Male C57BL/6J mice (6-week-old) were purchased from Gempharmatech (Nanjing, Jiangsu, China), and maintained under standard housing conditions (12-h light/dark cycle, 22 $\pm$ 1 °C, 50–60% humidity) with ad libitum access to food and water. The use of male mice in this initial study was to control for potential confounding associated with hormonal fluctuations during the female estrous cycle. After a one-week acclimatization period, the mice were randomly assigned to two experimental groups. The control group (Con) was fed a standard chow diet (10% of kcal from fat), while the diet-induced obese (DIO) group was maintained on an HFD (D12492). The fat in the D12492 diet is provided primarily by lard and soybean oil. This formulation was established to reliably induce obesity and metabolic syndrome, with 60% of kcal derived from fat (Research Diets, New Brunswick, NJ, USA; catalog No. D12492). Mice were fed D12492 for 20 weeks to induce obesity and depressive-like phenotypes. This period extended beyond the typical duration for induction of weight gain (8–12 weeks), ensuring the development of chronic metabolic dysfunction and stable, robust depressive-like behaviors associated with sustained neuroinflammation and microglial activation, as previously described [12, 13]. Body weight and food intake were recorded weekly throughout the experimental period. Following the dietary intervention, a series of behavioral tests were conducted sequentially to assess depressive-like behaviors, including the sucrose preference test, tail suspension test, and forced swim test. Upon the completion of behavioral assessments, all mice were deeply anesthetized with sodium pentobarbital (50 mg/kg, i.p.; cat. no. P3761, Sigma-Aldrich, St. Louis, MO, USA). The depth of anesthesia was confirmed by the absence of response to a toe-pinch reflex. Subsequently, mice were transcardially perfused with phosphate-buffered saline (PBS) while under deep anesthesia. Brain tissues were promptly harvested and processed for subsequent histological, biochemical, and lipidomic analyses. The hippocampus regions were microdissected for glial cell culture, or immediately frozen in liquid nitrogen and stored at –80 °C for further analysis (Fig. 1). The cell line has been validated by short tandem repeat authentication and surface marker analysis. Mycoplasma Contamination Testing was performed using a polymerase chain reaction (PCR)-based mycoplasma detection kit. The tests yielded negative results both at the beginning and at the end of the experimental period, confirming the absence of mycoplasma contamination.

Fig. 1.

Experimental design and timeline. Mice were fed a control or high-fat diet (HFD) for 20 weeks, followed by behavioral tests (Open Field Test, Sucrose Preference Test, Tail Suspension Test, Forced Swim Test). Hippocampal tissues were then collected for histological, biochemical, and lipidomic analyses.

To comprehensively evaluate the systemic metabolic phenotype of DIO mice, their body weight was monitored throughout the 27-week study period. Following the behavioral tests, mice were fasted for 6 h, after which fasting blood glucose was measured from the tail vein using an Elite Glucometer (Bayer, Mishawaka, IN, USA). Serum was also collected via retro-orbital bleeding for the quantification of fasting insulin levels using a mouse-specific enzyme-linked immunosorbent assay (ELISA) kit (JL48252, JONLNBIO, Shanghai, China). A glucose tolerance test (GTT) was then conducted by intraperitoneal injection of a glucose solution (2 g/kg body weight) into mice that had fasted for 6 hours. Blood glucose levels were monitored at 0, 15, 30, 60, 90, and 120 minutes post-injection. Total glycemic excursion was quantified as the area under the curve (AUC) using the trapezoidal rule.

