Sex Differences in Metabolite–Immune Circuits of Neuroinflammation

ABSTRACT

Sex is a fundamental yet underexplored determinant of human neuroinflammation. Across autoimmune, neurodegenerative, and post‐infectious neurological syndromes, males and females exhibit consistent differences in disease vulnerability, progression, and immune tone. While sex hormones and chromosomes strongly shape immune development and function in health and disease, they do not fully explain the magnitude or disease‐specific patterns of these disparities, nor do they provide sufficient mechanistic information for developing novel therapeutics. Emerging evidence suggests that sex‐defining factors interact with age and environment to shape downstream metabolite–immune circuits, networks in which metabolic enzymes, metabolites, and immune cells tune inflammatory set points. Pathways spanning purine metabolism, glycolytic remodeling, lipid sensing, mitochondrial stress, and nucleic‐acid sensing can recalibrate microglial activation thresholds, T‐cell cytokine programs, innate type I interferon antiviral responses, and shape overall CNS resilience in a sex‐dependent manner. Here, we synthesize mechanistic and human systems‐level studies to propose an integrated framework in which sex‐biased immunometabolism serves as a mechanistic bridge between biological sex and neuroimmune disease risk, progression, and responses to injury. We highlight key knowledge gaps and discuss how targeting metabolite–immune pathways may enable sex‐informed biomarkers and therapeutic strategies in neuroinflammatory disease.

1. Introduction

Neuroinflammation is a defining feature of many neurological and neurodegenerative diseases, playing a central role in reprogramming cellular states and shaping brain homeostasis, tissue injury, repair and long‐term clinical outcomes. Within the central nervous system (CNS), tissue‐resident and infiltrating immune cells act as both drivers and resolvers of inflammation, orchestrating the local immunological landscape, as demonstrated in mouse models and supported by human disease studies [1, 2]. Across neurological disorders as diverse as multiple sclerosis (MS) [3], amyotrophic lateral sclerosis (ALS) [4, 5], Alzheimer's disease (AD) [6, 7], Parkinson's disease (PD) [8], autoimmune encephalitis [9, 10], stroke [11, 12], traumatic brain injury [13] and post‐infectious or treatment‐related syndromes such as long COVID [14, 15], the CNS exhibits striking sexually dimorphic patterns of disease susceptibility, progression and inflammatory tone. Females disproportionately develop MS, AD, autoimmune encephalitis and many chronic pain and fatigue syndromes, whereas males more frequently experience severe disease progression in PD, ALS and ischemic injury. Collectively, these disparities point to fundamental sex‐specific differences in brain immunology shaped by sex chromosomal complement, hormones and environmental influences. How these upstream factors are translated into durable differences in neuroimmune function remains incompletely understood, pointing to the existence of downstream integrative pathways that bridge sex, environment and inflammation.

Despite robust evidence for sex bias in neuroimmune disorders, the mechanistic basis underlying these differences remains incompletely understood. Studies focused on the effects of sex chromosomes and gonadal hormones in experimental animal models have provided important insights into immune calibration and microglial maturation [16, 17], yet they do not fully account for the magnitude, persistence or disease‐specific patterns of sexual dimorphism observed across the CNS in human neurological disease [7, 18, 19]. At a mechanistic level, these upstream factors converge on distinct metabolic–immune circuits that regulate neuroimmune function across health and disease. Functional studies in both animal models and human immune cells show that purine and pyrimidine metabolism, lipid pathways and mitochondrial stress responses are all directly linked to immune cell and signaling modules as critical regulatory nodes through which sex differences can shape neuroinflammation [20, 21, 22]. This framework is particularly compelling given the CNS's exceptional energetic demand, limited metabolic flexibility and reliance on finely tuned interactions among microglia, astrocytes, neurons and infiltrating immune cells. Perturbations in nucleoside/nucleotide pools, lipid intermediates and other key metabolites can disproportionately alter microglial activation thresholds, antiviral responses, phagocytic behavior and synaptic maintenance, thereby destabilizing neuroimmune homeostasis [23, 24, 25].

Recent mechanistic studies, including work from our laboratory, define how bioactive metabolites and their associated regulatory enzymes intersect with immune cell signaling pathways involved in defense responses against pathogens [26, 27, 28]. Loss‐of‐function studies of the tumor‐suppressor kinase Death‐Associated Protein Kinase 3 (DAPK3), a metabolic stress–responsive regulator of cell death and autophagy expressed in neurons, endothelial cells and myeloid‐lineage cells including microglia, identified DAPK3 as an essential upstream activator of innate immune cyclic GMP–AMP synthase (cGAS)–Stimulator of Interferon Genes (STING) signaling in human and mouse cells [27]. In this pathway, cGAS senses cytosolic DNA danger signals of microbial origin or generated by genomic or mitochondrial stress, triggering STING‐dependent signaling to generate type I interferon and its associated immune response. Multiple studies, mostly conducted in mouse models, now implicate cGAS–STING signaling in CNS health and disease, showing that microbial, mitochondrial or nuclear DNA stress activates STING in microglia and other CNS‐resident cells during aging, stroke, traumatic injury and Alzheimer's disease, where it drives type I interferon responses associated with synapse loss, neuroinflammation and neurodegenerative progression [29, 30]. Mechanistically, DAPK3 is essential for STING activation, and directly interacts with STING and promotes its K63‐linked ubiquitination, enabling downstream IFNβ production and associated immune responses [27].

Functional characterization of the metabolic enzymes that regulate systemic purine nucleoside homeostasis revealed that the adenosine deaminases (i.e., ADA1 and ADA2) control a purine‐regulated innate immune circuit that determines the magnitude of the IFNβ immune‐inflammatory response [28, 31, 32]. In human cells, ADA deficiency and/or accumulation of deoxyadenosine and deoxyinosine metabolites perturb cellular methionine–S‐adenosylmethionine (SAM) metabolism, leading to decreased SAM trans‐methylation potential and epigenetic derepression of human endogenous retroviral elements (HERVs) encoded in cellular genome [28]. HERV‐derived transcripts engage innate sensors of double‐stranded RNA, including Toll‐like receptor 3 (TLR3) and Retinoic Acid Inducible gene I (RIG‐I), thereby priming basal IFNβ expression [28]. In humans, germline ADA2 deficiency (DADA2) manifests clinically as a syndrome of chronic vascular inflammation and stroke, implicating dysregulated IFNβ signaling in this disease pathology [31, 33]. Human cohort studies have demonstrated that ADA enzymatic activity is sexually dimorphic, with male subjects exhibiting higher ADA2 activity than female counterparts [34, 35]. Across groups of healthy individuals and patients with COVID‐19, lower ADA activity in females associated with enhanced expression of type I IFN–driven antiviral gene signatures in the lung [36, 37], situating purine metabolism within a broader framework of sex‐ and tissue‐specific immune regulation [34].

Additional work links mitochondrial purine metabolism to innate immune activation and neurological disease. Disruption of deoxyguanosine kinase (DGUOK), the mitochondrial enzyme responsible for phosphorylating deoxyadenosine to its monophosphate derivative, in human hepatocytes induces a robust IFNβ and interferon‐stimulated gene (ISG) program [35]. Given that DGUOK deficiency causes a rare disorder characterized by hepatic and neurological dysfunction, these findings implicate a previously underappreciated immunometabolic axis with direct relevance to neurological disease. Nucleoside metabolism also shapes adaptive immunity. A functional genetic screen in endogenous human Th1 memory CD4⁺ T cells, the long‐lived helper T cells that retain immunological memory of intracellular bacterial infections such as Mycobacterium tuberculosis , identified isochorismatase domain‐containing protein 1 (ISOC1) as a key regulator linking pyrimidine metabolism to IFNγ and IL‐17 production [35, 36]. Disruption of ISOC1 impaired effector cytokine output, which was rescued by supplementation with extracellular pyrimidines, highlighting direct metabolic control of human T‐cell effector function.

Collectively, these studies support a model in which metabolic enzymes (ADA1, ADA2, ISOC1, DGUOK) and innate immune sensors (TLR3, RIG‐I, and STING via DAPK3) operate as integrated circuits that tune the amplitude and chronicity of IFN‐driven inflammation in a tissue‐ and sex‐dependent manner. When considered alongside growing evidence that innate IFN signaling and purine metabolism shape microglial states and CNS injury responses [28, 29, 30, 32], these findings suggest that sex differences in neuroinflammation may arise, in part, from differential wiring of metabolic–immune pathways within microglia and other cells at the neuroimmune interface. Despite these advances, no unified framework currently synthesizes how sex, metabolism, and neuroimmune compartments converge to govern neuroinflammatory disease risk, progression, and burden. The field lacks comprehensive, sex‐stratified neuroimmune maps integrating transcriptomic, lipidomic, and metabolomic data with functional states of CNS‐resident and infiltrating immune populations across developmental stages and hormonal contexts. In this review, we outline current knowledge of sex differences in neuroinflammation, highlight emerging metabolic–immune pathways as key determinants of sexually dimorphic CNS inflammation, and integrate recent mechanistic studies as examples of this conceptual framework. We conclude by identifying unmet needs in experimental models, multi‐omic approaches, and therapeutic strategies aimed at targeting metabolic–immune circuits in a sex‐informed manner.

2. Sex Differences in Neuroinflammatory Diseases

Significant sex differences are extensively described across neurocognitive and neurodegenerative diseases in humans (summarized in Table 1). Females exhibit a substantially higher risk of autoimmune CNS disorders, most notably multiple sclerosis (MS). Large epidemiological studies demonstrate that MS incidence is approximately two to threefold higher in females than in males [38, 39]. Although MS is more prevalent in females, male patients tend to experience a more aggressive disease trajectory, characterized by faster disability accumulation and earlier conversion to secondary progressive MS [40]. Autoimmune encephalitis similarly displays pronounced sex bias. Anti–NMDA receptor encephalitis occurs predominantly in females, with approximately 80% of cases occurring in females and a female‐to‐male ratio of roughly 4:1 [40, 41]. While this disparity is partly attributable to ovarian teratomas, which may contain neural tissue expressing NMDA receptor subunits, it is also observed in teratoma‐negative females, indicating additional sex‐specific biological contributors beyond tumor‐associated immunity.

TABLE 1.

Summary of sex differences in human neurological diseases.

