Glial Cells in the Early Stages of Neurodegeneration: Pathogenesis and Therapeutic Targets

In addition to neurons, glial cells are major components of the central nervous system (CNS), with astrocytes, microglia, and oligodendrocytes together comprising ~50% of all cells [1,2,3]. In the peripheral nervous system (PNS), Schwann and satellite cells predominate. While neurons mediate signal transmission, nervous-system maintenance depends on neuroglia, which are increasingly implicated in neurodegenerative diseases such as Alzheimer's [3,4,5,6,7,8]. Glial functions include ion homeostasis, structural support, and regulation of the extracellular environment. Astrocytes, the largest glial population (~40%), modulate synaptic activity, support axonal growth, contribute to the blood–brain barrier (BBB), and form glial scars [8,9,10,11]. Microglia (~30%) act as resident immune cells, protecting against infection, clearing debris, and supporting neurogenesis and synaptic plasticity [3,5,12]. Both astrocytes and microglia secrete diverse signaling molecules—neurotransmitters, neuromodulators, hormones, growth factors, and inflammatory mediators—via vesicular transport [3,13,14,15,16].

One of the most widely accepted hypotheses in neurodegeneration proposes that microglia undergo a phenotypic switch from an M2 (Figure 1A), neuroprotective state to an M1, proinflammatory state (Figure 1B). This transition is driven by proinflammatory cytokines, including interferon-gamma (IFN-γ), lipopolysaccharide (LPS), and other immune mediators, which tip the balance toward an M1 profile [50] (Figure 1). In the context of neurodegenerative diseases, extensive tissue damage and microbial invasion further amplify signals that favor the M1 phenotype [51].

The resulting exaggerated inflammatory response releases neurotoxic factors such as reactive oxygen species, nitric oxide, and proinflammatory cytokines, which can directly injure neurons and exacerbate cell death [40,47,48,52]. Accordingly, chronically reactive microglia are now recognized as key contributors to both the initiation and progression of neurodegenerative processes [7,12,15,53,54,55].

A central regulator of this phenotypic shift is Toll-like receptor 4 (TLR4), a transmembrane pattern-recognition receptor expressed on microglia that activates the innate immune response. In the APP/PS1 mouse model of Alzheimer's disease, which overexpresses mutant human amyloid precursor protein (APP) and presenilin 1 (PS1) and accumulates β-amyloid, microglia display increased TLR4 expression compared with wild-type controls [56]. Engagement of TLR4 by LPS or endogenous damage-associated molecular patterns activates the myeloid differentiation primary response 88 (MyD88)-dependent pathway, leading to nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) translocation and transcription of M1-associated genes (Figure 1B) [56].

TLR4 is a type I transmembrane receptor with an extracellular leucine rich repeat (LRR) ligand-binding domain, a single transmembrane helix, and a cytoplasmic Toll/IL 1 receptor (TIR) signaling domain. Its activation requires dimerisation and accessory proteins such as myeloid differentiation factor-2 (MD-2) and cluster of differentiation 14 (CD14), which enable binding of LPS or endogenous damage associated molecular patterns. Structural studies show that ligand binding induces conformational changes in the LRR domain, promoting TLR4/MD-2 dimerization and juxtaposition of the TIR do-mains, which then recruit adaptor proteins (MyD88 or TRIF) to trigger NF κB and interferon regulatory factor (IRF) pathways [57].

Experimental inhibition of TLR4 attenuates NF-κB activation and promotes a return toward the M2 phenotype. Lang et al. [58] demonstrated in vitro that blocking LPS-induced TLR4 signaling reduced NF-κB activity and shifted microglia from M1 to M2. Complementary studies have shown that TLR4 blockade suppresses the MyD88/NF-κB/ NOD-like receptor family pyrin domain containing 3 (NLRP3) inflammasome axis, lowering NLRP3 levels and curbing the release of interleukin-1β and other proinflammatory mediators [56], (Figure 1B).

In summary, pathological stimuli in neurodegenerative diseases drive microglial polarization toward a neurotoxic M1 state, with TLR4 serving as a pivotal mediator. Therapeutic strategies aimed at inhibiting TLR4 or its downstream signaling pathways hold promise for preserving microglial homeostasis and mitigating neuronal injury. Nevertheless, because M1 polarization often follows initial tissue damage or infection, effective interventions may also need to address these upstream insults.

In addition to receptor-mediated signaling, the redox environment strongly influences microglial polarization. Enzymes such as superoxide dismutase (SOD), which require copper and zinc cofactors for catalytic activity, play a central role in detoxifying superoxide radicals. Alterations in metal availability or SOD function can exacerbate oxidative stress, thereby favoring NF-κB activation and a shift from the anti-inflammatory M2 phenotype toward the proinflammatory M1 state [59].

