Introduction

The pathophysiological role of lipid droplets (LDs) has been extensively studied in conditions such as obesity, non-alcoholic fatty liver disease, cardiovascular diseases, and cancer1. In the central nervous system (CNS), notably, LDs were first described over a century ago, referred to as “adipose saccules” in glial cells of Alzheimer’s disease (AD) patients by Alois Alzheimer in 19072. Despite this early observation, the role of LDs in glial cells remained largely unexplored for more than a hundred years.

Recent studies have renewed interest in this area, with growing evidence implicating LDs in brain aging and AD, particularly through the emergence of microglia enriched with LDs, such as lipid droplet-accumulating microglia (LDAM)3. Microglia, the resident immune cells of the CNS, play a critical role in maintaining brain homeostasis, surveilling for injury, and responding to disease4. In recent years, accumulating evidence has revealed that microglial dysfunction contributes to the progression of various neurodegenerative diseases (NDDs), including AD, Parkinson’s disease (PD), and amyotrophic lateral sclerosis (ALS)5.

While early microglia research often equated activation with neuroinflammation, it is now clear that microglial activation is far more diverse and dynamic than the simplified M1/M2 paradigm6,7. Multi-omics characterization and functional analyses have demonstrated that microglia respond in disease-specific ways. Targeting distinct microglial subpopulations—such as disease-associated microglia (DAM), microglia in neurodegenerative disease (MGnD), white matter-associated microglia (WAM), and LDAM—holds considerable promise as a therapeutic strategy for NDDs7,8. These subtypes exhibit unique transcriptional and functional profiles, often arising in response to aging or pathological stimuli. However, a major barrier to clinical translation is the current lack of precise tools to selectively manipulate these subpopulations in vivo.

Among these, LDs-laden microglia—commonly referred to as LDAM—stand out as a potentially tractable target due to their distinctive metabolic signatures, particularly LDs accumulation and impaired phagocytic activity3. Given that immune cells dynamically adapt to their microenvironment through intracellular metabolic reprogramming—a process known as immunometabolism—modulating LDs metabolism in microglia represents a promising therapeutic avenue9,10.

This review explores the emerging role of LDs in microglial biology, with a focus on their biogenesis, accumulation in neurodegenerative conditions, and potential as therapeutic targets. We highlight recent findings that link microglial lipid metabolism to neurodegeneration and discuss pharmacological strategies that may restore microglial homeostasis by modulating LDs dynamics. Furthermore, we propose a hypothetical mechanism by which promoting LDs turnover may lead to metabolic reprogramming and functional restoration of microglia. Understanding LDs turnover not only in microglia but also in other brain cells—such as neurons, astrocytes, and oligodendrocytes—may open new therapeutic avenues for slowing or reversing the progression of NDDs.

Microglial lipid droplet in NDDs

Lipid droplet biogenesis

LDs are dynamic intracellular organelles composed of a neutral lipid core—primarily triacylglycerols (TAGs) and cholesterol esters (CEs)—encased in a phospholipid monolayer embedded with regulatory proteins. Unlike conventional membrane-bound organelles, LDs biogenesis originates from specialized subdomains of the endoplasmic reticulum (ER), where TAGs and CEs are synthesized by diacylglycerol acyltransferases 1 and 2 (DGAT1/2) and acyl-coenzyme A: cholesterol acyltransferases 1 and 2 (ACAT1/2), respectively1, with the involvement of long-chain acyl-CoA synthetases (ACSLs). ACSLs play a pivotal role in lipid metabolism, contributing to the production of LDs components such as TAGs and CEs11,12. These lens-like lipid structures, defined as the initial neutral lipid aggregates that form between the two leaflets of the endoplasmic reticulum (ER) membrane during LD biogenesis, serve as critical hubs for lipid storage and adaptation to cellular stress. Their protein and lipid composition varies across cell types and conditions, influencing their localization, function, and metabolic fate.

Structural proteins, particularly perilipins (PLIN1–5), regulate LDs stability and accessibility. For example, PLIN2-coated LDs are more prone to degradation, whereas PLIN1 and PLIN5 facilitate lipolysis by recruiting or activating adipose triglyceride lipase (ATGL), a key lipolytic enzyme1. LDs grow through lipid trafficking and coalescence, processes mediated by ER-to-LD (ERTOLD) and cytoplasm-to-LD (CYTOLD) tethering proteins13.

Traditionally regarded as passive lipid reservoirs, LDs are now recognized as active organelles that buffer cellular stress. They mitigate lipotoxicity by sequestering excess free fatty acids (FFAs) and other reactive lipid intermediates into inert neutral lipid forms. During nutrient deprivation, autophagy liberates FFAs from membrane degradation, which are re-esterified by DGAT1 into TAGs and stored in LDs to prevent mitochondrial damage14. Oxidative and ER stress also stimulate LDs formation via enhanced lipogenesis and phospholipid turnover, involving MAPK-activated phospholipases such as cytosolic phospholipase A2α (cPLA2α). In addition, LDs compartmentalize polyunsaturated fatty acids (PUFAs), protecting them from peroxidation. Under chronic stress, lysosomal degradation of phospholipids may increase the fatty acid pool available for LDs biogenesis.

