Lysozyme M (LysM)-Cre mice (The Jackson Laboratory, Bar Harbor, ME, USA, #004781) in a C57BL/6 genetic background were crossed with Grn^f/f (The Jackson Laboratory, #013174) mice to produce macrophage-specific Grn depletion (Grn-MΦKO) mice. Grn^f/f littermates were used as Grn-WT mice. For macrophage labeling, LysM-Cre mice or Grn-MΦKO mice were bred with mice carrying a tandem dimer tomato (tdTomato) or enhanced green fluorescence protein (eGFP) reporter allele with an upstream loxP-flanked STOP codon (The Jackson Laboratory, #007914 or #007676). The animals were housed under controlled temperature conditions (22°C±1°C) and 12-hour light-dark cycle (lights on at 8:00 AM). The mice had ad libitum access to a standard chow diet (containing 12.5% of calories from fat; Cargill Agri Purina, Seongnam, Korea) and water, unless otherwise specified. To induce DIO, mice were fed a HFD containing 58% fat (Research Diets, New Brunswick, NJ, USA) starting at 7 weeks of age.

To confirm the Grn gene knockout in macrophages, blood samples (200 μL) were collected from the retro-orbital plexus of the mice (LysM-Cre;tdTomato or Grn-MΦKO;tdTomato) and then incubated with 1 mL of 1X red blood cell lysis buffer (Invitrogen, Waltham, MA, USA, #00-4333) for 10 minutes. After washing with 1X phosphate buffered saline (PBS) three times, the cells were centrifuged (300 ×g) at 4°C for 5 minutes and resuspended in 2% fetal bovine serum-containing PBS. The cells were then incubated with PGRN antibody (1:200, R&D Systems, Minneapolis, MN, USA, #AF2557) for 30 minutes followed by fluorescein Isothiocyanate (FITC)-conjugated anti-sheep secondary antibody (1:1,000, Thermo Fisher, Waltham, MA, USA, #A16042) for 20 minutes. The cells were subsequently subjected to fluorescence-activated cell sorting (FACS) using a BD FACS Aria II Cell Sorter (BD Biosciences, Franklin Lakes, NJ, USA), where they were sorted based on tdTomato and FITC fluorescence. For tdTomato fluorescence compensation, non-fluorescent blood samples were obtained from LysM-Cre littermates (Supplementary Fig. 1A). Successful Grn depletion in macrophages was confirmed by a significant depletion of PGRN⁺ blood myeloid cells (Supplementary Fig. 1). To ensure successful Grn depletion in blood monocytes, blood was collected from Grn-WT mice and Grn-MΦKO mice and processed as described above. The cells were then incubated with CD11b-allophycocyanin-cyanine 7 (APC-Cy7) (1:100, rat, BD Biosciences, #557657) and lymphocyte antigen 6 family member C (Ly6C)-phycoerythrin (1:100, rat, Thermo Fisher, #12-5932-80) antibodies for 30 minutes. Blood monocytes were then gated and sorted via FACS. The successful depletion of Grn in CD11b⁺ Ly6C⁺ monocytes was confirmed by Grn quantitative polymerase chain reaction (qPCR) analysis.

Food intake levels and body weights were monitored weekly after weaning until the indicated ages. Body compositions (lean mass and fat mass) were measured using dual X-ray absorptiometry (iNSiGHT VET DXA, OsteoSys, Seoul, Korea) on the day of energy expenditure (EE) measurement. EE and locomotor activity were determined using a comprehensive lab monitoring system (CLAMS) (Columbus Instruments, Columbus, OH, USA). Mice were placed in the CLAMS chambers for 48 hours to adapt to the conditions before measurement. During this period, day-night cycles were consistent with those during the initial housing of the animals, and food was provided as pellets on the floor of the CLAMS cages. All EE data were normalized to the lean mass values.

For the glucose tolerance test (GTT), D-glucose (1 g/kg, Sigma, St. Louis, MO, USA) was orally administered to overnight-fasted mice. For the insulin tolerance test (ITT), insulin (Humulin-R 0.25 U/kg, Eli Lilly, Indianapolis, IN, USA) was injected into the peritoneum of the overnight-fasted mice. Blood samples were obtained from tail veins for glucose measurements at 0, 15, 30, 60, and 120 minutes (for GTT) or 0, 30, 60, 90, and 120 minutes (for ITT) after injections. Blood glucose levels were measured using a glucometer (ACCU-CHEK, Aviva Plus System, Indianapolis, IN, USA).