To assess depressive-like behaviors, all mice underwent a battery of well-validated behavioral tests with a 48-h interval between each test to minimize potential carryover effects. The open field test was conducted to evaluate general locomotor activity and anxiety-like behavior. Each mouse was placed individually in a square arena (50 $\times$ 50 cm) and allowed to explore freely for 5 minutes under dim lighting conditions (20 lux). The number of rearing episodes (upright posture on hind legs) was quantified over the entire 5-minute session. Rearing is a validated measure of exploratory drive and risk-taking behavior, with a reduction in this behavior reflecting a depressive-like anhedonic state and/or heightened anxiety [14]. The number of rearing episodes was quantified over a 5-minute period. The sucrose preference test was performed to measure anhedonia, a core symptom of depression. Mice were first acclimated for 48 h to two bottles, one containing a 1% sucrose solution and the other tap water. Following food and water deprivation for 12 h, mice were individually housed and given free access to both bottles for 4 h [15]. Fluid intake was determined by weighing the bottles before and after the test. To ensure data validity, only mice with a total fluid intake exceeding 0.5 mL during the 4-h test were included in the analysis. All mice in the study met this criterion. Sucrose preference was calculated using the following formula: sucrose preference (%) = [sucrose intake / (sucrose intake + water intake)] $\times$ 100%. The tail suspension test was conducted to assess depression-like behavior. Each mouse was suspended by its tail using adhesive tape attached to a horizontal bar, with its head positioned approximately 30 cm above the surface. The test lasted 6 minutes, and the total duration of immobility during the final 4 minutes was recorded. For the forced swimming test, mice were placed individually into cylindrical containers (height: 25 cm; diameter: 18 cm) filled with water (23–25 °C) to a depth of 15 cm. After a 2-minute acclimatization period, the immobility time was measured over the subsequent 4 minutes. All behavioral sessions were video-recorded and analyzed by experimenters who were blinded to the group assignments.

After the behavioral tests, mice were deeply anesthetized and transcardially perfused with PBS followed by 4% paraformaldehyde (PFA, Shanghai Maokang Biotechnology Co., Ltd, Shanghai, China). The brains were post-fixed overnight at 4 °C and subsequently cryoprotected in 30% sucrose solution. Coronal hippocampal sections (20 µm thick) were prepared using a freezing microtome. For hematoxylin and eosin (H&E) staining, sections were first incubated in H&E solutions (catalog No. K1142, APExBIO, Houston, TX, USA), dehydrated through an alcohol series, and then mounted. For Nissl staining, sections were incubated in Cresyl Violet Stain Solution (1%) (catalog No. ab246817, Abcam, Cambridge, UK), differentiated in alcohol, and a coverslip was added. To assess hippocampal neuronal damage, the number of intact neurons per mm² was quantified across three non-overlapping fields in the ventral CA1 subregion. Analyses were performed on sections located between approximately AP –2.9 mm and AP –4.2 mm from Bregma (Paxinos and Franklin’s mouse brain atlas), with a representative imaging field centered at AP –3.5 mm, ML $\pm$ 2.5 mm, DV –3.0 mm. Neuronal integrity was assessed using established morphological criteria. Healthy neurons exhibited a large, pale nucleus with a visible nucleolus and abundant Nissl substance. Damaged neurons were identified by characteristic pyknosis (nuclear shrinkage and hyperchromasia) and a marked loss of Nissl bodies (chromatolysis).

Upon completion of the behavioral tests and metabolic phenotyping, mice were deeply anesthetized with sodium pentobarbital (50 mg/kg, i.p.). Blood samples were collected via retro-orbital bleeding. The blood was allowed to clot for 30 min at room temperature and then centrifuged at 3000 $\times$g for 15 min at 4 °C to obtain serum. Subsequently, mice were transcardially perfused with ice-cold PBS, and the brains were rapidly removed. Hippocampi were dissected on a chilled plate, snap-frozen in liquid nitrogen, and stored at –80 °C until analysis. For cytokine analysis, hippocampal tissues were homogenized in cold PBS containing 1$\times$ protease inhibitor cocktail. The homogenates were centrifuged at 12,000 $\times$g for 20 min at 4 °C, and the supernatants were collected. The levels of pro-inflammatory cytokines interleukin-6 (IL-6, PI326, Beyotime, Shanghai, China), interleukin-1$\beta$ (IL-1$\beta$, PI301, Beyotime), and tumor necrosis factor-$\alpha$ (TNF-$\alpha$, PT512, Beyotime) in hippocampal homogenates and monoamine neurotransmitters (5-HT, NA) in serum were quantified using commercially available ELISA kits, as per the manufacturers’ instructions. All samples were analyzed in duplicate, and absorbance was measured at 450 nm using a microplate reader (ELx800, BioTek, Santa Clara, CA, USA).