Disease	Sex bias in prevalence	Sex bias in presentation	Key publications and reviews
Alzheimer's disease (AD)	Higher prevalence in women (~2:1 F:M)	Women have greater lifetime risk; faster cognitive decline reported in several cohorts	Castro‐Aldrete L et al. PMID:37065460, Rosende‐Roca M et al. PMID:39844303, Lin KA et al. PMID:26451386
Amyotrophic lateral sclerosis (ALS)	Slight male predominance (~1.5:1 M:F)	Men present younger with more spinal‐onset disease; shorter survival in several cohorts	Fontana A et al. PMID:33630135, Manjaly ZR et al. PMID:20225930, Grassano M et al. PMID:38568048, McCombe PA, et al. PMID: 21195356
Autoimmune encephalitis (AE)	Female‐predominant overall (~4:1 F:M, but observed sex bias is disease subtype‐specific)	Anti‐NMDA receptor encephalitis has a female: male ratio of roughly 4:1. By contrast, other AE subtypes (e.g., LGI1 or CASPR2 antibody encephalitis) show male predominance (often older men)	Dalmau J et al. PMID:31326280, Shan et al. PMID: 33679765, Gold SM et al. PMID: 30361800
Multiple sclerosis (MS)	~2–3:1 female‐to‐male prevalence	Males show faster progression and worse disability accumulation	Riley Bove et al. PMID:23608496, Marrie RA et al. PMID: 33239364, Bove et al. PMID: 23608496
Parkinson's disease (PD)	Higher incidence in men (~3–4:1 M:F)	Women have later onset and slower progression; estrogen implicated as protective (evidence mixed)	Cerri S et al. PMID:31282427, Zirra A et al. PMID:36699001
Stroke	Higher mid‐life incidence in men	Women experience stroke at older ages and have poorer post‐stroke functional outcomes	Vyas MV et al. PMID:33493057, Yun SM et al. PMID:36786321

Sex bias is also evident in neurodegenerative diseases. Among the approximately 5.1 million Americans over the age of 65 living with Alzheimer's disease (AD), nearly 3.2 million are females [42]. Although increased female longevity may contribute to the imbalance, sex‐specific immune responses, microglial activation states, and interactions with APOE‐ε4 are increasingly recognized as additional biological factors [43]. In contrast, Parkinson's disease (PD) is significantly more prevalent in males, with meta‐analyses reporting a male‐to‐female incidence of approximately 1.5–2.0, with some cohorts approaching ratios of 3–4 across specific age groups [44]. In addition to higher male incidence, females with PD often exhibit slower motor progression, including reduced rates of UPDRS‐III decline and better preservation of dopaminergic imaging markers [45]. Males not only develop amyotrophic lateral sclerosis (ALS) more frequently than females, with a male‐to‐female ratio of approximately 1.3–1.6, but also show shorter survival, driven largely by more rapid respiratory decline and faster deterioration in forced vital capacity [46, 47]. Conversely, several studies report that females with ALS may experience more rapid decline in limb motor function [48]. These observations underscore that sex differences in ALS are domain‐specific rather than uniformly favoring one sex.

Acute neurological disorders also show marked sex‐specific patterns. Stroke incidence is higher in males at younger ages, particularly between 55 and 75 years, with males demonstrating a 33%–50% higher age‐adjusted risk in these strata [49, 50]. However, females, especially postmenopausal females, experience worse functional recovery, including higher 90‐day modified Rankin Scale scores, greater long‐term disability and lower rates of functional independence following ischemic stroke [51, 52]. These differences persist after adjustment for age and comorbidities, suggesting biological contributors beyond social or healthcare‐access factors. Together, epidemiological and clinical patterns indicate that sex differences in neuroinflammatory disease are robust, domain‐specific and mechanistically heterogeneous, reflecting the involvement of interactions between immune regulation, hormonal milieu, metabolic state, and genetic susceptibility. Epidemiological data thus clearly establish sex as a major determinant of disease vulnerability and outcome, motivating a deeper examination of the biological mechanisms underlying these disparities.

An important but largely underexplored question is whether males and females respond differently to disease therapies. Sex‐stratified analyses are still limited in clinical trial data, constraining definitive conclusions. In MS, males generally exhibit worse prognosis; however, available data suggest that first‐line disease‐modifying therapies such as IFNβ supplementation and glatiramer acetate show broadly comparable efficacy between sexes, although rigorous, adequately powered comparisons are still lacking [53]. Sex differences may influence treatment tolerability and optimal dosing through pharmacokinetic and metabolic effects. In PD, females experience higher rates of dyskensias after first‐line levodopa therapy, potentially reflecting differences in drug metabolism or body composition, whereas males may derive less benefit from certain interventions such as deep brain stimulation [54]. After stroke, females are less likely to receive acute thrombolysis and often demonstrate poorer rehabilitation outcomes [55, 56]. Despite these observations, sex is rarely incorporated into therapeutic decision‐making, representing a significant gap in clinical practice. Incorporating sex‐specific immune and metabolic differences into trial design and treatment strategies may improve personalization of neuroinflammatory disease management.

Despite growing recognition of sex differences in neuroinflammation, mechanistic understanding remains incomplete. In particular, how sex, metabolism, and immune signaling intersect within the CNS is unresolved. While robust sex differences in peripheral immune responses are well documented, comparatively few studies have examined sex‐specific neuroimmune interactions within the brain. As a result, key mechanistic questions remain unanswered. For example, how do female and male microglia differ in their metabolic reprogramming during neurodegeneration, and do these differences influence disease trajectory or severity? How do sex hormones interface with metabolic regulators and innate immune sensors in the aging brain? Converging hormonal, genetic, metabolic, and immune pathways are likely to act in combination rather than isolation, yet integrative studies addressing their interaction are scarce. Emerging areas such as immunometabolism and epigenetic regulation in the context of sex differences remain particularly underexplored in neuroinflammatory diseases. Most observations of sex bias derived from clinical cohorts or animal models do not identify causal mechanisms. For instance, although females develop MS more frequently than males, the drivers of this bias within the CNS remain poorly defined beyond differences in circulating immune populations.

Addressing these gaps will require coordinated approaches, including transcriptomic, epigenetic, and metabolic profiling of male and female patient‐derived CNS cells; systematic investigation of sex chromosome effects on brain‐resident immune cell development; and evaluation of sex‐specific responses to emerging immunotherapies. Elucidating how sex, immune signaling, and metabolism converge within the CNS will not only clarify disease pathogenesis but also inform the development of sex‐informed therapeutic strategies. Sex differences therefore represent a fundamental dimension of neuroinflammatory disease biology, and resolving these gaps—particularly at the level of metabolic and innate immune regulation—will be essential for advancing neuroimmunology and improving patient care.

2.1. Biological Contributors to Sex Differences in Neuroinflammation

Sex differences in neuroinflammation arise from a complex interplay of biological factors, most prominently sex hormones and sex chromosome complement. Together, these factors contribute to intrinsic immune response biases between males and females and shape neuroimmune function across the lifespan. Below, we consider each of these contributors.

2.1.1. Hormonal Influences on Neuroinflammation

Sex steroid hormones exert profound effects on immune responses within both the CNS and the periphery. Estrogens (including estradiol), progesterone, and androgens signal through nuclear and membrane‐associated receptors (e.g., GPER1) expressed on microglia, astrocytes, neurons, dendritic cells, T cells and B cells [57, 58, 59]. Activation of these receptors modulates cytokine production, activation thresholds, and transcriptional programs central to neuroimmune regulation [60, 61, 62]. Estrogen's immunoregulatory and neuroprotective effects are among the most extensively characterized. Estradiol suppresses NF‐κB signaling, reduces microglial production of pro‐inflammatory cytokines such as TNFα, IL‐1β and IL‐6, and enhances anti‐inflammatory mediators including IL‐10 [63, 64, 65]. In parallel, estrogen promotes neuronal survival by enhancing mitochondrial efficiency, upregulating brain‐derived neurotrophic factor (BDNF), and stabilizing synaptic structure [66, 67, 68].

Testosterone modulates microglial and lymphocyte function through distinct, pathway‐specific mechanisms. Although often broadly characterized as immunosuppressive, androgen signaling selectively shapes inflammatory and regulatory programs rather than uniformly suppressing immunity [69, 70]. In microglia, testosterone signaling via androgen receptors reduces TNFα and IL‐1β production, decreases MHC class II expression, and dampens phagocytic activity [71, 72]. Testosterone also suppresses Th1 and Th17 differentiation while promoting regulatory T cell expansion, leading to reduced IFNγ and IL‐17 secretion [70, 73]. In B cells, androgen signaling limits maturation and antibody production [74]. Clinically, androgen deficiency in males is associated with heightened inflammatory signatures and exaggerated T cell responses, which can be reversed by testosterone replacement therapy [74, 75]. Conversely, lower circulating testosterone levels in females are associated with stronger adaptive immune responses, consistent with the female‐biased prevalence of antibody‐ and T cell–mediated autoimmune diseases [75, 76].

Taken together, current evidence supports the general principle that females tend to mount more vigorous innate and adaptive immune responses, whereas males exhibit stronger immunoregulatory biases in peripheral immunity. However, robust human data defining sex‐biased baseline and inflammatory states within specific CNS‐resident cell types remain limited and, in some cases, in conclusive [77, 78]. Hormone‐dependent immune modulation thus represents a critical axis influencing susceptibility, disease activity, and recovery in neuroinflammatory conditions, but claims regarding sex‐biased baseline neuroinflammation require further mechanistic validation in human CNS tissue.

2.1.2. Sex Chromosomes and Genetic Factors

Beyond hormonal influences, sex chromosome complements (XX, XY and variations) exert cell‐intrinsic effects on immune and neuroimmune function. The X chromosome is particularly enriched for immune‐related genes, including TLR7, TLR8, IRAK1, BTK, CXCR3, and FOXP3. Although one X chromosome undergoes inactivation in females, ~15% of X‐linked genes escape inactivation and remain biallelically expressed [79], immune‐related genes among them. The most studied escapees are TLR7 and TLR8, encoding pattern‐recognition receptors that play central roles in antiviral and inflammatory signaling [80, 81]. Because these genes lack Y‐linked homologs, females effectively carry a higher functional gene dosage of these immune genes, resulting in enhanced basal and ligand‐induced TLR7/8 signaling and stronger type I IFN responses [81, 82].

Elevated TLR7 activity in females promotes plasmacytoid dendritic cell activation, increased IFNα/β production and enhanced B cell maturation and antibody class switching [77, 83, 84, 85, 86]. These mechanisms contribute directly to female‐biased autoimmunity, particularly systemic lupus erythematosus (SLE), where gain‐of‐function variants or duplications of TLR7 drive severe disease [87]. TLR7‐driven interferon programs are also activated in neuroinflammatory contexts and microglial antiviral responses [88], although CNS‐specific consequences remain under active investigation. Females also carry two copies of immunoregulatory genes such as FOXP3, although escape from X‐inactivation appears to be cell‐type dependent [88]. In addition, the X chromosome encodes ~118 microRNAs, compared with only four on the Y chromosome [89]. Many of these microRNAs are immune‐modulatory, regulating cytokine signaling, antigen presentation and T cell activation, thus expanding post‐transcriptional immune regulatory capacity in females.

Despite its limited gene content, the Y chromosome encodes regulators with emerging roles in epigenetic processes relevant to neurobiology and immunity. UTY (Ubiquitously Transcribed Tetratricopeptide Repeat–Containing Protein, Y‐Linked; also known as lysine demethylase 6C, KDM6C), the Y‐linked member of the KDM6 family of histone H3 lysine‐27 demethylases, exhibits markedly reduced catalytic activity relative to its X‐linked paralog KDM6A (UTX) but is thought to exert important non‐catalytic regulatory functions in chromatin organization and gene expression [90, 91]. KDM5D (JARID1D/SMCY) is a functional H3K4me3 demethylase that represses transcription at target promoters [92]. Recent in vivo studies demonstrate that disruption of the Uty locus in mouse hematopoietic cells alters chromatin accessibility and gene expression in cardiac monocytes and macrophages, skewing them toward profibrotic phenotypes and exacerbating heart failure, phenocopying effects of mosaic loss of the Y chromosome [93]. While these findings establish that Y‐linked epigenetic regulators can shape myeloid cell states in vivo, their roles in classical inflammatory polarization and CNS immunity remain incompletely defined.

