribute to white matter injury during neurodegeneration via CXCL10 and interferon signaling.

Alterations in the neuroinflammatory response are not limited to resident CNS cells; infiltration of activated peripheral immune cells, in particular T cells, occurs in chronic neurodegeneration via bidirectional trafficking across the BBB and glymphatic system [34]. Higher levels of autoantibodies to major myelin proteins, including MBP and PLP, have been detected in the blood of AD patients compared to healthy controls [45]. In line with T cells potentially recognizing specific CNS myelin antigens, clonally expanded T cells can be isolated from human AD brains [46]. In PD, T cells isolated from patients recognize alpha synuclein (αSyn) peptides, which is associated with preclinical and early PD, as well as interferon gamma (IFNγ) production, which is known to directly damage OLs [47-49]. Furthermore, in aging, CD8+ T cells induce interferon-responsive OLs that results in white matter degeneration in aged mice [50]. Infiltration of inflammatory Th1 and Th17 CD4+ T cells, similar to what has been observed in the demyelinating disorder multiple sclerosis (MS), is seen in the CNS of late-stage ALS patients, where they can potentially contribute to OL injury [51]. This chronic neuroimmune activation initiates a cascade of deleterious events that results in misfolded/aggregated proteins that directly interfere with OPC and OL function and induce further disruption through lipid dysregulation and organellar stress.

Misfolded proteins and protein aggregates

The accumulation of misfolded proteins and/or protein aggregates is seen in AD (Aβ plaques), PD (αSyn), and ALS (SOD1 and TDP43) and has been shown to contribute to the neuroinflammatory environment as well as damage to the OL lineage (Figure 2). Findings on the deleterious impact of Aβ oligomers on OPCs and OLs are conflicting. In isolated rat OPCs, Aβ oligomer treatment promoted the local translation of MBP in distal OL processes and promoted OPC differentiation in cerebellar slices via integrin β1 receptor and Src-family kinase Fyn signaling [52]. In another study, treatment of purified mouse OPCs with Aβ oligomers or injection of oligomers into the lateral ventricle of mice did not alter any OPC or OL marker [53]. In contrast to these two papers, the induction of Aβ seeds in the corpus callosum of 5xFAD mice resulted in demyelination and surprisingly fewer proliferative OPCs in response to the white matter loss [54]. The discrepancy these studies and others may be due to differences in Aβ oligomer species, duration of treatment, and model utilized. Work conducted in AD patient tissue and APP/PS1 double transgenic mice showed that OPCs near Aβ plaques exhibited a senescence-like phenotype, possibly preventing these cells from responding to injury and differentiating into myelinating OLs [55]. Additionally, in response to early cortical amyloid deposition in humans, OLs upregulate genes associated with β-amyloid production, which peaked during the earliest stages, suggesting they may be an unlikely source of Aβ plaques [37].

Figure 2. Misfolded proteins and protein aggregates impair homeostatic OPC and oligodendrocyte functions.

AD, PD, and ALS are characterized by the presence of protein aggregates and misfolded proteins. A) In AD and PD, protein aggregates (amyloid-β plaques and αsynuclein) can directly impair OPC maturation and induce alternative phenotypes within OPCs, including antigen presentation and senescence. The appearance of these phenotypes also prevents the ability of OPCs to differentiate into OLs, though the mechanisms underlying this observation are not known. Additionally, in response to cortical amyloid deposition, oligodendrocytes upregulated genes associated with β-amyloid production, suggesting they may be an unlikely source of amyloid plaques [37]. B) Accumulation of mutant SOD1 or TDP43 in ALS decreases the oligodendrocyte-specific lactate transporter, MCT1, and impairs axonal energy metabolism [62,64]. Aggregation of myelin proteins, specifically MBP, have been identified and this was associated with an increase in citrullination, a post-translational modification that inhibits the ability of MBP to bind the plasma membrane [63]. Abbreviations: AD, Alzheimer’s disease; ALS, amyotrophic lateral sclerosis; MCT1, monocarboxylate transporter 1; mtSOD1, mutant superoxide dismutase 1; PM, plasma membrane. Figure created with BioRender.com.

