Unlocking the Sugar Code: Implications and Consequences of Glycosylation in Alzheimer’s Disease and Other Tauopathies

Early studies revealed that tau is aberrantly N-glycosylated in AD brains, whereas normal adult tau is typically not glycosylated [7]. The addition of N-linked glycans occurs at asparagine residues and alters the conformation of tau, reducing its affinity for microtubules while increasing aggregation propensity [8]. N-glycosylated tau has been detected within paired helical filaments and neurofibrillary tangles, supporting a pathogenic role [4].

Importantly, aberrant N-glycosylation often precedes or facilitates hyperphosphorylation. Glycosylated tau shows increased susceptibility to kinases such as glycogen synthase kinase-3β (GSK3β), leading to pathogenic phosphorylation pattern. This sequence highlights glycosylation as an upstream event in tauopathy pathogenesis [7]. Glycosyltransferases—enzymes that catalyze the transfer of sugar moieties to proteins or lipids—mediate the covalent addition of glycan structures that can modulate tau folding, stability, and aggregation [6,7].

Another critical modification is O-GlcNAcylation, the addition of O-linked N-acetylglucosamine to serine or threonine residues. Unlike N-glycosylation, O-GlcNAcylation appears protective. Increased O-GlcNAcylation reduces tau phosphorylation and aggregation by competing with phosphorylation sites [8]. Animal studies show that enhancing O-GlcNAcylation stabilizes tau, improves neuronal survival, and ameliorates memory deficits [5].

However, AD brains consistently show reduced O-GlcNAcylation, correlating with increased phosphorylation and tangle burden [9]. This reduction may result from impaired glucose metabolism and flux through the HBP [3].

In FTLD, PSP, and CBD, tau aggregates share common modifications. Abnormal glycosylation of tau has been demonstrated across these conditions, despite disease-specific differences in glycosylation site occupancy and glycan composition [10]. In PSP, tau shows increased high-mannose N-glycans, while in CBD, complex-type N-glycans are enriched, suggesting distinct enzymatic dysregulation [11].

Moreover, O-GlcNAcylation deficits are not restricted to AD but also appear in FTLD brain tissue, further linking impaired glucose metabolism to tau pathology across tauopathies [9].

While all tauopathies share abnormal tau aggregation, key differences exist in isoform composition, filament structure, and regional vulnerability. In AD, neurofibrillary tangles contain both 3-repeat (3R) and 4-repeat (4R) tau isoforms that assemble into paired helical filaments (PHFs), typically distributed from the entorhinal cortex to the neocortex in a hierarchical pattern [9]. In contrast, PSP and CBD predominantly accumulate 4R-tau, forming straight filaments and involving basal ganglia, brainstem, and motor cortex [8,9]. FTLD-tau subtypes vary—Pick's disease features 3R-tau inclusions (Pick bodies) within frontotemporal regions, whereas globular glial tauopathies exhibit mixed isoforms with astrocytic or oligodendroglial predominance [10]. These distinctions suggest that the biochemical context of tau, including PTMs such as glycosylation, may differ by isoform and brain region. Evidence indicates that AD-associated tau tends to be more extensively N-glycosylated and exhibits reduced O-GlcNAcylation, while PSP and CBD display more restricted N-glycan complexity and relatively conserved O-GlcNAc profiles [11]. Thus, differential tau glycosylation may contribute to disease-specific aggregation patterns, cellular tropism, and progression rates among tauopathies [9,10,11].

Perhaps the most critical insight is the interplay between glycosylation and phosphorylation. N-glycosylated tau is a better substrate for phosphorylation, while O-GlcNAcylation prevents phosphorylation at adjacent residues [12]. This creates a pathogenic imbalance, reduced O-GlcNAcylation and increased N-glycosylation synergistically drive tau hyperphosphorylation and aggregation [13].

This dynamic crosstalk suggests that therapies aimed at restoring O-GlcNAcylation or preventing aberrant N-glycosylation could rebalance tau homeostasis and slow disease progression. Indeed, small-molecule inhibitors of O-GlcNAcase (OGA), the enzyme removing O-GlcNAc, are under investigation as disease-modifying treatments [6].