The Y chromosome also influences neuroimmune function indirectly through neural pathways. The sex‐determining gene SRY is expressed in catecholaminergic regions of the brain in rodents and humans, where it regulates transcription of enzymes controlling dopamine and noradrenaline synthesis [94, 95, 96]. Because catecholamines modulate both microglial behavior and peripheral immune responses [97, 98], SRY‐dependent differences in neurotransmitter production represent a plausible, though as yet untested, mechanistic route linking Y chromosome biology to neuroimmune regulation. Evidence from sex chromosome aneuploidies further supports a role for chromosome dosage in shaping immune risk. Individuals with Klinefelter syndrome (47,XXY) exhibit increased prevalence of several female‐predominant autoimmune diseases, including SLE, rheumatoid arthritis, Sjögren's syndrome, and autoimmune thyroid disease [99]. Notably, males with Klinefelter syndrome display approximately a 14‐fold increased risk of SLE compared with 46,XY males, approaching the risk observed in 46,XX females [100]. In contrast, data on autoimmune risk in Turner syndrome (45,X) are limited, whereas individuals with 47,XXX are overrepresented among patients with SLE and Sjögren's syndrome, consistent with an X chromosome gene‐dosage effect [101]. Functionally, neutrophils and monocytes from 47,XXY individuals produce higher levels of IL‐6, IL‐8, TNFα, and IL‐1β following TLR stimulation than cells from 46,XX females and often 46,XY males, indicating heightened innate inflammatory responsiveness [102].

Together, these data support a model in which X‐linked escape genes, Y‐linked regulators such as UTY, KDM5D, and SRY, and circulating sex hormones jointly shape sex‐biased immune and neuroinflammatory phenotypes. While sex chromosomes introduce intrinsic differences in microglial activation, antiviral signaling, and cytokine production, chromosomal complement alone cannot explain the full diversity of sex‐based phenotypes observed in neurocognitive and neuroinflammatory disorders. Instead, these phenotypes emerge from the integration of chromosomal effects with hormonal, epigenetic, metabolic, and environmental influences.

3. Metabolic Orchestrators of Neuroinflammatory Circuitry

A third determinant of neuroimmune divergence has come into focus: metabolism, the biochemical engine that governs how immune cells sense danger and execute inflammatory programs. Mounting evidence indicates that metabolic pathways, ranging from mitochondrial stress responses to lipid and nucleotide remodeling, are not only sexually dimorphic but actively shape male–female differences in neuroinflammatory vulnerability and disease progression. Metabolism does not merely support immune function; it fundamentally influences immune trajectory and persistence. Inflammatory responses impose rapid energetic and biosynthetic demands, requiring coordinated metabolic adaptation to sustain cellular homeostasis.

Pattern‐recognition receptors, including TLRs, TREM2, and the cGAS–STING pathway, converge on NF‐κB and HIF‐1α signaling, thereby directly linking innate immune sensing to metabolic reprogramming [103, 104, 105, 106, 107, 108, 109, 110, 111]. Upon activation, immune cells such as macrophages, dendritic cells, and microglia undergo coordinated metabolic remodeling to support effector functions, cytokine production, and stress responses [112, 113, 114, 115, 116].

Importantly, these immunometabolic programs are strongly shaped by sex hormones and sex chromosome complement, resulting in distinct baseline metabolic and immune states in males and females [117]. Systems‐level analyses in humans consistently show that females exhibit stronger type I interferon (IFN‐I) and antiviral signatures, suggesting an IFN‐linked metabolic bias toward glycolytic and antiviral programs [118]. In contrast, several studies have observed males more frequently display inflammatory and lipid‐associated signatures in human macrophages and mouse models, consistent with greater reliance on lipid utilization and inflammasome activation [119]. These differences are particularly relevant in the brain, which comprises only ~2% of total body mass yet consumes ~20% of resting energy to maintain ion gradients, synaptic transmission, and Na⁺/K⁺‐ATPase activity [120]. Neurons depend on metabolic coupling with astrocytes, oligodendrocytes, and microglia across mammalian systems, creating a tightly integrated energetic ecosystem that intersects directly with local immune regulation [121].

Immune activation in the CNS therefore carries a high metabolic cost. Upon activation, microglia and astrocytes shift toward glycolysis‐driven energy production and mobilize lipid stores to generate inflammatory mediators, remodel membranes, and sustain cytokine synthesis across mammalian systems [122, 123]. In parallel, extracellular metabolites—including purines, lactate, and bioactive lipids—fluctuate and feed back into microglial signaling pathways through receptors such as P2X7, adenosine receptors, and TREM2 in preclinical models, with emerging relevance in human disease [124, 125, 126, 127, 128, 129]. Even modest metabolic perturbations can thus have outsized neuroimmune consequences. Disrupted purine metabolism, for example, can amplify cytokine release and drive chronic neuroinflammation in preclinical models of AD and MS, with converging evidence from human disease studies [130, 131, 132].

Against this backdrop, biological sex emerges as a critical modifier of neuroimmune metabolism. Sex hormones and chromosomal complement shape both baseline metabolic programs and immunometabolic responses to stress, injury, and infection in the brain. These differences likely contribute to the well‐established sex biases observed across neuroinflammatory disorders in humans, including MS, AD, and autoimmune encephalitis [117, 133]. In the sections below, we examine how sex influences three central axes of neuroimmune metabolism—glycolytic reprogramming, lipid sensing, and purine metabolism—each of which differentially tunes inflammatory versus reparative programs in the CNS.

3.1. Glycolytic Reprogramming

Innate immune activation is tightly coupled to metabolic reprogramming. In response to pathogens, damaged neurons, or misfolded proteins, innate immune cells, including macrophages, dendritic cells, and microglia, undergo coordinated metabolic shifts to meet the energetic and biosynthetic demands of inflammation. Engagement of PRRs such as TLRs, TREM2, and the cGAS–STING pathway activates NF‐κB and HIF‐1α signaling, leading to upregulation of glucose transporters and glycolytic enzymes across experimental systems [107, 134, 135, 136]. This promotes a transition from oxidative phosphorylation (OXPHOS) to aerobic glycolysis, a Warburg‐like state that enables rapid ATP production and provides intermediates required for cytokine synthesis, redox balance, and membrane remodeling across immune cells and experimental systems [134, 137]. Glycolytic reprogramming is now recognized as a hallmark of inflammatory microglia and has been observed in both mouse models and human studies of neurodegenerative disease, including PD [138, 139, 140], AD [141, 142, 143], and ALS [144, 145]. Notably, early glycolytic engagement precedes overt inflammatory commitment in mouse microglia, occurring before cytokine amplification and morphological transformation [146, 147, 148].

Metabolic intermediates generated during glycolysis also function as signaling nodes that shape inflammatory output. Succinate stabilizes HIF‐1α and promotes IL‐1β production [149, 150]; lactate modulates immune tolerance and chronic inflammation [151, 152]; and itaconate, derived from the TCA cycle intermediate citrate, acts as a counter‐regulatory brake by suppressing IFN‐I and NF‐κB signaling in mouse and human immune cells [153, 154]. Through these metabolites, glycolysis both fuels and fine‐tunes innate immune responses, influencing the amplitude and duration of CNS inflammation.

Sex differences further modulate glycolytic programming and its immunological consequences. Multi‐omics studies demonstrate that female human immune cells exhibit higher basal IFN‐I and antiviral signatures, consistent with a glycolytic‐leaning metabolic bias [62, 118, 155, 156, 157]. In contrast, male myeloid cells preferentially engage lipid‐catabolic and inflammasome‐associated pathways in both mouse models and human studies, reflecting a tendency toward oxidative or lipid‐driven metabolism [119, 158, 159, 160]. In the CNS, these systemic tendencies may translate into sex‐biased microglial activation thresholds in experimental models, as well as emerging support from human studies, with females more prone to IFN‐linked glycolytic reprogramming and males favoring NLRP3‐dependent, lipid‐coupled inflammation [161, 162, 163, 164]. Together, these findings position glycolytic metabolism as a central nexus linking sex, immunity, and neuroinflammatory susceptibility.

3.2. Lipid Sensing and Metabolic Inflammation

Innate immune cells in the CNS, particularly microglia, are highly sensitive to changes in the lipid milieu. This sensitivity is mediated by a broad repertoire of lipid‐sensing receptors, including TLRs, CD36, TREM2, prostaglandin receptors, phospholipid receptors, and endocannabinoid receptors in mouse models, with corroborating evidence from human studies [165, 166, 167]. Through these receptors, microglia detect lipids derived from pathogens, damaged neurons, and myelin debris. In neurodegenerative disease models such as AD, lipid‐associated signals, including cholesterol crystals, oxidized phospholipids, and amyloid–lipid complexes, drive inflammatory activation by promoting inflammasome assembly, NF‐κB–dependent cytokine production, and impaired phagocytic clearance [168, 169, 170]. Beyond acting as inflammatory triggers, lipid metabolites actively instruct immune trajectories across murine and human immune systems. Oxidized LDL and cholesterol crystals promote inflammasome activation [171, 172]; eicosanoids such as prostaglandins and leukotrienes shape cytokine production and leukocyte recruitment [173]; and specialized pro‐resolving mediators (SPMs), including resolvins and maresins, actively terminate inflammation and promote tissue repair [174]. In mouse microglia, defective lipid handling, as exemplified by impaired TREM2 signaling, induces metabolic stress, defective phagocytosis, and chronic pro‐inflammatory states that exacerbate neurodegeneration [170, 171, 172]. Collectively, lipid sensing integrates environmental danger cues with metabolic state, positioning lipid pathways as central regulators of microglial inflammatory function.

Sex differences further shape how microglia interpret and respond to lipid cues. Sex hormones and chromosomal complement modulate the activity of multiple lipid‐sensing receptors, including TLRs, CD36, TREM2, and components of the endocannabinoid system, resulting in sex‐biased inflammatory tendencies across murine and human immune systems [175, 176, 177]. In human AD, interactions between APOE‐ε4 and TREM2 produce stronger transcriptional and pathological consequences in females [178, 179], suggesting differential tuning of key lipid‐sensing checkpoints. Lipid mediator biosynthesis is also sexually dimorphic in humans: androgens enhance leukotriene pathways and modulate prostaglandin production [180, 181], implicating eicosanoid balance itself as a contributor to sex‐specific inflammatory trajectories. Together, these findings indicate that sex shapes both lipid sensing and downstream lipid‐driven immunometabolic output, influencing vulnerability to neuroinflammation and neurodegeneration.

