The blood–brain barrier (BBB) forms a highly selective interface that governs molecular exchange between the circulation and neural tissue (Aborode et al., 2025). Astrocytic endfeet, integral components of the neurovascular unit, envelop the cerebral microvasculature and contribute to both barrier function and the regulation of cerebral blood flow (Preininger and Kaufer, 2022). Glucose transport into the brain is mediated predominantly by GLUT1, which is abundantly expressed on the luminal and abluminal surfaces of endothelial cells. Astrocytic endfeet also express high levels of GLUT1, establishing a coordinated glucose delivery route from the vasculature to astrocytes (Koepsell, 2020; Schiera et al., 2024).

Astrocytes serve as the principal reservoir of glycogen in the adult brain (Beltran-Velasco, 2025). Mobilization of this glycogen provides a rapid energy buffer during periods of heightened neuronal demand and supplies lactate through the ANLS to sustain synaptic transmission (Chen Z. et al., 2023). This capacity positions astrocytes as key regulators of energy substrate availability under both physiological and metabolically stressful conditions.

Metabolic coupling between astrocytes and neurons is essential for maintaining cerebral energy homeostasis, with the ANLS representing its core mechanism (Kim et al., 2025). Astrocytes express MCT1 and MCT4 together with LDHA to export glycolytically derived lactate, whereas neurons rely on MCT2 and LDHB to import and oxidize lactate for ATP generation (Coca et al., 2025; Lee J. S. et al., 2025; Kim et al., 2025). Through this division of labor, astrocytes function as metabolic gatekeepers that match neuronal energy supply to activity demands. Nevertheless, the functional significance of ANLS remains a subject of debate. Measurements of cerebral oxygen consumption using BOLD fMRI and related approaches have suggested that glucose remains the indispensable substrate for brain energy metabolism, thereby questioning the efficiency of lactate-based energy transfer and the quantitative contribution of lactate transport and oxidation under physiological conditions (Schurr, 2025).

Beyond metabolic support, astrocytes exert broad regulatory control over neural circuit development and function. They secrete glia-derived neurotrophic factors and lipid precursors such as cholesterol, which are essential for oligodendrocyte precursor differentiation and myelination, and form perisynaptic astrocytic processes that directly participate in synapse formation, maturation and elimination across the synaptic life cycle (Chung et al., 2024; Tan et al., 2023). The tripartite synapse exemplifies this integrative neuron–astrocyte signaling unit. Glutamate released during synaptic activity activates NMDA and AMPA receptors to drive Ca²⁺ influx and long-term potentiation (LTP), thereby encoding learning and memory (Basavaraju and Priyadarshini, 2025). However, excessive extracellular glutamate induces excitotoxicity and neuronal injury (Lee et al., 2022). Astrocytes prevent this by rapidly clearing synaptic glutamate via EAAT1/2 and converting it to glutamine through glutamine synthetase, thus completing the glutamate–glutamine cycle (Andersen and Schousboe, 2023a; Todd and Hardingham, 2020).

Importantly, glutamate uptake is tightly coupled to astrocytic glucose metabolism (Pellerin and Magistretti, 1994). The Na⁺ influx accompanying EAAT-mediated transport activates Na⁺/K⁺-ATPase, increasing ADP/AMP levels and subsequently stimulating glycolysis, partly via AMPK-dependent pathways (Guillem et al., 2022; Muraleedharan and Dasgupta, 2022; Steinberg and Hardie, 2023). This metabolic activation enhances lactate production and directly fuels the ANLS, forming a coordinated “glutamate uptake–glycolysis–lactate release” mechanism that links neurotransmission to neuronal energy provision. In parallel, astrocytes channel glucose into the pentose phosphate pathway to maintain glutathione (GSH) pools, thereby supporting neuronal antioxidant defense.

As the principal metabolic intermediary between the BBB and neurons, astrocytes rely on efficient glucose uptake to sustain neuronal energy homeostasis. In AD, this capacity deteriorates early and profoundly.

