Rethinking Spinocerebellar Ataxia Type 1: A Glial Perspective Beyond Purkinje Cell Degeneration

1. Introduction

Spinocerebellar ataxia type 1 (SCA1) is one of nine polyglutamine (polyQ) diseases, a group of inherited neurodegenerative disorders that also includes Huntington’s disease (HD), spinal and bulbar muscular atrophy (SBMA), dentatorubral-pallidoluysian atrophy (DRPLA), and several other spinocerebellar ataxias (SCA2, SCA3, SCA6, SCA7, and SCA17) [1]. Although these disorders share prominent motor deficits, they differ in the neuronal populations and central nervous system (CNS) regions that are most vulnerable to degeneration [1, 2]. SCA1 is an autosomal-dominant neurodegenerative disorder caused by a CAG repeat expansion in the ATXN1 gene, resulting in production and accumulation of polyQ-expanded ataxin-1 protein [1, 3]. Whereas wild-type ATXN1 alleles contain 4-36 CAG repeats, disease-causing alleles typically contain more than 39 uninterrupted repeats. Longer repeat expansions are associated with earlier disease onset [3]. Historically, SCA1 has been considered primarily a cerebellar disorder, characterized by selective Purkinje cell (PC) vulnerability and progressive motor impairment. Accordingly, decades of research have focused on PC-intrinsic pathogenic mechanisms. More recent studies demonstrate that SCA1-associated cellular and molecular abnormalities extend beyond cerebellar PCs, affecting multiple non-neuronal cell types and extracerebellar brain regions [4-6]. Because no disease-modifying treatments are currently available, defining the contribution of glial cells to SCA1 pathogenesis is essential. Oligodendroglia, astroglia, and microglia have emerged as potential therapeutic targets across a range of neurodegenerative and neurological conditions. Although glial contributions to polyQ disease pathophysiology have been most extensively characterized in HD [7], their roles in rarer polyQ ataxias, including SCA1, remain less well understood. In this review, we examine the molecular mechanisms through which glia contribute to SCA1 pathogenesis. We discuss the cell-autonomous and non-cell-autonomous roles of each glial subtype in disease progression and identify potential mechanisms shared across glial populations, therby highlighting new directions for research in polyQ disorders.

2. Oligodendroglia

2.1. White Matter Matters: Tracking Tract Abnormalities in SCA1

White matter (WM) consists primarily of axonal tracts ensheathed by myelin, which is generated and maintained by oligodendrocytes (OLs). WM volume loss was first identified as a pathological feature of polyQ disorders in HD in 1988 [8]. MRI subsequently identified WM abnormalities in SCA1 in 2001 [9]. Later quantitative MRI studies established WM degeneration as a prominent feature of SCA1. For example, diffusion tensor imaging (DTI) revealed significant microstructural abnormalities in the middle cerebellar peduncles (MCPs) and cerebellar hemispheric WM that correlated with clinical disease severity [10]. Voxel-based morphometry (VBM) further demonstrated symmetric macroscopic volume loss in peridentate WM and the MCPs, which correlated with disease duration and International Cooperative Ataxia Rating Scale (ICARS) scores [11].

Whole-brain tract-based analyses have also shown that WM damage in SCA1 extends beyond the cerebellum, affecting the corticospinal tracts and other supratentorial structures [12]. Comparative imaging studies of cerebellar ataxias, including SCA1, SCA2, SCA3, and multiple system atrophy with cerebellar features (MSA-C), have identified WM pathology as a shared feature across several ataxias [13, 14]. Within this shared pattern, patients with SCA1 and SCA2 exhibited more severe global WM atrophy than those with SCA3. In addition, cerebellar peduncle damage was most strongly associated with clinical severity in patients with SCA1 [14].

Longitudinal studies further suggest that WM pathology in SCA1 is spatially and temporally dynamic. Although the cerebellar peduncles show the most pronounced abnormalities, longitudinal diffusion MRI has demonstrated early, stage-dependent deterioration of corticospinal and pontine crossing tracts, followed by cortical microstructural changes [15]. Importantly, WM deficits can precede clinical onset. In preataxic individuals, cerebellar WM atrophy strongly predicted conversion to overt ataxia and outperformed clinical measures, including the Scale for the Assessment and Rating of Ataxia (SARA), in predicting subsequent disease progression [16-20].