Beyond lipid buffering, LDs participate in protein quality control, redox regulation, and organelle integrity15,16,17. In physiological states, they support membrane homeostasis and cell viability, and have also been implicated in viral replication and tumor survival1. For further details, LDs biogenesis has been comprehensively reviewed in previous papers1,18. The LD's synthetic mechanism is depicted briefly in Fig. 1.

Fig. 1: Lipid droplets biogenesis.
figure 1

Lipid droplets (LDs) originate from the endoplasmic reticulum (ER) through the enzymatic esterification of neutral lipids. Monounsaturated and polyunsaturated fatty acids (MUFAs and PUFAs) are first converted to acyl-CoA by long-chain acyl-CoA synthetases (ACSL). Acyl-CoA serves as a substrate for the synthesis of triacylglycerol (TAG) via diacylglycerol acyltransferase (DGAT) from diacylglycerol (DAG), and for cholesterol ester (CE) synthesis via acyl-CoA:cholesterol acyltransferase (ACAT) from free cholesterol. These neutral lipids accumulate between the leaflets of the ER membrane, where they coalesce and bud off into the cytoplasm as LDs coated with perilipins (PLINs). This process enables the storage of energy-dense lipids and regulates lipid homeostasis within the cell. Created with Biorender.com.

Full size image

Microglial LDs accumulation: lipid uptake from other cells, self cues, and non-self cues

LDs accumulation in glial cells, including microglia, has been documented for over a century. Historically linked to inflammation and cellular stress, LDs formation is now recognized as a broader metabolic response that may also reflect cellular adaptation to environmental cues19.

LDs accumulation in microglia arises from an imbalance between lipid supply and utilization. This elevation in intracellular lipids can result from exogenous inputs, such as intercellular fatty acid transfer, or endogenous requirement by oxidative stress and pro-inflammatory signaling—reflecting an adaptive mechanism to sequester excess fatty acids and buffer lipotoxic stress19. Neurons possess limited capacity for fatty acid oxidation (FAO) due to weak intrinsic antioxidant defenses, rendering them susceptible to the accumulation of fatty acids and subsequent lipid peroxidation20. PUFAs, which are enriched in neuronal membranes, are especially vulnerable to peroxidation by reactive oxygen species (ROS)21. In NDDs, including tauopathies, excessive neuronal lipid burden prompts intercellular lipid transfer to microglia as a compensatory mechanism to mitigate lipotoxic stress22. Similarly, conditioned media from lipid-laden astrocytes enhanced microglial chemotaxis via a CCR2–MCP1 axis and transferred LDs to microglia, suggesting that LDs-associated astrocytic inflammation may propagate neuroinflammation by recruiting and reactivating microglia23,24. However, while initially adaptive for CNS protection, excessive LDs accumulation or defective lipolysis in microglia may amplify oxidative stress, exacerbating neuronal injury and driving neuroinflammation.

Progranulin (GRN) is a key regulator of lipid droplet biogenesis in microglia25,26. Haploinsufficiency of GRN, a known genetic risk factor for frontotemporal dementia27, markedly enhances LDs accumulation in microglia. Notably, monocyte-derived microglia from patients carrying GRN mutations exhibit a pronounced increase in intracellular LDs, highlighting a direct link between GRN deficiency and microglial lipid dysregulation28.

Under stress conditions, microglia may shift energy metabolism away from FAO toward glycolysis, reducing lipid consumption. For example, inflammatory stimuli such as lipopolysaccharide (LPS) induce LDs formation in primary microglia29,30. Although not specific to microglia, viral infections such as HSV-1 and Zika virus are known to transiently increase LDs accumulation in human THP-1 monocytes. This process, driven by viral nucleic acids and EGF signaling, subsequently promotes type I interferon (IFN) secretion31. BV2 cells (murine microglial cell line) exposed to plasma derived from non-diabetic (control) mice (db/m) and leptin receptor-deficient (db/db) mice revealed a marked increase in BODIPY- and PLIN2-positive LDs following treatment with hyperglycemic db/db plasma, even after removal of lipids and EDTA, mirroring the in vivo phenotype32. Microglial LDs formation arises from diverse lipid sources, including the phagocytosis of apoptotic cells33, uptake of myelin debris34, and internalization of lipoprotein particles35. Lipoprotein lipase (LPL) facilitates the breakdown of TAGs derived from myelin debris or apoptotic cells and promotes lipoprotein uptake by interacting with lipoprotein receptor36. Actually, oleic acid treatment can induce LDs formation in microglia37,38. In addition, hypoxia also induces LDs formation in primary microglia39.

Direct evidence in microglia is lacking, but converging data implicate endogenous protein aggregates in LDs accumulation in microglia. TDP-43 C-terminal inclusions increase LDs in astrocytes40, PD-linked α-synuclein mutants show enhanced aggregation via LDs interactions in cells41, and A53T α-synuclein elevates neuronal TAG and LDs content—together establishing that aggregates can induce LDs formation in neurons with plausible relevance to microglia. Mechanistically, aggregates can elicit ER/lysosomal stress and inflammatory signaling, activating sterol regulatory-element binding protein (SREBP) and thereby driving lipogenesis and LDs biogenesis42,43,44,45,46. Reciprocally, LDs surfaces nucleate liquid-like protein condensates, suggesting phase separation helps organize LDs interfaces47. Together, these findings support a model in which endogenous protein aggregates drive LDs biogenesis in microglia, suggesting that these aggregate–LDs positive-feedback loop likely operates.