Astrocytic glucose uptake is mediated largely by GLUT1, which is abundantly expressed at astrocytic endfeet and endothelial interfaces (Koepsell, 2020). Studies have demonstrated that astrocytes differentiated from induced neural progenitor cells (iNPCs) derived from patients with AD, as well as astrocytes from AD mouse models, exhibit marked reductions in GLUT1 protein abundance and aberrant membrane localization, directly impairing glucose uptake efficiency (Bell et al., 2025; Hendrix et al., 2021). Using an abdominal surgery model in aged wild-type mice, Chen Y. et al. (2023) further showed that GLUT1 downregulation disrupts cerebral glucose metabolism and robustly precipitates postoperative cognitive dysfunction, thereby establishing a causal link between impaired GLUT1 expression and cognitive decline. Notably, reductions in cerebral GLUT1 can be detected at early stages of AD, and its functional insufficiency accelerates Aβ pathology and disease progression (Winkler et al., 2015). However, despite these observations, the direct contribution of astrocyte-specific GLUT1 loss to AD pathogenesis remains insufficiently explored. Impaired cerebral insulin signalling represents a second major pathogenic axis. Widespread brain insulin resistance is a defining feature of AD (Kshirsagar et al., 2021), and astrocytes display a pronounced, cell-type-specific attenuation of insulin receptor (IR) activation together with reduced phosphorylation and signalling efficiency of downstream effectors, including PI3K and Akt (Bentivegna et al., 2025). This signalling deficit compromises insulin-dependent regulation of GLUT1 (Barthel et al., 1999) and further exacerbates defects in glycolysis and mitochondrial function, thereby amplifying metabolic vulnerability in the AD brain (Chen W. et al., 2023).

Cerebrovascular abnormalities—including amyloid angiopathy, chronically reduced cerebral blood flow, and impaired neurovascular coupling—further restrict glucose delivery (Custodia et al., 2023; Gao et al., 2025; Lourenco et al., 2017; Thal and Gawor, 2023). Because astrocytic endfeet closely monitor and respond to vascular substrate supply, vascular insufficiency synergizes with intrinsic transporter deficits to impair metabolic input (Thieren et al., 2025; Sheng et al., 2025). Approaches that restore cerebrovascular function, such as remote ischemic conditioning, can enhance GLUT1 expression and may offer metabolic benefits (Ma et al., 2024).

After entering astrocytes, glucose is partitioned into cytosolic glycolysis and mitochondrial oxidative metabolism. In AD, both pathways exhibit profound and multifaceted disruption.

Glycolytic dysfunction: Glycolytic regulation in AD astrocytes displays a biphasic trajectory. Single-cell transcriptomic studies reveal early upregulation of glycolytic genes—possibly reflecting a compensatory response—followed by their downregulation at later stages (Serrano-Pozo et al., 2024). Excessive glycolysis during reactive astrocytosis may facilitate a detrimental phenotype (Bell and Mortiboys, 2026; Sanchez et al., 2025). Aβ-mediated inhibition of the key glycolytic regulator PFKFB3 impairs stress-responsive glycolytic upregulation, increasing astrocyte vulnerability and exacerbating plaque deposition (Almeida et al., 2023; Ahmad et al., 2023). Integrated analyses of AD brain transcriptomes further reveal downregulation of core glycolytic enzymes, including GPI, PFKM and LDHA, changes that correlate with cognitive decline (Jia et al., 2025). Consistently, astrocytes differentiated from sporadic and familial AD (sAD/fAD) patient-derived iNPCs display impaired glycolytic capacity accompanied by reduced HK1 expression (Bell et al., 2025). Notably, experimental enhancement of aerobic glycolysis in astrocytes has been shown to improve neuronal metabolic support, promoting neuronal survival and axonal outgrowth (Zheng et al., 2021). Nevertheless, both excessive and insufficient glycolytic activity appear detrimental, underscoring that dysregulated—rather than directionally altered—glycolysis is a driver of AD pathogenesis.