Metabolic alterations in WM may also contribute to SCA1 progression. Magnetic resonance spectroscopy (MRS) identified significantly reduced total N-acetylaspartate (tNAA), a marker of axonal integrity, in the parietal WM of patients with SCA1 [21]. Lower baseline tNAA and glutamate levels in cerebellar WM also predicted more rapid worsening of SARA scores [22]. Neuropathological and biochemical analyses of postmortem SCA1 tissue corroborate these imaging findings, demonstrating widespread WM degeneration throughout the cerebellum and brainstem, as well as reduced numbers of mature OLs and myelinated axons in the cerebellum [5, 23]. Single-nucleus RNA sequencing (snRNA-seq) of postmortem SCA1 cerebellum further revealed transcriptomic dysregulation in both oligodendrocyte progenitor cells (OPCs) and OLs, including reduced expression of key myelin-related genes in OLs [5]. Together, these findings establish WM pathology as a core feature of SCA1, alongside PC degeneration.

2.2. Oligodendroglial Lineage Development, Myelination, and Functional Heterogeneity

The oligodendroglial lineage comprises OPCs, immature OLs, and mature myelinating OLs. This lineage arises through a tightly regulated developmental program that begins during embryogenesis and continues throughout postnatal life [24]. Neural progenitor cells (NPCs) generate OPCs following induction of the transcription factors OLIG2 and SOX10, which remain expressed throughout lineage progression. OPCs are further characterized by expression of markers such as PDGFRα and CSPG4 (NG2) [25]. These cells are proliferative and migratory, are distributed throughout the CNS, and serve as the principal source of new OLs for developmental myelination, adult remyelination, and adaptive myelination [26, 27].

As OPCs differentiate, they downregulate progenitor-associated programs and upregulate markers including GalC, O4, TCF7L2, and MYRF as they transition to immature OLs [26]. TCF7L2, a Wnt/β-catenin co-effector, has a particularly important role at this stage. Wnt/β-catenin signaling through TCF/LEF family transcription factors can either delay or promote OL differentiation depending on developmental timing and brain region. Thus, endogenous Wnt signaling is highly context dependent and must be precisely regulated to support normal oligodendroglial lineage progression [28].

Mature OLs generate and maintain the multilamellar myelin membrane that wraps internodal axonal segments [29, 30]. Major myelin proteins, including MBP, PLP1, MAG, MOG, and CNP, turn over slowly; for example, PLP1 has a half-life of approximately six months [31-33]. Myelin provides mechanical and biochemical protection to axons and enables rapid saltatory conduction by restricting ionic currents to the nodes of Ranvier [29, 30]. OLs also regulate conduction velocity by modifying internode length and myelin thickness, commonly measured by the g-ratio; both properties vary across CNS regions and neuronal subtypes [34].

In addition to myelination, OLs provide critical metabolic support to axons. They convert glucose to lactate and export it through monocarboxylate transporters (MCT1/MCT2) to support axonal mitochondrial metabolism in an activity-dependent manner [29, 30, 35]. OLs can also enhance axonal ATP production by delivering SIRT2-containing exosomes to axon [36]. Furthermore, they help maintain the periaxonal environment by regulating ionic flux and providing trophic factors that support axonal caliber and mitochondrial density [29, 30].

Single-cell transcriptomics studies have revealed temporally and spatially distinct oligodendroglial populations [37-40]. OLs acquire region-specific identities that help shape local myelin architecture, and disruption of this spatial organization can produce substantial functional impairment [41]. In disease, OLs undergo marked transcriptomic and compositional changes [42-45] and may adopt disease-associated oligodendrocyte (DAO) states that contribute to tissue injury and repair in Alzheimer’s disease (AD) and multiple sclerosis (MS) [46, 47].

2.3. Oligodendroglia Dysfunction and Myelin Pathology in SCA1 Mouse Models

Consistent with the WM abnormalities observed in patients with SCA1, longitudinal snRNA-seq of the Atxn1^154Q/2Q knock-in (SCA1-KI) mouse model revealed that oligodendroglial dysfunction and impaired myelination are early transcriptomic events [5]. In this model, polyQ-expanded ataxin-1 is expressed endogenously in all cerebellar cell types [48]. Transcriptomic abnormalities in OPCs and OLs preceded PC degeneration, and OLs showed robust downregulation of myelin-related genes as early as 12 weeks of age [5]. At the same age, SCA1-KI mice exhibited progressive biochemical and ultrastructural demyelination of PC axons and impaired differentiation of OPC into OLs. These changes coincided with the onset of motor phenotypes but preceded overt PC loss [5].

Studies of primary OPC cultures from SCA1-KI mice, as well as an oligodendroglial lineage-specific conditional knock-in model (OL-SCA1-cKI), indicate that deficits in OL differentiation and myelination can occur through cell-autonomous mechanisms [5, 49]. In contrast, the B05 transgenic model (Pcp2-ATXN1^82Q), which selectively expresses mutant ataxin-1 in PCs, develops severe PC pathology without early demyelination at 15 weeks of age [5]. Although B05 mice show reduced WM area at later disease stages, this finding likely reflects overall cerebellar atrophy or the loss of myelinated axons rather than cell-autonomous OL dysfunction [50].