NLRP3 is a member of the nucleotide-binding oligomerization domain (NOD), leucine-rich repeat (LRR), and pyrin domain-containing (NLRP) receptor family within the broader NOD-like receptor superfamily. These cytosolic receptors detect pathogen-associated or damage-associated molecular patterns delivered via phagocytosis or membrane pores. In microglia, NLRP3 senses danger signals—such as extracellular adenosine triphosphate (ATP), uric acid crystals, or elevated extracellular K+ from damaged cells—and assembles into the NLRP3 inflammasome, a multiprotein complex composed of the NLRP3 sensor, the ASC adaptor (apoptosis-associated speck-like protein containing a CARD), and the effector protease caspase-1 (ICE) (Figure 1B). Activation of caspase-1 drives the maturation and release of IL-1β and interleukin-18 (IL-18) and triggers pyroptosis, a form of proinflammatory cell death mediated by gasdermin D pores [60]. Beyond its structural assembly, the NLRP3 inflammasome exerts its function through the catalytic activity of caspase-1. Caspase-1 is a cysteine protease whose active site contains a critical cysteine residue that performs nucleophilic attack on peptide bonds, enabling cleavage of pro-IL-1β and pro-IL-18 into their mature, bioactive forms. This cysteine-dependent mechanism is essential for inflammasome-driven cytokine release and pyroptotic cell death, thereby linking structural activation of the NLRP3/ASC/caspase-1 complex to downstream neuroinflammatory processes [52,61,62,63].

In neurodegenerative disease, M1-polarized microglia may amplify NLRP3 activation (Figure 1B). Extracellular ATP engages P2X purinergic receptor 7 (P2X7) on microglia, whose expression rises in late-stage Alzheimer's disease [64,65]. P2X7 stimulation promotes K+ efflux and NLRP3 assembly. Moreover, microglia release ASC-containing exosomes that propagate inflammasome activation in neighboring cells [66]. In the APP/PS1 Alzheimer's mouse 196 model, genetic deletion of NLRP3 or caspase-1 reduces cognitive deficits, shifts microglia toward an M2 phenotype, and enhances β-amyloid clearance [67]. Because P2X7 receptors require high (millimolar) ATP concentrations for activation—far above the micromolar levels that activate other P2X subtypes—robust NLRP3-mediated signaling occurs only after extensive membrane [43,68] (Figure 1B). However, it is important to note that P2X7R-expressing microglia can become markedly activated in response to elevated extracellular ATP levels, thereby exacerbating neuroinflammation and contributing to neurotoxicity [69]. Thus, NLRP3 inflammasome engagement is likely a late event in Alzheimer's pathogenesis. Nonetheless, therapeutically targeting P2X7R or components of the NLRP3/caspase-1 axis holds promise for dampening microglia-driven inflammation and preventing pyroptotic neuronal loss.

A series of recent studies highlights the critical role of cross-talk among astrocytes, microglia, and neurons in the pathogenesis of neurodegenerative diseases. According to Li et al. [70], reactive astrocytes and microglia establish a proinflammatory positive feedback loop that dysregulates and amplifies the neuroinflammatory response (Figure 2, 1–4). As noted above, in neurodegenerative diseases, microglia transition to a neurotoxic, M1 (proinflammatory) phenotype at relatively late stages of disease progression. M1-polarized microglia secrete inflammatory mediators that activate A1 astrocytes, eliciting a secondary inflammatory response; mediators released by A1 astrocytes then reinforce M1 polarization of microglia. This self-perpetuating interaction drives chronic astrogliosis and microglial activation within the CNS, culminating in neuronal damage [3,12,37,71,72,73,74].

Microgliosis is a non-specific reactive response characterized by microglial proliferation and hypertrophy. It typically precedes astrogliosis and involves recruitment of border-associated macrophages (BAMs) from the meninges, choroid plexus, and perivascular spaces, together with resident microglia, to sites of injury [75]. Astrogliosis (astrocytosis) is defined by astrocyte proliferation, altered molecular signatures, and morphological changes in response to neuronal injury. Severe astrogliosis leads to glial scar formation that impedes axonal regeneration, increases blood–brain barrier permeability, and permits inflammatory T-cell infiltration into the parenchyma [3,10,72,76]. Reactive astrocytes and microglia also damage oligodendrocytes, impairing axonal myelination and further compromising neural circuitry [73,77,78,79]. Given that microglia respond to local pathological stimuli before astrocytes, promoting microglial polarization toward an M2 (anti-inflammatory) phenotype is a logical first step to restore glial homeostasis. This shift can be induced by (see the summary in Figure 2, left):Inhibiting nuclear factor κB (NF-κB), which regulates genes involved in immune and inflammatory responsesBlocking the janus kinase 2–signal transducer and activator of transcription1/3 (JAK2–STAT1/3) pathway, overexpression of triggering receptor expressed on myeloid cells 2 (TREM2) on microglial membranes recruits the DNAX-activating protein 12 (DAP12) adaptor to suppress JAK2–STAT1/3 signaling while enhancing phagocytosis, migration, lipid processing, proliferation, lysosomal degradation, and metabolism [80,81,82]Activating alternative pathways (phosphoinositide 3-kinase—protein kinase (PI3K/AKT), neurogenic locus notch homolog protein (Notch), peroxisome proliferator-activated receptor gamma (PPAR-γ), or adenosine monophosphate-activated protein kinase (AMPK)) to drive M2 polarization [83].