Mitochondrial dysfunction: Mitochondrial abnormalities constitute a persistent antecedent of AD pathology (Maruthiyodan et al., 2024). Astrocytes from patients with AD and from AD mouse models exhibit pronounced defects in mitochondrial dynamics, characterized by increased fragmentation and disrupted cristae architecture, culminating in progressive mitochondrial dysfunction (Pradeepkiran and Reddy, 2020; Zysk et al., 2023). The tricarboxylic acid (TCA) cycle, a central metabolic pathway within the mitochondrial matrix, is markedly suppressed in astrocytes from 5xFAD mice, with concomitant reductions in NADH and FADH₂ levels, thereby impairing downstream oxidative phosphorylation (OXPHOS) and enhancing oxidative stress (Andersen et al., 2021c; Mohan and Kumar, 2025). Mitophagy serves as a critical quality-control mechanism by selectively eliminating depolarized, damaged or aged mitochondria to prevent their accumulation (Wang et al., 2026). However, studies in 5xFAD mice demonstrate that astrocyte-specific insulin signalling deficiency reduces autophagic flux and perturbs mitophagy, thereby aggravating mitochondrial dysfunction and accelerating disease progression (Chen W. et al., 2023). In parallel, downregulation of astrocytic GLUT1 markedly limits glucose availability for OXPHOS, leading to impaired ATP production and excessive reactive oxygen species (ROS) accumulation (D'Alessandro et al., 2025). Accordingly, astrocytes derived from both sporadic and familial AD patient iPSCs exhibit elevated mitochondrial ROS levels and aberrant complex I activity (Bell et al., 2025).

Impaired glycogen metabolism: As the brain’s principal glycogen reservoir, astrocytes depend on glycogen synthase to maintain emergency energy stores. In AD, glycogen synthase activity is inhibited and glycogen levels fall (Alberini et al., 2018; Bak et al., 2018), weakening a critical metabolic buffer.

The ANLS is central to metabolic coupling (Kim et al., 2025). In AD, multiple components of this metabolic output system fail. In AD, however, multiple components of this metabolic output system are disrupted. In both 3 × Tg-AD and APP/PS1 mouse models, astrocytes exhibit reduced basal lactate production derived from aerobic glycolysis, accompanied by marked downregulation of the lactate transporters MCT1, MCT2, and MCT4 relative to controls (Le Douce et al., 2020; Zhang et al., 2018). Consistently, individuals at high risk for AD show decreased MCT4 expression in the posterior cingulate cortex as early as young adulthood (Xu et al., 2016), resulting in a significant reduction in lactate availability to neurons (Wang et al., 2024). Neuronal lactate utilization is likewise compromised due to reduced LDHB expression and decreased pyruvate dehydrogenase (PDH) activity, both of which correlate with dementia severity (Jiang et al., 2023; Sakimura et al., 2024; Zhang et al., 2018).

By contrast, reactive astrocytes may, under certain conditions, upregulate glycolysis and produce excessive lactate, which can act as a signalling metabolite to amplify neuroinflammatory responses (Xiong et al., 2022). These seemingly divergent observations likely reflect differences in disease stage, experimental model or analytical methodology. Collectively, they underscore the importance of conceptualizing the ANLS as a dynamic and context-dependent pathway, rather than a static metabolic circuit, for understanding metabolic dysregulation in AD.

The glutamate–glutamine cycle, which under physiological conditions ensures efficient neurotransmitter recycling between astrocytes and neurons, is likewise driven by astrocytic glucose metabolism. Astrocytes supply both the carbon skeletons and the energetic input required for glutamine synthesis through glycolysis and the TCA cycle (Andersen and Schousboe, 2023b). In AD, this process is profoundly disrupted, resulting in impaired glutamine synthesis and pathological glutamate accumulation. The TCA cycle provides α-ketoglutarate as a precursor for glutamate and glutamine biosynthesis; notably, reduced TCA cycle activity and diminished glutamine production have been observed in hippocampal astrocytes from 5xFAD mice (Andersen et al., 2021b; Andersen et al., 2021c). Moreover, astrocytes exposed to amyloid-β (Aβ) display decreased expression and activity of glutamine synthetase (GS) (Andersen et al., 2022). Glutamate is the principal excitatory neurotransmitter in the central nervous system and is essential for learning and memory. In AD, Aβ enhances presynaptic glutamate release while concurrently downregulating astrocytic EAAT1/2 transporters, leading to excessive extracellular glutamate accumulation and excitotoxic neuronal injury (Andersen et al., 2022; Wu et al., 2025). Energy deficits weaken astrocytic glutamate clearance (Andersen et al., 2021b), disrupting the “glutamate uptake–lactate production” axis that normally drives ANLS coupling (Barros et al., 2023).

These failures in lactate shuttling and neurotransmitter cycling constitute a collapse of the metabolic output systems essential for neuronal survival.