Importantly, findings from OL-SCA1-cKI mice suggest that OLs contribute to SCA1 pathogenesis beyond demyelination. Intact OLs and myelin are required for normal PC development and function [51-54], whereas SCA1-associated OLs adversely affect PC health through non-cell-autonomous mechanisms. OL-SCA1-cKI mice developed progressive PC axonal pathology, including reduced axonal caliber reduction before the onset of demyelination and, subsequently, axonal torpedoes in the granule cell layer (GCL), without overt PC somata [49]. PC torpedoes are a characteristic pathological feature of SCA1 in humans [5, 23].

OL dysfunction also induced early non-cell-autonomous transcriptomic changes in PCs. snRNA-seq of OL-SCA1-cKI cerebella identified increased expression of genes involved in postsynaptic organization in PCs [49], a signature also observed in constitutive SCA1-KI mice [5]. In addition, PCs in OL-SCA1-cKI mice exhibited increased dendritic length [5, 49]. This phenotype may represent aberrant or compensatory synaptic remodeling in response to OL dysfunction, consistent with evidence that OLs can influence PC dendritic arborization [51] and spine turnover [55]. Notably, dendritic lengthening was absent in posterior cerebellar lobules [49], regions that are preferentially affected in SCA1 [56, 57]. This findings raise the possibility that insufficient region-specific compensatory remodeling in response to OL dysfunction contributes to the selective vulnerability of posterior cerebellar regions.

OL-SCA1-cKI mice also accumulated myelin debris in the cerebellum and hippocampus. Myelin debris can destabilize nodes of Ranvier and exacerbate axonal injury [49, 58]. Although overt cerebellar astrogliosis and microgliosis were not detected in OL-SCA1-cKI mice, snRNA-seq revealed microglial transcriptomic dysregulation, suggesting a more subtle immune response [49]. The accumulation of myelin debris in particularly notable in light of a reported case of co-occurring SCA1 and myelin oligodendrocyte glycoprotein (MOG) antibody-associated disease (MOGAD) [59]. This observation raises the possibility that persistent myelin debris contributes to secondary immune activation. Moreover, wild-type ataxin-1 regulates B-cell function and protects against demyelination in the experimental autoimmune encephalomyelitis (EAE) model [60-62]. Thus, polyQ-expanded ataxin-1 may affect both OL biology and adaptive immunity, potentially amplifying demyelination through a positive-feedback loop.

2.4. Unraveling the Molecular Mechanisms Driving Oligodendrocyte Dysfunction in SCA1

Transcriptomic analyses of SCA1 cerebellum have identified broad dysregulation of genes involved in myelination and axonal maintenance, as well as altered OL subtype composition. These changes suggest impaired neuron-OL communication and diminished neuroprotective signaling [5, 49]. Upstream-regulator analyses have implicated TCF7L2 and huntingtin (HTT) as candidate drivers of these transcriptional abnormalities, pointing to two potentially convergent mechanisms: (1) ataxin-1-specific dysfunction and (2) polyQ-dependent toxicity shared across diseases.

Evidence supports an intrinsic role for wild-type ataxin-1 in oligodendroglial function and WM integrity. Although long CAG expansions (≥39 repeats) cause fully penetrant SCA1 [63], intermediate expansions (32-39 repeats) are associated with an increased risk of sporadic amyotrophic lateral sclerosis (ALS) [64, 65]. In addition, intronic variants that reduce ATXN1 expression have been associated with increased risk of AD [66, 67] and MS [61, 68]. Because OL dysfunction and myelination deficits are also prominent features of these disorders [42, 69-73], these genetic associations suggest that ataxin-1 may have broader roles in myelin maintenance beyond SCA1.

Within the OL lineage, Wnt/β-catenin signaling may be a key ataxin-1-specific mechanism. Both wild-type and polyQ-expanded ataxin-1 physically interact with TCF7L2 [57] (Figure 1), a central transcriptional effector of Wnt/β-catenin signaling and an important regulator of the transition from OPCs to immature OLs [28]. TCF7L2 activity is predicted to be repressed in SCA1 oligodendroglia [49], implicating disruption of the ATXN1-TCF7L2 axis in impaired OL differentiation.

More broadly, OL dysfunction occurs across several polyQ diseases, including SCA2, SCA3, HD, and SBMA, suggesting that polyQ-dependent mechanisms may act across disease contexts independent of the host protein [8, 74-81]. Consistent with this possibility, TCF7L2 and HTT have also been identified as upstream regulators of OL dysfunction in HD and SCA3 [82, 83]. Together, these findings raise the possibility that expanded polyQ tracts converge on a shared transcriptional pathway. Whether this convergence reflects direct effects within OLs, however, remains an open question.

