Adipocyte metabolic state regulates glial phagocytic function

SUMMARY

Excess dietary sugar profoundly impacts organismal metabolism and health, yet it remains unclear how metabolic adaptations in adipose tissue influence other organs, including the brain. Here, we show that a high-sugar diet (HSD) in Drosophila reduces adipocyte glycolysis and mitochondrial pyruvate uptake, shifting metabolism toward fatty acid oxidation and ketogenesis. These metabolic changes trigger mitochondrial oxidation and elevate antioxidant responses. Adipocyte-specific manipulations of glycolysis, lipid metabolism, or mitochondrial dynamics non-autonomously modulate Draper expression in brain ensheathing glia, key cells responsible for neuronal debris clearance. Adipocyte-derived ApoB-containing lipoproteins maintain basal Draper levels in glia via LpR1, critical for effective glial phagocytic activity. Accordingly, reducing ApoB or LpR1 impairs glial clearance of degenerating neuronal debris after injury. Collectively, our findings demonstrate that dietary sugar-induced shifts in adipocyte metabolism substantially influence brain health by modulating glial phagocytosis, identifying adipocyte-derived ApoB lipoproteins as essential systemic mediators linking metabolic state with neuroprotective functions.

In brief

An obesogenic diet alters lipid metabolism in Drosophila adipocytes, remotely disrupting glial phagocytic function. Alassaf et al. identify adipocyte-derived ApoB and receptor LpR1 as crucial regulators of glial Draper expression and neuronal debris clearance, establishing a link between peripheral lipid metabolism and neuroprotection.

Graphical Abstract

INTRODUCTION

Obesity and type 2 diabetes significantly increase the risk for dementia and neurodegenerative diseases.^1–4 Although mechanisms linking metabolic disorders to cognitive decline remain incompletely defined, disruption of the adipose tissue-brain axis is strongly implicated.^5–8 Once considered merely an energy storage depot, adipose tissue is now recognized as a dynamic endocrine organ secreting diverse molecules that influence brain function, including lipid metabolites and adipokines.⁹ These adipocyte-derived factors modulate neuroinflammation, oxidative stress, and synaptic plasticity.^6,8,10 Thus, uncovering mechanisms of adipose-brain communication holds promise for therapeutic strategies against neurodegeneration.

Although the majority of brain lipids are synthesized locally,¹¹ a fraction is derived from peripheral tissues, especially adipose stores.^12–14 Lipoproteins facilitate inter-organ lipid transport, primarily through apolipoproteins acting as lipid chaperones. Apolipoprotein B (ApoB) delivers peripheral lipids into the brain by binding to low-density lipoprotein receptors (LDLRs) at the blood-brain barrier.^15–18 Dysfunctional ApoB signaling is linked to obesity, diabetes,^19–21 and neurodegeneration^22,23; however, direct causative roles remain unexplored.

Lipoproteins, composed of apolipoproteins surrounding a lipid core, facilitate lipid transport between tissues. In Drosophila, lipid transport relies on lipophorins (Lpp), scaffolded by apolipophorins structurally analogous to mammalian apolipoproteins^24,25. The fly ApoB ortholog, Apolpp, is processed in adipocytes into two major lipoproteins—ApoLI and ApoLII (collectively ApoB-Lpp)—which primarily circulate diacylglycerol (DAG) and phosphatidylethanolamine (PE).^24–27 ApoB-Lpp is essential for lipid delivery to peripheral tissues and developing organs, similar to mammalian ApoB’s role in systemic lipid distribution.²⁸

Despite the presence of a restrictive blood-brain barrier, the Drosophila brain acquires lipids from circulating ApoB-Lpp.²⁹ Loss of adipocyte ApoB dramatically reduces brain lipid stores, notably triacylglycerol, underscoring the critical role of ApoB-Lpp in brain lipid supply.³⁰ ApoB-Lpp delivers lipids through LDL receptor homologs, LpR1 and LpR2, which mediate lipoprotein internalization analogous to mammalian LDLR.^28,31–33

Within the brain, neuron-glia lipid trafficking is mediated by ApoD/ApoE homologs, Glaz and Nlaz.^34–36 Overexpression of Glaz confers neuroprotection by enabling glial lipid accumulation and reducing oxidative damage.^36,37 Similarly, in humans, the neuroprotective function of ApoE depends on its lipid-handling capacity, with ApoE4 impairing and ApoE2 enhancing glial lipid storage. ^35,38 While ApoB-Lpp signaling influences systemic insulin signaling and feeding behavior,^39–41 its role in regulating glial function and neuroprotection remains unknown.

Neurodegenerative disorders are characterized by the accumulation of toxic protein aggregates and cellular debris, leading to neuronal damage and death.⁴² Brain-resident phagocytes, microglia, protect neuronal integrity by efficiently clearing debris.⁴³ In Drosophila, ensheathing glia fulfill similar roles, clearing degenerating neuronal material via the conserved engulfment receptor Draper, homologous to mammalian MEGF10.^44–46 Draper-mediated signaling enables glial cells to detect, engulf, and degrade neuronal debris, maintaining neuronal health.^44,47–52 Indeed, in Drosophila Alzheimer’s disease models, ensheathing glia clear human β-amyloid aggregates through Draper,⁵³ analogous to microglial function in mammals.⁵⁴ However, age and metabolic stress impair glial phagocytic capacity, contributing to neuronal vulnerability.^50,55

Previously, we have demonstrated that a high-sugar diet (HSD) impairs Draper signaling and glial phagocytosis by inducing glial insulin resistance.⁵⁵ Given evolutionary conservation in adipocyte-brain signaling^56,57 and glial biology,^47,58 we hypothesized that the adipocyte metabolic state influences glial function via lipid-based signals, specifically focusing on adipocyte-derived ApoB lipophorins as potential mediators.

Here, we identify a novel adipocyte-glial metabolic coupling linking dietary sugar exposure to impaired glial phagocytic function, potentially underlying the elevated neurodegeneration risk observed with obesity. We show that a prolonged HSD induces shifts in adipocyte lipolysis, mitochondrial dynamics, metabolism, and antioxidant responses, remotely influencing Draper-mediated glial phagocytosis. Specifically, we propose two mechanisms through which adipocyte dysfunction impacts glial biology. First, altered adipocyte mitochondrial metabolism and increased ketogenesis disrupt glial Draper expression. Second, adipocyte-derived ApoB lipoproteins are critical for the glial response to neuronal injury. Together, these data uncover how the adipocyte metabolic state remotely modulates glial phagocytic capacity, highlighting potential pathways by which peripheral adipose dysfunction contributes to neurodegeneration risk.