Epididymal adipose tissues were dissected, chopped, and then incubated in collagenase buffer (0.1 M 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid [HEPES], 0.125 M NaCl, 5 mM KCl, 1.3 mM CaCl₂, 5 mM glucose, 1.5% [w/v] glucose, and 0.1% [w/v] collagenase I) for 20 minutes at 37°C with shaking. After incubation, the mixture was centrifuged, and the pelleted stromal vascular fractions (SVFs) were collected for flow cytometry analysis. SVFs were separated from red blood cells by adding lysis buffer (155 mM NH₄Cl/0.1 M Tris-HCl, pH 7.65, in a 9:1 ratio). Following washing with 1X PBS, SVFs were stained with monoclonal antibodies against CD11b (1:250, BD Pharmingen, San Diego, CA, USA, #557396), F4/80 (1:100, BD Horizon, Franklin Lakes, NJ, USA, #563900), CD11c (1:250, Invitrogen, #12-0114-81), and CD206 (1:250, BD Pharmingen, #565250) for macrophage analysis using a FACS Canto II (BD Biosciences).

Formalin-fixed paraffin-embedded epididymal adipose tissue samples were sectioned to 5 µm thickness, deparaffinized in xylene and then rehydrated using a gradient of alcohol before undergoing hematoxylin and eosin staining. Images were captured using a Leica bright-field microscope (Leica, Wetzlar, Germany) at a ×20 magnification. Adipocyte sizes were analyzed using Image J software (National Institutes of Health, Bethesda, MD, USA).

Mice were anesthetized with an intraperitoneal injection of 40 mg/kg Zoletil^® (Virbac, Carros, France) and 5 mg/kg Rompun (Bayer AG, Leverkusen, Germany) and then perfused with 50 mL saline followed by 50 mL 4% paraformaldehyde (PFA) via the left ventricle of the heart. The obtained brains were post-fixed with 4% PFA for 16 hours at 4°C and then dehydrated in a 30% sucrose solution until the tissues sank to the bottom of the container. Coronal brains, including the hypothalamic area (bregma –1.22 to –2.30 mm), were sectioned at a thickness of 30 µm using a cryostat (Leica). One out of every five slides was collected and measured. For Iba1 and glial fibrillary acidic protein (GFAP) staining, hypothalamic slices were blocked with 3% donkey serum for 1 hour and then incubated with Iba1 (1:400, goat, Abcam, Cambridge, UK, #5076) and GFAP (1:1,000, rabbit, Millipore, Burlington, MA, USA, #AB5804) at 4°C for 48 hours. For microtubule associated protein 2 (MAP2) and green fluorescence protein (GFP) staining, hypothalamic slices were permeabilized in 0.5% phosphate buffered saline-tween 20 (PBST) for 5 minutes, blocked with 3% donkey serum for 1 hour, and then incubated with MAP2 (1:1,000, Chicken, Abcam, #5392) and GFP (1:1,000, rabbit, Abcam, #6556) antibodies at 4°C overnight. After washing, slides were incubated with Alexa Fluor-conjugated secondary antibodies (1:1,000, Invitrogen) at room temperature for 2 hours. For nuclear staining, slides were treated with 4’,6-diamidino-2-phenylindole (DAPI) (5 mg/mL, 1:10,000, Sigma, #20718-90-3) for 10 minutes before mounting. Immunofluorescence images were obtained using a confocal microscope (Carl Zeiss 710, Oberkochen, Germany).

A stainless steel cannula (26 gauge) was implanted into the third ventricle (3V) of the mice using stereotaxic coordinates: 1.4 mm caudal to the bregma and 5.5 mm ventral to the sagittal sinus. Following a 7-day recovery period, the correct positioning of each cannula was confirmed by a positive dipsogenic response from administration of 50 ng of angiotensin-II. Animals with a negative drinking response to angiotensin-II were excluded from the study.

To evaluate insulin signaling in the adipose tissues, either insulin (Humulin-R 0.25 U/kg) or saline was administered intraperitoneally to mice receiving an HFD for 8 weeks following an overnight fast. Fifteen minutes after injection, epididymal adipose tissues were collected for Akt phosphorylation analysis. To assess leptin signaling in the hypothalamus, leptin (1 μg) or saline was injected intracerebroventricularly following an overnight fast in 8-week HFD fed mice. Food intake was monitored for 24 hours after intracerebroventricular (ICV) injections. Three days later, animals were sacrificed 45 minutes after the ICV injections of leptin and mediobasal hypothalamic blocks were collected for analysis of signal transducer and activator of transcription 3 (STAT3) phosphorylation.