The gut microbiota functions as a major upstream regulator of brain metabolism, communicating with the CNS through immune, endocrine, and neural pathways (Loh et al., 2024; Wang et al., 2023). Both astrocytes and the microbiota act as highly sensitive metabolic sensors, integrating systemic metabolic cues to maintain cerebral energy balance (Zhao et al., 2021). Gut microbes modulate expression of ANLS-related genes and influence astrocytic metabolism (Margineanu et al., 2020). Probiotic supplementation restores GLUT1, GLUT3, and IGF1Rβ expression in 3 × Tg-AD mice, enhancing glucose uptake and demonstrating microbiota-driven regulation of brain glucose metabolism (Bonfili et al., 2020).

Microbial metabolites—particularly short-chain fatty acids (SCFAs)—support BBB integrity, insulin sensitivity, astrocytic glycolysis, mitochondrial respiration, and MCT1/4 expression (Jung et al., 2024; Silva et al., 2020; Zhang L. et al., 2025). SCFA supplementation rescues glutamate–glutamine cycle deficits in APP/PS1 mice (Zhang L. et al., 2025). In AD, dysbiosis alters SCFA levels and diminishes these neuroprotective effects (Silva et al., 2020). Dysbiosis-induced metabolites such as TMAO and LPS penetrate the BBB, activate toxic astrocytic programs (e.g., Smurf2/ALK5 and TLR4/NF-κB pathways), impair insulin signaling, and promote systemic inflammation (Abildinova et al., 2024; Peng et al., 2021; Haripriya et al., 2024; Su et al., 2021; Zhao et al., 2021). These changes destabilize cerebral glucose supply and further impair astrocytic metabolism. A self-reinforcing cycle emerges: disrupted gut metabolites impair astrocytes, while metabolically compromised astrocytes weaken neuronal support, promote inflammation, damage the intestinal barrier, and exacerbate microbial imbalance (Zhang L. et al., 2025). This systems-level breakdown contributes to the global collapse of the astrocytic metabolic cascade.

Astrocytic glucose metabolism in AD is disrupted at every level—from systemic gut–brain regulation to metabolic substrate input, intracellular processing, and metabolic output. This progressive breakdown shapes a pathological axis of energy deficiency that intersects with Aβ accumulation, tau pathology, synaptic failure, and neuronal degeneration.

Synaptic dysfunction lies at the core of cognitive decline in Alzheimer’s disease (Mecca et al., 2022). Astrocytic metabolic dysfunction precipitates synaptic loss and excitotoxicity through intertwined mechanisms encompassing energy deprivation, and neurotransmitter dysregulation. In the hippocampus, reduced GLUT1 expression in hippocampal astrocytes diminishes local energy supply, leading to decreased dendritic spine density and impaired synaptic plasticity (Li Z. Y. et al., 2025), underscoring the direct impact of metabolic insufficiency on synaptic integrity.

In AD, impaired astrocytic glycolysis and the consequent reduction in lactate production compromise synaptic support. The disruption of the ANLS deprives neurons of adequate ATP. This energy deficit undermines key presynaptic processes, including the ATP-dependent proton gradient necessary for synaptic vesicle cycling and neurotransmitter loading (Faria-Pereira and Morais, 2022). Insufficient ATP also impairs Na⁺/K⁺-ATPase function, contributing to aberrant neuronal excitability (Guerra et al., 2025). Additionally, lactate acts as a signaling molecule to regulate the expression of plasticity-related genes (e.g., ARC, BDNF); its reduction may thus impair synaptic plasticity and memory consolidation (Beard et al., 2021; Suzuki et al., 2011). Astrocytic metabolic dysfunction disrupts neurotransmitter homeostasis. In AD mouse models, impaired metabolic capacity leads to reduced glutamine synthesis, thereby compromising neurotransmitter recycling and ultimately precipitating synaptic dysfunction (Andersen et al., 2021a). The synthesis of gliotransmitters such as D-serine relies on the glycolytic intermediate 3-phosphoglycerate. Downregulated L-serine levels in AD models, linked to weakened glycolysis, may impair NMDA receptor function and LTP (Orzylowski et al., 2021). Notably, elevated D-serine levels have been detected in postmortem AD brains, which may relate to dynamic changes in glycolysis during disease progression (Le Douce et al., 2020; Orzylowski et al., 2021). Astrocytes also mediate rapid clearance of excitatory neurotransmitter glutamate from the synaptic cleft via EAAT1/2 transporters, a process highly ATP-dependent (Guillem et al., 2022). ATP deficiency combined with reduced EAAT1/2 expression leads to glutamate accumulation, excessive NMDA receptor activation, calcium influx, enhanced long-term depression (LTD), and impaired LTP, collectively compromising synaptic function (Andersen et al., 2022; Popugaeva et al., 2017). Restoration of normal astrocytic glucose metabolism is therefore pivotal for maintaining synaptic integrity, supporting neuronal communication, and mitigating AD pathology.