RESULTS

Prolonged exposure to an HSD shifts adipocyte metabolism toward mitochondrial oxidative phosphorylation

Cellular energy production balances two major pathways: glycolysis and oxidative phosphorylation (OxPhos).⁵⁹ We examined HSD effects in adult Drosophila adipocytes by assessing glycolytic enzyme expression. qPCR analysis from abdominal segments of flies on a normal diet (ND) or HSD for 3 weeks showed significant downregulation of HexA and PyK, essential enzymes for glycolytic flux (Figures 1A and 1B). HexA converts glucose to glucose-6-phosphate, while PyK converts phosphoenolpyruvate to pyruvate,^60,61 consistent with reduced glycolytic activity in adipose tissue under a prolonged HSD.

Figure 1. An HSD induces a metabolic shift in fly adipose tissue.

(A) Schematic of glycolysis pathway enzymes (HexA, PyK, and Ldh).

(B and K) Fold change (qPCR) of the indicated mRNA relative to α-tubulin in the adipose tissue of flies fed either an ND or an HSD for 3 weeks. Student’s t test with Welch’s correction. N = 3 technical replicates of cDNA collected from 30 fly abdominal segments/treatment.

(C and D) Confocal images (C) and quantification (D) of adipocyte Ldh-GFP intensity. Scale bar: 10 μm. Student’s t test (Welch’s). Each dot represents an individual fly.

(E) Schematic: mitochondrial fission-fusion balance.

(F) TEM images of adipocyte mitochondria (white outline), ND vs. HSD at 1 week. Scale bar: 1 μm.

(G–I) Confocal mito-GFP images (G), mitochondrial circularity (H), and elongation (I). Scale bar: 10 μm (top) and 5 μm (bottom). Student’s t test; dots represent individual flies.

(J) Mitochondrial pyruvate uptake.

(L and L′) Confocal live imaging (L) and quantification (L′) of mitochondrial pyruvate uptake (mitoPyronicSF) at baseline and pyruvate addition (5 and 10 mM). Scale bar: 10 μm. Student’s t test.

All experiments except (F): 3 weeks of diet (ND or HSD). Statistical significance was determined using Student’s t test (with Welch’s correction where appropriate). Exact p values are indicated; all error bars represent ±SD. See also Figure S1.

Using a GFP-tagged lactate dehydrogenase (Ldh) reporter, we monitored glycolytic activity.⁶² Confocal imaging revealed a marked reduction in Ldh-GFP levels after HSD feeding (Figures 1C and 1D), with decline detectable at 2 weeks (Figure S1). Glycolysis in the brain remains unchanged at this time point,⁵⁵ suggesting that adipose tissue is more sensitive to HSD-induced metabolic shifts.

We examined whether reduced glycolytic activity coincides with mitochondrial morphology changes (Figure 1E). Transmission electron microscopy (TEM) of adipose tissue showed that HSD exposure increases lipid droplet size⁴⁰ and results in elongated, irregularly shaped mitochondria (Figure 1F). Using mitochondrion-targeted GFP (mitoGFP) under the adipocyte-specific Lpp promoter, we found a shift toward elongated mitochondria in HSD-fed flies (Figures 1G, 1G′, and 1H). This correlated with downregulation of the fission gene fis1 and upregulation of the fusion gene opa1, suggesting adaptive remodeling of mitochondrial networks under an HSD (Figure 1I).

Elongated mitochondria have greater OxPhos capacity compared to circular counterparts.^63,64 We assessed the expression of key OxPhos genes and mitochondrial substrate carriers (Figure 1J). CoxIV, a crucial complex IV component required for mitochondrial respiration, was significantly elevated in adipose tissue following an HSD (Figure 1K). Mitochondrial pyruvate carriers (MPC1 and MPC2) showed marked upregulation, consistent with enhanced mitochondrial pyruvate transport under an HSD (Figure 1K).

To directly assess mitochondrial metabolic activity, we employed genetically encoded mito-Pyronic sensors,⁶⁵ in which a circularly permuted GFP (cpGFP) fused to the bacterial PdhR transcription factor, which binds pyruvate. Pyruvate binding causes a fluorescence intensity shift that enables real-time tracking of mitochondrial pyruvate uptake⁶⁵ (Figure 1G). The mitoPyronicSF sensor measures mitochondrial pyruvate concentration, transport, and flux in real time through a cpGFP fused to the bacterial PdhR transcription factor.⁶⁵ Using the Lpp promoter to express mitoPyronicSF in fat tissue, we found significantly increased mitochondrial pyruvate uptake in HSD-fed flies at baseline compared to ND-fed flies (Figure 1L). Upon adding 5 mM pyruvate, mitochondrial activity increased in both groups; however, while ND flies exhibited a plateau between 5 and 10 mM pyruvate, HSD-fed flies showed a sustained significant increase in sensor activity (Figures 1L and 1L′), suggesting that HSD-fed mitochondria retain elevated capacity for pyruvate uptake or utilization.

Our findings indicate that HSD enhances mitochondrial pyruvate uptake and activity, supporting a shift toward mitochondrial metabolic activity under HSD stress. Collectively, adipose tissue undergoes a metabolic shift toward increased mitochondrial activity following prolonged HSD exposure, evidenced by decreased glycolysis, altered mitochondrial morphology, upregulation of OxPhos-related genes, and enhanced mitochondrial engagement in response to substrate availability.

HSD shifts adipose metabolism toward fatty acid oxidation and ketogenesis

In obesity and type 2 diabetes, chronic hyperinsulinemia induces insulin resistance, disrupting lipid homeostasis and increasing lipolysis and fatty acid oxidation (FAO).^66,67 FAO generates acetyl-coenzyme A (CoA), which fuels the mitochondrial TCA cycle to support OxPhos.⁶⁸ Given that HSD reduces glycolysis (Figures 1B–1D) and increased mitochondrial activity in fly adipose tissue (Figure 1L), we hypothesized a compensatory increase in FAO (see the schematic in Figure 2A).

Figure 2. Molecular characterization of metabolic changes in the adipose tissue.

(A) Schematic: metabolic shift toward FA oxidation (FAO) induced by an HSD.

(B and C) Confocal images (B) and quantification (C) of adipocyte Plin1 intensity in w1118 fat bodies; ND vs. HSD. Scale bar: 20 μm. Student’s t test (Welch’s). Each dot represents an individual fly.

(D, E, and I) qPCR fold change of lipid metabolism (D), FAO enzyme genes (E), and antioxidant markers (I) in adult fat body explants; ND vs. HSD (3 weeks). Student’s t test (Welch’s) and one-way ANOVA; N = 3 replicates (each replicate contains 30 fat explants).

(F and G) Metabolite quantification (μM): acetylcarnitine, α-HB, and β-HB, in whole flies, ND vs. HSD (2 weeks). Student’s t test (Welch’s); N = 3 replicates. 10 flies per replicate.