N-glycosylation appears to promote tau hyperphosphorylation and aggregation, whereas O-GlcNAcylation exerts a protective role [11]. However, the relationship between glycosylation and phosphorylation is highly dependent on site specificity and structural context. Tau protein contains over 80 potential phosphorylation sites and at least 10 experimentally confirmed glycosylation sites, many of which are spatially proximal or overlapping [12]. N-glycosylation typically occurs on asparagine residues (e.g., Asn167, Asn359), whereas O-GlcNAcylation affects serine and threonine residues (e.g., Ser400, Thr123, Thr245, Ser356) that often coincide with major phosphorylation motifs recognized by kinases such as GSK3β, cyclin-dependent kinase 5 (CDK5), and microtubule affinity-regulating kinase 2 (MARK2) [13]. These modifications act in a competitive and regulatory manner. N-glycosylation changes the tertiary structure of tau, reducing its microtubule affinity and exposing neighboring serine and threonine residues to kinases, thereby facilitating hyperphosphorylation. In contrast, O-GlcNAcylation directly competes with phosphorylation at the same or adjacent residues, inhibiting excessive phosphorylation and stabilizing tau–microtubule interactions [12]. Consequently, decreased O-GlcNAcylation in AD—often linked to impaired glucose metabolism—removes this protective brake, leading to pathological phosphorylation and tangle formation [11,12,13]. In AD, both 3R and 4R tau isoforms exhibit extensive N-glycosylation and reduced O-GlcNAcylation, driving PHF assembly [10]. In other tauopathies such as PSP and CBD, tau aggregates are composed predominantly of 4R isoforms with more restricted glycosylation patterns and distinct phosphorylation site usage (e.g., Ser262, Ser396) [10,11]. These biochemical differences suggest that glycosylation and phosphorylation site occupancy contribute to the diversity of filament morphology and regional vulnerability observed across tauopathies [10,11,13]. Future work should address these site-specific interactions, emphasizing that N-glycosylation (Asn167/359) acts as an early structural modifier that primes tau for pathogenic phosphorylation, while O-GlcNAcylation (Ser400, Thr245, Ser356) counteracts aggregation by masking phosphorylation motifs [11,13]. This comparative framework could enhance understanding of tau pathology across AD and non-AD tauopathies and guide the design of site-directed therapeutic interventions.

APP is a heavily N-glycosylated type I transmembrane protein. N-glycans at asparagine residues near its luminal domain regulate APP folding, trafficking through the secretory pathway, and stability at the cell surface [14]. Disruption of these N-glycans alters APP sorting, diverting it toward endosomal compartments enriched in β-site β-secretase enzyme, APP-cleaving enzyme 1 (BACE1), thereby increasing amyloidogenic cleavage [15].

Experimental studies confirm that loss of APP N-glycosylation enhances Aβ production, whereas stabilizing N-glycan structures reduces amyloidogenic processing [16]. These findings indicate that altered N-glycosylation in the AD brains may bias APP processing toward Aβ generation.

In addition to N-glycosylation, APP and its secretases undergo O-glycosylation, particularly O-GalNAc modification, which influences protein trafficking and enzymatic activity. O-glycosylation of APP modulates its endocytosis rate, thereby controlling access to β- and γ-secretases [5]. Reduced O-glycosylation correlates with increased amyloidogenic cleavage and elevated extracellular Aβ accumulation [8].

Secretases themselves are glycoproteins. BACE1 contains multiple N-glycans that regulate folding, transport, and stability. Inhibition of BACE1 glycosylation reduces its enzymatic activity and impairs Aβ generation [17]. Similarly, γ-secretase components such as nicastrin are highly glycosylated. The glycan structures near the substrate-binding pocket determine APP cleavage patterns. These modifications fine-tune Aβ species ratios, including the pathogenic Aβ42/40 balance [18].

Emerging evidence suggests that Aβ peptides themselves can undergo glycosylation or interact with glycans. Modified Aβ shows altered aggregation kinetics and toxicity. Glycosylated Aβ peptides aggregate faster and form more neurotoxic oligomers compared to unmodified Aβ [6]. Additionally, glycans on neuronal membranes interact with Aβ, promoting deposition and impairing clearance [1].

Glycosylation also modulates Aβ clearance via receptor-mediated endocytosis. For instance, the receptor for advanced glycation end products (RAGE) binds both glycated proteins and Aβ, facilitating their transcytosis across the blood–brain barrier [2]. Enhanced RAGE glycosylation increases its affinity for Aβ, promoting deposition in the brain parenchyma [19].