In AD, Aβ/tau pathology and metabolic dysfunction mutually reinforce each other. Studies show that inhibiting the key glycolytic regulator PFKFB3 in astrocytes (e.g., with 3PO/PFK15) increases Aβ accumulation and plaque formation (Fu et al., 2015); whereas inhibiting the astrocytic pyruvate carrier in 3xTgAD mice reduces amyloid and tau deposition (Ceyzeriat et al., 2024). These findings suggest that astrocytic glucose metabolism can drive AD pathological progression, likely through indirect mechanisms involving the exacerbation of neuronal energy crisis and oxidative stress (Figure 2).

Impaired astrocytic glycolysis and the consequent disruption of the ANLS lead to neuronal ATP deficiency and an energy crisis. This energetic state directly influences amyloid precursor protein (APP) processing: under sufficient energy, APP is primarily cleaved via the non-amyloidogenic pathway (α/γ-secretase) into soluble fragments; whereas under energy deficit, it tends to be processed through the amyloidogenic pathway (β/γ-secretase), generating full-length Aβ peptides (Blonz, 2017). The energy crisis also activates AMPK. While AMPK activation may help reduce Aβ production, it can also directly or indirectly (via GSK-3β activation) lead to tau hyperphosphorylation, promoting neurofibrillary tangle formation (Marinangeli et al., 2016; Khandelwal et al., 2022; Yu et al., 2025). Glycolytic products such as lactate exert bioactivity via post-translational modifications; for instance, lactylation of APP at K612 attenuates its interaction with BACE1, reducing Aβ generation. In AD patients and mouse models, APP-Kla levels are markedly decreased, and exogenous lactate supplementation can suppress Aβ production (Tian et al., 2025), an effect attenuated in advanced disease stages due to diminished glycolysis and lactate availability. Astrocytic antioxidant capacity is also compromised in AD, with reductions in GSH levels. Neuronal GSH synthesis relies on cysteine and glycine derived from astrocytic GSH degradation, as well as glutamine provision by astrocytes (Dringen and Arend, 2025). In AD, downregulation of the key transcription factor Nrf2 further impairs astrocytic GSH production, while glucose metabolic deficits reduce glutamine availability (Andersen et al., 2022; Chu et al., 2024; Osama et al., 2020). Accumulating ROS further deplete astrocytic GSH reserves, impairing neuronal antioxidant defenses and contributing to oxidative stress intimately linked to AD pathology (D'Alessandro et al., 2025). ROS elevation enhances β- and γ-secretase cleavage of APP, favoring aggregation-prone Aβ species (D'Alessandro et al., 2025; Huang et al., 2020). Oxidative stress often precedes Aβ and tau pathology, and mitochondrial ROS promotes tau oligomerization and accumulation (Du et al., 2022; Roy et al., 2023). Excess ROS inhibit protein phosphatase 2A (PP2A), a primary astrocytic tau phosphatase, while activating GSK-3β, collectively driving tau hyperphosphorylation and NFT formation (D'Alessandro et al., 2025). The energy crisis and metabolic dysregulation also impair Aβ clearance mechanisms. Insufficient ATP supply may disrupt the autophagy-lysosomal pathway (Falace et al., 2024). In 5xFAD mice, loss of insulin signaling in astrocytes enhances tau phosphorylation and impairs Aβ clearance (Chen W. et al., 2023). Hyperglycemia and insulin resistance resulting from impaired insulin signaling competitively inhibit Aβ degradation by insulin-degrading enzyme (IDE) and downregulate the expression of the key astrocytic Aβ clearance receptor LRP1, collectively reducing Aβ clearance (Faissner, 2023; Sedzikowska and Szablewski, 2021; Li W. et al., 2025; Shinohara et al., 2017).