While αSyn aggregation is typically observed only in neurons in PD, there is evidence that neuronal synucleinopathies can influence the myelin deficits observed in PD. In two mouse models of αSyn accumulation (Prp-A53T αSyn and Thy-1 wt-αSyn), an age- and disease-dependent loss of MBP in striatal striosomes (patches) occurred, which was associated with lower P25α levels, a marker for myelinating OLs [56]. Interestingly, prior to this age-related decrease in myelin, phospholipid content of myelin in these mutant mice was higher, suggesting that neuronal αSyn may be involved in the regulation and/or maintenance of myelin phospholipid content [56]. Direct in vitro treatment of purified rat OPCs with αSyn inhibited OL maturation and the formation of membranous myelin sheets [56]. Early phenotypic and pathologic changes have been observed in PD patient O4+ OL lineage cells derived from iPSCs, where OL maturation was substantially impaired and the transcription of myelin genes was decreased [57]. Additionally, these cells had changes associated with an immune-reactive phenotype characterized by increased expression of complement component C4b, HLA, and immunoproteasome subunits, which may, in part, explain why these cells fail to mature [57]. A similar phenotype, characterized by upregulation of C4b, as well as serine (or cysteine) peptidase inhibitor, clade A, member 3N (SERPINA3A), was observed in the 5xFAD mouse model of AD; furthermore, in vitro treatment of a human oligodendrocyte cell line with oligomerized Aβ was sufficient to induce C4 expression [58]. These immune-reactive phenotypes mirrors a recently discovered function of OPCs to present antigens, a function that may involve the phagocytosis of extracellular material other than axons and synapse, another recently observed OPC phenotype [59-61].

Mutations and aggregation of superoxide dismutase 1 (SOD1) is one of the most common causes of familial ALS [24]. Accumulation of mutant SOD1 (mtSOD1) in mature OLs in zebrafish resulted in disrupted myelin sheaths and downregulation of the essential lactate transporter, monocarboxylate transporter 1 (MCT1), in OLs; this eventually resulted in spinal motor neuron degeneration despite no accumulation of mtSOD1 in those cells [62]. However, in two mouse models of ALS (SOD1G9A and PFN1C71G), protein aggregates containing both PLP and MBP, and rarely SOD1 and PFN, were observed in spinal cord white matter; this was associated with an increase in protein citrullination, a post-translational modification that has been shown to reduce the positive charge of MBP and decrease its binding to the negatively charged plasma membrane [63]. Intracellular inclusions of transactive response DNA binding protein 43 (TDP43) are present in almost familial ALS cases and low-level overexpression of wild-type TDP43 in OLs results in age-dependent myelin degeneration and OL loss; however, it is important to note that TDP-43 is essential for homeostatic OL function and influences OL cholesterol metabolism [64-66]. Across neurodegenerative diseases, aggregated proteins are not only important mediators of cellular stress through initiation of organellar dysfunction and lipid dysregulation, but also perpetuate chronic neuroinflammation due to their failure to be cleared by microglia and astrocytes.

Lipid dysregulation

The myelin sheath, in contrast to most biological membranes, contains remarkably high lipid levels, which accounts for 70-85% of its dry weight [67]. Depletion of myelin-associated lipid species, including sulfatides and galactosylceramides, in 5xFAD mice was associated with amyloid plaques and myelin sheath degeneration; additionally, accumulation of the lipid and cholesterol transporter APOE in plaques was found to be in line with sulfatide depletion, suggesting a potential relationship between the two [68]. Myelin sulfatide depletion in transgenic mice is sufficient to induce activation of disease-associated microglia and astrocytes, increase AD risk gene expression, and induce mild cognitive impairment [69]. Thus, a feedback loop between sulfatide loss, amyloid plaque/APOE accumulation, and neuroinflammation may contribute significantly to OL dysfunction and cognitive impairment in AD. Sulfatide disruption has also been observed in a 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) non-human primate model of PD within motor-related brain regions [70]. In the motor cortex of ALS patients, myelin protein content was unaltered compared to controls; however, lipid levels were dysregulated, including increases in cholesterol esters and triglycerides and decreased in sphingomyelins, ceramides, and phospholipids, suggesting altered myelin composition occurs in ALS [71].

The E4 allele of APOE (APOE4) is the most prevalent genetic risk factor for AD and this polymorphism interferes with APOE’s cholesterol- and lipid-transport ability [72]. In human post-mortem PFC from APOE4 carriers, OLs had an increase in canonical enzymes responsible for cholesterol synthesis and additional work in APOE4 iPSCs showed that cholesterol accumulates within OLs with reduced localization to the plasma membrane, increased formation of cholesterol esters and droplets, and upregulation of ATF6-mediated endoplasmic reticulum (ER) stress response [73]. In contrast to this, another group reported that endogenous human APOE4 (hAPOE4) had no effect on OL maturation in cultured mouse OPCs, but OL differentiation was attenuated in cholesterol-depleted Apoe^−/− OPCs cocultured with hAPOE4-expressing astrocytes compared to hAPOE3 astrocytes [74].