Total RNA was extracted from samples using TRIzol reagent (catalog No. 15596018CN, Sigma-Aldrich). Complementary DNA (cDNA) was then synthesized from the isolated RNA with a PrimeScript RT reagent kit (catalog No. RR037A, Takara, Kusatsu, Japan). Quantitative real-time PCR (qPCR) was performed on a 7300 Plus Real-Time PCR System (catalog No. JM1966-018481, Thermo Fisher Scientific, Waltham, MA, USA) using SYBR Premix Ex Taq I (Takara, RR820A). The 20 µL reaction mixture contained 1 µL of cDNA template, 10 µL of PCR master mix, 5 pmol of each primer, and nuclease-free water (catalog No. AM9937, Ambion™, Waltham, MA, USA). The PCR protocol involved an initial denaturation at 95 °C for 13 s, followed by 40 cycles of 94 °C for 10 s and 60 °C for 20 s. Gene expression levels were calculated using the comparative $\mathrm{\Delta }$$\mathrm{\Delta }$Ct method, with GAPDH serving as the internal reference control. All primer sequences are shown in Table 1.

Glial cell-enriched isolates were obtained from the brains of 16-week-old C57BL/6J mice using a combined enzymatic and mechanical dissociation protocol. In brief, mice were transcardially perfused with ice-cold PBS. The brains were rapidly excised, and the cortices and hippocampi were carefully dissected. The tissues were enzymatically digested in Hank’s balanced salt solution (HBSS) (catalog No. 14185052, Thermo Fisher Scientific) supplemented with 2 mg/mL papain (catalog No. LS003126, Worthington, Lakewood, NJ, USA) and 100 U/mL DNase I (catalog No. D5025, Sigma-Aldrich) for 30 minutes at 37 °C. The resulting cell suspension was triturated, filtered through a 70-µm cell strainer, and centrifuged on a 30%/70% Percoll (catalog No. 17089101, GE Healthcare, Chicago, IL, USA) density gradient. Fractions enriched with glial cells were harvested and used as glial cell-enriched isolates from the interphase layer, washed, and subsequently seeded in DMEM/F-12 complete medium for further culture.

For immunofluorescence staining, brain sections or cultured primary microglia were fixed with 4% PFA, permeabilized with 0.3% Triton X-100 (catalog No. T8787, Sigma-Aldrich), and blocked with 5% bovine serum albumin (BSA, catalog No. A3156-25G, Sigma-Aldrich). Sections or cells were incubated overnight at 4 °C with the following primary antibodies: anti-ionized calcium-binding adapter molecule 1 (Iba1) (1:100, catalog No. ab283319, Abcam) and anti-inducible nitric oxide synthase (iNOS, 1:100, catalog No. ab210823, Abcam), or appropriate combinations for co-staining. Bodipy™ 493/503 (1:600, Thermo Fisher Scientific, Cat. No. D3922) was applied for the detection of LDs. Following three washes with PBS, samples were incubated with Alexa Fluor® 488-conjugated goat anti-rabbit IgG (1:500, A11017, Thermo Fisher Scientific) and Alexa Fluor® 594-conjugated goat anti-mouse IgG (1:500, A-11020, Thermo Fisher Scientific) secondary antibodies for 1 h at room temperature. Nuclei were counterstained with 4’,6-diamidino-2-phenylindole (DAPI) (catalog No. D9542-1MG, Sigma-Aldrich). Fluorescent images were acquired using a Nikon A1R HD25 confocal microscope (Nikon Corporation, Tokyo, Japan) equipped with a Plan Apo $\lambda$ 20x/0.75 NA objective and CFI Plan Apo $\lambda$ 60x/1.40 NA oil immersion objective. The following laser and filter sets were used: DAPI (excitation 405 nm, emission 450/50 nm), Alexa Fluor 488 (excitation 488 nm, emission 525/50 nm), and Alexa Fluor 594 (excitation 561 nm, emission 595/50 nm). Images were captured at a resolution of 1024 $\times$ 1024 pixels. For co-localization analysis, Z-stacks were collected with a 1.0 µm step size, and the pinhole was set to 1 Airy Unit for all channels. To ensure comparability, the imaging settings (laser power, gain, offset) were kept identical for all samples within the same staining batch.