3. Astroglia

3.1. Astroglia Diversity in the Cerebellum: Bergmann Glia in Focus

Astroglia are, named for their star-like morphology, are commonly classified by their location in CNS WM or gray matter (GM). The cerebellum, however, contains several morphologically distinct astroglial subtypes, including radial Bergmann glia (BG) and Fañanas cells, fibrous astrocytes, velate astrocytes, and deep cerebellar nuclei (DCN) astrocytes [84-86]. Although a detailed discussion of cerebellar astroglial markers is beyond the scope of this review, these evolutionarily conserved astroglial populations can be identified histologically using ALDH1L1 and GFAP, together with markers such as S100β, glutamine synthetase (GS), and SOX9 [87]. These subtypes occupy distinct cerebellar microenvironments. BG and Fañanas cells are located in the Purkinje cell layer (PCL) and molecular layer (ML); velate astrocytes localize to the GCL; fibrous astrocytes reside in the WM; and DCN astrocytes occupy the DCN. Because BG are closely associated with PCs, the neuronal population most affected in SCA1, they are well positioned to influence SCA1 pathogenesis.

3.2. Mechanisms of Astroglial Involvement in SCA1 Pathogenesis

In 2006, Custer et al. showed that selective overexpression of mutant ataxin-7 in astroglia was sufficient to induce SCA7-like motor deficits, including hindlimb clasping and impaired performance on an accelerating rotarod [88]. This work, together with studies in other neurodegenerative disorders, helped shift the view of astroglia from passive support cells to potential drivers of disease pathogenesis [88, 89]. Astroglia have since been studied extensively in SCA1, particularly in the SCA1-KI and B05 mouse models [48, 90]. In both models, reactive astrogliosis, typically assessed by increased GFAP expression, occurs early, at 8 weeks in SCA1-KI mice and at 4 weeks in the PC-specific B05 model [6]. Because astroglial reactivity precedes overt PC neurodegeneration in both models [6, 48, 90, 91], astroglia may actively contribute to degeneration rather than simply respond to neuronal dysfunction. However, this possibility requires direct testing. Newly developed ON/OFF systems that enable cell type-specific manipulation of mutant ataxin-1 [49, 92] may help define the contribution of astroglia to SCA1 progression and pathology.

3.2.1. Bergmann Glia as the Cerebellar Glutamate-Recycling Hub

A core homeostatic function of astroglia is the clearance of synaptic glutamate through the transporters GLAST and GLT-1 [93]. In the glutamate-glutamine cycle, astroglial take up glutamate and convert it to glutamine via glutamine synthetase. Glutamine is then returned to presynaptic neurons to sustain neurotransmission [94]. Disruption of astroglial glutamate transporters or glutamine synthetase can therefore promote glutamate spillover, excitotoxicity, and neurodegeneration [93,95].

Reduced GLAST expression has been observed in both an astroglia-conditional ON SCA7 mouse model and in the SCA1 B05 model, suggesting that impaired glutamate recycling may arise through both cell-autonomous and non-cell-autonomous mechanisms [88, 93, 95, 96]. Because GLAST expression is enriched in the PCL and ML [97], glutamate-handling deficits in SCAs are primarily attributed to BG rather than other cerebellar astroglial populations. In vivo GLAST deletion is sufficient to produce mild ataxia, including rotarod impairment [93], and disrupts parallel fiber and climbing fiber innervation of PC dendrites [98].

Although astroglia mediate most glutamate uptake in the cerebellum, PCs also express the glutamate transporter EAAT4, which may partially compensate for glutamate spillover [99]. In addition, PC-expressed Down syndrome cell adhesion molecule (DSCAM) interacts with GLAST to support proper parallel fiber and climbing fiber synaptic organization [100], highlighting close functional coupling between PCs and BG. Loss of DSCAM and associated deficits in excitatory synaptic innervation can be partially rescued by riluzole, an enhancer of glutamate transporter function [101]. Troriluzole a riluzole prodrug [102], activates GLAST and GLT-1 and has advanced to clinical testing for SCAs. In Study 206-RWE, Troriluzole was reported to slow disease progression by 50-70% relative to untreated controls [103].