Neuronal mitochondrial dysfunction and ROS induce LDs accumulation in glia via neuronal c-Jun NH2-terminal kinase (JNK) and SREBP activation in Drosophila, leading to neurodegeneration48. Moreover, oxidative stress and mitochondrial dysfunction—hallmarks of the aging brain—can shift microglial metabolism toward lipid storage, promoting LDs accumulation3. Microglia harboring an apolipoprotein E (APOE4) allele showed altered cellular metabolism, increasing intracellular toxic LDs accumulation12. In this study, conditioned media from LDs-laden microglia induce Tau phosphorylation and neurotoxicity in an APOE-dependent manner. Microglia form LDs in response to amyloid-β (Aβ), with LDs accumulation increasing in proximity to amyloid plaques in both human AD brains and 5xFAD mouse models12. Plaque-associated human microglia accumulate LDs in a chimeric model of AD49. Interestingly, microglia in aged brains or in proximity to toxic protein aggregates—such as Aβ—often exhibit senescent features and elevated LDs content, likely due to suppressed autophagic activity and disrupted metabolic homeostasis50.

Regarding the metabolic pathway of LDs, LDs are metabolized through lipolysis and lipophagy (LDs autophagy)18,51. In lipolysis, TAGs are sequentially hydrolyzed by ATGL, hormone-sensitive lipase (HSL), and monoacylglycerol lipase (MAGL), releasing FFAs and glycerol. FFAs are then oxidized in mitochondria for ATP production52. In parallel, lipophagy mediates LDs degradation via autophagic pathways51. Unlike adipose tissue, which releases lipolysis products into the circulation, non-adipose tissues typically channel them into mitochondrial or peroxisomal β-oxidation for acetyl-CoA production53. Lipophagy mediates the trafficking of LDs to lysosomal compartments for degradation, a process orchestrated by Autophagy-related (ATG) proteins. Although lipophagy and classical lipolysis represent mechanistically distinct pathways for lipid mobilization, emerging evidence suggests functional crosstalk between them. The entire mechanism of LDs turnover was described in Fig. 2.

Fig. 2: Mechanisms of excessive lipid droplet accumulation and metabolic dysfunction in microglia.
figure 2

Microglia accumulate excessive lipid droplets (LDs) due to both intrinsic and extrinsic cues. Intrinsic triggers include ER stress, toxic protein aggregates, inflammatory mediators, metabolic stressors (e.g., high glucose, hypoxia), and genetic mutations (e.g., GRN, APOE4), all of which disrupt lipid homeostasis and promote LDs biogenesis. Extrinsic factors such as lipid or LDs transfer from neurons and astrocytes, and possibly peripheral lipids, further exacerbate LDs accumulation. This LDs-laden microglial state is associated with impaired nuclear-cytoplasmic transport (NCT), unresolved DNA damage, and mitochondrial dysfunction characterized by reduced fatty acid oxidation and ATP production. Maybe, as a compensatory response, GLUT1 expression is increased to support glucose uptake. Created with Biorender.com.

Full size image

LDs accumulation in CNS with NDDs

Lipid metabolism plays a central role in brain homeostasis, and its dysregulation is increasingly recognized as a contributing factor in NDDs. The brain, rich in lipids and highly dependent on energy homeostasis, becomes vulnerable to imbalances in lipid handling, especially within glial cells.

LDs accumulation progresses with age across multiple brain regions, including the pia mater, cortex, and striatum54. Ultrastructural analysis reveals LDs deposition along the blood–brain barrier (BBB), contributing to age-associated thickening of the basement membrane55. These findings indicate that aging modulates both the burden and anatomical distribution of LDs within the brain. In mice, aged microglia exhibit a significant rise in LDs content, as evidenced by increased BODIPY-positive cells at 20 months compared to 3 months3. This age-related trend is mirrored in human postmortem tissue, where PLIN2⁺Iba1⁺ microglia are more abundant in aged individuals. In this section, we address the potential involvement of microglial LDs in NDDs, including AD, PD, and ALS.

Microglial LDs in AD

Although first noted by Alois Alzheimer in 1907 as “adipose saccules” in glia, their significance was largely overlooked for over a century. However, recently, in AD, the role of LDs is gaining increasing attention. Recent studies reveal that LDs accumulate early in AD pathogenesis—preceding Aβ plaque and tau tangle formation—and contribute to disease progression12. LDs are particularly abundant in ependymal cells and glia of transgenic AD mouse models and human AD brains56. LDs accumulation in male microglia impairs tau clearance and exacerbates tauopathy and neuroinflammation, highlighting microglial LDs as key modulators of disease progression37.