Isolated glial cell-enriched isolates were fixed with 4% PFA for 15 minutes and rinsed with PBS. Cells were subsequently incubated with 60% isopropanol (catalog No. 563935, Sigma-Aldrich) for 1 minute to permeabilize the membranes, followed by staining with freshly prepared Oil Red O working solution (catalog No. O0625, Sigma-Aldrich) for 15 minutes at room temperature. After staining, cells were thoroughly washed with distilled water and counterstained with hematoxylin (catalog No. H9627, Sigma-Aldrich) for 1 minute. Brightfield images were captured using a Nikon Eclipse Ni light microscope (ECLIPSE Ni-L, Nikon Corporation). The extent of lipid droplet accumulation was quantified by measuring the integrated optical density (IOD) of Oil Red O staining across five randomly selected fields per sample using ImageJ software (Version 1.41, NIH, Bethesda, MD, USA). ImageJ analysis of immunofluorescence involved automated thresholding (“Analyze Particles”) for quantification of the Percent Area/Integrated Density and plugin-based co-localization (Manders’ Coefficients), with all analyses performed blinded.

For the analysis of apoptosis, glial cell-enriched isolates were stained using the Annexin V-FITC/PI Apoptosis Detection Kit (catalog No. 556547, BD Biosciences, San Jose, CA, USA) as per the manufacturer’s instructions. Briefly, cells were resuspended in binding buffer and incubated with Annexin V-FITC (catalog No. 51-65874X, BD Biosciences) and propidium iodide (PI) for 15 minutes in the dark. The percentage of apoptotic cells was quantified using a BD FACSVerse flow cytometer (catalog No. 651155; BD Biosciences). For the measurement of intracellular reactive oxygen species (ROS), cells were incubated with 10 µM DCFH-DA (catalog No. S0033S, Beyotime Biotechnology, Shanghai, China) at 37 °C for 30 minutes, washed with PBS, and analyzed immediately by flow cytometry. Fluorescence intensity was detected in the FITC channel. All data were analyzed using FlowJo software (BD Biosciences).

To assess the neurotoxic effects of microglia on hippocampal neurons in a transwell co-culture system, neuronal viability and cytotoxicity were evaluated using the CCK-8 and LDH assays, respectively. For the CCK-8 assay, hippocampal neurons were seeded in the lower chamber of 24-well plates. After 24 h of co-culture with microglia in the upper chamber, the CCK-8 reagent (catalog No. CK04, Dojindo Laboratories, Kumamoto, Japan) was added to each well, and the plates were incubated for 2 h at 37 °C. Absorbance was then measured at 450 nm using a microplate reader. Neuronal viability was expressed as a percentage relative to the control group. For the LDH release assay, culture supernatant was collected from the lower chamber after 24 h of co-culture. LDH activity was measured using a Cytotoxicity Detection Kit (catalog No. 11644793001, Roche Diagnostics, Mannheim, Germany) according to the manufacturer’s instructions. Absorbance was read at 490 nm. The percentage of LDH release was calculated as follows: (LDH activity in supernatant / total LDH activity) $\times$ 100%.