More than 50 cytokines and growth factors—including interferons and interleukins—participate in the JAK/STAT cascade, which governs hematopoiesis, immune homeostasis, tissue repair, inflammation, apoptosis, and adipogenesis [84].

Cytokine signaling in the crosstalk among microglia, astrocytes, and neurons is mediated not only by receptor engagement but also by specific phosphorylation events that propagate intracellular responses. Binding of cytokines such as IL-6 to their receptors activates JAK2, which catalyzes phosphorylation of tyrosine residues within the receptor's intracellular domain. These phosphotyrosine motifs serve as docking sites for STAT3, which is subsequently phosphorylated on a critical tyrosine residue (Tyr705) and, in some contexts, on serine residues (Ser727) to modulate transcriptional activity. The dual Tyr/Ser phosphorylation of STAT3 regulates its dimerization via the Src homology 2 (SH2) domain, nuclear translocation, and transcription of genes involved in inflammation, gliosis, and neuronal survival. Thus, the chemistry of tyrosine and serine phosphorylation provides a mechanistic link between cytokine receptor activation and the transcriptional programs that shape microglial and astrocytic responses in neurodegeneration [84,85,86].

A notable strategy for modulating pathological microglia is delivering anti-inflammatory cytokines such as interleukin-4 (IL-4) (Figure 2), interleukin-10 (IL-10), and transforming growth factor-β (TGF-β), which:IL-4 drives proliferation and activation of microglia;IL-10 suppresses proinflammatory cytokine production;TGF-β regulates cell proliferation, differentiation, and apoptosis

Inhibition of TNFα and IL-1β in BV2 cells treated with soluble amyloid-β oligomers attenuates neuronal cytotoxicity. In APN/5xFAD mice, elevated TNFα and IL-1β correlate with widespread microglial activation, reduced clustering around fibrillar plaques, and abnormal microglial morphology [40,83,87,88,89,90,91].

Chuang et al. [92] showed that a small molecule protects rat hippocampal neurons from Aβ-induced toxicity and attenuates LPS-driven neuroinflammation by reducing nitric oxide production, downregulating inducible nitric oxide synthase (iNOS) and IL-1β, suppressing IL-6 and TNFα secretion, and blocking NF-κB activation in BV-2 cells.

Resveratrol, a plant-derived polyphenol, activates sirtuin-1 (SIRT1), a nicotinamide adenine dinucleotide–dependent deacetylase that regulates energy metabolism, aging, oxidative-stress responses, and inflammatory pathways, thereby inhibiting microglial M1 transformation [93,94].

The microglial landscape includes multiple disease-associated phenotypes—activated response microglia (ARM), disease-associated microglia (DAM), microglial neurodegenerative phenotype (MGnD), lipid droplet–accumulating microglia (LDAM), white matter–associated microglia (WAM), and dark microglia (DM)—each defined by distinct localization, markers, and functions. A refined classification of these subtypes will inform targeted therapeutic strategies [52,95,96,97].

In summary, bidirectional signaling between microglia and astrocytes critically modulates CNS homeostasis and pathology. While shifting microglia toward an M2 phenotype suppresses inflammation and protects neurons in model systems, its efficacy across diverse phenotypes and its ability to reverse established neurodegeneration before the therapeutic window closes remain open questions.

This hypothesis posits interactions between beta-amyloid (Aβ) and microglia. Aβ activates microglia via Toll-like receptors (TLRs), triggering the release of neuroinflammatory mediators [30,98,99,100,101].

Under physiological conditions, microglia efficiently phagocytose and clear Aβ. In Alzheimer's disease, however, pathological microglia engulf Aβ and tau but fail to degrade these proteins fully, resulting in their intracellular accumulation (Figure 3A). These aggregates are packaged into microglial extracellular vesicles (EVs), which in turn promote EV release. Furthermore, the lipid constituents of large EVs can facilitate the conversion of Aβ into more neurotoxic species [101,102,103,104,105]. In a mouse model, Aβ plaque–associated microglia exhibited increased secretion of tau-containing EVs, contributing to the propagation of tau pathology in AD [47,101,105,106].