The interplay between glycosylation and Aβ toxicity extends beyond processing and clearance. Aberrant glycosylation of synaptic proteins, receptors, and ion channels exacerbates Aβ-induced dysfunction. For example, N-methyl-D-aspartate (NMDA) receptors require N-glycosylation for proper surface expression as loss of this modification sensitizes neurons to Aβ-mediated excitotoxicity [20]. Similarly, glycosylation of prion protein (PrP^C) modulates its interaction with Aβ oligomers, influencing synaptotoxic signaling cascades [17,21].

Furthermore, microglial and astrocytic receptors involved in Aβ clearance—such as triggering receptor expressed on myeloid cells 2 (TREM2) and cluster of differentiation 33 (CD33)—are heavily glycosylated. Altered glycan structures modulate receptor affinity and downstream inflammatory responses [22]. This suggests that glycosylation contributes to the balance between protective clearance and harmful neuroinflammation.

Given the centrality of glycosylation in APP metabolism and Aβ biology, targeting glycosylation pathways represents a promising therapeutic approach. Experimental strategies include inhibiting BACE1 glycosylation, enhancing protective O-glycosylation, and modulating glycosyltransferases that remodel glycans on APP and its processing enzymes [16,17]. Additionally, blocking RAGE glycosylation has been proposed to limit Aβ transport into the brain and reduce Aβ deposition within the cerebral parenchyma and vasculature [2,19,23,24,25].

Neuronal communication depends on the proper glycosylation of neurotransmitter receptors and cell adhesion proteins. NMDA receptors, critical for synaptic plasticity, require N-glycosylation for assembly, trafficking, and surface expression. In AD models, aberrant glycosylation disrupts NMDA receptor localization, sensitizing neurons to Aβ-induced excitotoxicity [20].

Similarly, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors rely on N-glycans for stability and gating properties. Altered glycosylation reduces synaptic strength and long-term potentiation, processes essential for learning and memory [15].

Adhesion molecules such as neural cell adhesion molecule (NCAM) and L1CAM are highly glycosylated, and polysialylation of NCAM regulates neurite outgrowth and synaptic remodeling. In addition, reduced polysialylation has been detected in AD hippocampus, correlating with impaired synaptic connectivity [28].

Microglia, the brain's resident immune cells, rely on glycosylated receptors for recognition and clearance of pathological proteins [22]. TREM2 contains multiple N-glycosylation sites that regulate folding and surface expression [26]. Mutations altering TREM2 glycosylation reduce receptor stability, impair microglial phagocytosis of Aβ, and increase AD risk [22,26,27].

Another example is CD33, a sialic acid-binding receptor implicated in AD genetic risk [6,8,19]. CD33 glycosylation governs ligand binding and inhibitory signaling while aberrant glycosylation enhances its inhibitory effect, reducing Aβ clearance and exacerbating inflammation [12].

Astrocytes also display glycosylated receptors that influence neuroinflammation. For instance, glial fibrillary acidic protein (GFAP)-interacting proteins undergo glycan remodeling in reactive astrocytes, altering cytokine release and glial scar formation [5,29].

Inflammatory mediators in the central nervous system (CNS) are heavily glycosylated. Interleukin-6 (IL-6), a pro-inflammatory cytokine elevated in AD, requires N-glycosylation for secretion and receptor binding [29]. Altered IL-6 glycosylation patterns modulate its bioactivity and half-life in the extracellular space [29,30,31]. Similarly, tumor necrosis factor-alpha (TNF-α) and interleukin-1β (IL-1β) undergo glycosylation changes in AD, influencing receptor affinity and downstream signaling [32,33,34].

Chemokine receptors such as CCR5 and CX3CR1 are also glycosylated, with changes in glycan branching modulating microglial migration and response to injury [11]. In AD, aberrant glycosylation of these receptors contributes to chronic non-resolving inflammation [11,28].

The complement cascade, a major mediator of synaptic pruning and neuroinflammation, is highly influenced by glycosylation [7]. C1q, the initiator of the classical pathway, carries sialylated glycans that regulate binding to immune complexes and synaptic elements [7,11]. Loss of sialylation enhances complement activation and drives synapse elimination in AD models [9,35].

Other complement proteins, including C3 and factor H, show altered glycosylation in AD brain and CSF, correlating with disease severity [5,8,14,23,26]. This suggests that glycosylation defects amplify maladaptive immune responses.