(H and H′) Confocal images (H) and quantification (H′) of mitochondrial oxidation (MitoTimer) in adipocytes. Scale bar: 20 μm.

Student’s t test (Welch’s); exact p values are indicated; dots represent individual flies; all error bars represent ±SD). See also Figure S2.

Perilipin 1 (Plin1/Lsd1 in flies^69,70) restricts lipolysis.^71–73 Immunohistochemistry showed reduced Plin1 levels in HSD-fed flies, consistent with increased lipolysis (Figures 2B and 2C). qPCR revealed upregulation of hepatocyte nuclear factor 4 (hnf4), a regulator of lipid mobilization and FAO, and carnitine palmitoyltransferase 1 (CPT1A/whd in flies⁷⁴), which facilitates mitochondrial fatty acid transport (Figure 2D). HSD downregulated acyl-coA carboxylase (ACC), potentially enhancing mitochondrial fatty acid uptake. Mitochondrial FAO enzymes, including acetoacetyl-CoA thiolase (Acat1), were upregulated (Figure 2E), supporting a shift toward FAO.

Acat1 catalyzes the first step of ketogenesis.⁷⁵ Metabolite profiling confirmed increased acetylcarnitine in HSD-fed flies (Figure 2F), reflecting increased acetyl-CoA availability during active FAO.^76,77 We observed increased succinylcarnitine, succinate, and NAD⁺—metabolites that feed into the TCA cycle and OxPhos⁷⁸ (Figure S2). Acetylcarnitine serves as an acetyl donor for ketogenesis, with levels rising during increased FAO.⁷⁹ Metabolomics analysis detected increased α-hydroxybutyrate (α-HB) and β-hydroxybutyrate(β-HB) in whole flies after 2 weeks of HSD feeding (Figure 2G) and in the circulation after 1 week (Figure S2). These findings suggest that HSD promotes FAO and ketone body accumulation in adipose tissue.

Since FAO-driven OxPhos generates ROS, we investigated whether HSD-induced mitochondrial expansion was associated with oxidative stress using MitoTimer. ^80,81 After 3 weeks on HSD, flies exhibited a lower red-to-green fluorescence ratio, suggesting reduced ROS levels in adipose mitochondria (Figures 2H and 2H’). One explanation is increased antioxidant defenses. β-HB, elevated in HSD-fed flies, enhances antioxidant responses.⁸² qPCR showed increased expression of nuclear factor erythroid 2-related factor 2 (nrf2), a regulator of antioxidant pathways (Figure 2I).

These findings demonstrate that an HSD induces a metabolic shift in adipose tissue, favoring FAO and ketogenesis while upregulating antioxidant defenses to mitigate ROS accumulation. This adaptation maintains energy homeostasis under conditions of impaired glycolysis and insulin resistance.

Adipocyte mitochondrial and lipid metabolism remotely impacts the glial phagocytic state

Given the established role of lipid-derived metabolites in systemic metabolic communication,^76,78 we investigated whether HSD-induced changes in adipocyte mitochondrial and lipid metabolism influence glial function. Glial cells rely on lipid metabolism⁸³ and maintain brain health during metabolic challenges.^84,85 Aging and an HSD reduce microglial efficiency, leading to chronic inflammation.^50,55 Drosophila ensheathing glia serve as the brain’s resident phagocytes.⁴⁸ Central to glial phagocytosis is Draper, the Drosophila homolog of the MEGF10 family of receptors.^51,86 Draper is essential for pruning axons during development^87,88 and responding to axon injury⁵¹ and clearance of degenerating axons.⁸⁷ We have shown previously that an HSD impairs ensheathing glial phagocytic function by inducing glial insulin resistance, downregulating Draper.⁵⁵ We observed that an HSD alters mitochondrial metabolism and disrupts lipid homeostasis in adipose tissue (Figures 1 and 2). Hence, we explored whether HSD-induced changes in adipocyte-glia metabolic signaling contribute to defects in glial phagocytosis.

We examined how genetic modulation of adipocyte metabolic pathways affects Draper expression in ensheathing glia surrounding olfactory neuropil in the antennal lobe (Figure 3A, schematic). A representative image of the adult fly brain stained with Draper is shown in Figure 3A’. We genetically manipulated key genes involved in lipid metabolism (Figure 3B) and mitochondrial function (Figure 3C) and then assessed the basal glial phagocytic state.

Figure 3. Adipocyte lipid metabolism and mitochondrial dynamics regulate Draper in ensheathing glia.

(A–C) Schematics illustrating the experimental context.

(A) Adult fly brain schematic highlighting antennal lobes (ALs) and ensheathing glia.

(A′) Representative confocal image showing Draper staining in ALs; the inset defines the analyzed region of interest (ROI) in (D)–(J). Scale bar: 20 μm.

(B and C) Lipid metabolism and mitochondrial dynamics pathways, indicating adipocyte-specific genetic manipulations analyzed below.

(D–J) Representative confocal z stack projections showing Draper staining in ALs ensheathing glia from flies with adipocyte-specific gene knockdown (bottom) vs. matched controls (top). Scale bar: 20 μm.

(D′–J′) Quantification of Draper fluorescence intensity in ROIs (white boxes). Circles represent individual flies.

Student’s t test with Welch’s correction; exact p values are shown; all error bars represent ±SD. See also Figure S3.

Previously, we have shown that a prolonged HSD reduced fatty acid (FA) levels.⁴⁰ Here, we find that a prolonged HSD upregulates FA oxidation in adipose tissue (Figure 2). We investigated whether altered adipocyte FA homeostasis impacts glial phagocytic receptor expression by genetically manipulating FA metabolism in adipocytes (Figure 3B). Plin1 restricts lipolysis and FA production. We hypothesized that adipocyte-specific alterations in Plin1 and the lipase brummer (bmm)^89,90 would alter FA homeostasis and impact glial phagocytic receptor expression. Overexpression of human Plin1⁷⁰ in Drosophila adipocytes led to increased glial Draper levels (Figures 3D–3D’), whereas knockdown of bmm lipase reduced Draper expression (Figures 3E and 3E’). Knocking down Plin1 decreased Draper expression (Figure S3A). We tested CPT1A, required for mitochondrial FA uptake.⁷⁴ Pharmacological inhibition of CPT1A reduced FA levels,⁹¹ and we found that adipocyte-specific knockdown of CPT1A reduced FA levels (Figure S3B). CPT1A knockdown significantly reduced Draper expression in ensheathing glia (Figures 3F and 3F’). In summary, perturbations in FA homeostasis (Figures 3D–3F) in adipocytes alter the basal glial phagocytic state.