Impaired microglial clearance of amyloid β (Aβ) is strongly influenced by the biochemical properties of Aβ itself. It stems from peptides' hydrophobic motifs in central/C-terminal regions, promoting β-sheet formation, fibrillization, and rendering aggregates resistant to proteolysis. These structural features facilitate ligand interactions with cell surface receptors and hinder efficient phagocytosis. In addition, metal ions such as zinc can bind to histidine residues within Aβ, stabilizing fibrillar assemblies and further reducing solubility. Serine protease-mediated degradation (e.g., neprilysin, insulin-degrading enzyme) fails against fibrillar Aβ, leading to impaired proteolytic clearance. Together, the interplay of hydrophobicity, β sheet structural motifs, and metal coordination chemistry contributes to microglial dysfunction and persistence of neurotoxic Aβ deposits [107,108].

The chemokine C-X3-C motif ligand 1 (CX3CL1), expressed by neurons and glial cells, is a proposed target for modulating microglial phagocytic, degradative, and clearance functions. Like other chemokines, CX3CL1 orchestrates chemotaxis and the migration of diverse cell populations, including microglia [109,110]. Another potential target is the neuronal membrane glycoprotein OX-2 (CD200). Both CX3CL1 and CD200 suppress the proinflammatory M1 microglial response via their respective surface receptors. In neurodegenerative diseases, the function of these receptors is compromised, and levels of CD200 and CD200 receptor 1 (CD200R)—and the strength of their interaction—are markedly reduced [3,73,111,112] (Figure 3A).

Impaired fractalkine signaling via the CX3CL1– CX3C motif chemokine receptor 1 (CX3CR1) axis may promote the emergence of disease-associated microglia (DAM), a distinct microglial phenotype [95], yet consensus regarding CX3CR1 is lacking. In 5xFAD mice, Puntambekar et al. [113] demonstrated that CX3CR1 deficiency impairs clearance of fibrillar Aβ, leading to accumulation of neurotoxic Aβ oligomers, neuronal loss, cognitive deficits, and activation of a neurodegenerative microglial phenotype [114,115]. Alternative findings indicate that CX3CR1 deficiency enhances microglia-mediated Aβ clearance but also triggers microglial hyperactivation, elevates IL-6 and IL-1α secretion, and impairs memory in APP/PS1 transgenic mice. In contrast, Cho et al. [114] and Pawelec et al. [112] demonstrated that CX3CR1 deficiency prevents neuronal loss in the 3xTg-AD mouse model. CX3CR1 has also been implicated in promoting tau protein clearance [116].

Another promising therapeutic target is the microglial receptor TREM2, which governs transitions among distinct microglial phenotypes (Figure 3B) [38,117]. In neurodegenerative conditions, TREM2 expression is elevated in both DAM and microglial neurodegenerative (MGnD) phenotypes, and some studies have described a distinct TREM2-high microglia subtype [52,95,97]. TREM2 overexpression in a mouse model of vascular dementia improves spatial memory performance and reduces neuronal loss. It suppresses microglial M1 polarization by decreasing iNOS expression and proinflammatory cytokine secretion while promoting M2 polarization and elevating anti-inflammatory cytokine levels. In the 5XFAD mouse model, chronic TREM2 activation enhances microglial recruitment to amyloid plaques, diminishes amyloid deposition, and ameliorates memory [118,119,120]. In the APP/PS1 model, increasing TREM2 levels reduces synaptic and neuronal loss and upregulates molecular markers of the M2 phenotype [73,81,120,121,122,123]. TREM2 signals via the spleen tyrosine kinase (SYK) pathway and the hematopoietic cell signal transducer (HCST, also known as DNAX-activating protein 10 (DAP10)) pathway [124] (Figure 3A,B). SYK plays a central role in adaptive immune receptor signaling and participates in diverse regulatory processes [125]. DAP10 also contributes to immune regulation [126,127]. DAP10 deficiency alters phosphorylation of protein kinase B (AKT) and glycogen synthase kinase-3β (GSK-3β), impairing microglial proliferation, survival, and phagocytic activity (Figure 3B). SYK-deficient microglia fail to phagocytose Aβ plaques, accelerating amyloid pathology. Moreover, SYK loss disrupts the PI3K–AKT– peroxisome proliferator-activated receptor gamma coactivator 1-alpha (GSK-3β)–mammalian target of rapamycin (mTOR) signaling cascade required for microglia to acquire the DAM phenotype and bind Aβ proteins [124]. Conversely, one study proposes that suppressing TREM2 activation in neurodegenerative disease may be protective, since TREM2 signaling through the TREM2–Apolipoprotein E (ApoE) axis drives microglia toward the MGnD phenotype and promotes neuronal loss [90,128] (Figure 3A).