Glycosylation also extends its influence to pattern recognition receptors (PRRs) that detect damage-associated molecular patterns (DAMPs) [2]. RAGE is heavily glycosylated, and its modification enhances binding to glycated proteins and Aβ, facilitating inflammatory cascades [2,26]. Glycosylated RAGE amplifies NF-κB signaling, perpetuating neuroinflammation [3].

In addition, toll-like receptors (TLRs) rely on N-glycans for stability and ligand recognition [35]. Dysregulated TLR glycosylation biases microglia toward pro-inflammatory phenotypes, further fueling pathology [35,36,37,38,39].

Given its central role in receptor biology and inflammatory signaling, modulating glycosylation could rebalance immune responses. Potential strategies include glycoengineering of anti-inflammatory cytokines, inhibition of aberrant receptor glycosylation, or restoring protective glycosylation patterns on complement regulators [2,12,40].

Proteomic studies demonstrate profound changes in glycosylation within AD brains. Comparative analyses of postmortem cortical tissue show altered N-glycan branching, sialylation, and fucosylation across multiple glycoproteins [1,44,45]. Specific glycosylation signatures distinguish AD from age-matched controls, implicating disrupted glycosyltransferase and glycosidase activity [3,33].

In CSF, glycoproteomics reveals increased bisected N-glycans and altered glycoforms of tau and APP fragments, correlating with disease severity [46,47]. Serum analyses have also detected disease-specific glycosylation patterns, including reduced sialylation of acute-phase proteins and altered IgG Fc (Immunoglobulin G—fragment crystallizable) glycosylation, linking systemic inflammation to neurodegeneration [10,14,24,32,47,48].

The diagnostic potential of glycosylation is underscored by studies demonstrating that altered glycosylation of tau precedes overt phosphorylation changes [49]. Detection of glycosylated tau in CSF or plasma could thus provide early disease biomarkers. Likewise, O-GlcNAcylation status of tau correlates inversely with disease progression, suggesting its utility as a prognostic marker [46,50,51].

Moreover, glycan-based profiling has shown promise in differentiating AD from other dementias. For instance, patients with FTLD and PSP display distinct glycosylation patterns compared to AD, enabling differential diagnosis [16,20].

Recent advances in MS have revolutionized glycoproteomics. Methods such as electron-transfer dissociation (ETD) and higher-energy collisional dissociation (HCD) provide site-specific glycan characterization with high sensitivity [42]. Enrichment strategies using lectins, hydrophilic interaction chromatography (HILIC), and nanoporous materials further enhance glycopeptide detection [5,12,52].

Novel probes and nanocomposites allow in situ visualization of glycosylation in brain tissue, thereby bridging biochemical analyses with histopathology. These tools have revealed regional differences in glycosylation within hippocampus, cortex, and subcortical structures, aligning with clinical phenotypes of memory loss, executive dysfunction, and motor impairment [53,54].

Given the complexity of tau PTMs, including phosphorylation and glycosylation, the development of specific and validated antibodies remains a significant challenge. Most commercially available anti-tau antibodies are designed against phosphorylated epitopes (e.g., pSer202, pThr231, pSer396), many of which have been extensively validated for immunohistochemistry (IHC), Western blot, and enzyme-linked immunosorbent assay (ELISA) applications in brain and CSF samples [12]. In contrast, antibodies specifically targeting glycosylated tau are scarce and often lack full validation due to the structural diversity and low abundance of glycoforms [53]. Only a limited number of monoclonal antibodies raised against N-glycosylated or O-GlcNAc-modified tau have been reported, with most data derived from research-grade reagents used in immunoprecipitation or lectin-based assays rather than standardized diagnostic kits [12,53].

Validation of tau glycosylation-specific antibodies for clinical biofluids such as CSF or serum remains preliminary, primarily constrained by low antigen concentration and conformational masking of glycan epitopes [52,53]. While recent studies have employed antibody–lectin hybrid assays and antibody enrichment followed by liquid chromatography (LC)–MS/MS to enhance specificity, no single antibody currently demonstrates broad cross-platform reliability [5,12]. For IHC detection, phospho-tau antibodies remain highly reliable and routinely used to stage pathology (e.g., AT8, PHF-1, CP13), whereas glycosylation-targeting antibodies are still in experimental phases [54]. As glycoproteomic characterization advances, development of rigorously validated monoclonal antibodies recognizing specific glycoepitopes on tau will be critical for translating glycosylation research into clinical biomarker applications [5,12,52,53,54].