3.3. Purine Metabolism and the ADA Axis

Alongside lipid‐mediated pathways, purine metabolism represents a second major sex‐biased axis through which microglia integrate metabolic stress with innate immune activation. denosine deaminases catalyze the deamination of adenosine and deoxyadenosine to inosine and deoxyinosine, thereby shaping intracellular and extracellular purine pools. In humans, ADA1 is broadly expressed and primarily intracellular, whereas ADA2 is secreted by myeloid cells and functions predominantly in the extracellular space [28, 182, 183]. In the CNS, microglia and perivascular macrophages are repeatedly exposed to purine‐rich microenvironments generated by neuronal injury, demyelination, and metabolic stress in human inflammatory disease contexts, positioning ADA activity as a key determinant of microglial inflammatory set point [184]. Microglia express a diverse repertoire of purinergic receptors, including P2X7, P2Y receptors, and adenosine receptors (A1–A3), that enable detection of nucleotides and nucleosides released during tissue damage across murine and human immune systems [185, 186, 187]. ATP–P2X7 signaling promotes NLRP3 inflammasome activation and IL‐1β release, whereas adenosine receptor signaling can suppress or sustain inflammation depending on receptor subtype and ligand concentration, defined in murine immune cells and conserved in humans [188, 189, 190]. ADA1 and ADA2 sculpt this extracellular landscape: ADA1 maintains intracellular purine homeostasis, while ADA2 clears extracellular adenosine and deoxyadenosine [182, 183]. Reduced ADA2 activity biases signaling toward persistent purine exposure, favoring inflammasome activation, heightened IFN‐I output, and chronic neuroinflammation [28, 31].

In human cells, loss of ADA2 profoundly rewires purine metabolism and engages antiviral‐like innate immune programs [28]. Accumulated extracellular deoxyadenosine is alternatively deaminated by ADA1, increasing deoxyinosine levels, which competitively inhibit methionine adenosyltransferase (MAT) and reduce S‐adenosylmethionine (SAM) [28]. SAM depletion drives global hypomethylation and derepression of human endogenous retroviruses (HERVs), whose transcripts and nucleic‐acid intermediates are sensed by RIG‐I/MDA5 and the cGAS–STING pathway, culminating in robust IFNβ induction [28]. In the CNS, repeated fluctuations in ATP and nucleoside levels during injury and degeneration may predispose microglia with impaired ADA2 activity to HERV derepression and IFN‐I activation in murine models [185, 191]. Consistent with this model, HERV transcripts and proteins are elevated in human AD and MS [192, 193, 194], HERV‐W envelope protein contributes to inflammation in human MS [195], and IFN‐I–responsive microglial states, defined in experimental models and supported by human studies, are increasingly implicated in synapse loss and neurodegeneration in AD and ALS [196, 197].

Emerging evidence further indicates that purine metabolism and ADA activity are sexually dimorphic. Systems‐immunology analyses show that females more frequently exhibit lower ADA/ADA2 activity accompanied by higher basal HERV expression, elevated interferon‐stimulated gene (ISG) signatures, and enhanced IFN‐I programs in both healthy individuals and patients with COVID‐19 [194]. These findings align with the ADA2–HERV–IFNβ circuit described above and suggest that modest reductions in ADA activity may preferentially prime antiviral‐like programs in females. More broadly in humans, females display greater sensitivity to systemic metabolic stress—including glycemic, vascular, and hepatic perturbations—with distinct sex‐specific metabolite signatures across metabolic states [159, 198]. Reviews of sex‐biased immunometabolism similarly indicate that female immune systems more readily couple metabolic fluctuations, such as altered lipid or nucleotide flux, to stronger inflammatory outputs [78, 199].

These sex‐biased immunometabolic programs parallel disease‐specific patterns across neurodegenerative disorders. In AD, which disproportionately affects females, female microglia in both APP/PS1 mice and human tissue exhibit a pronounced shift toward aerobic glycolysis (elevated ECAR, PFKFB3 expression, and lactate production), accumulation of pro‐inflammatory succinate, impaired phagocytosis, and increased amyloid burden [160]. By contrast, male microglia more frequently maintain oxidative metabolism and phagocytic function [160, 200]. In human MS, more prevalent in females, glycolytic and IFN‐skewed immune programs may underlie heightened autoimmunity, whereas males more often develop chronic‐active lesions with iron‐laden microglia consistent with sustained inflammasome activation [201, 202]. Conversely, human PD shows male predominance, where estrogen's neuroprotective and mitochondria‐stabilizing effects may confer female resilience, while male‐biased lipid and inflammasome pathways amplify disease risk [19, 203, 204]. In ALS, recent multi‐omic profiling reveals greater disruption of immune, extracellular matrix, and mitochondrial pathways in male patients compared with females [205]. Together, these findings support a model in which sex‐specific metabolic programs—from PFKFB3‐driven glycolysis in females to androgen‐associated lipid signaling in males—shape divergent neuroinflammatory trajectories across AD, PD, MS, and ALS.

4. Metabolic Control of Innate Immune Signaling in Neuroinflammation

Metabolism and innate immune signaling have traditionally been framed as parallel but independent regulatory layers. Recent studies support an integrated model in which metabolic enzymes, nucleotide flux, and post‐translational signaling checkpoints directly shape inflammatory identity across human immune cells. These discoveries—spanning nucleotide‐controlled T‐cell cytokine fate [183], a kinase‐dependent licensing switch for cyclic cGAS–STING activation [203], and a purine‐driven endogenous retrovirus–type I interferon circuit with sex‐biased features [28, 32], collectively position metabolism as a first‐order determinant of immune response amplitude and persistence. Importantly, these mechanisms intersect with central neuroimmune pathways, including meningeal and brain‐border T‐cell activity, microglial DNA sensing, nucleoside‐reactive interferon programs, and sex‐biased inflammatory set points that influence neurodegenerative risk. In this section, we synthesize these mechanistic nodes and outline how they form an integrated, and potentially targetable, framework for understanding maladaptive CNS inflammation.

4.1. ISOC1 Couples Nucleotide Metabolism to Inflammatory Fate in Human Memory T Cells

Kushnareva et al. [204] investigated how gene programs shape effector function in human Th1‐like memory CD4⁺ T cells, a subset enriched in IFN‐experienced individuals. Using CRISPR/RNAi perturbations integrated with transcriptomics and targeted metabolomics, the study identified ISOC1 (Isochorismatase Domain–Containing Protein 1) as a previously underappreciated metabolic gatekeeper. ISOC1 knockdown blunted IFNγ and IL‐17 production, and metabolomic profiling revealed broad defects in pyrimidine and purine metabolism, including depleted nucleotide pools and altered redox balance. Mechanistically, ISOC1 deficiency constrained glycolytic and mitochondrial support for nucleotide biosynthesis, functionally limiting cytokine output upon recall stimulation. These findings position ISOC1 at the intersection of nucleotide anabolism and pro‐inflammatory effector programming, emphasizing that immune fate depends not only on canonical transcription factors but also on metabolic enzymes that license nucleotide availability. Beyond defining a single gene, this work established a generalizable pipeline integrating functional genomics and metabolomics to connect metabolic control points to human T‐cell function.

Insights from ISOC1‐dependent metabolic control in memory T cells offer a lens through which to interpret how nucleotide‐stress pathways might shape neuroinflammatory circuitry. Increasing evidence supports the relevance of metabolic gatekeepers identified in peripheral immune cells for understanding CNS inflammation [206]. In particular, ISOC1‐dependent control of IFNγ and IL‐17 production has implications for meningeal and blood–brain barrier (BBB) compartments that influence microglial state and synaptic integrity. Conditions that impose nucleotide stress at CNS interfaces, including aging, vascular dysfunction, chronic infection, hypoxia, mitochondrial injury, or BBB leakage, could plausibly skew T‐cell cytokine programs. Heightened IFNγ and IL‐17 in these settings may reinforce microglial transitions into disease‐associated or interferon‐responsive states described in neuroinflammatory disease and aging, thereby potentiating microgliosis and aberrant synaptic remodeling.

More broadly, the central principle emerging from this study, that nucleotide metabolism licenses inflammatory fate, aligns with core features of microglial biology. Microglia integrate extracellular ATP and adenosine to regulate chemotaxis, phagocytosis, and inflammasome activation [207]. Thus, limited nucleotide availability, diminished biosynthetic capacity, and redox stress may act as conserved metabolic signals that shape inflammatory behavior across immune cell lineages.

Disease contexts further illustrate translational relevance. In the APP/PS1 mouse model of AD, IFNγ–driven immune signaling promotes microglial activation during early stages of pathology, and Th1‐like cytokines exacerbate synaptic dysfunction [208], features that could be intensified by ISOC1‐tuned cytokine output at the BBB. In MS patients, IL‐17 and IFNγ are key drivers of lesion pathogenesis [209], raising the possibility that nucleotide‐regulated effector programs contribute to inflammatory fluctuations during relapse. In PD and ALS, pervasive mitochondrial dysfunction and metabolic stress [210] may similarly constrain nucleotide availability, potentially biasing both T‐cell and glial responses toward chronic, low‐grade inflammation.

The methodological blueprint established in the ISOC1 study, integrating CRISPR perturbations with metabolomics, is readily applicable to human iPSC‐derived microglia to test whether ISOC1 or adjacent metabolic enzymes tune IFN‐I programs, phagocytic capacity, synapse engulfment, and interactions with purinergic and cGAS–STING pathways. From a translational perspective, ISOC1 highlights nucleotide metabolism as a potentially modifiable regulator of neuroinflammation that could be targeted without broad immunosuppression. In neurodegenerative cohorts, boosting nucleotide synthesis through salvage‐pathway precursors or one‐carbon donors could, in principle, normalize pathologic IFNγ/IL‐17 output at CNS borders and reset microglial states. Conversely, in cytokine‐rich environments such as active MS lesions, fine‐tuning nucleotide availability or reactions proximal to ISOC1 may constrain excessive Th1/Th17 effector output while preserving antiviral defenses. Nucleotide–redox signatures in CSF or blood may provide practical pharmacodynamic biomarkers of pathway engagement [211]. In combination with purinergic interventions (e.g., P2X7 or A2A modulation) linked to microglial reactivity, ISOC1‐informed strategies could enable multi‐node regimens that target both intracellular nucleotide constraints and extracellular nucleotide signaling.

Direct evidence for sex‐based differences in ISOC1 expression is currently limited, but several lines of evidence suggest potential modulation. Rodent studies link ISOC1 to testis (and eye) development [212, 213], consistent with possible hormonal regulation. More broadly, sex hormones and sex chromosomes shape immune gene networks, and disease risk and inflammatory programs diverge in males and females. For example, IFNγ variants affect MS risk in patients in a sex‐dependent manner [214], whereas in AD, which is more prevalent in females, female microglia in APP/PS1 mice adopt a glycolytic, pro‐inflammatory phenotype [160]. If ISOC1 limits nucleotide supply for cytokine synthesis, these contexts could plausibly modulate inflammation in a sex‐dependent manner. Accordingly, ISOC1‐targeted or nucleotide‐boosting strategies may show sex‐differential efficacy, motivating direct testing in male versus female immune cells and models.