Another potential target for modulating microglial function is the transmembrane receptor CD33 (siglec-3), expressed on myeloid-lineage cells and functionally linked to TREM2. In AD models, CD33 inhibition has yielded beneficial effects [129,130,131]. For instance, CD33 knockout in 5xFAD mice improved cognition and reduced Aβ burden; these effects were abrogated by concurrent TREM2 knockout [131]. A combined strategy of CD33 inhibition and TREM2 activation may represent an optimal therapeutic approach [132] (Figure 3B). Reports reveal an isoform-dependent dual role for CD33. In 3BV2 mouse microglial cells, elevated expression of the long isoform hCD33M correlates with reduced phagocytic activity. In contrast, predominance of the short isoform hCD33m enhances phagocytosis and migration while diminishing cell adhesion [83].

Progranulin (PGRN) is a relatively under-investigated yet highly promising target for microglial modulation. As the precursor protein of granulin (GRN), PGRN is expressed by both microglia and neurons, where it promotes neuronal survival, maintains lysosomal homeostasis, and regulates inflammatory processes. Progranulin deficiency causes neuronal ceroid lipofuscinosis—characterized by pathological accumulation of toxic lipofuscin pigments that impair lysosomal function—and frontotemporal dementia, and is associated with an increased risk of AD and other neurodegenerative diseases [133,134]. In a transgenic mouse model of AD, PGRN overexpression inhibited Aβ deposition, reduced cerebral amyloid burden, and prevented spatial memory deficits and hippocampal neuronal loss [135], (Figure 3A,B).

Although the literature reports conflicting findings regarding these targets [52,83], there have also been clear successes in restoring microglial function and modulating phenotypes. These discrepancies may arise from differences in molecular forms of the same proteins (e.g., soluble versus insoluble) and variations in disease stage (e.g., pre- versus post-microgliosis) [52,83,95,97]. It is essential to consider combined influences: for example, optimal interventions may involve activating TREM2 while concurrently restoring SYK function and inhibiting CD33. Modulation of TREM2 impacts the regulatory kinase GSK-3β, which has a distinct role in the pathogenesis of various neurodegenerative diseases [136], and is associated with the PI3K–AKT–GSK-3β–mTOR signaling axis, a critical pathway for cellular metabolism, growth, and proliferation (Figure 3A,B).

Effective microglial targeting requires not only restoring homeostatic function or skewing toward the M2 phenotype but also maintaining the balance between non-pathological M1 and M2 states, which is critical in neurodegenerative diseases [12,24,137]. This balance can be modulated via the P2X purinergic receptor 7 (P2X7R), a member of the ATP purinoceptor family expressed on microglia. Transient elevations in extracellular ATP activate P2X7R, trigger autophagy, and promote an M1/M2 hybrid phenotype. Maintaining a dynamic equilibrium between M1 and M2 states in microglia is essential for DAM emergence [83].

The hypothesis that astrogliosis, induced by neurodegenerative processes, leads to glial scar formation is fundamentally distinct from previously discussed models. Although scar formation is more commonly associated with peripheral tissues, emerging evidence suggests that similar processes may occur in the central nervous system (CNS) following chronic cellular and tissue injury. This hypothesis posits that glial scars may impair neuronal regeneration by obstructing axonal growth post-injury, while simultaneously contributing to tissue repair and structural stabilization [138,139,140,141].

Astrocyte migration toward the injury site plays a key role in scar formation and can be modulated by several molecular pathways (Figure 4). Aquaporin-4 (AQP4), a water channel protein expressed in astrocyte membranes, facilitates migration by maintaining water–electrolyte balance and osmotic pressure (Figure 4A). Matrix metalloproteinase-9 (MMP-9), a zinc-dependent enzyme involved in extracellular matrix degradation, also regulates astrocyte motility. In murine models, both MMP-9 inhibition and AQP4 knockout significantly reduce astrocyte migration (Figure 4B) [142,143].

Additional contributors to glial scarring include type III intermediate filament proteins such as vimentin and glial fibrillary acidic protein (GFAP). Dual knockout of these proteins in mice results in diminished post-traumatic glial scar formation [143]. Con-nexins—transmembrane proteins that form intercellular gap junctions—also represent potential therapeutic targets. Astrocytes express high levels of connexin 30 (Cx30) and connexin 43 (Cx43), although their expression is downregulated in reactive astrocytes [143]. The JAK–STAT3 signaling pathway, previously discussed, promotes astrocyte proliferation at sites of injury (Figure 4). Inhibition of STAT3 via JAK signaling has been shown to suppress both proliferation and migration of reactive astrocytes following spinal cord injury, suggesting a similar mechanism may operate elsewhere in the CNS (Figure 4B, C). Notably, STAT3 regulates the transcription of GFAP, Cx43, and AQP genes.