Beyond CNS, peripheral immune system glycosylation offers insights into neurodegeneration. Altered IgG Fc glycosylation has been documented in AD, with reduced galactosylation and sialylation associated with pro-inflammatory phenotypes [20,47,55]. These changes may reflect systemic inflammation or feedback from neuroinflammatory cascades.

Additionally, plasma glycoproteomic profiling identified altered glycosylation of acute-phase reactants such as haptoglobin and α1-acid glycoprotein, linking metabolic stress to disease progression [56,57,58].

Glycosylation-based biomarkers should not be considered in isolation but integrated with imaging, fluid, and genetic markers. Combining CSF tau glycosylation profiles with positron emission tomography (PET) amyloid imaging improves diagnostic accuracy over either modality alone [10,51]. Similarly, serum glycoproteomic signatures complement plasma phosphorylated tau (phospho-tau181, phospho-tau217) measurements, enhancing early detection potential [11,46].

Despite major progress, glycoproteomics faces challenges. Glycan heterogeneity complicates quantification and standardized reference libraries are limited [59]. Moreover, interindividual variability in glycosylation due to genetics, diet, and comorbidities complicates biomarker validation [1,59,60,61].

However, these same features offer opportunities. Disease-specific glycan "fingerprints" could yield highly specific biomarkers and integration with machine learning approaches promises improved classification [3,62,63]. In addition, longitudinal studies are beginning to reveal how glycosylation changes track disease trajectory, paving the way for prognostic applications [62].

Table 1 and Table 2 present analytical limitations in tau glycoproteomics and emerging solutions.

N-acetylglucosaminyltransferases (GnTs) regulate branching and extension of N-glycans. Upregulation of GnT-III increases bisected N-glycans on tau and APP, promoting pathological processing [16]. Similarly, elevated GnT-V activity has been detected in AD brains, producing β1,6-GlcNAc branching that alters synaptic receptor function [32,64,65].

Sialyltransferases, which attach sialic acid residues, are dysregulated in AD, leading to reduced sialylation of neuronal adhesion molecules such as NCAM and impaired synaptic remodeling [18]. Altered sialylation also contributes to immune dysfunction by modifying microglial receptor interactions with sialylated ligands [36,66].

In tauopathies beyond AD, such as PSP and CBD, distinct changes in glycosyltransferase expression drive unique glycosylation signatures. Comparative analyses reveal elevated fucosyltransferase activity in PSP, whereas CBD brains show increased sialyltransferase expression, contributing to disease-specific tau glycoforms [11,16,67].

Glycosidases remove glycans from proteins and modulate glycan turnover. In AD, elevated β-N-acetylglucosaminidase (β-N-OGA) activity reduces protective O-GlcNAcylation of tau, facilitating its hyperphosphorylation and aggregation [20,59,68]. Conversely, inhibition of β-N-OGA increases O-GlcNAcylation and reduces neurodegeneration in tauopathy mouse models [69,70].

Neuraminidases (sialidases) regulate the removal of sialic acids. Increased neuraminidase activity has been linked to loss of polysialylation on NCAM in AD hippocampus, impairing synaptic plasticity [18,32,71]. Additionally, altered lysosomal glycosidase activity disrupts degradation of glycoproteins, contributing to protein aggregation and lysosomal storage pathology observed in AD [3]. Abnormal N-glycosylation of tau promotes hyperphosphorylation and aggregation, while reduced O-GlcNAcylation removes a protective brake on phosphorylation. N- and O-glycosylation of APP and secretases regulate Aβ generation, with aberrant patterns favoring amyloidogenic processing [16,71]. Synaptic receptors and adhesion molecules require proper glycosylation for stability and plasticity [72]. Their dysregulation contributes to synaptic failure [73]. Immune receptors (TREM2, CD33, and RAGE) and cytokines are glycosylated, shaping microglial and astrocytic activation, while complement protein glycosylation regulates synapse pruning. Finally, altered activity of glycosyltransferases, glycosidases, and the HBP underlies these shifts. Together, these processes highlight glycosylation as both a pathogenic mechanism and therapeutic target [72,73]. A schematic overview of glycosylation is presented in Figure 1.