Since glia utilize ketone bodies as an alternative energy source,⁹² we examined whether ketogenesis influences glial Draper expression. β-hydroxybutyrate shifts brain metabolism from glycolysis toward OxPhos.⁹³ Given that glial activation requires a glycolytic shift,^94,95 we hypothesized that ketone bodies suppress glial activation and reduce Draper expression. Supporting this, an HSD induces a shift toward OxPhos at the expense of glycolysis in the brain.⁵⁵ To test whether adipocyte-derived ketone bodies regulate Draper expression, we knocked down Acat1, a key enzyme in ketogenesis. Suppressing ketone body production significantly increased Draper expression, indicating that adipocyte-derived ketone bodies negatively regulate glial Draper levels (Figures 3G and 3G’).

To assess whether HSD-induced mitochondrial elongation affects glial Draper expression, we knocked down the mitochondrial fission factor Fis1 in adipose tissue, resulting in reduced glial Draper expression (Figures 3H and 3H’). We investigated whether adipocyte reactive oxygen species (ROS) levels influence glial Draper signaling. Since HSD-fed flies exhibit reduced mitochondrial oxidation and elevated antioxidant activity (Figures 2H and 2I), we reasoned that increasing adipocyte ROS should enhance Draper expression. RNAi-mediated knockdown of the mitochondrial antioxidant Sod2 in adipocytes led to increased glial Draper expression (Figures 3I and 3I’). Conversely, overexpression of Catalase reduced Draper levels, phenocopying effects in HSD-fed flies (Figures 3J and 3J’). These findings indicate that adipocyte mitochondrial metabolism and redox balance regulate glial Draper expression.

Hence, in Drosophila, adipocyte lipid metabolism and mitochondrial dynamics exert remote influence over the glial phagocytic state.

Adipocyte-derived ApoB lipoprotein signals regulate glial phagocytic response to neuronal injury

Leptin and its Drosophila ortholog, Upd2, regulate satiety via JAK/STAT signaling in the brain.^57,96,97 Given JAK/STAT’s role in maintaining Draper levels in ensheathing glia during axonal injury,⁹⁸ we tested whether upd2 regulates the glial phagocytic state. However, Draper levels remained unchanged in upd2Δ deletion mutants,⁹⁹ indicating that adipo-glial coupling occurs through another adipocyte-derived signal (Figure S4).

Adipose tissue distributes lipids to multiple organs via ApoB lipoproteins (ApoB-Lpps). Drosophila ApoB undergoes post-translational cleavage, producing ApoLI and ApoLII.^27,30,39 ApoLII contains the lipid-binding domain,²⁷ for which we have previously generated and validated reagents.⁴⁰ For clarity, hereafter we refer to ApoLII as ApoB.

An HSD reduces ApoB delivery to the brain, confirmed by western blot analysis of adult fly brains showing decreased endogenous ApoB levels compared to ND controls (Figure 4A). Previous studies have shown that ApoB primarily delivers PE-rich lipids to the brain.²⁷ Consistently, hemolymph lipid analysis showed a decrease in PE-rich phospholipids, the major component of Drosophila lipophorins²⁷ (Figure S5; Tables S1 and S2), supporting reduced ApoB-mediated lipid transport. Under ND conditions, ApoB localized to glial regions expressing Draper, while an HSD significantly reduced this co-localization (Figures 4B–4B’), suggesting impaired ApoB signaling in glia.

Figure 4. Adipocyte-derived ApoB modulates baseline glial phagocytic competence and injury response.

(A and A′) Western blot (A) and quantification (A′) of brain ApoII levels in control (w1118) flies after 3 weeks of ND or HSD. ApoII was normalized to tubulin; dots represent 10 pooled brains each (3–4 replicates). Student’s t test (Welch’s correction); exact p value is shown.

(B and B′) Confocal images (B) and quantification (B′) of ApoB (purple) and Draper (cyan) co-localization (Manders’ coefficient) in ALs from control flies (w1118, ND vs. HSD). Insets show magnified merged and single-channel views. Dots represent single z-slices. Student’s t test (Welch’s); scale bars: 50 μm (top) and 20 μm (insets).

(C–F) Confocal z stacks are shown and quantificatied. Basal (C and E) and injury-induced (D and F) Draper staining in antennal lobes after adipocyte ApoB-RNAi or glial LpR1-, LpR2-RNAi vs. controls (Luc-RNAi).

(G–H′) Olfactory axon degeneration assay.

(G and G′) Schematic of unilateral antennal ablation in ORN22-CD8GFP flies.

(H and H′) Confocal z stacks (H) and quantification (H′) of GFP intensity in injured vs. uninjured axons after glial RNAi (Luc, LpR1, and LpR2), assessed 1 day post injury.

Dots represent individual flies; Student’s t test (Welch’s); exact p values are shown; scale bar: 20 μm; all error bars represent ±SD. See also Figures S4–S6 and Tables S1 and S2.

We hypothesized that reduced glial Draper under an HSD resulted from diminished ApoB signaling. Adipocyte-specific ApoB knockdown significantly reduced basal Draper expression (Figures 4C and 4C’). Following antennal ablation,⁵¹ Draper upregulation occurred in controls but failed in ApoB-RNAi flies (Figure 4D), mirroring HSD effects. These results establish that adipocyte-derived ApoB supports both basal and injury-induced Draper expression. Significantly, adipocyte-derived ApoB knockdown did not affect Drosophila insulin-like peptide 5 levels (Figure S6), suggesting that ApoB regulation occurs independent of insulin signaling.

The LDLRs LpR1 and LpR2²⁸ facilitate ApoB-chaperoned lipid uptake. Ensheathing glia-specific knockdown of these receptors non-significantly reduced basal Draper levels (Figures 4E and 4E’). However, post-injury Draper upregulation failed in both LpR1- and LpR2-RNAi flies (Figure 4F), phenocopying adipocyte-specific ApoB knockdown. This suggests that ApoB-chaperoned lipophorins maintain glial phagocytic competency post injury via LpR1/2 signaling.

To assess functional consequences, we knocked down these receptors in ensheathing glia expressing membrane-tagged GFP in olfactory neurons. Post antennal ablation (Figure 4G), LpR1 knockdown significantly impaired phagocytosis, shown by increased GFP retention (Figures 4H and 4H’). Surprisingly, LpR2 knockdown did not significantly impact debris clearance despite similar effects on Draper signaling. This suggests that downstream Draper pathways are specifically disrupted in LpR1 knockdowns but may be unaffected or compensated for in LpR2 knockdowns (discussion). These findings establish that adipocyte-to-glia LpR1-mediated ApoB signaling supports glial phagocytic function by regulating Draper expression post injury.