AD risk genes, including triggering receptor expressed on myeloid cells 2 (TREM2), SREBP2, APOE4, ATP-binding cassette transporters A1 and A7 (ABCA1/7), phosphatidyl-inositol-binding clathrin assembly protein (PICALM), and low-density lipoprotein receptor-related protein 1 (LRP1), strongly suggest that disruption of cellular lipid homeostasis, which are lipid sensor, lipid synthesis, and lipid trafficking, is central to disease pathophysiology56,57,58. Analysis of RNA-seq data from the Accelerating Medicines Partnership–Alzheimer’s Disease (AMP-AD) cohorts consistently demonstrates elevated expression of lipid metabolism-related genes in AD brains. Given that cerebral glucose uptake is reduced in AD59, this upregulation may represent a compensatory adaptation to impaired glucose metabolism22. Given that cerebral glucose uptake is reduced in AD59, this upregulation may represent a compensatory adaptation to impaired glucose metabolism. Notably, a recent study showed that tauopathy neurons overproduce unsaturated lipids in both a Drosophila model and human AD brains, transferring these lipids to microglia22. This lipid transfer leads to LDs accumulation in microglia, driving neuroinflammation. TREM2, CD36, and LPL are implicated in lipid uptake from AD-related neurons into microglia60,61,62. Furthermore, APOE mutations, which impair lipid transport and clearance, can exacerbate LDs accumulation in microglia12,38. The APOE4 allele, the strongest genetic risk factor for late-onset AD, is less efficient at lipid transport compared with the protective isoforms APOE2 and APOE3. APOE-deficient glia accumulate more LDs even in the absence of neurons, whereas ABCA1 overexpression reduces tau pathology in P301S/APOE4 mice63. Additionally, neuronal adenosine monophosphate-activated protein kinase (AMPK) activation limits LDs formation cell-autonomously and suppresses lipid transfer to microglia, thereby mitigating microgliosis and neurotoxicity22.

Microglial LDs in PD

PD, the second most prevalent neurodegenerative disorder, is characterized by motor symptoms such as tremor, rigidity, and bradykinesia, as well as non-motor features including cognitive decline and sleep disturbances. Its neuropathological hallmarks include the progressive loss of dopaminergic neurons in the substantia nigra and intracellular accumulation of misfolded and insoluble α-synuclein protein. Histological analysis using the fluorescent lipid probe BODIPY revealed substantial intracellular lipid accumulation in dopaminergic neurons and midbrain microglia within the substantia nigra of PD brains, whereas neighboring astrocytes exhibited a comparatively reduced lipid burden64.

α-Synuclein is mainly cytoplasmic but also binds dynamically to membranes, including LDs41. α-Synuclein physiologically binds to phospholipid membranes, and the lipid composition of these membranes influences its aggregation propensity65. Accumulating evidence indicates that α-synuclein has significant reciprocal effects on cellular lipid metabolism66, reinforcing the growing notion that lipid-related defects play a central role in the pathogenesis of PD and other synucleinopathies67. Notably, α-synuclein interacts with the phospholipid monolayer on LDs, where it suppresses lipolysis and promotes LDs accumulation41,68. In addition, α-synuclein interacts with LDs-associated proteins to inhibit lipolysis, and its binding to LDs has been shown to promote pathogenic misfolding and aggregation of α-synuclein within neurons69.

Microglia not only rapidly engulf and degrade fibrillar α-synuclein transferred from neurons via tunneling nanotubes70, but also buffer excessive neuronal α-synuclein, thereby reducing the intracellular burden in neurons71. In addition, microglia-specific overexpression of α-synuclein leads to severe dopaminergic neurodegeneration by phagocytic exhaustion and oxidative toxicity72. Because LDs interact with α-synuclein and modulate its aggregation propensity, the presence of LDs-laden microglia is likely to contribute to PD progression. Moreover, α-synuclein–induced LDs accumulation may accelerate microglial phagocytic dysfunction and exacerbate inflammatory responses. Figure 3 provides a comparative illustration of LDs formation in AD, PD, and ALS, together with associated risk factors, synthesized from currently available research.

Fig. 3: LD dynamics in neurodegenerative diseases (AD, PD, ALS).
figure 3

Schematic of pathways proposed to promote microglial lipid-droplets (LDs) accumulation. AD: risk genes (TREM2, SREBP2, APOE4, ABCA1/7, PICALM, LRP1) disrupt lipid homeostasis; Aβ/ER stress drive DGAT2-dependent FFA → TAG conversion; neurons transfer lipids to microglia via TREM2/CD36. PD: neuronal α-syn aggregates transfer to microglia (e.g., via tunneling nanotubes); LDs-bound α-syn is linked to reduced phagocytosis and increased oxidative stress; lipolysis may further modulate toxicity. ALS: neuronal mitochondrial defects and astrocytic LDs remodeling (e.g., SOD1G93A, PDK2 silencing) increase lipid flux; mutations (VAPB/ALS8, SPG11) and DGAT2/TDP-43 dysregulation associate with altered LDs metabolism. Central cell depicts a microglion with increased LDs load. Solid arrows indicate reported interactions; dashed arrows denote inferred links. Created with Biorender.com.

Full size image

Microglial LDs in ALS

LDs have been linked to motor neuron diseases, including ALS73. ALS pathology includes significant alterations in lipid metabolism, particularly in glial cells74. However, although lipid dysregulation in ALS is well established, the role of LDs accumulation in microglia remains largely unexplored so far, with most evidence to date focusing on LDs formation in astrocytes. For example, ALS-related SOD1G93A mice exhibit abnormal mitochondrial function in motor neurons and LDs accumulation in aberrant astrocytes of the spinal cord75. Silencing pyruvate dehydrogenase kinase (PDK2) in SOD1G93A astrocytes restored mitochondrial bioenergetics and promoted a more interconnected mitochondrial network. This intervention also reduced LDs content, suggesting a metabolic shift toward enhanced oxidative capacity76.