Glial cell‑enriched isolates were fixed with 2.5% glutaraldehyde (catalog No. G916054-25ml, Macklin, Shanghai, China) overnight at 4 °C and subsequently post-fixed in 1% osmium tetroxide (catalog No. BD9014, Exonlab, Guangzhou, China). Following dehydration through a graded ethanol series, the cells were embedded in EPON resin (catalog No. 02334-500, Polysciences, Warrington, PA, USA). Ultrathin sections (70 nm) were prepared and double-stained with uranyl acetate (catalog No. 22405, Electron Microscopy Sciences, Hatfield, PA, USA) and lead citrate (catalog No. 17800, Electron Microscopy Sciences). The morphology of LDs was examined using a transmission electron microscope (catalog No. HT7800, Hitachi High-Tech Corporation, Tokyo, Japan) at 80 kV. The number and average diameter of LDs were quantified from randomly selected micrographs using ImageJ software.

Lipids were extracted from glial cell-enriched isolates using a methyl-tert-butyl ether (MTBE)/methanol-based extraction method. Lipid profiling was carried out using ultra-high-performance liquid chromatography (UHPLC, Waters Corporation, Milford, MA, USA, ACQUITY UPLC I-Class) coupled with tandem mass spectrometry (MS/MS, Thermo Scientific, Q-Exactive HF-X). Chromatographic separation was performed on a C18 column (catalog No. 186005296, Waters Corporation) with a gradient elution program involving acetonitrile and isopropanol. Mass spectrometry data were acquired in both positive and negative ionization modes.

Data processing was conducted using LipidSearch™ software (v4.2) (catalog No. OPTON-30880, Thermo Fisher Scientific) for lipid identification and quantification. Multivariate statistical analyses, including principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA), were performed using SIMCA-P (v16.0). Significantly altered lipid species were selected based on a variable importance in projection (VIP) value $>$1.0 and a p-value 0.05.

HFD induced a significant increase in the body weight of DIO mice, as well as elevated glucose and insulin levels. After 20 weeks of HFD consumption, the body weight of mice in the DIO group was significantly higher than in the control group (p 0.001, Fig. 2A). Furthermore, fasting blood glucose levels in DIO mice were markedly increased compared to controls (p = 0.001, Fig. 2B), as well as the fasting insulin levels (p 0.001, Fig. 2C). Following an intraperitoneal glucose challenge, blood glucose levels in DIO mice were substantially higher at all time points measured (0, 15, 30, 60 and 120 minutes), resulting in a markedly elevated glycemic curve (Fig. 2D). Quantitative analysis of the total glycemic response, represented by the AUC, confirmed a significant increase in the DIO group (p 0.001, Fig. 2E). These data clearly indicate the development of glucose intolerance in DIO mice.

Fig. 2.

DIO mice exhibit systemic metabolic dysfunction and glucose intolerance. (A) Terminal body weight of control (Con) and diet-induced obese (DIO) mice after 20 weeks of dietary intervention (n = 8). (B) Fasting blood glucose (FBG) levels (n = 8). (C) Fasting serum insulin concentrations (n = 8). (D) Blood glucose levels during the intraperitoneal glucose tolerance test (GTT) (n = 8). (E) Quantification of the total glycemic response shown as the area under the curve (AUC) for the GTT (n = 8). **p 0.005, ***p 0.001.