4.2. DAPK3 As a cGAS–STING Licensing Kinase: Mechanistic Insights and Neuroimmune Implications

Death‐Associated Protein Kinase 3 (DAPK3) has been identified as an essential upstream kinase for cGAS–STING activation that supports antitumor innate immunity [209]. Using a 1001‐gene loss‐of‐function screen in human umbilical vein endothelial cells (HUVEC), DAPK3 emerged as a positive regulator of the cGAS–STING axis. Mechanistically, DAPK3 maintains basal STING stability by inhibiting K48‐linked polyubiquitination, thereby limiting proteasomal degradation. Upon DNA sensing (e.g., 2′,3′‐cGAMP stimulation), DAPK3 promotes STING K63‐linked polyubiquitination and assembly of the STING–TBK1 signalosome. Phosphoproteomics identified LMO7 as a DAPK3 substrate; phosphorylation of LMO7 was required for LMO7–STING interaction and the K63‐linked ubiquitin editing that licenses downstream IFNβ transcription. In murine syngeneic tumor models, DAPK3 loss suppressed interferon‐stimulated gene programs, impaired recruitment of CD103⁺ dendritic cells and cytotoxic T cells, accelerated tumor growth, and blunted responses to chemo‐immunotherapy. Collectively, these data position DAPK3 as a dual‐mode STING rheostat: preserving STING at baseline while catalyzing activating ubiquitin‐state transitions during innate sensing.

Increasing evidence implicates cGAS–STING signaling in CNS aging and neurodegeneration, including AD [215]. In aging and AD human samples and mouse models, STING activation in microglia has been linked to reactive transcriptional states, synaptic loss, neurodegeneration, and cognitive decline [216], whereas genetic or pharmacologic disruption of cGAS/STING in AppNL‐G‐F/hTau double‐knock‐in mice ameliorates pathology and improves behavior [217]. Human AD tissue has been reported to show elevated STING immunoreactivity in proximity to plaques and microvasculature, consistent with sterile DNA stress (e.g., mitochondrial DNA leakage, micronuclei, or endogenous retroelements) fueling innate activation [215]. Within this framework, the DAPK3 study adds a mechanistic lever: a kinase‐dependent licensing step that controls STING stability and activation through post‐translational ubiquitin editing (K48 versus K63) via LMO7.

In the brain, where chronic, low‐grade DNA danger signals may persist, a DAPK3‐dependent licensing step could set the threshold for microglial STING activation. Higher DAPK3 activity could facilitate STING–TBK1 assembly and sustain IFN‐I programs, whereas reduced DAPK3 activity could destabilize STING and attenuate signaling. This bidirectional control aligns with evidence that tonic IFN‐I signaling in murine microglia worsens AD‐like phenotypes and that IFNAR blockade can rescue synaptic and cognitive measures in models [218]. Beyond AD, both ALS patient‐derived cells and a mouse model of C9orf72 repeat‐associated ALS also feature DNA damage and mitochondrial stress, conditions favorable for cGAS–STING engagement [219], suggesting that a DAPK3‐tuned STING checkpoint could generalize across neurodegenerative contexts. In parallel, work implicating purine metabolism as an upstream driver of IFN‐I (the ADA2 axis) in humans positions DAPK3 as a convergent checkpoint on the STING arm of the same interferon‐centered circuitry [216].

These observations motivate the hypothesis that DAPK3 and its substrate circuit (LMO7‐dependent ubiquitin editing) could represent therapeutic entry points in contexts where cGAS–STING–IFN‐I signaling is pathogenic. If excessive STING activity contributes to AD/ALS/aging phenotypes, DAPK3 inhibition could, at least in principle, reduce K63‐linked licensing and dampen IFN‐I programs, potentially offering finer control than direct STING antagonism. Conversely, augmenting DAPK3‐mediated stabilization might be relevant in settings requiring enhanced antiviral responses, though such approaches are less likely to be desirable in chronic neurodegeneration. The defined phosphorylation site on LMO7 and the K48‐to‐K63 ubiquitin switch suggest tractable intervention points (protein–protein interactions or E3‐ligase pathways) that could modulate STING without globally suppressing innate immunity. Potential pharmacodynamic markers include phospho‐LMO7, the ratio of K63‐ubiquitylated to total STING, and DAPK3 kinase activity in peripheral myeloid cells or iPSC‐derived microglia.

Sex‐stratified analyses of DAPK3 in human neurodegenerative disease remain limited; however, several lines of evidence suggest that DAPK3 signaling is hormonally and transcriptionally modulated in a sex‐dependent manner in rodent and human sytems [220, 221, 222]. DAPK3 has been described as androgen responsive and can enhance androgen receptor–mediated transcription as a co‐activator within AR signaling complexes [222]. Consistent with androgenic regulation, DAPK3 expression was higher in males in a transcriptomic survey of mitochondrial and apoptotic genes in the hearts of F344 rats, clustering within a male‐biased stress response program [223]. Although peripheral, these data provide precedent for sex‐linked regulation of DAPK3 expression and motivate investigation in CNS‐resident immune compartments.

Integrating sex‐biased epidemiology and immune programs with DAPK3–STING biology suggests testable predictions. For example, if DAPK3 expression or activity differs by sex in CNS myeloid cells, DAPK3 inhibition could yield sex‐differential effects on STING‐dependent IFN‐I outputs, potentially producing steeper IFN‐I normalization in IFN‐high contexts. Such hypotheses require direct testing in human brain tissue and in sex‐stratified iPSC‐microglia models. Biomarker strategies for DAPK3–STING modulation should therefore be sex stratified, incorporating phospho‐LMO7, STING ubiquitin‐state measures, and interferon signatures as pharmacodynamic readouts.

Finally, accumulating evidence indicates that the pathological consequences of STING signaling in the CNS are shaped by cellular context, mitochondrial stress, and interferon sensitivity, not kinase control alone [215]. Several of these contextual determinants differ by biological sex. AD exhibits female‐biased interferon‐responsive microglial states, whereas PD and ALS show greater male vulnerability to mitochondrial and oxidative stress [215, 224, 225, 226, 227]. Notably, APOE ε4 and female sex have been reported to amplify tau‐induced microglial cGAS–STING–IFN‐I responses in APOE4‐R47H human P301S mutant tau mice, providing direct evidence that STING pathway output is sex modulated in neurodegeneration [228]. In this framework, DAPK3‐dependent stabilization and licensing of STING may represent a necessary but insufficient step for IFN‐I induction, with downstream amplitude further tuned by sex‐biased immunometabolic programs.

4.3. The ADA Axis Links Purine Nucleosides to IFN‐I and Sex Bias

A novel purine‐to‐interferon pathway was recently described in which extracellular deoxyadenosine and deoxyinosine accumulate when ADA2 activity is reduced in HUVEC [28]. These nucleosides are taken up and metabolized in a manner that inhibits methionine adenosyltransferase, lowering S‐adenosylmethionine (SAM), reducing DNA methylation, and derepressing human endogenous retroviruses (HERVs). HERV‐derived transcripts engage RNA‐sensing pathways (RIG‐I/MDA5–MAVS), inducing IFNβ and interferon‐stimulated genes. Genetic or pharmacologic reduction of ADA2 activity, as well as nucleoside supplementation, was sufficient to trigger IFN‐I programs, whereas restoring ADA activity or disrupting nucleic‐acid sensing curtailed responses [28]. This work established ADA2 as a metabolic brake on an innate antiviral circuit rooted in purine catabolism, epigenetic regulation, and HERV sensing.

Subsequent human cohort analyses extended this framework to sex differences [227]. Across cohorts, females exhibited lower ADA activity and higher HERV/ISG signatures, consistent with stronger basal IFN‐I priming. Together, these studies position ADA1/ADA2 as upstream tuners of tissue interferon set points and provide a biochemical rationale for sex‐biased innate immune activation across tissues, with direct relevance to neuroinflammation.

CNS‐resident glia are highly responsive to purine cues [185]. During neuronal stress, extracellular ATP activates P2X7 and promotes NLRP3 inflammasome signaling, whereas ectonucleotidases CD39 and CD73 regulate adenosine levels and, in turn, ADA1 and ADA2 metabolize nucleosides [229, 230]. These findings suggest that when ADA2 is insufficient, extracellular deoxyadenosine/deoxyinosine accumulation may prime IFN‐I via HERV derepression, potentially shifting microglia toward interferon‐responsive states observed in aging and AD. In MS, tonic IFN‐I can be beneficial or harmful depending on context; ADA–nucleoside signaling at CNS borders could contribute to this bidirectionality by modulating baseline interferon tone. In PD and ALS, mitochondrial stress and nucleic acid instability could synergize with ADA‐low states to sustain aberrant IFN programs that impair neuronal survival and microglial function. Critically, the observed sex bias that is, lower ADA activity and higher IFN priming in females, offers a plausible mechanistic contributor to female‐skewed AD and MS risk and to sex‐differential therapeutic responses.

From a translational perspective, three complementary avenues emerge. First, restoring the ADA2 brake on nucleoside accumulation could, in principle, lower deoxyadenosine/deoxyinosine levels and attenuate IFN‐I signatures. Candidate approaches include recombinant ADA2, gene therapy strategies informed by DADA2, or small‐molecule ADA‐enhancing strategies, initially tested peripherally with evaluation of central effects through interferon‐state profiling and clinical outcomes. Given consistently lower ADA activity in females, sex‐aware dosing and stratification would likely be required. Second, if direct ADA restoration proves impractical, intervention along the HERV–sensor axis remains an alternative, including modulation of RNA‐sensing pathways (RIG‐I/MDA5), downstream IFNAR/JAK–STAT signaling, or HERV biogenesis, ideally with CNS‐tuned delivery to avoid broad immunosuppression. These strategies could be complemented by purinergic modulators (P2X7 or A2A) that shape the ATP/adenosine signals co‐regulating microglial activation. Third, stratification frameworks incorporating ADA activity, deoxyadenosine/deoxyinosine levels, HERV transcripts, and ISG signatures, contextualized by sex and hormonal status, may guide eligibility and pharmacodynamic monitoring in early‐phase efforts targeting ADA restoration or HERV sensing pathways, with the expectation that ADA‐low/IFN‐high states (more common in females) may represent particularly responsive subgroups.

5. Immunometabolic Drivers of Neuroinflammation

The studies synthesized above highlight a recurrent theme: innate immune activation in neuroinflammatory disease is strongly shaped by underlying metabolic programs and is modulated in sex‐specific ways. Collectively, these findings underscore the value of examining upstream metabolic cues that bias immune cell fate and inflammatory trajectories. In keeping with this broader immunometabolic landscape, we next focus on bioactive lipids, particularly lysophosphatidylcholine (LPC) species, as regulators of neutrophil behavior and systemic inflammation. Recent work from our laboratory indicates that LPC dysregulation emerges in human post‐viral inflammatory syndromes, including neurocognitive long COVID (unpublished), and during autoimmune‐like immune‐related adverse events (irAEs) following immune checkpoint blockade (ICB) therapy [231]. In these settings, declines in LPC 18:2 correlate with heightened neutrophil activation and persistent inflammatory tone. Preliminary data in AD models further suggest that circulating LPC 18:2 declines in aged AD mice (unpublished) coincident with rising peripheral neutrophil counts, consistent with a lipid–immune imbalance that could amplify early neuroinflammation [232]. In addition, intraperitoneal LPC administration in 5xFAD and 3xTg‐AD mice has been reported to reduce Aβ plaque burden and attenuate gliosis in 5xFAD mice [233]. Together, these observations across distinct disease contexts support the idea that LPC species may function as broader immunometabolic indicators of inflammatory load, with emerging evidence linking circulating LPC levels to peripheral immune activation and early features of AD in human cohorts [234, 235]. Although direct causal links between circulating LPC changes and microglial state transitions remain to be established, these associations raise the possibility that lipid–immune signatures intersect with early microglial remodeling.