Collectively, our results reveal a previously unrecognized role of adipose tissue in maintaining glial phagocytic function via lipoprotein-mediated signaling. Importantly, we demonstrate that a prolonged HSD disrupts adipo-glial coupling, impairing glial debris clearance.

DISCUSSION

The adipose tissue-brain axis controls glial response to neuronal injury

The adipose-brain axis is increasingly recognized for its role in brain health, with adipocytes secreting diverse signals influencing brain function. Our study demonstrates that dietary sugar-induced obesity disrupts adipocyte lipid homeostasis, impairing glial phagocytic activity via a novel metabolic coupling involving adipocyte-derived ApoB lipoproteins and glial LpR1.

Recent mammalian studies highlight adipocyte-derived leptin influencing glial metabolism during peripheral nerve injury.¹⁰⁰ However, our findings suggest that adipocyte-glia metabolic coupling in Drosophila does not depend on the leptin homolog Upd2 but, rather, involves ApoB-Lpp signaling. This highlights the complexity of adipocyte-glia interactions, indicating distinct signaling pathways regulating different glial populations. Importantly, our work reveals that adipocyte-derived signals extend beyond traditional adipokines to include lipid-based signaling via ApoB lipoproteins.

Starvation amid nutritional abundance: Adipocyte metabolic adaptation in diet-induced insulin resistance

During starvation, adipocytes adapt by enhancing lipolysis, releasing FAs to provide alternative energy sources and increasing FAO to compensate for reduced glycolysis. Remarkably, metabolic changes observed in insulin resistance resemble the body’s response to starvation.¹⁰¹ Consistent with this, we found that prolonged exposure to an HSD reduces adipocyte glycolytic enzymes (HexA, PyK, and Ldh) while simultaneously increasing key FAO enzymes. These adaptations may initially sustain energy homeostasis, but chronic FAO dependence can elevate lipid metabolites, potentially causing lipotoxic effects in distant tissues, including the brain.^102,103 Indeed, a prolonged HSD increased ketone bodies, typically elevated during nutrient scarcity, supporting the concept that prolonged sugar intake paradoxically induces starvation-like metabolic responses in adipocytes.

A prolonged HSD alters mitochondrial dynamics in adult adipocytes

We observed that HSD-induced mitochondrial elongation and antioxidant upregulation in adipocytes remotely regulate glial Draper levels. Elongated mitochondria generally enhance OxPhos efficiency^63,64 and reduce ROS; however, the exact signaling from mitochondria to glia remains unclear. Potential mechanisms include antioxidant-induced shifts in adipocyte-secreted metabolites that modulate glial Draper signaling. Additionally, reduced mitochondrial oxidation in adipocytes after a prolonged HSD, driven by increased Nrf2-mediated antioxidant responses, may represent a hormetic adaptation, initially triggered by transient oxidative stress. This response mirrors observations in mouse heart tissue under obesogenic stress.¹⁰⁴ Another plausible explanation for enhanced antioxidant defense includes increased adipose lipid droplet storage, which buffers ROS accumulation.^105,106

Moreover, heightened FAO typically enhances mitochondrion-lipid droplet interactions, efficiently transferring FAs into mitochondria and preventing cytosolic oxidation, thereby reducing oxidative stress.^107,108 Although adipocyte glycolysis declines during HSD feeding, elevated mitochondrial pyruvate uptake suggests alternative pyruvate sources, potentially from amino acid catabolism, glycerol metabolism, or TCA cycle intermediates replenished via FAO.⁷⁶ Future investigations of these metabolic shifts will provide deeper insights into adipocyte metabolic adaptation under chronic dietary sugar stress.

ApoB-Lpp signals through glial LpR1 to regulate Draper expression and phagocytosis

Our findings establish ApoB-Lpp as critical systemic regulators of glial Draper expression and phagocytic function. ApoB-Lpp are known to deliver lipids to peripheral tissues through LDL receptor-mediated internalization.²⁸ The knockdown of LDLR homologs LpR1 or LpR2 individually did not alter basal Draper levels but impaired injury-induced Draper upregulation, suggesting redundancy in maintaining basal Draper expression but critical roles during injury responses. Interestingly, LpR1 knockdown impaired glial phagocytosis of neuronal debris after injury, whereas LpR2 knockdown had no such effect. Similar context-specific receptor roles occur during lipoprotein delivery to ovaries,²⁸ supporting the idea that LpR1 may specifically mediate injury-induced glial responses. Indeed, activity-dependent LpR1 upregulation influences dendritic morphogenesis in larvae,¹⁰⁹ raising the possibility that neuronal injury similarly induces LpR1, enhancing Draper expression and phagocytic capacity. Future studies should clarify how ApoB-LpR1 signaling precisely regulates Draper-mediated phagocytosis.

ApoB primarily transports PE-rich lipids,²⁷ which directly bind Draper.¹¹⁰ In our prior work, we have shown that prolonged HSD exposure reduces PE levels in whole-fly lysates,⁴⁰ and in this study, we showed that circulating PE levels are also impacted. Draper, a multi-liganded receptor, binds PE-rich lipid membranes through its EMI and NIM domains, possibly activating an auto-regulatory feedback loop that maintains basal phagocytic activity.^98,110–112 Thus, reduced PE-lipid availability under an HSD could directly impair basal Draper levels independent of LpR signaling. Recent studies also suggest that systemic lipid signals, especially ApoB, regulate phagocytic receptor expression in Drosophila immune cells.¹¹³ Thus, ApoB-Lpp-mediated lipid signaling could represent a conserved mechanism regulating phagocytic competence across cell types.

In conclusion, our findings reveal direct communication from adipocyte lipid metabolism and mitochondrial state to glial phagocytic function via ApoB lipoproteins. Given clinical correlations linking obesity and cognitive decline,^1–4 this work provides critical insight into how dysfunctional adipocyte metabolism can impact brain health.

Limitations of the study

While our study provides strong evidence linking adipocyte mitochondrial metabolism and ApoB-LpR1 signaling to glial phagocytic function, several limitations should be considered. First, although Drosophila is a powerful genetic model with conserved metabolic pathways, translation of these findings to mammalian systems, particularly humans, requires further validation. Second, the precise molecular identities of lipid metabolites and reactive intermediates responsible for remote signaling between adipocytes and glia remain unclear. The proposed mediators (PE-rich lipids and ketone bodies) require explicit validation in future in vivo studies. Finally, while our findings demonstrate correlations between mitochondrial morphology and glial signaling, additional work is necessary to pinpoint specific factors that communicate the adipocyte mitochondrial metabolic state to the brain, particularly under obesogenic stress conditions.

RESOURCE AVAILABILITY

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Akhila Rajan (akhila@fredhutch.org).

Materials availability

All reagents generated in this study are available from the lead contact without restriction.

Data and code availability

• This paper does not report original code.