Overexpression of TDP-43, which is a ubiquitous protein encoded by the ALS10 gene, increased fat accumulation and induced LDs growth with adipocyte hypertrophy77. LDs content is increased in the cytoplasm of TDP-43-expressing astrocytes40. On the other hand, loss of TDP-43 induces an apoptosis-independent cell death in HeLa cells and a build-up of LDs78. Mutations in VAPB/ALS8, which are associated with ALS, are involved in LDs accumulation in the muscle of Drosophila79. Mutations in SPG11, which is also responsible for rare forms of Charcot-Marie-Tooth (CMT) disease and progressive juvenile-onset ALS, promote lipid accumulation in lipofuscin-like structures in neurons80. A mutation in DGAT2—an enzyme responsible for converting DAG to TAG during LDs formation—has been identified as a causative factor in a family with autosomal-dominant early-onset axonal Charcot–Marie–Tooth disease (CMT)81.

Section “LDs accumulation in CNS with NDDs” has covered studies on LDs accumulation in other brain cells in addition to microglia. While research on microglial LDs in NDDs has focused mainly on AD and brain aging, studies in PD and ALS have largely examined astrocytes and other cell types. However, given the role of microglia in brain aging and their propensity to form LDs around toxic protein aggregates, further investigation into microglial metabolic abnormalities and LDs formation is warranted not only in AD but also in other NDDs.

A Hypothetical model on functional restoration of microglia via lipid droplets turnover

Recently approved monoclonal antibodies against amyloid aggregates in AD—such as lecanemab, aducanumab, and donanemab—have been shown to restore microglial phagocytic function82, suggesting that enhancing microglial phagocytosis holds therapeutic potential across NDDs characterized by toxic protein aggregates. While various steps of the phagocytic pathway are being explored as druggable targets, metabolic reprogramming of LDs-laden microglia offers a compelling therapeutic approach. Metabolic reprogramming is the process by which abnormalities in cellular metabolic pathways are corrected, enabling cells to regain or enhance beneficial functions. In this context, metabolic reprogramming of LDs-laden microglia can be achieved by suppressing LDs synthesis and enhancing LDs degradation. At the signaling pathway level, this can be mediated by AMPK, PPARs, CPT1A, SREBP, LPL, and TFEB, which will be discussed in Section 4. Microglia with excessive LDs accumulation—frequently observed in aged brains and AD—exhibit severely impaired phagocytic capacity3. Since protein aggregates such as Aβ and tau promote LDs formation in microglia, strategies aimed at reversing this lipid-laden state may restore microglial function and provide disease-modifying benefits.

Notably, LDs-accumulating microglia—termed LDAM—are enriched in the aging brain and might share molecular and phenotypic features with senescent microglia8. Senescent microglia exhibit nuclear-cytoplasmic transport (NCT) dysfunction, DNA damage, and secretion of senescence-associated secretory phenotypes (SASPs), including cytokines, ROS, and nitric oxide (NO)83. In addition, senescent glia arise in response to neuronal mitochondrial dysfunction and drive lipid accumulation in neighboring non-senescent glia, a phenomenon also observed in senescent human fibroblasts. Targeting these cells reduces senescence markers, extends lifespan and health span, and prevents lipid buildup, but increases brain oxidative damage and does not restore neuronal mitochondrial function. These findings position senescent glia as a link between mitochondrial dysfunction and lipid dysregulation in aging84. Recent studies suggest that LDs accumulation itself may serve as a hallmark of cellular senescence85. Disrupted fatty acid metabolism—commonly observed in senescent cells—further drives LDs formation, reinforcing a pathological cycle of impaired mitochondrial bioenergetics and chronic inflammation86.

Microglial phagocytic capacity depends on lipase-mediated release of fatty acids from intracellular lipid stores. ATGL deficiency impairs TAG hydrolysis, resulting in reduced intracellular FFA levels and LDs accumulation. Consequently, ATP production declines, leading to compromised phagocytosis87,88,89. Disrupted fatty acid metabolism—a hallmark of senescence—thus promotes LDs formation and establishes a vicious cycle in microglia characterized by inefficient mitochondrial energy production and persistent inflammatory responses86.

In 5xFAD AD mice, microglia upregulate the glycolytic enzyme hexokinase 2 (HK2), yet this metabolic shift fails to meet energy demands for effective phagocytosis and contributes to an inflammatory response86. Interestingly, HK2 inhibition reprograms microglial metabolism toward lipid utilization, enhancing ATP production and delaying disease progression. This shift reduces glycolytic intermediates such as glucose-6-phosphate (G6P) and fructose-6-phosphate (F6P), which otherwise repress lipid metabolism gene expression. The result is enhanced lipid uptake, partly mediated by upregulation of LPL61. However, for increased lipid uptake to effectively support energy production, imported lipids must be directed toward FAO rather than being stored as LDs. Although it remains unclear whether glycolysis inhibition alone is sufficient to induce a complete metabolic switch to FAO, available evidence suggests that reactivating lipid metabolism is closely associated with restored phagocytic function.

Taken together, we hypothesize that targeting LDs in microglia could interrupt the pathological energy metabolism cycle and promote functional restoration. Here, functional restoration refers to the normalization of key features observed in LDs-laden microglia, including elevated ROS, pro-inflammatory phenotype, impaired phagocytosis, and disrupted metabolic balance.