LPCs are single–acyl‐chain phospholipids generated by partial hydrolysis of phosphatidylcholine, typically via phospholipase A₂ activity [236]. LPC 18:2 contains a linoleic acid (18:2) chain and is a prominent circulating LPC species in both humans and mice [237, 238, 239, 240, 241]. LPCs are enriched in oxidized lipoproteins and inflamed tissues, where they act as signaling molecules and exert context‐dependent immunomodulatory functions [236, 242, 243]. LPCs can function as chemoattractants for murine T cells, human monocytes, and human neutrophils [244] and can enhance endothelial expression of adhesion molecules such as ICAM‐1 and VCAM‐1 in human cells, potentially facilitating leukocyte recruitment [244, 245]. In monocyte‐derived dendritic cells, LPCs promote maturation, upregulate CD83/CD86, and enhance antigen presentation, functioning as endogenous adjuvant‐like signals that support CD4⁺ and CD8⁺ T‐cell priming [246, 247] LPCs can also activate NLRP3 inflammasomes in human macrophages in certain contexts [248], yet in other settings can promote pro‐resolving macrophage programs as demonstrated in an in vivo mouse model system [249]. Thus, LPC species function as lipid cues that can either amplify or restrain inflammation depending on tissue context, acyl composition, and immune activation state.

5.1. LPC 18:2 as a Systemic Regulator of Neutrophil Persistence During ICB‐irAEs

LPC 18:2 has been identified as an endogenous suppressor of severe immune‐related adverse events (irAEs) in patients receiving immune checkpoint blockade therapy (ICB) [231]. irAEs are inflammatory toxicities that arise from loss of self‐tolerance when checkpoint inhibition reinvigorates antitumor T cells, enabling collateral immune‐mediated damage to healthy tissues, often in the setting of inflammatory cytokines such as IL‐6 and TNFα [250]. irAEs can affect nearly any organ system; common manifestations include dermatitis/pruritus, diarrhea/colitis, and endocrinopathies [251]. Across trials and real‐world cohorts, a majority of patients experience any‐grade irAEs, while a substantial subset develop severe grade III/IV toxicities [252]. Although fatal irAEs are uncommon overall, specific syndromes such as ICB‐associated myocarditis can carry high mortality [253].

Longitudinal lipidomic analyses across multiple ICB‐treated patient cohorts revealed that individuals who later developed grade III/IV irAEs exhibited a marked post‐treatment decline in plasma LPC 18:2 [231]. Across both ICB cohorts and orthogonal control datasets, lower LPC 18:2 correlated strongly with higher neutrophil counts, and in irAE cases LPC 18:2 inversely tracked with irAE‐associated neutrophilia [231]. Mechanistically, IL‐6 suppressed LPC 18:2 biosynthesis during inflammatory activation [231]. Importantly, in mouse models of CTLA‐4 blockade and colitis, LPC 18:2 supplementation reduced neutrophil expansion and colonic inflammation without impairing tumor rejection [231].

Ongoing work further suggests a cellular mechanism by which LPC 18:2 may restrain tissue‐damaging inflammation. In preliminary studies, LPC 18:2 promoted caspase‐3/7–dependent human neutrophil apoptosis with minimal induction of NETosis, consistent with an immunologically “silent” resolution program (unpublished) [254]. This provides a plausible mechanistic rationale for the observed reduction in neutrophil‐driven colonic inflammation in ICB‐induced colitis models and supports the broader view that LPC 18:2 can act as a metabolic brake on neutrophil persistence. Collectively, these findings position LPC 18:2 as a context‐dependent regulator of innate inflammatory load during immune checkpoint therapy.

Beyond cancer immunotherapy, converging evidence from the broader LPC literature suggests that LPC biology represents a unifying metabolic–immune axis relevant to chronic inflammation, post‐viral syndromes, and neurodegenerative disease. Circulating LPCs decline across diverse inflammatory states in humans, including sepsis, cardiovascular disease, inflammatory bowel disease, atherosclerosis, and acute SARS‐CoV‐2 infection [239, 240, 255, 256, 257, 258, 259], and lower systemic LPC levels frequently associate with heightened innate immune activation and adverse outcomes. Both long COVID and AD are characterized by persistent innate immune activation, altered systemic lipid metabolism, endothelial dysfunction, and impaired resolution programs [260, 261, 262, 263, 264], contexts in which dysregulated LPC‐mediated pathways could plausibly amplify neurocognitive symptoms. In long COVID, sustained neutrophil activity, mitochondrial stress, and endothelial injury provide a biological setting where LPC‐driven apoptotic programming may be particularly relevant [265, 266, 267]. In AD, early declines in circulating LPCs have been reported alongside exaggerated neuroinflammatory signatures [235, 268], raising the possibility that LPC deficiency may lower the threshold for both peripheral and CNS immune activation.

Finally, LPC biology intersects directly with CNS lipid homeostasis. Docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) are long‐chain omega‐3 fatty acids with well‐established cardioprotective and neuroprotective functions [239, 269, 270, 271]. Notably, LPC is a principal esterified carrier that transports DHA and EPA across the blood–brain barrier via MFSD2A [239, 272, 273, 274]. Multiple studies show that, in mice, LPC is required to enrich DHA/EPA in the brain, whereas supplementation with unesterified (free) DHA/EPA does not reliably increase brain levels [273, 274]. Consistent with this, LPC‐mediated DHA enrichment in mice improves spatial learning and memory in experimental systems, whereas free DHA has limited effect in the absence of LPC transport [274]. Because DHA and EPA are largely diet derived and not synthesized in the brain at meaningful levels in humans [275], LPC availability may be a key determinant of intracerebral omega‐3 pools. Supporting this model, mutations in MFSD2A/Mfsd2a impair brain uptake of LPC‐bound lipids and cause neurodevelopmental and cognitive defects in humans and mice [276, 277, 278, 279]. Together, these findings position LPC as a lipid mediator linking peripheral lipid metabolism to CNS resilience and neuroinflammatory set points [275, 280].

5.2. A Lipid–Immune Disconnect in Neurocognitive Long COVID

Following acute SARS‐CoV‐2 infection, a substantial fraction of survivors experience at least one unresolved symptom lasting ≥ 3 months, meeting criteria for long COVID [281, 282]. Long COVID is highly heterogeneous and can involve more than 200 symptoms spanning multiple organ systems [283]. The biological mechanisms remain incompletely understood, and no FDA‐approved therapeutics exist. Given this heterogeneity, focusing on individuals with a defined symptom cluster may help clarify mechanistic drivers.

To address this, a longitudinal study of healthcare workers in Los Angeles was initiated during the first wave of the COVID‐19 pandemic. From this cohort, 135 age‐ and sex‐matched participants were selected and grouped into: (i) never infected, (ii) COVID‐19 recovered, or (iii) experiencing one or more neurocognitive symptoms of long COVID. Serological analysis showed that both the neurocognitive long COVID and COVID‐recovered patients showed higher circulating spike IgG than never‐infected controls, consistent with prior exposure to SAR‐CoV‐2 (unpublished data). Notably, the male neurocognitive long COVID participants exhibited significantly higher spike IgG titers compared to recovered individuals, suggesting sex divergence in post‐infection immune regulation in settings of long COVID disease.

Untargeted lipidomics revealed a distinct lipid signature in neurocognitive long COVID participants compared to COVID‐recovered participants, even after multivariate modeling accounting for age and sex. Relative to both never‐infected and COVID‐recovered groups, neurocognitive long COVID was characterized by lower circulating levels of inflammation‐resolving lipid mediators, including specialized pro‐resolving mediators (SPMs) such as maresin 1 (MaR1) and 5‐HEPE (EPA‐derived). These SPMs are central to active resolution programs, including in neuroinflammatory contexts [284, 285]. Neurocognitive long COVID participants also showed reduced circulating LPC 18:2 relative to both comparator groups.

Together, these data suggest that neurocognitive long COVID may involve impaired lipid‐mediated resolution capacity rather than broadly elevated cytokinemia. Consistent with this, we did not observe uniform increases in canonical pro‐inflammatory cytokines in our neurocognitive long COVID subgroup. Differences across long COVID studies likely reflect cohort heterogeneity and sampling context [283, 286, 287]. Our cohort was drawn from the first pandemic wave (vaccine‐naïve, infection‐naïve), and participants were selected for neurocognitive symptoms, factors that may shape immune and cytokine profiles compared with cohorts enriched for recurrent infection or vaccination. These findings support the idea that cytokine measures alone may not reliably distinguish neurocognitive long COVID from recovery and instead highlight lipid mediators—including LPC 18:2 and SPMs—as candidate mechanistic drivers and therapeutic targets in this symptom‐defined subgroup [283, 288]. As many of these lipids are influenced by diet and systemic metabolism, a testable implication is whether targeted nutritional or pharmacologic strategies that restore resolution mediators can improve neurocognitive outcomes.

5.3. Metabolites, Microglia, and Sex: Early Divergence Points in Alzheimer's Disease

LPC 18:2 has gained attention not only for immunoregulatory properties but also for inverse associations with cognitive decline and dementia risk in humans [239, 289, 290]. Epidemiological studies report negative associations between LPC 18:2 and subclinical atherosclerosis [290], a vascular phenotype linked to dementia risk [291], positioning LPC 18:2 as a candidate metabolic indicator of neurocognitive vulnerability. Complementing these findings, APOE‐ε4 carriers exhibit broad disruptions in glycerophospholipid and ether‐lipid metabolism that track with amyloid burden, tau pathology, and cognitive decline [292, 293, 294], reinforcing the concept that lipid dysregulation intersects with core pathways in AD progression. In line with this, our preliminary data indicate that circulating LPC 18:2 declines in aged AD mice (unpublished) coincident with rising peripheral neutrophil counts, consistent with a lipid–immune imbalance that may amplify early neuroinflammation and synaptic dysfunction in vivo [232].