To investigate the impact of diet-induced obesity on depression-related pathology, mice were first subjected to a series of behavioral tests. Compared to the control group, DIO mice exhibited behaviors consistent with depressive-like phenotypes (Fig. 3A). This was evidenced by increased immobility time in both the tail suspension test (p 0.001) and forced swim test (p 0.001), both of which are well-validated paradigms for assessing depression-like behavior. Furthermore, DIO mice exhibited significantly less rearing in the open field test (p = 0.005), indicating heightened anxiety-like behavior, as well as a significant decrease in sucrose preference (p 0.001), reflecting anhedonia, a core symptom of depression. We next assessed neuronal integrity in the hippocampus, a brain region critically involved in mood regulation. Histological analysis revealed substantial neuronal damage in the hippocampus of DIO mice. Representative images of H&E staining showed marked morphological changes in the DIO group, including pyknotic nuclei (indicated by arrows in Fig. 3B), which are a hallmark of neuronal injury. Quantitative analysis of H&E stained sections confirmed a significant increase in the number of neurons with pyknotic nuclei in DIO mice compared to controls (p 0.001; Fig. 3C). Representative Nissl staining images demonstrated a noticeable reduction in Nissl bodies and a disordered arrangement of neurons in the hippocampus of DIO mice (Fig. 3D). Quantitative analysis of Nissl staining confirmed a significant decrease in the number of healthy Nissl-positive neurons in the DIO group, indicating severe neuronal damage (p 0.001; Fig. 3E). To investigate the potential neurochemical alterations underlying these behavioral and histological deficits, we measured the levels of key monoamine neurotransmitters in the serum. The concentrations of both 5-hydroxytryptamine (5-HT, serotonin; p = 0.016; Fig. 3F) and norepinephrine (NA; p = 0.025; Fig. 3F) were found to be significantly lower in DIO mice compared to the control group.

Fig. 3.

Depressive-like phenotypes and neuronal damage in DIO mice. (A) Comparison of depressive-like behaviors between groups (n = 8). (B) Representative hematoxylin and eosin (H&E) stained images of the ventral hippocampal CA1 subregion. Pyknotic cells are indicated by black arrows. Scale bars = 40 µm (overview) and 10 µm (inset/enlarged view). (C) Quantification of pyknotic cells in the hippocampus. Damaged neurons are indicated by black arrows (n = 4). (D) Representative Nissl-stained images of the hippocampus. Scale bar = 20 µm. (E) Quantification of hippocampal neuronal damage (n = 4). (F) Abundance of 5-hydroxytryptamine (5-HT) and norepinephrine (NA) in serum (n = 4). *p 0.05, **p = 0.005, ***p 0.005.

To determine whether DIO triggers a pro-inflammatory response in the brain, we assessed the polarization state of microglia in the hippocampus. Immunofluorescence double-labeling for Iba1 (a microglial marker) and iNOS (a marker for the M1 pro-inflammatory phenotype) revealed markedly increased co-localization in the hippocampus of DIO mice compared to controls, suggesting a shift towards the M1 polarized state (Fig. 4A). Quantitative analysis confirmed a significant increase in the percentage of iNOS-positive microglia in the DIO group (p = 0.002; Fig. 4B). To further validate this pro-inflammatory shift, glial cells from the brains of both groups were isolated. Representative images of the extracted glial cells are shown in Fig. 4C. Analysis of these cells revealed a significantly enhanced inflammatory profile in the DIO group. At the transcriptional level, mRNA expression of classic M1 markers, including TNF-$\alpha$ (p = 0.008) and iNOS (p 0.001), was significantly upregulated in glial cell-enriched isolates from DIO mice, as measured by qPCR (Fig. 4D). Consistent with the gene expression data, ELISA measurements demonstrated substantially increased secretion of key pro-inflammatory cytokines, namely IL-6 (p = 0.011), IL-1$\beta$ (p = 0.007), and TNF-$\alpha$ (p = 0.002), in the DIO group compared to the control group (Fig. 4E). Collectively, these results indicate that microglia in DIO mice adopt a pronounced pro-inflammatory state.

Fig. 4.