3.2.2. Stage-Specific Roles of NF-κβ Signaling in SCA1 Astroglia

Beyond synaptic homeostasis, astroglia respond to their microenvironment through signaling pathways such as nuclear factor-κβ (NF-κβ). NF-κβ is a family of transcription factors expressed broadly in CNS cell types, including astroglia, microglia, and OLs [104, 105]. Its activation is commonly triggered by injury or disease and has been investigated as a regulator and therapeutic target of astrogliosis [106, 107]. In the SCA1 B05 model, astroglia isolated from 8-week-old cerebella showed increased NF-κβ phosphorylation, indicating inflammatory pathway activation in response to altered PC health [108]. However, the effects of astroglial NF-κβ modulation appear to be disease- and stage-dependent. Although NF-κβ suppression can reduce astrogliosis and improve outcomes in models of spinal cord injury, MS, glaucoma, vascular dementia [106, 107, 109, 110], timing is critical in SCA1. Conditional NF-κβ silencing in BG at 8 weeks of age in B05 mice (ATXN1[82Q]; IKKβF/F; Slc1a3-cre/ERT) reduced astrogliosis but accelerated ataxia and worsened PC pathology, ML thinning, and climbing fiber innervation of PC dendrites [108] (Figure 2). In contrast, silencing NF-κβ in BG at 16 weeks partially improved motor phenotypes and PC neurodegeneration [108] (Figure 2). Together, these findings suggest a biphasic role for astroglial NF-κβ signaling: it may be protective early in disease abut detrimental at later stages, consistent with observations in ALS and AD models [111,112].

3.2.3. JNK Activation in SCA1 Astroglia

Reactive astrogliosis is also regulated by c-Jun N-terminal kinase (JNK) signaling. The relationship between JNK activation and GFAP accumulation, a hallmark of reactive astrogliosis, was initially investigated in Alexander disease (AxD), an astroglial protein-aggregation disorder [114]. In COS-7 cells, expression of AxD-associated GFAP^R239C [115] resulted in intracellular GFAP inclusions and increased JNK phosphorylation [116]. Similarly, activation of upstream JNK kinases, including MLK2, MLK3, and ASK1, increased GFAP accumulation, supporting a reciprocal relationship between JNK signaling and GFAP dysregulation [116].

In the cerebellum, increased c-Jun, a key JNK target, has been observed in mouse models of PC degeneration (pcd) and granule cell degeneration (weaver) [117-119]. In postmortem cerebellar tissue from individuals with SCA1, SCA2, and SCA7, phosphorylated c-Jun accumulated prominently in the PCL and ML [113]. In SCA1-KI mice, c-JUN phosphorylation at serine 63, a JNK target residue, was particularly enriched in BG, as shown by S100β co-staining [113] (Figure 2). Treatment of SCA1-KI mice with the blood-brain barrier-penetrant JNK inhibitor SP600125 reduced BG phospho-c-Jun staining and GFAP fluorescence intensity, improved accelerating rotarod performance, and preserved PC dendritic arborization [113] (Figure 2). These findings implicate JNK signaling in BG reactivity and suggest that JNK inhibition may mitigate both glial and neuronal phenotypes. However, because SP600125 broadly suppresses JNK signaling, cell type-specific genetic studies are needed to determine whether JNK inhibition in BG alone is sufficient to rescue SCA1-associated PC neurodegeneration.

3.3. Developmental Origins of Astroglial Dysfunction in SCA1

Although astroglial reactivity is a hallmark of SCA1 pathogenesis, cerebellar astroglia are also essential for normal cerebellar development. SCA1 onset and severity are closely associated with polyQ tract length, with longer CAG repeats correlating with earlier onset and more severe disease in both patients and mouse models [3, 120]. In addition to intergenerational repeat expansion, the SCA1 CAG repeat is somatically unstable across brain regions and cell types [121], as observed in other polyQ disorders, including HD [122, 123]. These findings underscore the need to examine how CAG-repeat instability influences disease across the lifespan, rather than focusing exclusively on post-symptomatic pathology. Consistent with this view, transcriptomic changes occur across cerebellar cell types in SCA1 models before overt PC degeneration [5, 124, 125]. At postnatal day 7 (P7), SCA1-KI mice show increased cerebellar neural stem-cell proliferation, including elevated Ki67⁺/Nestin⁺ progenitor populations relative to wild-type controls [126]. This phenotype is absent in Atxn1-null mice, indicating that it depends on mutant ataxin-1 expression [126].

Although the contribution of these early progenitor changes to later disease onset remains unclear, normal cerebellar development depends on functional BG, positioning astroglia as potential determinants of early SCA1 vulnerability. Rakic’s landmark work established BG as migratory scaffolds for granule cells. In the weaver mouse model, BG hypertrophy precedes abnormal granule-cell migration [127], as well as later neurodegeneration and ataxia [118, 128].