Other promising molecular targets include mothers against decapentaplegic homolog 3 (SMAD3), a signal transducer in the transforming growth factor-β (TGF-β) pathway, and cluster of differentiation 36 (CD36), an integral membrane protein expressed on various cell types. SMAD3 knockout significantly reduces glial scar formation [142,143], while CD36 deficiency in stroke model mice attenuates astrocyte proliferation and delays wound-gap closure. CD36 knockout mice also exhibit reduced scar formation [144] (Figure 4B,C).

Despite these findings, the glial scar hypothesis remains contentious. Some researchers argue that glial scars may exert neuroprotective effects by limiting the spread of damage and inflammation [145]. Moreover, many current approaches focus on suppressing reactive astrocyte activity rather than restoring astrocytes to a physiologically normal state. Therapeutic strategies aimed at normalization may offer more effective mitigation of scarring-related damage.

Recent studies in Acomys (spiny mouse), a species with exceptional regenerative capacity, support this notion. These rodents are capable of regenerating spinal cord fibers without forming astrocytic scars at the injury site during recovery [146,147].

In recent years, research has increasingly focused on reprogramming astrocytes and other glial cells into mature, functional neurons to replenish lost neuronal populations, rather than merely targeting their pathological phenotypes. For instance, Yin and coauthors [148] demonstrated that human cortical astrocytes reprogrammed in vitro remained viable for over seven months and exhibited robust functionality, including synaptic burst activity. Furthermore, neurons derived from astrocytes have been shown to form synaptic connections with endogenous cortical neurons in adult mice in vivo [149].

Guo and coauthors [150] reported that reactive astroglia—following brain injury or in Alzheimer's disease—can be reprogrammed into functional glutamatergic neurons. Their study also showed that NG2 glia can be converted into both glutamatergic and GABAergic neurons via expression of the neuronal transcription factor NeuroD1. The overarching goal is to restore functional neuronal populations lost due to injury or disease. NG2 cells, also known as polydendrocytes, typically serve as precursors to oligodendrocytes. However, evidence suggests that they can differentiate into proliferating reactive astrocytes within glial scars [143].

NG2 glia have demonstrated neuroprotective properties in prion-infected mouse models, where their depletion exacerbates neurodegeneration and accelerates prion pathology [151]. Additionally, both NG2 glial cell density and differentiation capacity are reduced in Alzheimer's disease patients [152], with similar declines observed in APP23 transgenic mice [153].

To date, a substantial body of research has explored astrocyte reprogramming in the context of neurodegenerative diseases and other pathological conditions. This approach holds promise not only for neuronal replacement but also for attenuating proinflammatory astroglia phenotypes, offering a dual therapeutic benefit in the treatment of various neurodegenerative disorders [154,155,156].

Neurodegenerative diseases (NDDs) disrupt microglial mitochondrial function, potentially acting as an early trigger for pathological microglial activation. Abnormalities in mitochondrial DNA (mtDNA), impaired bioenergetics, and defects in the mitochondrial quality-control machinery all contribute to this dysfunction [157].

Interactions between amyloid-β (Aβ) and the TREM2 and P2X7 receptors further exacerbate microglial mitochondrial damage (Figure 5A) [158]. In 5×FAD mice lacking TREM2, mTOR signaling—which governs autophagy—was impaired, leading to reduced mitochondrial mass and lower ATP levels in microglia. This implicates TREM2 deficiency as a critical driver of microglial mitochondrial dysfunction [158].

Accumulation of damaged mitochondria elevates reactive oxygen species (ROS) production, which in turn inflicts oxidative damage on neighboring neural and glial cells [159,160,161].

Activation of the AMP-activated protein kinase (AMPK) pathway shifts microglia toward an anti-inflammatory M2 phenotype, stimulates PGC-1α, and increases nicotinamide adenine dinucleotide (NAD) levels (Figure 5B,C). Elevated NAD activates the deacetylase Sirtuin 1 (Sirt1), collectively promoting mitochondrial biogenesis, enhancing oxidative phosphorylation, and facilitating mitophagy [4,83] (Figure 5B,C).

An alternative mechanism of microglial mitochondrial deficiency is excessive Drp1/Fis1-mediated fission. Driven by dynamin-related protein 1 (Drp1) and mitochondrial fission protein 1 (Fis1), this pathological fission fragments mitochondria, triggering innate immune signaling, inducing A1-phenotype astrocytes, and culminating in neuronal apoptosis [162].