Organellar stress

The above-mentioned pathways can result in the disruption of homeostatic organelle function and this organellar stress can also affect the neuroinflammatory environment and lipid regulation (Figure 3). The lysosome is the primary degradative organelle but also plays essential roles in metabolic signaling, gene regulation, and plasma membrane repair [75]. OLs from TREM2 risk-modifying variant individual post-mortem tissue were highlighted by a dysfunctional autophagy-lysosomal pathway with a higher expression of the master lysosomal biogenesis regulator transcription factor EB (TFEB) [76]. TFEB has been shown to coordinate OL programmed cell death during development and its overexpression suppresses myelination at different stages [77,78]. The transmembrane lysosomal protein TMEM106B, which is a major dementia risk allele, has been proposed to play a role in the clearance of lipids [79]. Indeed, in post-mortem tissue of TMEM106B risk allele carriers, levels of myelin-enriched sulfatides and hexosylceramides were significantly lower in the hippocampus with no changes in myelin protein expression [80].

Figure 3. Organellar dysfunction and lipid dysregulation are critical mediators of oligodendrocyte injury in neurodegeneration.

Depletion of myelin-specific lipids has been observed across neurodegenerative conditions, including in response to amyloid plaques and the major dementia risk allele TMEM106B. Organellar stress also occurs downstream of neuroinflammation and aggregated proteins. Increased expression of the master lysosomal biogenesis transcription factor, TFEB, is observed in post-mortem tissue from patients with TREM2 risk-modifying alleles; TFEB has been shown to coordinate oligodendrocyte cell death during development and represses myelination. Endoplasmic reticulum stress, which halts global translation and leads to the selective transcription of stress-related genes via ATF4 translocation, is associated with cholesterol accumulation in APOE4 carriers. Failure to resolve chronic ER stress becomes deleterious over time, and is characterized by persistent upregulation of GADD34. Overactivation of Drp1 on mitochondria activates the NLRP3 inflammasome and activates pyroptotic cell death and myelin loss. Loss of glycolytic and ketolytic pathway genes within oligodendrocytes has also been observed in Alzheimer’s disease post-mortem tissue. Abbreviations: Atf4; activating transcription factor 4; Drp1, dynamin-related protein 1; eIF2α, eukaryotic transcription factor 2 alpha; ER, endoplasmic reticulum; GADD34, growth arrest and DNA damage-inducible protein 34; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; TFEB, transcription factor EB; TMEM106B, transmembrane protein 106B. Figure created with BioRender.com.

Mitochondria are hubs of cellular activity, immune responses, and programmed cell death [81]. Overactivation of dynamin-related protein 1 (Drp1), a key regulator of mitochondrial fission, in mature OLs is seen in AD patients and 5xFAD mice which results in inhibition of the glycolysis initiation enzyme hexokinase 1, activation of NLR family pyrin domain containing 3 (NLRP3)-dependent inflammatory injury, and myelin loss [82]. Glycolytic stress in OLs, as well as impairment in ketolytic pathways, has been confirmed by another group who conducted RNA-seq on post-mortem AD tissue; interestingly, ketolytic pathway dysfunction was not observed in any other cell type assessed [83].

ER stress has been identified as a hallmark of chronic neurodegeneration conditions; however, how this impacts the OL lineage during disease, outside of MS, has not been thoroughly explored. As mentioned earlier, APOE4 OLs accumulate lipid droplets, which is a hallmark of ER stress, and a majority of the cell’s cholesterol was retained in the ER [73]. ER stress activates the PKR-like endoplasmic reticulum kinase (PERK) which in turns phosphorylates eukaryotic initiation factor 2α (elF2α) and suppresses global translation [84]. While this process is adaptive in the short-term, chronic activation in the context of neurodegeneration and inflammation may have deleterious effects, particularly in long-lived cells like OLs that need to turnover and maintain their elaborate myelin sheaths [85]. Upregulation of growth arrest and DNA damage-inducible protein (GADD34), a negative regulator of eIF2α phosphorylation, in mature OLs was observed in AD brains and E693Δ APP mouse; this increase in GADD34+ positivity was seen early in these mice and maintained for up to 2 years, suggesting that the stress response was unable to resolve [86].