In NDDs, persistent cellular stress triggers maladaptive responses in microglia, leading to imbalanced LDs turnover and excessive LDs accumulation. This, in turn, impairs NCT by physically obstructing macromolecule trafficking between the nucleus and cytoplasm90. In addition, LDs can physically associate with the nucleus, potentially affecting nuclear morphology and integrity, and contributing to DNA damage and cellular stress. For example, oleic acid-induced LDs can physically indent and displace the nucleus, locally depleting lamin-B1, sometimes causing nuclear rupture, and triggering cGAS accumulation, DNA repair factor mislocalization, DNA damage, and delayed cell cycle in macrophage91. Regarding LDs acting as a physiological reservoir for nucleoporins, LDs accumulation impacts the conformation of NPCs and hence their function in nucleo-cytoplasmic transport, chromatin configuration, and genome stability92. Thus, we speculate that LDs formation itself is enough to impair the NCT system. Furthermore, LDs overload has been implicated in the structural and functional disruption of nuclear pore complexes (NPCs), further exacerbating nucleocytoplasmic transport deficits. Continuously accumulating, non-metabolized LDs fuel a vicious cycle of metabolic dysfunction, impairing homeostatic regulation and progressively diminishing microglial phagocytic and immunoregulatory functions. These dysfunctional microglia ultimately might contribute to chronic neuroinflammation, or inflammaging.

To interrupt this pathological cycle, we propose a hypothetical model for a therapeutic strategy centered on enforced LD degradation. Promoting LD turnover initiates a multifaceted metabolic reprogramming cascade. The proposed mechanisms include:

  1. (1)

    enhancement of LD metabolism via lipolysis or lipophagy18;

  2. (2)

    increased fatty acid supply to mitochondria52;

  3. (3)

    restoration of NCT through the removal of mechanical barriers (like LDs)83,93;

  4. (4)

    facilitation of DNA repair by improving nuclear import of DNA repair proteins91;

  5. (5)

    suppression of inflammatory mediator release;

  6. (6)

    functional restoration of microglia (Fig. 4).

Fig. 4: Hypothetical mechanism of microglial restoration through lipid droplet degradation.
figure 4

In LDs-laden microglia, pathological LD accumulation impairs mitochondrial metabolism, nuclear-cytoplasmic transport (NCT), and DNA repair, while promoting chronic inflammation. Enforcing LDs degradation through lipolysis or lipophagy initiates a metabolic reprogramming cascade (1), leading to increased fatty acid oxidation (FAO) (2), restoration of NCT (3), and enhanced DNA damage repair (4). These changes reduce the release of pro-inflammatory mediators such as IL-1β, IL-6, IL-8, MCP1, PGE2, and reactive oxygen/nitrogen species (ROS/NOX) (5). Together, these processes break the vicious cycle of microglial dysfunction by improving bioenergetics, reducing inflammation, and restoring nuclear homeostasis. The outcome is a functional restoration of microglia from a pro-inflammatory, LDs-laden phenotype to a homeostatic state, characterized by improved phagocytosis and reduced oxidative stress. Created with Biorender.com.

Full size image

This hypothetical framework may also apply to other LDs-laden brain cell types, such as neurons and astrocytes.

For LDs-targeted microglial restoration to be effective, several prerequisites must be met. Mitochondrial function must remain intact, as fatty acids released from LDs via lipolysis or lipophagy need to be fully oxidized in mitochondria to avoid additional metabolic burden. In the context of mitochondrial dysfunction, such as in Mfn1KO cells, FAs are redirected back into LDs or exported from the cell rather than oxidized94. Furthermore, excessive FFAs can induce lipotoxicity if not properly metabolized. While increased FAO can support energy production, it may also raise mitochondrial ROS levels, necessitating balanced activation of antioxidant defense systems to prevent oxidative damage95. Therefore, the metabolic prerequisites for safe and effective LDs clearance include sufficient mitochondrial oxidative capacity, competent FAO, and robust ROS-buffering systems.

Because lipids serve as a principal energy source for microglia, regulating LDs turnover—not indiscriminate LDs elimination—is essential. Moreover, it remains unclear whether promoting LDs degradation in microglia without significant lipid accumulation confers any functional benefit because LDs physiologically protect the cell under oxidative stress. These considerations underscore the need for precision strategies that account for the metabolic state and resilience of microglia. Enhancing LDs metabolism in a context-dependent manner may offer the most effective route to restore microglial homeostasis and mitigate neurodegeneration.

Possible therapeutics targeting lipid droplets

Caloric restriction (CR), intermittent fasting (IF), and exercise have been consistently associated with lifespan extension and neuroprotection in various models of NDDs96,97,98, effects partly captured by aging clocks98,99,100,101,102. Their shared therapeutic mechanisms across aging-related NDDs are strongly associated with autophagy and lipid metabolism, particularly through the regulation of key signaling pathways such as mTOR, AMPK, and TFEB. For example, CR and IF reshape hepatic lipid droplet proteomes and diacylglycerol species, protecting diabetes-prone New Zealand Obese mice from disease103. In humans, IF and CR modulate lipid metabolism markers in skeletal muscle104, while acute CR triggers a metabolomic shift from carbohydrate to fat utilization, marked by increased lipolysis and β-oxidation105. Exercise similarly improves LDs metabolism, as shown in a mouse model of non-alcoholic fatty liver disease via PLIN2–lysosomal acid lipase (LIPA) axis-mediated lipophagy106. Notably, exercise also reprograms aged microglia toward a youthful gene expression profile, and although microglial LDs were not directly measured, their known accumulation in the aged brain suggests that exercise may help reduce LDs burden107. Although the systemic nature of these interventions complicates mechanistic dissection, it has been reported that nutrient deprivation robustly enhances autophagic flux and facilitates LDs turnover52. Interestingly, in lipid-rich microglia, transient starvation promotes efficient ATP generation via mitochondrial FAO, although prolonged deprivation or depletion of LDs can result in lipotoxicity by mobilizing membrane-derived fatty acids52.