HBP provides UDP-N-acetylglucosamine (UDP-GlcNAc), the substrate for O-GlcNAcylation [60]. Reduced glucose flux into the HBP, due to impaired brain glucose metabolism in AD, diminishes O-GlcNAcylation levels [60,72,73]. This metabolic defect links systemic insulin resistance and type 2 diabetes to enhanced risk of AD and tauopathies [74,75,76].

Experimental restoration of HBP flux via glucosamine supplementation increases O-GlcNAcylation and reduces tau phosphorylation in animal models, highlighting its therapeutic potential [68]. Similarly, caloric restriction and metabolic interventions that enhance glucose utilization may indirectly restore protective glycosylation patterns [68,77,78].

The interplay between glycosylation and phosphorylation is tightly controlled by enzymes. Reduced O-GlcNAc transferase (OGT) activity decreases tau O-GlcNAcylation, thereby exposing phosphorylation sites for kinases such as GSK3β [30,57,79,80]. Conversely, enhancing OGT activity shifts the balance toward protective O-GlcNAcylation [6,56].

This enzymatic tug-of-war suggests that targeting glycosylation enzymes may indirectly modulate tau phosphorylation, providing a dual mechanism for therapeutic intervention [34,68,81].

Aberrant expression of glycosylation enzymes is detectable in CSF and serum, offering biomarker potential [3]. Elevated OGA and reduced OGT levels correlate with higher CSF tau and worse cognitive outcomes [3,29,82,83]. Similarly, altered glycosyltransferase expression patterns in peripheral blood mononuclear cells reflect disease stage and may serve as accessible biomarkers [39,76].

Emerging evidence indicates that aberrant tau glycosylation patterns are not only neuropathological hallmarks but may also correlate with measurable clinical outcomes. In AD, elevated N-glycosylated tau in brain tissue and CSF has been associated with advanced Braak stages and greater neurofibrillary tangle burden, reflecting a molecular shift toward aggregation-prone conformers [29,39]. The Braak staging system assigns cases to six stages based on the presence of Lewy pathology in the medulla oblongata (I), pons (II), mesencephalon (III), amygdala and other limbic structures (IV), association neocortex (V), and first order sensory association neocortex (VI). Conversely, reduced O-GlcNAcylation levels in tau correlate with accelerated cognitive decline and lower Mini-Mental State Examination (MMSE) scores, suggesting a loss of neuroprotective regulation [76]. Longitudinal cohort studies have shown that specific CSF glycoforms of tau—particularly those carrying high-mannose N-glycans or lacking O-GlcNAc at Ser400 and Thr245—predict progression from mild cognitive impairment (MCI) to Alzheimer's dementia over 2–4 years [76,82]. Furthermore, glyco-profiling in CSF and plasma using lectin-based assays or glycoproteomic MS has demonstrated potential for distinguishing AD from other tauopathies such as PSP and CBD, based on differences in glycan branching and sialylation patterns [82], as shown in Table 3. Collectively, these findings suggest that the glycosylation state of tau serves as a dynamic biochemical correlate of disease stage and may complement phosphorylated tau and Aβ42 as part of a next-generation biomarker panel for early detection and disease monitoring [83].

Enzymes regulating glycosylation are increasingly viewed as druggable targets. OGA inhibitors, such as thiamet-G, are under preclinical investigation for enhancing tau O-GlcNAcylation [39,57,66,84]. By preserving polysialylation, sialidase inhibitors may protect synaptic plasticity [81]. Small molecules targeting glycosyltransferases are less developed but hold potential for rebalancing glycosylation networks in neurodegeneration [8,81,85,86,87]. The glycosylation processes involved in AD and other tauopathies are summarized in Table 4.

One of the persistent debates in AD research concerns the sequence of pathological events. To this end, glycosylation research offers new perspectives. Evidence suggests that N-glycosylation of tau occurs early, priming the protein for hyperphosphorylation and aggregation [64,66,90]. In parallel, aberrant glycosylation of APP biases processing toward amyloidogenic cleavage [5,91]. Thus, glycosylation abnormalities may precede both amyloid and tau pathologies, positioning them as an upstream "common denominator".