Mice were kept under freely-fed conditions and sacrificed from 9:00 AM to 10:00 AM. Mediobasal hypothalamus (MBH), brown adipose tissue (BAT), and both inguinal and epididymal white adipose tissue (WAT) were then harvested, snap frozen in liquid nitrogen, and stored at –70°C. Total RNA was subsequently extracted from these tissues using TRIzol in accordance with the manufacturer’s protocol. 5 μg RNA aliquots were then reverse transcribed to generate cDNA. The threshold cycle (C_T) values were normalized to that of glyceraldehyde 3-phosphate dehydrogenase (GAPDH). The primers used for qPCR analysis are listed in Supplementary Table 1.

Plasma IL-6 levels (50 μL) were assayed using mouse IL-6 enzyme-linked immunosorbent assay (ELISA) kits (R&D Systems, #M6000B-1) following the manufacturer’s protocol.

Protein lysates extracted from the mouse MBH and adipose tissues (20 µg protein each) were subjected to immunoblotting. Briefly, following blocking in 3% skim milk for 1 hour at room temperature, the membranes were incubated overnight at 4°C with primary antibodies targeting total STAT3 (1:1,000, Cell Signaling, Danvers, MA, USA, #4904S), phospho (Y705)-STAT3 (1:1,000, Cell Signaling, #9145S), total Akt (1:1,000, Cell Signaling, #9272S), phospho (Ser473)-Akt (1:1,000, Cell Signaling, #9271S), and β-actin (1:1,000, Santa Cruz Biotechnology, Santa Cruz, CA, USA, #sc-47778).

All data are presented as the mean±standard error of the mean. Statistical analyses were performed using Prism 9.1.2 (GraphPad Software Inc., San Diego, CA, USA). Cell numbers and immunofluorescence intensities were quantified using Image J and Photoshop version CS6 (Adobe Systems, San Jose, CA, USA). Statistical significance among the study groups was tested using one-way or two-way analysis of variance (ANOVA) followed by a post hoc Tukey test or an unpaired Student’s t-test if appropriate. Analysis of covariance (ANCOVA) (SPSS version 28, IBM Co., Armonk, NY, USA) was used to compare the group’s EE. Statistical significance was defined as P<0.05.

All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC, #2015-11-214) of the Asan Institute for Life Sciences (Seoul, Korea).

To investigate the metabolic roles of macrophage-derived PGRN, we generated a mouse model in which the PGRN-encoding gene, Grn, was specifically depleted in bone marrow-derived macrophages. We achieved this by crossbreeding mice expressing Cre recombinase under the control of the LysM promoter (LysM-Cre mice) with Grn-floxed mice, resulting in mice referred to as Grn-MΦKO. To verify the Grn deletion in macrophages, we labeled blood myeloid cells with tdTomato by mating LysM-Cre mice or Grn-MΦKO mice with tdTomato reporter mice. tdTomato-labeled blood myeloid cells were isolated using flow cytometry (Supplementary Fig. 1A) and sorted into PGRN⁺ cells or PGRN⁻ cells. The PGRN⁺ cell population was reduced in the Grn-MΦKO;tdTomato mice compared to the LysM-Cre;tdTomato mice (Supplementary Fig. 1B). We further found that the Grn expression level in CD11b⁺ Ly6C⁺ blood monocytes was significantly lower in Grn-MΦKO mice than in Grn-WT mice (Supplementary Fig. 1C), thus confirming the successful depletion of macrophage Grn.

Some neurons have been shown to be labeled in LysM-Cre reporter mice [21] To address this concern, we generated LysM-Cre-eGFP mice and tested the eGFP expression in hypothalamic neurons. The colocalization of eGFP and neuronal marker MAP2 expression was very low in the ARC, paraventricular nucleus, and dorsomedial nucleus (Supplementary Fig. 2) indicating that hypothalamic neuronal Grn depletion was minimal in the Grn-MΦKO mice.