Astrocytes also exhibit profound mitochondrial dysfunction in NDDs. Uptake of Aβ and α-synuclein aggregates impairs their respiratory capacity [163]. In ischemia models, a loss of healthy mitochondria in astrocytic processes correlates with neuronal death, while mitochondrial damage activates astrocytes to release proinflammatory cytokines and disrupt Ca²⁺ homeostasis (Figure 5). During ischemic injury, astrocytes can transfer healthy mitochondria to neurons, but if dysfunctional mitochondria are transferred—as may occur in NDDs—neuronal viability is compromised [164]. In brains of 18-month-old mice, astrocytes show pronounced mitochondrial fragmentation accompanied by a tenfold upregulation of Drp1 [165].

Mitochondrial defects in astrocytes also impair oxidative phosphorylation, essential for fatty-acid oxidation and lipid homeostasis. Such defects contribute to synaptic loss, neuroinflammation, cognitive decline, and reactive astrogliosis [166].

The interaction between the extensive endoplasmic reticulum (ER) network of astrocytes and mitochondria plays a particularly important role in astrocytic function. Regions of close apposition and membrane contact between the ER and mitochondria (mitochondria-associated membranes, MAMs) maintain precise coordination between these organelles and likely determine ATP production, synthesis of proteins and membrane components, stress responses, induction of mitophagy, and initiation of apoptosis. A marked reduction in glucose metabolism in astrocytes is already observed at early stages of neurodegenerative diseases. With extracellular accumulation of Aβ aggregates, astrocytes adopt a reactive phenotype that, to some extent, initially counteracts neurodegeneration. However, the emergence of proinflammatory phenotypes at more advanced disease stages accelerates, rather than delays, the progression of dementia. In particular, ATP deficiency and increased reactive oxygen species (ROS) production contribute to this process (Figure 5A). Studies have also reported ER stress and impaired protein synthesis in glial cells [167,168]. Under these conditions, cytosolic Ca²⁺ levels are elevated while mitochondrial Ca²⁺ is reduced (Figure 5A). Although the precise mechanism remains incompletely understood, evidence suggests that abnormally enhanced ER–mitochondria interactions—resulting from shortened intermembrane distances and elongation of the MAM—may underlie these changes. Dematteis and colleagues have proposed that efficient Ca²⁺ transfer from the ER to mitochondria requires an interorganelle distance of approximately 15–25 nm, whereas in proinflammatory astrocytes this distance is reduced to 10 nm or less [169]. This mechanism could account for both cytosolic Ca²⁺ elevation and mitochondrial Ca²⁺ depletion, leading to an uncompensated bioenergetic deficit.

Although the precise molecular cascades remain to be fully elucidated, converging evidence indicates that early mitochondrial dysfunction in both microglia and astrocytes represents a pivotal risk factor for the onset and progression of neurodegenerative diseases.

The connection between glial cells and neurodegenerative pathology is broader than the mechanisms discussed in detail here. Our focus has been on glial dysfunction as a primary driver, with neuronal effects mediated through glia. However, several additional processes directly impact neurons and synaptic transmission, with only secondary involvement of glia. While not the central scope of this review, they warrant brief mention.

Impact of redox-active metals. A key mechanism implicated in astrocyte and microglia activation during neurodegeneration is the disruption of metal homeostasis, particularly involving iron and copper. Redox-active Fe²⁺ and Cu⁺ ions can catalyze Fenton-type reactions, generating hydroxyl radicals that drive oxidative stress and lipid peroxidation. These processes contribute to regulated cell death pathways such as ferroptosis and cuproptosis, which impair mitochondrial integrity and energy metabolism. The ensuing redox imbalance activates transcriptional regulators including NF-κB, thereby amplifying the production of proinflammatory mediators such as nitric oxide (NO) and tumor necrosis factor-α (TNF-α). Sustained release of these mediators compromises the neuroprotective functions of glial cells, fostering a cycle of chronic neuroinflammation and neuronal injury. Thus, metal-catalyzed oxidative reactions represent a critical intersection between glial dysfunction, oxidative stress, and progressive neurodegeneration [170,171,172,173].

Neuromelanin accumulation. Neuromelanin, a dark pigment that progressively accumulates with age in dopaminergic neurons of the substantia nigra, has been increasingly recognized as a key factor in neuroinflammation and neurodegeneration. Beyond serving as a byproduct of catecholamine metabolism, neuromelanin can act as a microglial activator: when released from degenerating neurons, neuromelanin granules are phagocytosed by microglia, triggering proinflammatory responses and the release of cytokines and reactive oxygen species. This process contributes to a self-perpetuating cycle of neuroinflammation that exacerbates neuronal vulnerability. Such mechanisms are under active investigation as potential contributors to Parkinson's disease pathogenesis [174,175]. Recent studies further suggest that neuromelanin may serve as a nidus for α-synuclein aggregation and synucleinopathy, linking pigment accumulation to protein misfolding and progressive dopaminergic cell loss [176,177]. Collectively, these findings highlight neuromelanin as both a biomarker of aging and a potential driver of microglia-mediated neuroinflammatory cascades in Parkinson's disease.