Although partly speculative, these findings provide a rationale for restoring microglial metabolic homeostasis by promoting catabolism of pre-existing LDs. In lipid-laden microglia, a shift toward glycolytic metabolism is often insufficient to meet bioenergetic demands, particularly under inflammatory stress. Therefore, re-engaging FAO through enforced LD degradation may represent a precise strategy to restore microglial function.

A primary approach to preventing excessive LDs accumulation is to limit neutral lipid synthesis. DGAT1 and 2, which catalyze the final step in TAG synthesis, are central to LDs biogenesis. For example, pharmacological inhibition of DGATs reduces LDs accumulation in astrocytes under nutrient stress14 and in iPSC-derived APOE4 microglia, where it also downregulates pro-inflammatory and cytokine response pathways108. In AD and 5xFAD brains, microglial LDs accumulation near amyloid plaques impairs Aβ phagocytosis through a DGAT2-driven shift from FFA to TAG. Pharmacological DGAT2 inhibition restores Aβ clearance, reduces plaque burden, and mitigates neuronal damage—providing the most direct evidence to date for our hypothesis that targeting LDs metabolism can restore microglial function in AD109. Furthermore, evidence also indicates that reducing LDs accumulation in brain macrophages—including both microglia and border-associated macrophages (BAMs)—is beneficial in the treatment of AD110, suggesting DGAT2 as an LDs-related target in AD111.

Similarly, chronic administration of the ACAT inhibitor CP-113,818 in 6-month-old hAPPFAD mice for two months significantly reduced amyloid plaque burden and improved spatial memory in the Morris Water Maze test112. Genetic deletion of FIT2, an ER-resident protein that facilitates LDs formation, also decreased LDs load and enhanced Aβ phagocytosis in microglia110. In PD models, ACSL4 inhibition ameliorates disease phenotypes by reducing lipid peroxidation and ROS production113. Interestingly, in the context of acute inflammation, inhibition of ATGL-mediated lipolysis in microglia attenuates LPS-induced neuroinflammation and behavioral deficits114.

These observations suggest that, while LDs formation acts as a protective response under certain stress conditions, its chronic inhibition may lead to cellular damage. Conversely, enhancing LDs metabolism in cells already burdened with excess LDs exerts cytoprotective effects, improving cellular function and reducing inflammatory outputs.

Despite promising preclinical evidence, several challenges hinder the clinical translation of LDs-targeting strategies in the CNS. First, many LDs-modulating small molecules developed for peripheral disorders lack the ability to cross the BBB. Second, LDs serve important physiological roles—even in healthy states—such as buffering lipid toxicity and maintaining membrane integrity. As such, therapeutic strategies should avoid completely blocking LDs formation. Instead, enhancing LDs turnover—mobilizing stored lipids for mitochondrial oxidation—may be a more effective and safer approach, particularly in microglia, which rely heavily on lipid metabolism for energy. Thus, a second approach to consider is promoting the catabolism of pre-existing LDs.

LDs catabolism is regulated through two main pathways: lipolysis and lipophagy. AMPK, a central regulator of energy homeostasis, stimulates both processes while inhibiting anabolic lipid synthesis115. Metformin, a well-established AMPK activator116, reduces LDs burden in multiple cell types. Although context-dependent effects on LDs formation have been noted, metformin’s capacity to enhance autophagy may contribute to the restoration of microglial lipid homeostasis, potentially explaining its neuroprotective effects observed in various NDD models. Chaperone-mediated autophagy (CMA) regulates lipid metabolism by selectively degrading LDs-associated proteins, specifically PLIN2 and PLIN3, in lysosomes via heat shock cognate 71 kDa protein (HSC70)117. This degradation occurs prior to lipolysis and facilitates the recruitment of lipases like ATGL and macrolipophagy-related proteins to LDs. In this study, nutrient deprivation activates CMA, enhancing PLIN2/3 degradation and promoting lipid oxidation, resulting in lipolysis and FAO. The effects of CMA-inducing small molecules NDDs, have been discussed in a previous review article118.

Transcription factor EB (TFEB) is another promising therapeutic target. As a master regulator of lysosomal biogenesis and autophagy119, TFEB is dephosphorylated under stress or starvation and translocates into the nucleus, where it activates genes involved in autophagy, lysosomal function, and lipid degradation. Pharmacological activators of TFEB—including metformin and resveratrol—act through AMPK-mediated dephosphorylation. Rapamycin, a mTOR inhibitor, similarly facilitates nuclear TFEB translocation. Other small molecules such as gemfibrozil and all-trans retinoic acid (ATRA) enhance TFEB expression and nuclear localization, producing anti-inflammatory effects in neurons, astrocytes, and microglia120,121,122. Moreover, a synthetic curcumin analog activates TFEB via an mTOR-independent pathway, reducing tau pathology in vivo123,124. Supplementary Table 1 summarizes agents currently in clinical trials that may modulate LDs across multiple diseases.