• All data generated or analyzed during this study are included in this published article and its supplemental information. The source data underlying the metabolite profiling has been deposited in the General Data Repository Fig Share. The links to the public datasets are provided in the key resources table.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Chicken anti-GFP	Abcam	Cat#ab13970; RRID:AB_300798
Mouse anti-Draper	Developmental Studies Hybridoma Bank (DSHB)	RRID: AB_2618105 (clone 5D14)
Mouse anti-Draper	DSHB	RRID: AB_2618106 (clone 8A1)
Rabbit anti-Apo-II	Rajan Lab	Brent and Rajan114
Rabbit anti-Dilp5	Rajan Lab	Brent and Rajan114
Rabbit mito-mCherry	Abcam	Cat#ab167453, RRID: AB_2571870
Donkey anti-Chicken Alexa 488	Jackson ImmunoResearch	Cat#703-545-155; RRID: AB_2340375
Donkey anti-Rabbit Alexa 594	Jackson ImmunoResearch	Cat#711-585-152; RRID: AB_2340621
Donkey anti-Rabbit Alexa 488	Jackson ImmunoResearch	Cat#711-545-152; RRID: AB_2313584
Donkey anti-Mouse Alexa 594	Jackson ImmunoResearch	Cat#715-585-150; RRID: AB_2340854
Mouse anti-Tubulin	Sigma-Aldrich	Cat#T5168; RRID: AB_477579
Chemicals, peptides and recombinant proteins
Pyruvate	Agilent	Cat#103578-100
HALT Protease Inhibitor Cocktail	ThermoFisher	Cat#87786
5x RIPA Buffer	ThermoFisher	Cat#J62524.AD
Laemmli Sample Buffer	Bio-Rad	Cat#161047
TCEP Bond Breaker	Thermo Scientific	Cat#77720
Precision Plus Protein™ Dual Color Standards	Bio-Rad	Cat#1610374
Starting Block™ Blocking Buffer	Thermo Scientific	Cat#37538
Enhanced Chemiluminescence (ECL) detection system	Abcam	Cat#AB65623
TriReagent	Sigma-Aldrich	Cat#T9424
iScript RT supermix	Bio-Rad	Cat#1708841
ssoAdvanced SYBR green master mix	Bio-Rad	Cat#1725270
Critical commercial assays
Free Fatty Acid Assay Kit	Sigma-Aldrich	Cat#MAK466
Pierce™ BCA Protein Assay Kit	Thermo Scientific	Cat#23228
Direct-zol RNA microprep kit	Zymo Research	Cat#R2060
Deposited data
Whole Fly Lysate Aqueous Metabolomic datasets for ND, HSD (7 and 14 days)	Figshare	https://doi.org/10.6084/m9.figshare.28765562.v1
Hemolymph Aqueous Metabolomic datasets for ND, HSD (7 and 14 days)	Figshare	https://doi.org/10.6084/m9.figshare.28765067.v2
Hemolymph Lipidomics datasets for ND, HSD (14 days)	Figshare	https://doi.org/10.6084/m9.figshare.28765064.v1
Experimental models: Organisms/strains
UAS-Plin1:GFP	Kuhnlein Lab	Mathias115
Ldh-GFP	Tennessen Lab	Brent and Rajan114
ppl-Gal4	Leopold Lab	N/A
Lpp-Gal4	Eaton Lab	N/A
UAS-mitoTIMER	Bloomington Drosophila Stock Center (BDSC)	BDSC#57323
UAS-mitoPyronicSF	BDSC	BDSC#94536
UAS-Fis1-RNAi	BDSC	BDSC#63027
UAS-SOD2-RNAi	BDSC	BDSC#36871
UAS-Apolpp-RNAi	BDSC	BDSC#33388
UAS-LpR1-RNAi	BDSC	BDSC#27249
UAS-LpR2-RNAi	BDSC	BDSC#31150
UAS-Catalase	BDSC	BDSC#24621
UAS-whd-RNAi	BDSC	BDSC#34066
UAS-Acat1-RNAi	BDSC	BDSC#51785
Or22a-mCD8GFP	BDSC	BDSC#52620
Ensheathing glia-Gal4	BDSC	BDSC#39157
UAS-Luc-RNAi	BDSC	BDSC#31603
UAS-bmm-RNAi	Vienna Drosophila Resource Center (VDRC)	VDRC#37880
upd2-del	Zeidler Lab	N/A
Oligonucleotides
Primers	See Table S3	See Table S3
Software and algorithms
ImageJ	NIH	https://imagej.nih.gov/ij/
GraphPad PRISM	GraphPad Software Incorporated	https://www.graphpad.com/scientific-software/prism/
AB Sciex Analyst 1.6.3	AB Sciex	https://sciex.com/products/software/analyst-software
AB Sciex MultiQuant 3.0.3	AB Sciex	https://sciex.com/products/software/multiquant-software
Other
Zeiss LSM 800 Confocal Microscope	Zeiss	N/A
Leica Stellaris Confocal Microscope	Leica Microsystems	N/A

STAR★METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Drosophila melanogaster strains and husbandry

The Drosophila strains used in this study are listed in the key resources table. Flies were housed at 25°C, and experiments utilized 7–10-day-old adult male flies. The flies’ diet consisted of 15 g yeast, 8.6 g soy flour, 63 g corn flour, 5 g agar, 5 g malt, and 74 mL corn syrup per liter. Diet formulations were normal diet (ND) or high sugar diet (HSD) with added sucrose (300 g/L).

METHOD DETAILS

Antennal nerve injury

As adapted from prior studies by Freeman & Logan et al., ^{44,51,52,116–118} Briefly, flies were anesthetized using CO₂, and antennal nerve injury was accomplished by unilaterally removing the third antennal segment using forceps. Flies were then placed back into the vial post-injury until dissection 24 h later.

Immunostaining

Adult brains and fat bodies were dissected in ice-cold PBS, fixed overnight in 0.8% paraformaldehyde (PFA) in PBS at 4°C. Fixed brains were washed 5 times in PBS with 0.5% BSA and 0.5% Triton X-100 (PAT), blocked for 1 h in PAT +5% NDS, incubated overnight with primary antibodies, followed by 5 washes with PAT, re-blocking for 30 min, and secondary antibody incubation for 2–4 h at RT. Finally, the brains were washed 3 times in PAT and mounted using Slow fade gold antifade.

Image analysis

Images were acquired with a Zeiss LSM 800 and Leica Stellaris confocal system. Images were analyzed using ImageJ. All images within each experiment were acquired with the same confocal settings. z stack summation projections at 0.5 μm intervals were generated, and a region of interest (indicated in the fig) was used to measure the integrated density values of each fluorescent tag. A maximum-intensity projection of Z-stacks that covered the full depth of the antennal lobe was used for ImageJ analysis to measure mitochondrial morphology.