Hippocampal microglia in DIO mice show enhanced M1 polarization and a pro-inflammatory signature. (A) Representative double immunofluorescence staining of Iba1 and iNOS in hippocampal sections. The white arrow represents an M1-polarized pro-inflammatory microglial phenotype. Scale bars = 100 µm (overview) and 10 µm (inset/enlarged view). (B) Bar chart showing the proportion of double-labeled cells relative to microglia (n = 4). (C) Immunofluorescence of glial cell cultures. Scale bar = 10 µm. (D) mRNA levels of M1 phenotypic markers in glial cells from the two groups (n = 4). (E) Levels of pro-inflammatory cytokines (n = 4). *p 0.05 , **p 0.01 , ***p 0.005.

A transwell co-culture system was established to directly investigate the functional impact of DIO microglia on neuronal health. As schematically illustrated in Fig. 5A, this system allows the exchange of soluble factors between glial cells cultured in the upper chamber and hippocampal neurons cultured in the lower chamber, without direct cell-to-cell contact. We first assessed the viability of hippocampal neurons after exposure to conditioned media from the different microglial populations. The CCK-8 assay revealed the viability of hippocampal neurons co-cultured with DIO microglia was significantly reduced compared to those co-cultured with control microglia (p = 0.031; Fig. 5B). Concurrently, the release of LDH, an indicator of neuronal cytotoxicity and membrane damage, was markedly elevated in the co-culture system containing DIO microglia (p = 0.005; Fig. 5C). Representative flow cytometry plots (Fig. 5D) and their quantitative analysis confirmed a significantly higher rate of apoptosis in hippocampal neurons exposed to factors derived from DIO microglia (p = 0.005; Fig. 5E). Given the established link between inflammation and oxidative stress, we also measured the levels of intracellular ROS. Representative flow cytometry histograms (Fig. 5F) and their quantification revealed a significant increase in ROS accumulation in neurons co-cultured with DIO microglia (p 0.001; Fig. 5G). Taken together, these results demonstrate that soluble factors released by microglia derived from DIO mice can directly induce neurotoxicity, as shown by reduced viability, increased cytotoxicity, elevated apoptosis, and oxidative stress in hippocampal neurons.

Fig. 5.

Neurotoxic effects of DIO microglial conditioned media on hippocampal neurons in a co-culture system. (A) Schematic diagram of a transwell co-culture system. (B) CCK-8 assay results for cell viability (n = 4). (C) Measurement of lactate dehydrogenase (LDH) activity (n = 4). (D) Analysis of apoptosis by flow cytometry. (E) Bar graph showing apoptosis of hippocampal neurons (n = 4). (F) Flow cytometric analysis of intracellular reactive oxygen species (ROS). (G) Bar graph showing the relative fluorescence intensity of ROS (n = 4). *p 0.05, **p 0.01, ***p 0.001.

To investigate the hypothesis that metabolic disturbances in DIO lead to lipid accumulation within microglia, we performed a morphological and ultrastructural analysis. Immunofluorescence staining for Iba1 and Bodipy (a neutral lipid dye) revealed markedly increased co-localization in the hippocampus of DIO mice compared to controls (Fig. 6A), indicating a substantial accumulation of neutral lipids within microglial cells in the DIO group. Quantitative analysis of the fluorescence intensity confirmed a significant increase in Bodipy signal within Iba1-positive cells in DIO mice (p = 0.015; Fig. 6B). To further validate these findings at a higher resolution, microglia were examined using transmission electron microscopy (TEM). Representative electron micrographs revealed the presence of large, electron-lucent LDs in the cytoplasm of microglia from DIO mice (Fig. 6C). Taken together, these data provide direct visual and quantitative evidence that DIO induces a state of significant lipid overload, as demonstrated by the accumulation of LDs in hippocampal microglia. Quantitative morphometric analysis of TEM images also revealed a significant increase in the average diameter of LDs within the hippocampal microglia of DIO mice compared to the control group (p 0.001, Fig. 6D). This data provides robust statistical support for the qualitative observation of marked LD accumulation in DIO microglia.

Fig. 6.