Notably, PET imaging indicates that in late-stage AD, glucose hypometabolism correlates with tau burden but not Aβ load (van Gils et al., 2024). While numerous studies link regions of metabolic impairment with Aβ and tau deposition, mechanistic delineation of how astrocytic metabolic dysfunction directly drives these pathologies remains limited and warrants further investigation.

In AD, astrocytes undergo reactive transformation, exhibiting high phenotypic and functional heterogeneity that extends beyond the simplistic A1 (neurotoxic)/A2 (neuroprotective) binary classification (Xiong et al., 2022). The A1 phenotype can directly damage neurons by releasing complement components (e.g., C3), pro-inflammatory factors, and reactive oxygen species, while losing their normal synaptic support functions (Liddelow et al., 2017; Xiong et al., 2022). Furthermore, reactive astrocytes may also influence neuronal network excitation-inhibition balance through GABA secretion (Garaschuk and Verkhratsky, 2019). Conversely, the A2 phenotype is capable of releasing neurotrophic factors and aiding in the clearance of pathological protein aggregates (Liddelow and Barres, 2017).

Current evidence indicates that augmented glycolysis in astrocytes is a pivotal driver of their reactive activation in AD, fostering a self-reinforcing pro-inflammatory feed-forward loop (Bell and Mortiboys, 2026; Sanchez et al., 2025). Notably, dual PET tracer imaging has demonstrated a close association between astrocytic activation and heightened glycolytic activity (Carter et al., 2019). Evidence indicates that Aβ-activated microglia indirectly induce A1 astrocyte formation, a process dependent on upregulated astrocytic glycolysis (Zhang et al., 2022). Glycolytic flux is critical for activating neurotoxic astrocytes, and its inhibition in APP/PS1 models suppresses A1 reactivity (Yang X. et al., 2024). The glycolytic regulator PFKFB3 sustains reactive astrocyte metabolism and function (Calì et al., 2024). Notably, increased glycolysis in AD can paradoxically fuel inflammation by promoting astrocytic activation, forming a pathogenic cycle. Studies in primary astrocytes confirm that NF-κB-driven inflammatory responses are glycolysis-dependent (Robb et al., 2020). PFKFB3 overexpression in reactive astrocytes drives a Warburg-like shift, leading to lactate accumulation that exacerbates neuroinflammation (Fang et al., 2024; Xiong et al., 2022). Excess lactate may also contribute to tau pathology, for instance by promoting lactylation at the K331 site, thereby exacerbating tau misfolding and aggregation. Inhibition of LDHA expression has been shown to ameliorate this process (Zhang X. et al., 2025). Furthermore, mitochondrial dysfunction participates in modulating the pro-inflammatory phenotype of astrocytes. The mitochondrial fission protein DRP1 promotes inflammatory factor release by driving mitochondrial fission and activating the NF-κB pathway (Cai et al., 2024; Sbai et al., 2023). Inhibition of DRP1-mediated mitochondrial fission has been proven to mitigate astrocyte-derived inflammation (Wu et al., 2020).

Microglial metabolic reprogramming further interweaves with astrocytic dysfunction (Sangineto et al., 2023). M1 microglia secrete IL-1α, TNF-α, and C1q, inducing resting astrocytes into neurotoxic A1 states (Singh, 2022), or directly stimulate A1 astrocytes via IL-1β, TNF-α, and IL-6 (Wu and Eisel, 2023). Reciprocally, activated astrocytes modulate microglial states, secreting IL-1β, CCL2, and CXCL10 to promote M1 microglial activation (Kwon and Koh, 2020; Wu and Eisel, 2023). This glial crosstalk fosters neuroinflammation, which in turn serves as a potent stimulus for glial activation, establishing a self-perpetuating vicious cycle. Pro-inflammatory stimuli (e.g., LPS, IL-1β/TNF-α) enhance glial activation and astrocytic glycolysis, and can induce lasting epigenetic changes in astrocytes (Linnerbauer et al., 2020; Pamies et al., 2021; Xingi et al., 2023). Together, this metabolic-inflammatory vicious cycle drives AD progression.