Wnt-1 (int-1) is also required for cerebellar development [129, 130]. During embryogenesis (E14.5), Wnt-1 is expressed in granule-cell progenitors, whereas by P10 its expression is restricted to BG [130]. Conditional disruption of Wnt/β-catenin signaling in BG through Cre-mediated deletion of floxed Ctnnb1 disrupts foliation and lamination, causes astroglial mislocalization, alters granule-cell and PC positioning, and produces ataxia [131, 132]. These findings identify BG-mediated Wnt signaling as a key regulator of cerebellar morphogenesis and a plausible contributor to the vulnerability of SCA1-relevant neuronal populations.

Aberrant Wnt/β-Catenin Activation in SCA1 Astroglia

Wnt/β-catenin signaling is abnormally increased in the SCA1 cerebellum [57]. Expression of the canonical Wnt target genes Ccnd1 and c-Myc increases with age in SCA1-KI cerebella [57] (Figure 1). In HeLa cells, co-expression of ataxin-1 and β-catenin increased Wnt/β-catenin reporter activity in a polyQ length-dependent manner: ataxin-1 [30Q] and ataxin-1 [82Q], but not ataxin-1 [2Q], significantly increased signaling [57]. Similarly, PC-specific SCA1 transgenic model (Pcp2-ATXN1^63Q) exhibit increased Ccnd1, c-Myc, and active β-catenin expression [57]. In Pcp2-ATXN1^63Q; TCF/Lef:H2B/GFP reporter mice, GFP signal was increased in PCs and neighboring PCL cells, implicating additional cell types in Wnt-pathway activation [57]. Importantly, PC-specific suppression of Wnt/β-catenin signaling through Ctnnb1 deletion failed to attenuate SCA1 phenotypes, whereas PC-specific pathway hyperactivation through Apc suppression was insufficient to induce disease-associated changes [57]. These findings point to a non-PC contributor to Wnt dysregulation. In contrast, conditional Wnt/β-catenin activation in astroglia under a Gfap promoter induced astroglial reactivity in the PCL and ML and reduced expression of BG-enriched GLAST [57] (Figure 1). Both are established astroglial features of SCA1 [6, 96]. Although it remains unclear whether BG-specific Wnt/β-catenin inactivation can rescue SCA1 neurodegeneration, Ctnnb1 siRNA knockdown in C8-S astrocytes expressing Flag-hATXN1 [82Q] reduced cleaved caspase-3 (CC3) levels in co-cultured N2A cells [57]. Together, these findings support further investigation of astroglial Wnt/β-catenin inhibition as a strategy to slow SCA1 progression and pathology.

4. Microglia

4.1. Microglia: Hero, Villain, or Bystander?

Microglia (MG) are highly heterogeneous, mesoderm-derived glial cells that serve as the resident macrophages of the CNS. Their functions differ from those of circulating monocytes and other tissue-resident macrophages [133]. MG originate from yolk sac progenitors, enter the CNS during early embryonic development, and self-renew locally throughout adulthood [133]. Under homeostatic conditions, they support CNS function through synaptic pruning [134, 135], phagocytosis [136], debris clearance [137], myelin maintainance [138], and regulation of neural circuitry [139]. However, MG also contribute to neuroinflammation following injury and in neurodegenerative disease [138, 140, 141]. Whether a microglial response is beneficial or harmful depends on the local microenvironment, disease stage, and brain region.

Microglial States Beyond the M1/M2 Binary

Historically, activated MG were classified using the M1/M2 polarization framework. M1 MG were associated with pro-inflammatory functions, including reactive oxygen species production [140], whereas M2 MG were associated with phagocytosis [141] and release of anti-inflammatory cytokines, such as IL-4, IL-10, IL-13, TGF-β [142-145]. However, snRNA-seq and single-cell RNA sequencing (scRNA-seq) have shown that this binary model does not adequately capture microglial diversity. MG adopt distinct transcriptional states across development, aging, injury, and disease [146-150] (Table 1).

For example, scRNA-seq of the healthy murine brain identified nine microglial subclusters present from E14.5 through adulthood (P540) [146]. With aging, the proportion of MG clusters expressing inflammatory genes, including Ccl3, Ccl4, and Il1b, increased [146]. Disease can further induce microglial states that differ from both homeostatic and age-associated populations.

Disease-associated microglia (DAM) were first described in the 5xFAD model of AD, where they were proposed to limit plaque accumulation and protect against neurodegeneration [150]. Subsequent studies showed that DAM functions are context-dependent. In early AD, both DAM and homeostatic MG may contribute to synaptic loss and plaque deposition, whereas later-stage DAM may exert protective effects [159-164]. Moreover, pro- and anti-inflammatory DAM-like states can coexist within the same disease stage [164]. Although DAM are often discussed in the context of AD, DAM-like transcriptional programs have also been reported in MS, PD, TBI, and HD. Together, these findings demonstrate that MG adopt disease-, region-, and stage-specific stages that are not captured by the traditional M1/M2 framework.