To measure the average mitochondrial circularity major and minor axes, we adapted methods from prior work.¹¹⁹ Maximum-intensity projection was inverted and automatic threshold completed before applying the ‘analyze particles’ function to measure the average mitochondrial circularity. Mitochondrial elongation (Aspect ratio) was measured from the same maximum intensity projection using the ImageJ plugin developed by work from Chaudry et al.,¹¹⁹ The size and dimensions of all ROIs were maintained consistently throughout each experiment.

Dilp5 levels were quantified using z stack summation projections to capture the full depth of the IPCs. A region of interest (ROI) around the IPCs was manually outlined with the freehand tool, and integrated density values were then measured.

For co-localization analysis, confocal images were acquired using Zeiss LSM 800 with a 63× oil immersion objective at 2X zoom. Co-localization analysis was performed using Mander’s overlap coefficient (MOC) to quantify the degree of signal overlap between ApoB and Draper. One measurement per z-slice for each image (N = 11) was performed, to ensure that the measurement is not affected by tissue depth and the resulting change in intensity.

MitoPyronic sensor live imaging

The pyruvate sensor expressing fly stock UAS-Mito-PyronicSF were procured from BDSC (Stock #94536). Flies expressing the sensor in adult fat were aged for 3 weeks on either normal diet (ND) or high sugar diet (HSD) at 25°C. Drosophila abdomens were dissected in HL3 buffer (70mM NaCl, 5mM KCl, 20mM MgCl₂, 10mM NaHCO₃, 115mM sucrose, 5mM trehalose, 5mM HEPES; pH 7.1), and mounted on a glass-bottom FluoroDish (WPI, Cat#FD3510–100). Image stacks were acquired with a Zeiss LSM 800 using a 20X air lens (Plan-Apochromat 20x/0.8) at 0.5μm intervals. Pyruvate (Agilent, Cat#103578–100) was superfused manually from the side, such that the dissected tissue was not disturbed. Imaging was done within 60 s of addition of each concentration of pyruvate. GFP signal intensity in the acquired images was measured using ImageJ.

Transmission electron microscopy (TEM)

Adipocyte explants were prepared for transmission electron microscopy (TEM) by fixation in a solution containing 1.25% formaldehyde, 2.5% glutaraldehyde, and 0.03% picric acid in 0.1M cacodylate buffer (pH 7.4). Following fixation, the samples were stained with osmium tetroxide and uranyl acetate, then progressively dehydrated through a graded ethanol series and propylene oxide. For embedding, the samples were infiltrated with a 1:1 mixture of EPON resin (Westlake) and propylene oxide at 4°C for 16 h before polymerization in pure EPON resin at 60°C for 24 h. The embedded samples were sectioned using standard ultramicrotomy techniques and examined with a JEOL 1200EX transmission electron microscope.

Quantitative PCR (qPCR)

Thirty flies were dissected in RNAlater, then placed in 30 μL of TriReagent and a scoop of beads in a 1.5 mL safelock tube. The abdominal segments dissected from flies were homogenized using a bullet blender. RNA was then isolated using a Direct-zol RNA microprep kit following the manufacturer’s instructions. Isolated RNA was synthesized into cDNA using the Bio-Rad iScript RT supermix for RT-qPCR, and qPCR was performed using the Bio-Rad ssoAdvanced SYBR green master mix. Primers were designed using DRSC’s FlyPrimer Bank¹²⁰ and are listed in Table S3. Relative mRNA quantification was performed using the comparative CT method and normalized to alpha-tubulin mRNA expression. Three technical replicates were used for each gene.

Western blotting

10–12 flies were homogenized in triplicates using 1mm zirconium beads (Cat#ZROB10, Next Advance) in a Bullet Blender Tissue homogenizer (Model BBX24, Next Advance) in 250 μL of a mixture containing 200 μL of diH2O and 2 μL of HALT Protease Inhibitor Cocktail (ThermoFisher, Cat#87786), along with 50 μL of 5x RIPA Buffer (ThermoFisher, Cat# J62524.AD) in 1.5 mL LoBind Eppendorf Tubes (Eppendorf, Cat#022431081). Samples were incubated on a rocking platform for 30–60 min at 4°C. Subsequently, the tubes were centrifuged at 10,000 × g at 4°C for 10 min, and the supernatant was carefully collected into separate 1.5 mL tubes. Protein quantification was performed on all three lysate samples using the Pierce BCA Protein Assay Kit (ThermoScientific Cat#23228, Cat#1859078). Samples were then prepared for western blotting by adjusting the concentration to 2 μg/μL using 4x Laemmli Sample Buffer (Bio-Rad Cat#161047) and TCEP Bond Breaker (Thermo Scientific Cat#77720), followed by heating at 95°C for 5–10 min.

For preparing protein lysates from adult Drosophila brains, 5 μL of lysis buffer (0.1% SDS in PBS with HALT) was added to a PCR tube. 10 brains were dissected and placed in the lysis buffer. 10 μL of Laemmli buffer was added per tube. Samples were boiled at 95°C for 10 min.

Equal amounts of protein (10 μg per lane for whole fly lysates; 12μL per lane for brain lysates) were loaded onto SDS-PAGE gels (Bio-Rad, Cat#4561103). Precision Plus Protein Dual Color Standards (Bio-Rad, Cat#1610374) were included as molecular weight markers. Following electrophoresis, proteins were transferred from the gel to a nitrocellulose membrane using a wet transfer system. The membrane was blocked in Starting Block Blocking Buffer (Thermo Scientific, Cat#37538) in TBS-T (Tris-buffered saline with 0.1% Tween 20) for 1 h at room temperature with gentle agitation to prevent non-specific binding. Primary antibodies against the target protein and loading control were diluted in blocking buffer and TBS-T; Mouse anti-Draper (1:250; DSHB 8A1 RRID: AB_2618106), Rabbit anti-apo2 (1:2500; generated by the Rajan Lab), Mouse anti-Tubulin (1:4000; Sigma Cat#T5168). and incubated with the membrane overnight at 4°C. Membranes were washed and incubated with HRP (horseradish peroxidase)-conjugated secondary antibodies diluted in blocking buffer and TBS-T for 1 h at room temperature. Protein bands were visualized using an enhanced chemiluminescence (ECL) detection system (Abcam, Cat #AB65623). Images were captured using a chemiluminescence imaging system.