However, achieving microglia-specific delivery of these compounds remains a major hurdle for translational medicine. An emerging strategy involves lipid nanoparticles (LNPs) engineered to cross the BBB and selectively deliver above mentioned targets to microglia125,126. A recent study developed an LNPs platform enabling efficient, low-toxicity RNA delivery to activated microglia both in vitro and in vivo via systemic or intracisternal injection127. Targeting the AD-risk gene PU.1 using siRNA-loaded LNPs reduced neuroinflammation in AD models, providing a promising proof-of-concept for microglia-focused RNA therapeutics. Similarly, an adeno-associated virus (AAV) vector containing a 466-bp fragment of the human IBA1 promoter has been used to achieve microglia-specific expression128. However, such gene therapy strategies currently face translational limitations for routine clinical use.

In summary, a growing body of evidence supports the therapeutic potential of modulating microglial LDs metabolism in NDDs. Strategies that fine-tune the balance between LDs formation and degradation—and reprogram intracellular lipid metabolism—hold promise for restoring microglial homeostasis, mitigating neuroinflammation, and slowing disease progression. Figure 5 illustrates the proposed mechanisms by which candidate therapeutics may promote LDs turnover and exert neuroprotective effects.

Fig. 5: Therapeutic candidates targeting LDs in microglia.
figure 5

This figure summarizes the molecular mechanisms governing LDs biogenesis and degradation, highlighting potential therapeutic targets and modulators. LDs degradation occurs via two major pathways: lipolysis and lipophagy. Lipolysis is mediated by ATGL, HSL, and MGL, sequentially hydrolyzing TAG into free fatty acids (FFAs) and monoacylglycerol. Lipophagy involves the engulfment of LDs by autophagosomes (LC3-associated) and subsequent degradation in lysosomes. Therapeutic interventions can target these pathways. Inhibitors of LDs formation include DGAT and ACAT inhibitors (e.g., CP-113,818), perilipin modulators, and ACSL inhibitors (e.g., Triacsin C). In contrast, enhancing LDs degradation can be achieved through lipolysis inducers, perilipin regulators, or lipophagy enhancers such as metformin, resveratrol, rapamycin, and TFEB activators. These strategies offer potential to restore cellular lipid homeostasis and improve microglial function in neurodegenerative diseases. Created with Biorender.com.

Full size image

Future directions and unanswered questions

Intervention timing for LDs–laden microglia in NDDs

LDs accumulation is a hallmark of microglia in the aging brain. However, few studies have delineated its correlation with clinical metrics in human neurodegenerative disorders. As LDs formation is not inherently pathological, the critical challenge lies in defining the temporal inflection point at which LDs transition from adaptive to maladaptive—characterized by failure in metabolic turnover and functional impairment. This window may represent an optimal stage for therapeutic intervention. Uncertainty remains as to whether sustained pharmacological manipulation of LDs dynamics yields long-term benefits once LDs burden is initially reduced. This ambiguity parallels systemic interventions such as, where transient energy restriction promotes cellular resilience and longevity, but chronic deprivation induces stress and dysfunction. Similarly, autophagy induction confers cytoprotection, yet its overactivation can IF culminate in membrane destabilization and cell death. These observations underscore the necessity for temporally calibrated and mechanistically informed strategies to therapeutically modulate LDs metabolism in microglia.

Mechanistic gaps in lipid handling pathways

Despite growing interest in microglial lipid biology, fundamental gaps remain in understanding how mitochondrial dysfunction, ER stress, and autophagy impairment converge to promote aberrant LDs formation. Genetic risk alleles linked to neurodegeneration—such as APOE4 and TREM2 variants—are known to affect microglial lipid sensing and trafficking, yet their roles in LDs biogenesis and catabolism are incompletely defined. Addressing these unresolved mechanisms will require integrative approaches that combine spatial transcriptomics, proteomics, and metabolomics with high-resolution live-cell imaging and functional metabolic profiling.

Therapeutic delivery and safety considerations

Translation of LDs-targeted therapies into clinical settings is limited by the challenge of delivering lipid-modulating agents selectively to microglia within the CNS. Efficient crossing of the BBB without perturbing peripheral immune populations or non-target neural cells remains a major obstacle. Furthermore, systemic targeting of lipid metabolic enzymes—such as TFEB modulators, DGAT inhibitors, or mTOR inhibitors—risks off-target effects on hepatic, adipose, and myeloid tissues. Consequently, the development of CNS- and microglia-specific delivery systems, including LNPs, engineered extracellular vesicles, or antibody–drug conjugates targeting TMEM119 or P2RY12, is critical for therapeutic success.

Conclusion

The emerging landscape of LDs biology in microglia lies at the intersection of neuroimmunology, metabolism, and aging. Rather than viewing LDs solely as pathological indicators, it is increasingly evident that their roles are context-dependent—shifting from protective energy reservoirs to drivers of dysfunction under chronic stress. Clarifying these functional dualities and identifying molecular switches that govern LDs fate will be pivotal. Future efforts must focus on establishing spatially and temporally defined therapeutic windows, coupled with in vivo tools to monitor and modulate LDs dynamics. With improved mechanistic insight and the advent of precision delivery technologies, targeting microglial lipid metabolism holds promise as a novel therapeutic avenue for NDDs and brain aging.