Lipid droplet accumulation in the hippocampal microglia of DIO mice. (A) Representative immunofluorescence images showing co-localization of Iba1 (green) and Bodipy (red) in the ventral hippocampal CA1 subregion (approximately Bregma –3.5 mm). White arrows indicate lipid droplet-positive microglia. Scale bars = 100 µm (overview) and 20 µm (inset/enlarged view). (B) Proportion of double-stained (Iba1⁺/Bodipy⁺) microglia (n = 4). (C) Representative electron micrograph showing lipid droplet size in microglia. Scale bar = 500 nm. LD, lipid droplet; Lys, lysosome. (D) Quantitative analysis of lipid droplet size in microglia (n = 8). *p 0.05, ***p 0.001.

To assess the intrinsic propensity for lipid accumulation in microglia under DIO conditions, we examined primary cells isolated from the DIO and Con groups. Bodipy staining of cultured microglia revealed the presence of LDs (Fig. 7A). Morphometric analysis of these images indicated a distinct pattern of lipid storage: the average size of LDs was significantly larger in microglia from DIO mice (p = 0.004; Fig. 7B), whereas the number of LDs per cell was comparable between the two groups (p = 0.123; Fig. 7B). This observation was confirmed using an alternative histological method. Oil Red O staining also demonstrated more substantial lipid accumulation in DIO microglia (Fig. 7C). Densitometric measurement of the staining intensity confirmed a significant increase in total neutral lipid content in the DIO group compared to controls (p 0.001, Fig. 7D). To obtain a biochemical correlate, we also measured intracellular triglyceride levels. This assay demonstrated a significantly higher triglyceride content in glial cells from DIO mice, consistent with the histochemical findings (p = 0.007; Fig. 7E). Collectively, these in vitro results indicate that microglia from DIO mice possess a cell-autonomous predisposition towards excessive lipid storage, characterized primarily by enlarged LDs and increased triglyceride deposition, even in a standardized cell culture environment.

Fig. 7.

Enhanced in vitro lipid accumulation and triglyceride content in glial cells from DIO mice. (A) Bodipy staining in glial cells. Scale bar = 100 µm. (B) Analysis of lipid droplet size and number in microglia (n = 4). (C) Oil Red O (ORO) staining in glial cells. Scale bar = 100 µm. (D) Normalized number of Oil Red O-positive puncta per cell (n = 4). (E) Normalized intracellular triglyceride (TG) level (n = 4). **p 0.01, ***p 0.001, ns, not significant.

To comprehensively investigate the impact of the DIO state on the lipid landscape of microglia, we performed an untargeted lipidomics analysis on glial cells isolated from Con and DIO mice. PCA demonstrated a clear separation between the two groups (Fig. 8A), indicating a distinct global lipidomic profile in DIO microglia. Unsupervised hierarchical clustering of the top 50 most differentially abundant lipids further highlighted pronounced differences, as visualized in a heatmap (Fig. 8B). A ranked metabolite plot detailed these specific alterations (Fig. 8C). Pathway enrichment analyses were performed to gain functional insights. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis revealed significant enrichment for several lipid metabolism-related pathways, including adipocytokine signaling, fat digestion and absorption, lipid degradation, sphingolipid signaling, glycerophospholipid metabolism, and biosynthesis of unsaturated fatty acids (Fig. 8D). Consistent with these findings, reactome pathway analysis confirmed significant perturbations in key biological processes such as fatty acid transport, sphingolipid metabolism, and overall lipid metabolism (Fig. 8E). Collectively, these data demonstrate substantial and broad rewiring of the lipid metabolic network in the microglia of DIO mice.

Fig. 8.

Lipidomic profiling of glial cells from DIO mice. (A) PCA analysis of untargeted lipidomics. (B) Heatmap analysis of differential metabolites. (C) Visualization of differential metabolites with a lollipop plot. (D) KEGG pathway enrichment analysis. (E) Reactome pathway enrichment analysis. PCA, principal component analysis; KEGG, Kyoto Encyclopedia of Genes and Genomes.