Ammonia detoxification. Astrocytes normally detoxify ammonia via glutamine synthetase (GS). In Alzheimer's disease (AD), GS activity is reduced, and hyperammonemia has been observed in both brain and blood. Elevated ammonia promotes Aβ42 accumulation within astrocytes, creating a vicious cycle: Aβ degradation generates ammonia, which fuels the astrocytic urea cycle but also produces excess gamma-aminobutyric acid (GABA) and hydrogen peroxide, impairing synaptic function and memory [178].

Potassium homeostasis. Astrocytes maintain extracellular K⁺ levels through inwardly rectifying potassium channel 4.1 (Kir4.1) channels. In AD, Kir4.1 expression and conductivity decline, particularly in the hippocampus, disrupting K⁺ clearance and impairing neuronal excitability and memory [179].

Zinc and copper regulation. Glial cells regulate synaptic zinc via transporters (ZnTs, ZIPs) and contribute to copper balance. In AD, dysregulated zinc transport elevates extracellular zinc, accelerating Aβ aggregation into neurotoxic fibrils. High zinc and copper levels within amyloid plaques and impaired synaptic zinc/copper homeostasis compromise plasticity and cognition [180].

In this review, we have summarized the pathological mechanisms underlying several neurodegenerative diseases and highlighted potential therapeutic targets that warrant further attention. Importantly, we provided, for the first time, a synthesis of the principal hypotheses linking glial cells to the early stages of neurodegeneration. Key limitations to consider include protein function being highly dependent on isoform, the existence of multiple microglial phenotypes for which the same therapeutic approaches may not be appropriate, and the need not only to shift phenotypes from M1/A1 to M2/A2 but also to maintain an appropriate balance between them.

Astrocytes and microglia are indispensable for neuronal support and protection, yet early dysfunction in these glial cells often initiates neurodegeneration. Their morphofunctional changes—microglial polarization, inflammasome activation, impaired phagocytosis, astrocytic scar formation, and mitochondrial stress—emerge before clinical symptoms and may drive pathology.

Mitochondrial deterioration in glia is a consistent early hallmark, underscoring the potential of therapies that restore mitochondrial health prior to irreversible neuronal loss. Likewise, the microglial shift from anti-inflammatory M2 to proinflammatory M1 states amplifies neuroinflammation. Preventing this transition through modulation of TLR4, JAK–STAT, NF-κB, cytokine signaling, or TREM2 pathways represents a promising strategy.

Unchecked M1 activation also promotes NLRP3 inflammasome assembly and induces neurotoxic A1 astrocytes, highlighting maladaptive glial–glial interactions as therapeutic targets. Restoring microglial clearance functions via CX3CL1/CX3CR1, CD200–CD200R, TREM2, CD33 inhibition, or progranulin may further mitigate disease progression. Balancing M1/M2 states through P2X7 signaling could help prevent chronic inflammation.

Astrocyte-directed therapies aim to limit scar formation (via AQP4, MMP-9, vimentin/GFAP, connexins) or reprogram reactive glia into neurons. Enhancing mitophagy and suppressing mitochondrial fission also show promise. Given the heterogeneity of glial phenotypes and targets, multitargeted approaches applied at the earliest stages are most likely to succeed.

Recent discoveries reveal astrocytes' roles in memory and microglia's involvement in forgetting and fear [181,182,183,184], emphasizing the need for continued investigation of glial biology. Early-phase insights may yield strategies to intercept neurodegeneration before it becomes irreversible.

Despite the encouraging therapeutic avenues outlined, several challenges remain. Many proposed molecular targets are pleiotropic, raising concerns about off-target effects and systemic toxicity. Some of the hypothetical activators, agonists, and antagonists we mentioned have not yet been developed or lack sufficient safety and specificity. We hope this review will draw the attention of researchers to them. The heterogeneity of glial phenotypes across brain regions and disease stages complicates the translation of preclinical findings into clinical therapies. Furthermore, most current studies rely on animal models that only partially recapitulate human neurodegenerative pathology, limiting predictive validity. Future work should prioritize the development of humanized models, longitudinal studies to capture dynamic glial changes, and combinatorial approaches that integrate mitochondrial protection, immune modulation, and synaptic support. Ultimately, therapeutic strategies must balance efficacy with safety, ensuring that modulation of glial activity does not disrupt their essential homeostatic functions.