Not all glycosylation is deleterious. O-GlcNAcylation of tau exemplifies a protective modification by competing with phosphorylation sites, which reduces aggregation and toxicity [16,77,92,93]. Conversely, reduced O-GlcNAcylation in AD brains correlates with worsened pathology [57,77,94]. Similarly, polysialylation of NCAM supports synaptic plasticity, but its reduction contributes to connectivity deficits [64,65,95,96]. The dualistic nature of glycosylation complicates therapeutic strategies. Hence, interventions must distinguish between protective and pathogenic glycosylation events [22,97].

A recurring theme is the integration of glycosylation with broader cellular networks. Metabolic dysfunction, particularly impaired glucose utilization, reduces flux through the HBP, limiting substrate availability for O-GlcNAcylation [3,98,99]. This explains why AD brains show reduced O-GlcNAc despite increased phosphorylation pressure. Moreover, enzymatic regulators such as OGT and OGA directly shape the phosphorylation landscape of tau by modulating glycosylation at competing sites [20,79,100], as presented in Figure 2.

Neuroinflammation plays a key role in the neurodegenerative process. Glycosylation critically modulates immune receptor signaling [27]. TREM2 requires N-glycosylation for stability. Mutations impairing this modification reduce microglial Aβ clearance [27,101,102]. Conversely, aberrant glycosylation of CD33 enhances its inhibitory signaling, further dampening microglial phagocytosis [66,103]. Glycosylation of complement proteins amplifies maladaptive synaptic pruning [9,104]. Collectively, these findings highlight glycosylation as a molecular amplifier of neuroinflammatory cascades.

Glycoproteomics offers promising avenues for biomarker discovery. Altered glycosylation of tau, APP, and inflammatory proteins has been consistently detected in CSF and serum [42,105,106,107]. These patterns not only differentiate AD from controls but may also discriminate between tauopathies such as PSP and CBD [67,108]. However, technical barriers remain. Glycan heterogeneity complicates quantification and interindividual variability poses validation challenges [1,109].

Therapeutic manipulation of glycosylation is a rapidly advancing field. OGA inhibitors that increase protective O-GlcNAcylation of tau are under active investigation [110]. Strategies targeting glycosyltransferases or glycosidases could theoretically rebalance glycosylation networks, though selectivity remains a major obstacle [5,16,111,112]. In addition, targeting glycosylation of immune receptors such as RAGE and TREM2 may offer immunomodulatory benefits [24,27,113].

Despite major progress, several controversies remain. It is unclear whether glycosylation changes are primary drivers of pathology or secondary adaptations to cellular stress. Some argue that abnormal glycosylation results from metabolic and inflammatory dysfunction rather than initiating pathology [27,114]. Others suggest that glycosylation shifts act as compensatory mechanisms, initially protective but maladaptive when chronic [9,115]. Clarifying these temporal relationships will be essential for therapeutic translation.

Glycosylation has emerged as a pivotal post-translational modification in neurodegeneration, shaping the trajectory of pathology through its multifaceted impact on protein structure, stability, trafficking, and signaling. Across this review, the following themes consistently emerge. Aberrant N-glycosylation primes tau for hyperphosphorylation and aggregation while reduced O-GlcNAcylation removes a critical brake against pathogenic modifications [91,116]. N- and O-glycosylation of APP and its processing enzymes profoundly influence amyloidogenic cleavage, with downstream effects on Aβ burden [5,117]. Glycosylation of receptors and adhesion molecules determines synaptic resilience, while glycosylation of immune receptors (TREM2, CD33, and RAGE) amplifies maladaptive neuroinflammation [19,27,118]. Enzymes such as OGT, OGA, and glycosyltransferases orchestrate glycosylation balance. Their dysregulation, often linked to impaired glucose metabolism and HBP flux, couples systemic metabolic dysfunction to neurodegeneration [3,119]. Glycoproteomic studies reveal distinct glycosylation signatures in brain, CSF, and serum, offering diagnostic and prognostic markers and providing disease-specific fingerprints that differentiate AD from other tauopathies [11,34,120].

Collectively, glycosylation abnormalities are not epiphenomena but core contributors to disease mechanisms, interwoven with phosphorylation, inflammation, and metabolism [44,121,122].

To help the field converge on comparable and clinically relevant measurements, we propose the following minimum reporting standards and decision rules for glycosylation in tauopathies, expanding beyond prior reviews by integrating site-resolved glycosylation with assay readiness and disease-specific signatures, as illustrated in Figure 3.

Table 5 presents recommended reporting standards, disease signatures, validation rules, and translation framework for tau glycoproteomics.