A key recent methodological advance has been the development of high‐resolution morphometric tools to quantify murine microglial phenotypes in situ. Using machine‐learning–assisted segmentation of Iba1⁺ microglia with masks capturing process architecture, these approaches distinguish reactive versus homeostatic morphologies with increased sensitivity [295]. Applied to 5xFAD mice, this analysis revealed evidence of early, sex‐biased microglial remodeling before overt plaque deposition. At very early ages (∼1.5–2 months), quantitative morphometry showed that female 5xFAD microglia had already begun to adopt reactive, amoeboid features (retracted/thickened processes and enlarged soma) compared with age‐matched males, particularly in hippocampal regions, despite minimal classical gliosis. By contrast, male 5xFAD microglia remained more hyper‐ramified at the same stage, and wild‐type mice of either sex did not show these shifts [295]. These findings support the possibility that female microglia enter a “primed‐to‐reactive” transition earlier, potentially contributing to sex‐biased trajectories of amyloid pathology and cognitive decline. Notably, the 5xFAD transgene is driven by a Thy1 promoter containing an estrogen response element, which may amplify Aβ expression in females [296] and should be explicitly considered when interpreting sex effects in this model.

Despite growing evidence that immune activation contributes to AD from early stages, reliable biomarkers of early neuroimmune remodeling remain limited. Canonical AD biomarkers (Aβ and tau) reflect proteinopathy but do not directly report neuroimmune state and can precede symptoms by years without indicating inflammatory activity [297, 298]. TSPO PET imaging has enabled in vivo glial monitoring but has recognized limitations, including low signal‐to‐noise for first‐generation ligands (e.g., [¹¹C]PK11195), genetic binding variability for newer tracers (e.g., TSPO Ala147Thr), and limited functional specificity because TSPO is expressed in multiple glial populations and does not discriminate among microglial states [299, 300]. Moreover, some disease‐associated microglial programs may not upregulate TSPO, raising the possibility that TSPO‐based readouts miss key early activation states [300]. Peripheral inflammatory markers show inconsistent predictive value, and CSF measures (e.g., soluble TREM2) require invasive sampling. Together, these limitations highlight an unmet need for minimally invasive biomarkers that sensitively track early microglial remodeling in prodromal AD.

Looking ahead, a critical next step is to test whether quantitative features of microglial remodeling (e.g., branching complexity, soma hypertrophy, CD68‐associated phagolysosomal activity) correlate with distinct plasma metabolomic profiles, particularly bioactive lipid species such as LPC 18:2. Establishing these relationships will help resolve whether peripheral lipid changes merely track, or potentially precede, early neuroimmune transitions. Future studies should also evaluate whether composite lipid–immune signatures, combining lipid mediators with neutrophil phenotypes and other innate immune readouts, improve prediction of emerging microglial activation relative to any single measure. Importantly, the sex‐biased microglial patterns observed in 5xFAD mice raise the possibility that biomarker thresholds and trajectories differ between males and females, motivating explicit sex‐stratified testing in both preclinical models and longitudinal human cohorts.

Finally, dissecting mechanistic links among LPC biology, neutrophil persistence, and microglial remodeling remains an important frontier. It remains unresolved whether LPC 18:2 loss facilitates neutrophil‐mediated inflammatory signaling that secondarily shapes microglial behavior, or whether both reflect broader systemic metabolic dysregulation. Addressing these questions will require integrated lipidomic, immunophenotypic, and spatial microglial analyses across time, ideally in models that capture human‐relevant risk factors such as APOE‐ε4. Together, these lines of work support a broader paradigm in which neurodegeneration is approached not solely through protein aggregation, but through immunometabolic resolution, cellular communication, and sex‐specific neuroimmune resilience.

6. Sex Matters in Neuroinflammation: Translational Barriers and Pathways Forward

A review of the biomedical literature spanning the last four decades reported that only ~12% of human studies discussed sex as a biological variable [301]. Historically, both neuroscience and immunology often under‐emphasized sex in study design and analysis [302, 303]. Male animals became the default experimental model under assumptions of “convenience and control,” while female hormonal cycling was frequently treated as an unwieldy source of variability. Early observations of estrous‐cycle–linked behavioral variation contributed to the perception, often incorrectly generalized, that females were inherently more variable [304]. Practical and institutional factors reinforced this bias: female rodents were commonly reserved for breeding and were more costly, leading many laboratories to rely on surplus males [302]. Regulatory history also played a role. While policies from the 1990s onward required inclusion of females in US human clinical trials [305], comparable mandates for sex‐balanced design in preclinical animal research emerged much later, allowing male‐only paradigms to persist.

The cumulative impact has been a major sex‐based data gap. One widely cited analysis reported a ~5.5:1 male‐to‐female bias in rodent studies, with > 80% excluding females; even among neuroscience papers that included both sexes, only ~15% formally tested for sex effects [302]. These long‐standing design choices have entrenched blind spots in female neurobiology and limited mechanistic generalization across sexes. Many foundational mechanisms in immune and metabolic regulation were characterized largely in male animals, leaving critical questions unresolved regarding female‐specific pathways and sex‐by‐environment interactions. Although funders and journals increasingly require sex‐disaggregated reporting [306], the legacy of male‐centric design remains embedded throughout the neuroinflammation literature. Sex differences in murine neurological disease models are summarized in Table 2.

TABLE 2.

Summary of sex differences in mouse models of human neurological disease.

Disease model	Mouse model name	Mouse model makeup	Disease pathology	Sex differences observed
Alzheimer's disease (AD)	APP/PS1	Human APP (Swedish) + PS1 mutation	Extracellular Aβ plaque deposition, gliosis	Increased expression of microglial activation genes in females (PMID: 34112929) Glycolytic, less phagocytic microglia in females associated with increased amyloidosis; amoeboid microglia in males (PMID: 34112929) Higher parenchymal Aβ burden, more severe cerebral amyloid angiopathy and subsequent microhemorrhage, higher levels of phosphorylated tau and proinflammatory cytokines, more severe astrocytosis and microgliosis, and greater neuronal and synaptic degenerations in females (PMID: 26707129) Greater amyloid burden and higher plaque number in females; greater Aβ40 and Aβ42 in age‐matched females (PMID: 14678749) Shorter female lifespan (PMID: 40138433)
	5xFAD	3 APP mutations (Swedish, Florida, London) + 2 PS1 mutations	Very early, aggressive Aβ plaque pathology, synaptic degeneration	Higher levels of Aβ42 and steady‐state transgenic APP in females; heightened inflammation in females (PMID: 25567526, PMID: 34864660) Greater amyloid deposition in the hippocampus and entorhinal cortex of females (PMID: 36899916) Sex‐based differences in cognitive performance (PMID: 34864660) Hyper‐ramified morphology in age‐matched male microglia; greater density and more reactive phenotype in female microglia (doi:10.1101/2025.10.07.681006)
	Tg2576	APP Swedish mutation	Age‐dependent Aβ plaque accumulation	Greater Aβ plaque load and occupation in age‐matched females (PMID: 11238065) Increased cognitive impairment in females (PMID: 30458163)
	3xTg‐AD	APP (Swedish) + PS1 + Tau (P301L)	Both Aβ plaques and tau neurofibrillary tangles	Greater Aβ burden and larger behavioral deficits in age‐matched females (PMID: 20934413) More aggressive Aβ pathology, increased beta‐secretase activity, and reduction of neprilysin in females (PMID: 18486110) More prominent amyloid plaques, neurofibrillary tangles, neuroinflammation and spatial cognitive deficits in females than in male mice (PMID: 30099679) Age‐dependent worsening of cognitive performance in females (PMID: 17659878)
	EFAD	5xFAD × human ApoE2/E3/E4	ApoE‐isoform–dependent modulation of Aβ burden and neuroinflammation	Increased cerebral cortex microbleeds in females (APOE4) (PMID: 26686669) Lower Iba1 plaque coverage and lower plaque compaction in females (APOE3) (PMID: 31113487) Lower TREM2 load (%Iba1) and lower TREM2 plaque coverage ratio in females (APOE3); fewer number of processes per microglia in the near plaque environment in females (APOE3) (PMID: 31113487)
	AppNL‐F	APP knock‐in with Swedish (NL) and Iberian (F) mutations; normal APP expression levels	Progressive Aβ42 increase, plaques, gliosis, and mild cognitive deficits	Worse Morris water maze (MWM) performance in males compared to females (PMID: 34897085) Higher female minimal longevity but lower female maximal longevity (PMID: 40138433) Age‐dependent increased plaque burden in females (PMID: 40138433) Sex‐specific differences in memory deficits (PMID: 39543985)
	P301S/PS19 tau transgenic mice	Overexpression of human P301S mutant tau under the Prnp promoter	Tau accumulation and tangles with neurodegeneration, motor decline, and cognitive impairment	Earlier sleep disruption in females; chronic sleep disruption‐based spatial memory issues in males (PMID: 38858068) Sex‐based differences in behavior and phenotype; differential cytokine profiles by sex (PMID: 32093751)
Parkinson's disease (PD)	MPTP	Dopaminergic neurotoxin	Selective degeneration of substantia nigra dopaminergic neurons	Stride length decreased in males (PMID: 20347863) Decrease of dopamine, 3,4‐dihydroxyphenylacetic acid (DOPAC), and homovanillic acid (HVA) more pronounced in females (PMID: 19631714) More pronounced striatal and midbrain tyrosine hydroxylase (TH) protein decrease in females; increase in striatal glial fibrillary acidic protein (GFAP) in females; greater decrease of striatal dopamine transporter (DAT) protein in females (PMID: 19631714)
	6‐OHDA	Dopamine neurotoxin (lesion model)	Targeted loss of nigrostriatal dopaminergic projections	Lower dopaminergic cell loss and higher behavioral recovery in female rats (PMID: 15698888, PMID: 12535954) Sex‐specific responses to amphetamine and apomorphine in rats (PMID: 40389167) Early non‐motor impairments present in male but not female mice (PMID: 34710535) Increased oxidative stress in male mesencephalic neurons (PMID: 20416276)
	A53T α‐synuclein	Human α‐synuclein with A53T mutation (transgenic)	α‐synuclein aggregation and motor deficits	Sex‐specific differences in motor and memory tests (PMID: 32477098) Anxiety‐ and depression‐like behavior in males; reduction in alcohol drinking in males (PMID: 38284431) Sex‐ and age‐specific changes in brain ceramide metabolism (PMID: 38284431) Sex‐specific differences in weight gain, anxiety‐like behavior, and oxidative stress after high‐fat diet consumption (PMID: 40602652)
Post‐COVID‐19	SARS‐CoV‐2 K18‐hACE2	Human ACE2 under K18 promoter; infected with live SARS‐CoV‐2	Severe viral pneumonia and alveolar lung damage; viral dissemination including brain	Lower susceptibility to infection and less serious disease in females; higher ACE2 expression in the lung (PMID: 37609639, PMID: 37259182, PMCID: PMC9665217) Increased lung immune cell infiltration, greater tissue damage, and greater alveolar space reduction in males (PMCID: PMC9665217) Differential pulmonary infiltration of immune cells and cytokine profile by sex depending on SARS‐CoV‐2 variant (PMID: 40686017) Upregulation of soluble thrombomodulin (sTM) in females; increase in von Willebrand Factor (VWF) in males (PMCID: PMC9665217)
Traumatic brain injury (TBI)	Controlled cortical impact (CCI)	Mechanical cortical impact	Focal cortical contusion and neuronal loss	Improved motor performance in male mice and worse motor performance in female mice following progesterone treatment after CCI (PMID: 25280093) Sex‐specific differences in gut microbiota composition following CCI (PMID: 38410824) Greater astrocytic hypertrophy and heme‐oxygenase‐1 (HO‐1) induction in female rats post‐CCI, while males had increased endothelial activation and expression of beta‐catenin; increase in the number of vessels and complexity in males compared to females (PMID: 29648973) Sex‐based differential microglial and macrophage phenotype post‐CCI (PMID: 28608978)
	Fluid percussion injury (FPI)	Dural fluid pulse	Diffuse axonal injury	More robust defensive behavior to white noise in females following FPI (PMID: 33324313) Estrogen administration prior to FPI protective in male rats but exacerbates brain injury in female rats (PMID: 849535) Smaller contusion volume in female rats and greater numbers of NeuN‐positive cortical neurons in female mice (PMID: 14769396) Posttraumatic hypothermia reduced contusion volume in male but not female mice and protected male but not female mice against loss of cortical neurons (PMID: 14769396)
Multiple sclerosis (MS)	EAE (MOG/PLP/MBP)	Autoimmune demyelination	Immune‐mediated CNS demyelination and paralysis	Greater spinal cord infiltrating cells and demyelination in females; increased inflammatory and regulatory cell types in males (PMID: 33190849) Sex differences in measures of spinal cord inflammation and plasticity (PMID: 30649100) Greater EAE severity in SJL and ASW females; greater EAE incidence in NZW females; more severe disease in B10.PL and PL/J males (PMID: 15081249) Behavioral signs of neuropathic pain and disruption of the estrous cycle in females (PMID: 24581045)
	Cuprizone	Copper chelator diet	Toxic demyelination (especially corpus callosum)	Disruption of female estrous cycle (PMID: 19746424) More severe motor coordination deficits in males in the horizontal bar and passive wire hang tests; moe severe motor coordiantion deficits in females in the motor skill sequence test (PMID: 36552147) More anxiety‐like behaviors in males in the elevated zero maze (PMID: 36552147) More severe demyelination in SJL mice (PMID: 19016742)
Amyotrophic lateral sclerosis (ALS)	SOD1‐G93A	Mutant SOD1	Progressive motor neuron degeneration	Longer survival of females with SJL background and hybrid B6/SJL background; delayed disease onset in female mice with the C57BL/6 background (PMID: 16024047, PMID: 26594635) Sex‐based differences in the mitochondrial intermembrane space mitochondrial unfoldeed protein response (MS‐UPRmt) (PMID: 28186560)
	TDP‐43	Mutant/overexpressed TDP‐43	TDP‐43 proteinopathy with motor neuron loss	Age‐dependent onset of motor symptoms more pronounced in male mice with human WT TDP‐43 overexpression (PMID: 40847737) Impaired electroretinogram (ERG) responses in retinas of young female TDP‐43M337V mice; sex‐specific differences in key regulators of mitochondrial dynamics and bioenergetics (PMID: 38143367)