Concluding remarks and future perspectives

Emerging evidence has repositioned the OL lineage in neurodegeneration from recipient of collateral damage to an active player that may be responsible for initiation of disease. In particular, clinical data from imaging and post-mortem analyses of human tissue have revealed that both deleterious and compensatory changes occur early in OPCs and OLs during neurodegeneration, even before the appearance of clinical symptoms. Ongoing and future studies should continue to determine whether therapeutic intervention targeted towards OPCs and OLs will be sufficient to restore cognitive, motor, and behavioral function in patients (see Outstanding questions).

Outstanding Questions.

• There has been a long history of focusing on neuronal loss as the initiator of clinical dysfunction in neurological disorders; however recent studies outlined here demonstrate that OL dysfunction occurs earlier. Is this important when developing therapeutic options for neuroinflammatory diseases?

• Will intervention at the OL level be sufficient to restore function or will therapies need to target different cell types within the CNS?

• Most mechanistic in vitro studies of OLs have been conducted on isolated rodent cells; do these accurately capture phenomena that rely on 1) aging and 2) necessary interactions with other cells of the CNS?

• Does OPC and OL heterogeneity play a role in determining susceptibility to neuroinflammation?

• How do non-canonical roles of OPCs, particularly antigen presentation, impact the neuroinflammatory milieu?

• Is myelin debris presented to T cells during chronic neurodegeneration, and if so, does it result in autoimmunity that could contribute to OL dysfunction?

• The interactions between OLs and other glia within the CNS (e.g. microglia and astrocytes) have been shown to be both beneficial and inhibitory towards differentiation and myelination; how could this be leveraged for therapy given the potential small therapeutic window?

Indeed, over the last several years, preclinical studies, primarily in AD and ALS, have begun to explore this by targeting different aspects of the OL lineage. Clemastine fumarate, a Food and Drug Administration approved muscarinic receptor 1 (M1R) antagonist, has shown clinical efficacy in patients with chronic demyelinating injury, suggesting that it may also have similar benefits in other chronic neurodegenerative conditions [87]. In two different studies utilizing mouse models of AD, clemastine enhanced myelin renewal, improved myelin-related learning tasks, and attenuated the transition of OPCs to a state of cellular senescence [88,89]. It is important, however, to consider that during developmental myelination, clemastine has been shown to reduce the population of promyelinating CD11b+ microglia [90]. Therefore, careful consideration of off-target effects, as well as the interaction between cell types in the CNS is required. Other OL-targeted therapies for AD and ALS have been investigated in mice, including inhibition of ERK1/2 signaling, facilitation of cholesterol transport, adeno-associated virus (AAV) delivery of MCT1, and antagonism of G-protein coupled receptor 17 (GPR17), with various success [73,91-93].

Another question that has been at the focus of important discussions is whether the widely utilized animal models accurately represent the complex neuropathology observed in patients in terms of temporal progression as well as cellular processes. Most rodent models used for AD, PD, and ALS research are based on overexpression of familial mutations that only account for a small portion of patient cases. Despite the strides that have been made in understanding the processes underlying OPC differentiation and OL function, there is still no approved therapy to restore myelination for chronic demyelinating and/or neurodegenerative conditions. Perhaps the biggest roadblock in these conditions is the presence of chronic neuroinflammation that initiates and perpetuates a cascade of deleterious and compensatory changes within the OL lineage. Most likely, a two-prong approach will be needed to combat activation of the immune system, as well as promote repair of OPCs and OLs to ameliorate cognitive, behavioral, and motor dysfunction across a spectrum of neurodegenerative conditions.

Highlights.

• White matter injury is a generalizable phenomenon across a spectrum of chronic, neurodegenerative conditions, including Alzhemier’s disease (AD), Parkinson’s disease (PD), and amyotrophic lateral sclerosis (ALS).

• The oligodendrocyte (OL) lineage is disrupted early during neurodegeneration with dysfunction appearing before the onset of clinical symptoms.

• Despite different etiologies in AD, PD, and ALS, there are common pathologic cellular and molecular pathways that converge on the OL lineage including neuroinflammation, protein aggregation, lipid dysregulation, and organellar stress.

• Future work is needed to explore whether therapeutically restoring OL function can ameliorate clinical signs in patients with chronic neurodegenerative conditions.