We next conducted metabolic phenotype analysis in Grn-MΦKO male mice in comparison with Grn-WT mice under NCD-fed conditions. No significant differences were found in body weight, fat mass, or lean mass levels between the two groups of mice (Fig. 1A and B). Consistently, monitoring of food intake and EE revealed no changes associated with macrophage Grn depletion (Fig. 1C and D). The respiratory exchange ratio (RER), an index of nutrient oxidation for fuel, was also unchanged in NCD-fed Grn-MΦKO mice (Fig. 1E). There was no detectable alteration in locomotor activity (Supplementary Fig. 3). We further evaluated glucose metabolism by conducting GTT and ITT. Changes in blood glucose levels after the glucose and insulin challenges did not differ between the two groups (Fig. 1F and G). These data implied that a PGRN deficiency in macrophages does not significantly affect energy and glucose metabolism under normal diet conditions.

We next investigated the metabolic phenotype of Grn-MΦKO mice under conditions of overnutrition. These mice were fed an HFD starting at 7 weeks of age until sacrifice. Grn-MΦKO mice gained less weight during this HFD challenge compared to wild littermates (Fig. 2A). The lower body weights of the Grn-MΦKO mice were due to a lower fat mass, as their lean mass was unchanged (Fig. 2B) and demonstrated a resistance to DIO. We further found that this obesity resistance was attributable to an enhanced EE, especially during nighttime, with no alterations in food intake and RER (Fig. 2C and D, Supplementary Fig. 4A). Locomotor activity was not significantly altered in the HFD-fed Grn-MΦKO mice compared with their diet-matched wild littermates (Supplementary Fig. 4B). We thus explored the possibility that increased EE might be caused by changes in thermogenesis. Indeed, The mRNA expression levels of thermogenic genes such as uncoupling protein 1 (Ucp1), PR domain containing 16 (Prdm16), and iodothyronine deiodinase 2 (Dio2) were increased in the BAT and inguinal WAT of Grn-MΦKO mice (Fig. 2E and F).

Glucose and insulin tolerance were also improved in HFD-fed Grn-MΦKO mice compared to the diet-matched wild controls (Fig. 2G and H). These findings suggested that a macrophage Grn deficiency protects against HFD-induced obesity and its metabolic complications.

As changes in the WAT are closely related to the development of insulin resistance and glucose intolerance under DIO conditions [2,3], we examined whether any WAT changes arose in the Grn-MΦKO mice upon HFD challenge. Histological analysis of epididymal WAT revealed that HFD feeding increased adipocyte sizes in Grn-WT mice, but that these HFD-induced changes were blunted in the Grn-MΦKO mice (Fig. 3A). Crown-like structures (CLS) in WAT are histological features characterized by macrophages surrounding dead or dying adipocytes [22], that indicate adipose tissue inflammation in conditions of obesity. As expected, CLS formation in WAT increased in HFD-fed Grn-WT mice compared to NCD-fed Grn-WT mice. Notably, HFD-induced CLS formation was reduced in Grn-MΦKO mice to the levels of NCD-fed Grn-WT mice (Fig. 3A). These histological findings suggested that Grn depletion in macrophages prevents adipocyte hypertrophy and ATM accumulation under conditions of chronic overnutrition.

Macrophages are categorized into two main polarization types: classically-activated (M1) and alternatively-activated (M2) [4,23]. During the progression of DIO, M1 macrophages increase while M2 macrophages decrease in the adipose tissues, contributing to adipose tissue inflammation and the development of insulin resistance [4]. We therefore investigated changes in macrophage polarization in our mouse models. Flow cytometry analysis of SVFs, obtained from epididymal adipose tissues, revealed that among the total ATMs (CD11b⁺ F4/80⁺), the proportion of M1 cells (CD11c⁺) was elevated, while the proportion of M2 cells (CD206⁺) was reduced in HFD-fed Grn-WT mice compared to NCD-fed Grn-WT mice (Fig. 3B). Strikingly, these HFD-induced changes in macrophage polarization were completely prevented in the HFD-fed Grn-MΦKO mice (Fig. 3B). Consistently, the M1/M2 ratio in adipose tissues was increased by HFD feeding, but this change was fully suppressed by a macrophage Grn deficiency (Fig. 3B). These findings suggested that macrophage PGRN may play a critical role in ATM polarization during chronic HFD consumption.