6.1. Roadblocks to Studying Sex Bias in Neuroinflammatory Disease

Despite growing awareness, structural and analytical barriers continue to limit progress in dissecting the effects of sex as a biological variable. Many cohort studies and clinical trials remain unstratified by sex or underpowered to detect sex effects, and large omics initiatives often lack sex‐balanced sampling [302, 307]. Analytical bias persists even when both sexes are included. One evaluation found that among clinical studies enrolling both sexes, fewer than half (42%) conducted or reported sex‐disaggregated analyses [308]. Similarly, surveys suggest that while many experiments now include both sexes, fewer than ~15% statistically analyze sex as a factor [309]. As a result, sex‐specific cellular programs, particularly in CNS‐resident immune populations, can be missed when data are pooled.

Confounding biological variables further complicate interpretation. Hormonal transitions (e.g., menopause), metabolic comorbidities (e.g., obesity and metabolic syndrome), and aging can interact with neuroinflammatory disease in sex‐specific ways [310, 311, 312], yet these interactions are often not modeled explicitly. On the translational side, sex‐agnostic reporting remains common even in pivotal therapeutic trials. In AD, re‐analyses of recent trials (e.g., lecanemab) have raised the possibility of sex‐dependent efficacy signals [313, 314], but most studies were not designed or powered to test sex‐specific outcomes. Without systematic incorporation of sex into trial design, power calculations, and analysis plans, including sex‐by‐age interaction models and careful adjustment for key confounders, true biological differences may remain obscured.

6.2. Limitations of Current Animal Models

Animal models aim to capture key features of human immune–metabolic neuroinflammatory disease (summarized in Table 2). Yet rodent immune systems and metabolic programs differ from humans in meaningful ways [315, 316, 317]. For example, while activated microglia in both mice and humans shift toward glycolysis, species can rely on different rate‐limiting enzymes and regulatory nodes (e.g., hexokinase‐centered programs in mice versus phosphofructokinase‐linked programs in human myeloid cells), underscoring nontrivial divergence in immunometabolic control points [318]. Rodents also do not model human endocrine aging trajectories well (e.g., menstruation, menopause, multi‐decade hormonal change) [319], limiting interpretability of sex effects across the human lifespan.

Moreover, commonly used transgenic and toxin‐based models often isolate a single pathological feature and compress chronic disease into accelerated timelines, for example, amyloid plaque focused models in AD, toxin‐induced dopaminergic loss in PD, or chemically induced autoimmunity in MS. These approaches can underrepresent gradual progression, mixed pathology, and interactions with systemic metabolism that are central to human disease. Persisting male bias in preclinical work further compounds these limitations. The thalidomide tragedy remains a canonical reminder of how sex‐ and pregnancy‐linked risks can be missed when safety evaluation fails to model relevant biology [320]. Together, limited biological fidelity and sex imbalance in experimental systems represent a major translational bottleneck for discovery and therapeutic development in neuroinflammatory disease.

6.3. Toward Next‐Generation Models and Study Designs

Future models must deliberately incorporate sex, age, and metabolic context. Preclinical studies should include both sexes across appropriate life stages and explicitly consider hormonal status (e.g., modeling menopause/senescence in females alongside comparable aging in males). Aged cohorts are essential. Genetic diversity should also be embraced to better mirror human heterogeneity. The AD‐BXD panel provides an instructive example. The BXD recombinant inbred strains, derived from C57BL/6J and DBA/2 J backgrounds, capture extensive natural genetic variation and enable mapping of modifiers of neurodegeneration and immunity. Introducing AD transgenes (e.g., 5xFAD or APP/PS1) into genetically diverse BXD lines yields stable strains that share a pathogenic driver but differ across thousands of naturally segregating alleles. This strategy reveals strain‐dependent differences in amyloid deposition, neurodegeneration, microglial activation, synaptic integrity, and cognitive decline [321, 322], including resilience in some strains despite the same initiating mutation [322]. Such trajectories map onto genetic modifiers enriched for immune, metabolic, and lipid‐handling pathways, strengthening the bridge between mechanistic neurobiology and human risk architecture.

Ideal future models will also incorporate human risk alleles (e.g., APOE variants), endocrine‐genetic factors (including variation in sex hormone receptors), and environmental modulators (dietary fat, microbiome composition) that reshape metabolism. These systems should be deeply phenotyped with multi‐omic approaches, paired single‐cell transcriptomics, proteomics, metabolomics/lipidomics, and immune profiling in CNS tissues to define sex‐specific pathways. Animal‐derived pathway signatures should be validated in human brain tissue and biofluids to ensure translational relevance.

7. Conclusion

Across this review, metabolic–immune axes emerge as central to sex‐biased neuroinflammatory vulnerability (Figure 1). Purine metabolism and adenosine signaling tune CNS immunity and intersect with sex‐biased interferon set points. Activated glia undergo glycolytic remodeling during inflammation, yet key control nodes can differ by sex and species. Lipid sensing pathways (including PPAR/LXR‐linked programs) shape microglial activation in sex‐dependent ways, consistent with known sex differences in lipid metabolism. The cGAS–STING–interferon axis exemplifies an immune–metabolic node with functional sex divergence: in one model, activating STING in spinal microglia relieved neuropathic pain in male mice but elicited counterproductive inflammatory responses in females [323]. Together, these circuits align with clinical patterns in which female‐skewed conditions (e.g., AD, MS) often feature prominent inflammatory and interferon‐linked signatures, while male‐skewed disorders (e.g., PD, ALS) exhibit distinct metabolic vulnerabilities.

FIGURE 1.

Sex‐biased immunometabolic programs shape microglial IFN signaling and neuroinflammatory vulnerability. This figure summarizes sex‐associated differences in microglial immunometabolic programs that shape innate immune signaling and neuroinflammatory outcomes across neurodegenerative and neuroinflammatory diseases. Male‐ and female‐associated features are shown as enriched tendencies across models and disease contexts, rather than mutually exclusive states. Male‐biased microglial programs are characterized by increased reliance on lipid catabolism and oxidative phosphorylation (OXPHOS), enhanced inflammasome‐associated signaling, and androgen‐modulated eicosanoid pathways. Higher ADA1/ADA2 activity promotes efficient purine clearance and is associated with a lower basal type I interferon (IFN‐I) tone. Functionally, these programs bias microglia toward lipid‐ and inflammasome‐driven inflammatory states, increased chronic lesion burden, and heightened sensitivity to mitochondrial and oxidative stress. Female‐biased microglial programs preferentially engage glycolytic and ECAR‐high metabolic states, coupled to stronger IFN‐responsive transcriptional programs. Reduced ADA1/ADA2 activity permits purine accumulation, facilitating SAM‐dependent derepression of endogenous retroelements and amplification of IFN‐β and interferon‐stimulated gene expression. These states are associated with early IFN priming, increased phagolysosomal activity, impaired plaque clearance, and enhanced synaptic engulfment or pruning. Together, these sex‐biased immunometabolic programs converge on the purine‐interferon axis, providing a conceptual framework for how biological sex modulates microglial IFN tone, inflammatory trajectories, and susceptibility to neurodegenerative pathology.

Progress now requires treating sex and metabolic state as core variables in study design, analysis, and translation. This includes powering cohorts for sex‐stratified inference; modeling sex‐by‐age and sex‐by‐metabolism interactions; building and profiling sex‐balanced, metabolically informed experimental systems; and developing therapies that target glycolysis, purine turnover, lipid signaling, and interferon pathways in sex‐aware ways. Only through a sex‐specific, immunometabolically grounded approach can the field uncover actionable mechanisms and develop effective interventions for both females and males.

Funding

The work discussed in this review was sponsored by generous donation of philanthropic funds from the following sources: The Clouse Family, The Shiley Foundation, The Shirley and Harry Beer Charitable Foundation, Nancy Vaughan, Nabil and Gayda Hanna, Brad and Rachel Greenwald, Bridget Cresto, James and Katya Hazel, the Prebys Foundation, and CZI grant DAF2019‐198153 from the Chan Zuckerberg Initiative DAF, and an advised fund of Silicon Valley Community Foundation. This work was also supported by National Institutes of Health (NIH) Grant numbers: R01CA273230 (S.S. and P.S) and U54AG065141 (S.S and P.S). AG was supported by the BioLegend Fellowship in Immunology.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

All data used in the preparation of the article will be made available by the corresponding author.