4.2. Molecular Mechanisms of Microglial Activation and Dysfunction

Although transcriptomic studies have greatly expanded our understanding of MG diversity, the functional consequences of specific migroglial states remain incompletely defined. MG may be protective in one disease context but harmful in another, and their effects can vary by brain region and disease stage [165]. Three pathways recurrently implicated in neurodegeneration are: (1) major histocompatibility complex class II (MHC-II)-mediated antigen presentation [166], (2) complement-dependent synaptic pruning [134], and (3) triggering receptor expressed on myeloid cells 2 (TREM2) signaling [163, 167, 168]. Each pathway can support homeostasis or promote pathology depending on context.

4.2.1. MHC-II-Mediated Antigen Presentation

MHC-II presents peptides derived from extracellular or phagocytosed material to CD4⁺ T cells [169]. Under homeostatic conditions, MG express low levels of MHC-II and are not considered major antigen-presenting cells (APCs), while border-associated macrophages serve this role more prominently [151]. However, inflammatory cytokines, tissue damage, or infiltrating immune cells can induce MHC-II expression in MG, enabling them to sample and present local antigens, including self-antigens. MHC-II-mediated antigen presentation contributes to normal CNS immune surveillance [170]. In autoimmune diseases such as MS, however, it may become pathogenic when MG present myelin-derived peptides to autoreactive T cells and locally amplify theimmune response [155, 171].

4.2.2. Complement-Dependent Synaptic Pruning

MG also participate in the classical complement cascade, which labels synapses for elimination during development and normal circuit refinement [134]. In neurodegenerative disease, this pathway can be aberrantly reactivated, resulting in excessive or inappropriate synaptic elimination in the adult CNS. In AD models, complement-mediated synapse loss has been proposed as an early contributor to cognitive decline that precedes neuronal death [172]. Thus, a physiological pruning mechanism may become pathogenic when inappropriately activated.

4.2.3. TREM2-Dependent and TREM2-Independent Signaling

TREM2 regulates MG survival, proliferation, phagocytosis, and metabolism and is required for the adoption of disease-responsive states, including DAM-like programs in AD models [151, 167]. In APP/PS1 mice, Trem2 deficiency reduces microglial clustering around plaques and impairs phagocytic clearance, worsening pathology [163]. These findings supports a protective role for TREM2 in plaque containment and clearance. TREM2 may also promote microglial survival through Wnt/β-catenin signaling [173]. TREM2 signals through DAP12 to activate PI3K/Akt, which phosphorylates and inactivates GSK3β. This stabilizes β-catenin and promotes expression of Wnt/β-catenin target genes, including Ccnd1, c-Myc, Bcl-2, that support proliferation and survival (Figure 3). In cultured Trem2-null MG, loss of TREM2 accelerated β-catenin degradation. Pharmacological activation of Wnt signaling with Wnt3a or inhibition of GSK3β with LiCl or TDZD-8 partially restored β-catenin levels, proliferation, and viability [173, 174].

Wnt signaling may also promote MG inflammation through non-canonical pathways that are independent of TREM2 and canonical β-catenin signaling. An early study proposed that Wnt3a binding to Frizzled receptors (FZD4/5/7/8) and LRP5/6 activated canonical Wnt/β-catenin signaling and induced expression of IL-6, IL-12, and TNFα [174]. More recently, Wnt3a was shown to enhance TNFα release in lipopolysaccharide (LPS)-primed MG through non-canonical JNK/ planar cell polarity (PCP) and PLC/Ca²⁺ pathways [175]. Inhibition of canonical Wnt signaling did not block this effect, whereas inhibition of either non-canonical pathway did [175]. Wnt3a alone had no effect in unprimed MG, suggesting that inflammatory priming is required [175].

Together, these findings suggest at least two Wnt-linked programs in MG: a TREM2-dependent pathway that promotes β-catenin-mediated survival and proliferation, and a potentially TREM2-independent pathway in which Wnt ligands amplify inflammatory cytokine release through non-canonical signaling. How these pathways interact in vivo, and whether TREM2 loss shifts MG toward inflammatory signaling, remain unresolved.