Free fatty acid measurements

We used the colorimetric Free Fatty Acid Assay Kit (Sigma Aldrich, Cat#MAK466) and followed the manufacturer’s instructions. Briefly, 8 flies were homogenized in 60μL of 5% iso-propanol and 2% Triton X- solution using a pestle in 1.5 mL LoBind Eppendorf Tubes (Eppendorf, Cat#022431081). Samples were centrifuged at 15,000g for 1 min at room temperature. The supernatant was carefully removed to separate 1.5 mL tubes. Triplicates of 10μL (neat) and 5μL (1:2 dilution, made up to 10μL in diH₂O) were used per well of a clear 96-well plate (Corning, Cat#3585). 90μL of working reagent was added (as per manufacturer’s instructions). Plates were incubated at room temperature for 30min. Absorbance (OD) was measured at 570nm in a microplate reader.

Hemolymph extraction

For hemolymph extraction, 30 adult flies were anesthetized on a CO2 pad and punctured in the thorax region with a tungsten needle. The flies were then transferred into a 0.5 mL Eppendorf tube with holes made in the bottom using an 18G needle. This 0.5 mL tube was placed inside a 1.5 mL Eppendorf tube containing 30 μL of PBS and centrifuged at 5000 RPM for 5 min. The samples were then flash-frozen in liquid nitrogen until ready to use.

Metabolomics sample preparation

Aqueous metabolites for targeted LC-MS profiling of whole flies and hemolymph samples were extracted using a protein precipitation method similar to the one described elsewhere^121,122 by the Seattle Northwest Metabolomics Research Center (NW-MRC).

Whole fly Samples:

Whole adult male flies were frozen in liquid nitrogen after 7 or 14 days on a normal diet or HSD. Ten flies were used per biological sample, and 3 biological replicates were used for each diet and time point. Samples were first homogenized in 200 μL purified deionized water at 4°C. Then 800 μL of cold methanol containing 124 μM 6C13-glucose and 25.9 μM 2C13-glutamate was added (reference internal standards were added to the samples to monitor sample prep). Afterward, samples were vortexed, stored for 30 min at −20°C, sonicated in an ice bath for 10 min, centrifuged for 15 min at 14,000 rpm and 4°C, and then 600 μL of supernatant was collected from each sample. Lastly, recovered supernatants were dried on a SpeedVac and reconstituted in 0.5 mL of LC-matching solvent containing 17.8 μM 2C13-tyrosine and 39.2 3C13-lactate (reference internal standards were added to the reconstituting solvent to monitor LC-MS performance). Samples were transferred into LC vials and placed into a temperature-controlled autosampler for LC-MS analysis.

Hemolymph Samples:

30 flies were used per biological sample, and 3 biological replicates were used for each diet and timepoint. Samples were thawed at 4°C for 60 min and vortexed for 10 s 50uL of each sample was transferred to a 2 mL Eppendorf tube, and 50 μL of 50%MeOH/50%Water containing 30 stable isotope-labeled internal standards (SILISs) was added. Afterward, 250 μL of cold MeOH containing two additional SILISs was added to each sample. Samples were vortexed, stored for 30 min at −20°C, centrifuged for 15 min at 14,000 rpm and 4°C, and then 250 μL of supernatant was collected from each sample. Lastly, recovered supernatants were dried on a SpeedVac and reconstituted in 0.5 mL of LC-matching solvent containing two more SILISs. 34 SILISs were added to the samples in various sample prep steps to monitor sample prep, assay performance, and determine absolute concentrations for the metabolites that had corresponding SILISs.

Metabolomics

Targeted LC-MS metabolite analysis was performed on a duplex-LC-MS system composed of two Shimadzu UPLC pumps, CTC Analytics PAL HTC-xt temperature-controlled auto-sampler, and AB Sciex 6500+ Triple Quadrupole MS equipped with ESI ionization source. UPLC pumps were connected to the auto-sampler in parallel and could perform two chromatography separations independently from each other. Each sample was injected twice on two identical analytical columns (Waters XBridge BEH Amide XP), where separations were performed in hydrophilic interaction liquid chromatography (HILIC) mode. While one column was performing separation and MS data acquisition in ESI+ ionization mode, the other column was getting equilibrated for sample injection, chromatography separation, and MS data acquisition in ESI− mode. Each chromatography separation was 18 min (total analysis time per sample was 36 min). MS data acquisition was performed in multiple-reaction-monitoring (MRM) mode. LC-MS system was controlled using AB Sciex Analyst 1.6.3 software. Measured MS peaks were integrated using AB Sciex MultiQuant 3.0.3 software. Up to 158 metabolites (plus 4 spiked standards) were measured across the fly samples study set, and up to 148 metabolites (plus 34 SILISs) were measured in the hemolymph sample set. For the hemolymph set, absolute concentrations of 30 metabolites were determined. In addition to the two study samples, two sets of quality control (QC) samples were used to monitor the assay performance and data reproducibility. One QC [QC(I)] was a pooled human serum sample used to monitor system performance, and the other QC [QC(S)] was pooled study samples. This QC was used to monitor data reproducibility. Each QC sample was injected per every 10 study samples. The median CV for the fly set was 2.8%, while for the blood samples was 11.1%.

Lipidomics

Frozen hemolymph samples from ND and HSD-fed flies were sent to the Northwest Metabolomics Research Center (NW-MRC) for targeted quantitative lipid profiling using the Sciex 5500 Lipidyzer as per methods established by NW-MRC.¹²³ The materials used include LC-MS grade methanol, dichloromethane, and ammonium acetate, all sourced from Fisher Scientific (Pittsburgh, PA). HPLC grade 1-propanol was obtained from Sigma-Aldrich (Saint Louis, MO). Milli-Q water was produced using an in-house Ultrapure Water System by EMD Millipore (Billerica, MA). The Lipidyzer isotope-labeled internal standards mixture, which contained 54 isotopes from 13 different lipid classes, was acquired from Sciex (Framingham, MA). HCER and LCER were not detected in Drosophila hemolymph, so we report 11 different lipid classes.

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical analysis was performed with GraphPad PRISM (GraphPad Software Incorporated). Data are expressed as the mean ± standard deviation (SD). Data normality was assessed via Shapiro-Wilk tests. Parametric analyses included two-tailed unpaired t-tests with Welch’s correction or two-way ANOVA with Holm-Sidak correction. Non-parametric analysis employed Mann-Whitney tests. Statistical significance was defined as p < 0.05.

ADDITIONAL RESOURCES

FlyBase (release FB2023_05): https://flybase.org/.

Supplementary Material

SUPPLEMENTAL INFORMATION

Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2025.115704.

Highlights.

• Obesogenic diet triggers a starvation-like metabolic response in adipose tissue

• Dysfunctional adipocyte mitochondrial and lipid metabolism impairs glial phagocytic function

• Adipocyte ApoB functions as a novel regulator of glial phagocytic competence

• LpR1, in ensheathing glia, is essential for proper glial response to axonal injury