Quantitative PCR analysis of inflammation-related genes further revealed that HFD-fed Grn-MΦKO mice exhibited a lower expression of proinflammatory cytokines (Il-6, Tnfα) and the chemokine C-C motif chemokine ligand 2 (Ccl2), along with increased expression of the anti-inflammatory cytokine Il-10, reflecting reduced adipose tissue inflammation (Fig. 3C). In addition, the mRNA expression levels of peroxisome proliferator-activated receptor γ (Pparγ), a key transcription factor in adipogenesis, and the adipocyte-derived hormone leptin were reduced in HFD-fed Grn-MΦKO mice (Fig. 3C). These changes were consistent with the observed reduction in adipose mass. A previous study has reported that a generalized Grn depletion blocked an HFD-induced elevation in plasma and adipose tissue IL-6 expression levels [17]. Likewise, we found that plasma IL-6 levels were elevated by HFD consumption in Grn-WT mice but that this increase was significantly blocked in Grn-MΦKO mice (Fig. 3D).

We further investigated adipose tissue insulin signaling, given that HFD-fed Grn-MΦKO mice showed improved insulin resistance (Fig. 2H). Insulin-induced Akt phosphorylation in epididymal WAT was reduced in HFD-fed Grn-WT mice compared to their NCD-fed counterparts, but recovered in HFD-fed Grn-MΦ mice (Fig. 3E). These findings suggest that a macrophage-specific Grn deficiency is sufficient to rescue HFD-induced insulin resistance in adipose tissue.

As hypothalamic macrophages are activated to sustain hypothalamic inflammation and disrupt metabolic hormonal signaling under HFD conditions [9], we tested for alterations in hypothalamic macrophages in HFD-fed Grn-MΦKO mice in comparison with NCDPGRN⁺and HFD-fed Grn-WT mice. We traced these cells by cross-bleeding of LysM-Cre mice or Grn-MΦKO mice with Cre-dependent GFP reporter mice. Consistent with the findings of a previous study [9], we found that 10 weeks of HFD feeding significantly increased the numbers of GFP⁺ hypothalamic macrophages in the ARC of LysM-Cre;GFP mice (Fig. 4A). In addition, the HFD conditions caused morphological changes in macrophages from a rod shape without cellular processes to an enlarged soma with multiple processes (Fig. 4A). The macrophage numbers and morphological changes in the HFD-fed Grn-MΦKO mice were remarkably attenuated, indicating less macrophage activation in these animals (Fig. 4A). We further examined changes in the Iba1⁺ hypothalamic microglia and GFAP⁺ astrocytes as their activation is a cardinal feature of hypothalamic inflammation in DIO mice [10]. HFD-induced increases in microglia and astrocyte numbers were significantly reduced in Grn-MΦKO mice on HFD feeding for 10 weeks (Fig. 4B and C). The soma sizes of Iba1⁺ microglia showed a decrease (Fig. 4B). These findings suggested a reduction in hypothalamic microglia and astrocyte activation in Grn-MΦKO mice despite consuming a similar amount of HFD to the controls.

Analysis of proinflammatory and anti-inflammatory cytokine expression in the hypothalamus revealed that chronic HFD consumption upregulated the Il-1β, Il-6, and Tnfα expression levels but downregulated Il-4 expression (Fig. 4D). Notably HFD-induced elevation in proinflammatory cytokine expression was completely reversed in the hypothalamus of Grn-MΦKO mice (Fig. 4D). These data suggested that macrophage PGRN actively participates in the hypothalamic inflammation triggered by fat-rich diets.

Leptin, a critical hormone for maintaining energy homeostasis, exerts its effects primarily on the hypothalamus [24]. We therefore investigated whether a macrophage Grn deficiency could influence hypothalamic leptin signaling and its anorexigenic actions. ICV leptin injection suppressed food intake for 24 hours in NCD-fed Grn-WT mice and the anorexigenic effect of leptin was blunted in 8-week HFD fed Grn-WT mice, suggesting the development of hypothalamic leptin resistance following chronic HFD consumption (Fig. 4E). Notably, Grn-MΦKO mice on an HFD for 8 weeks exhibited improved feeding response to leptin compared to diet-matched Grn-WT mice (Fig. 4E). We finally assessed leptin signaling by measuring hypothalamic STAT3 phosphorylation. Leptin-induced STAT3 phosphorylation in the hypothalamus was significantly reduced in 8-week HFD fed Grn-WT mice compared to NCD-fed Grn-WT mice (Fig. 4F). In contrast, the Grn-MΦKO mice fed HFD for the same duration exhibited improved STAT3 phosphorylation in response to leptin (Fig. 4F). Taken together, these findings indicated that a lack of Grn in hypothalamic macrophages protects against hypothalamic leptin resistance under the conditions of fat overconsumption.