4.3. Microglial Contributions to PolyQ Diseases

MG involvement in polyQ diseases, including HD and SCAs, remains understudied. However, emerging evidence supports both cell-autonomous and non-cell-autonomous roles. In HD, mutant HTT is expressed in neuronal and non-neuronal cells. Nuclear mutant HTT inclusions have been observed in MG from the frontal cortex of adult-onset HD and from the frontal cortex and striatum of juvenile-onset HD [176]. Primary MG isolated from mutant HTT (Hdh^175Q/175Q) knock-in mice induced apoptosis in embryonic stem cell-derived neurons cultured on wild-type astrocytes, whereas wild-type MG did not [176]. Similarly, mice expressing mutant HTT selectively in MG exhibited greater neuronal death under sterile inflammatory conditions than control mice [176].

Microglial toxicity may also be induced by expanded polyQ tracts independently of the host protein. Conditioned medium from BV2 MG-like cells treated with a 69Q peptide, but not a 15Q peptide, induced neurite retraction, reduced branch points, and impaired neurite outgrowth in PC12 cells [156]. These effects were MG-specific, as conditioned medium from polyQ-treated SH-SY5Y neuroblastoma cells did not reproduce the phenotype [156]. Thus, expanded polyQ sequences may be sufficient to induce MG-derived factors that compromise neuronal process integrity in vitro.

Microglial depletion studies in SCAs have yielded mixed findings. In a SCA3 mouse model, MG displayed activated morphology at a mid-disease stage [157]. Treatment with the CSF1R inhibitor PLX3397 reduced MG numbers by 42–59% in the cerebellum and brainstem but did not improve balance, coordination, gait, strength, or reflexes [157]. Importantly, PLX3397 did not normalize MG morphology in SCA3 mice and induced an activated-like phenotype in MG from wild-type mice. Clustering analyses grouped PLX3397-treated wild-type MG with SCA3 MG [157]. These findings complicate interpretation because CSF1R inhibition may alter the phenotype of residual MG rather than simply restore a homeostatic state.

In SCA1, MG function remains poorly characterized. Early postmortem studies reported gliosis in cerebellar regions but did not identify the predominant glial subtype [177, 178]. The first studies to examine MG in SCA1 assessed activation in SCA1-KI and B05 mice [6, 90]. At 8 and 12 weeks, SCA1-KI mice showed early microgliosis in the cerebellar cortex, including a 43.5% increase in Iba1⁺ cells relative to wild-type mice. In the PC-specific B05 model, robust MG activation was evident as early as 4 weeks [6]. Because mutant ataxin-1 expression in B05 mice is restricted to PCs, these studies suggest that PC dysfunction may be sufficient to trigger MG activation before overt neuronal death.

The functional impact of MG was later assessed by depleting 70-80% of MG in B05 mice with PLX3397 [158]. Sustained depletion partially improved motor performance at 3 months, as measured by rotarod and balance-beam tests, and reduced TNFα and monocyte chemoattractant protein-1 (MCP-1) levels. However, MG depletion did not reduce astrogliosis or prevent PC loss [158]. PLX3397-treated B05 mice also showed increased levels of the postsynaptic density protein PSD95 without a corresponding increase in the presynaptic protein synapsin [158]. This pattern may reflect selective preservation of postsynaptic structures and could contribute to modest behavioral improvement without restoring synaptic integrity or neuronal survival. These results should be interpreted cautiously. CSF1R inhibitors deplete MG regardless subtype, potentially eliminating both harmful and protective populations. In addition, their effects extend across the brain, which can obscure region-specific MG functions. Consistent with this concern, microglial activation in SCA1 appears region-specific. In a 2022 study of SCA1-KI motor cortex, Luttik et al. found no overall increase in microgliosis at 12 or 30 weeks, despite a trend toward increased MG numbers at 12 weeks [4]. This contrasts with significant microgliosis in the cerebellum and inferior olive at comparable time points [179]. These findings suggest that glial activation may parallel region-specific neuronal vulnerability in SCA1. They also highlight a limitation of relying on pan-microglial markers such as Iba1, which may fail to detect changes in discrete MG subpopulations.

Overall, studies in HD, SCA3, and SCA1 indicate that MG respond to polyQ disease and can influence functional outcomes, but t the molecular programs underlying distinct MG states remain unclear. In SCA1, reduced inflammatory cytokines and increased PSD95 after MG depletion may reflect changes in inflammatory, clearance-related, or synapse-associated MG functions [6, 158]. Whether these effects involve TREM2-dependent pathways, as reported in some AD models [159, 163, 164], remains unknown and should not be assumed given the strong disease-context dependence of MG signaling. A consistent conclusion is that MG activation in SCA1 is not uniform: robust microgliosis in the cerebellum and inferior olive contrasts with minimal microgliosis in the motor cortex at matched time points [4, 179]. Determining how protective and pathogenic MG programs vary across brain regions and disease stages will be essential for defining their role in SCA1.