USP25 deficiency suppresses diet-induced obesity via ubiquitination and degradation of PARP1 and Elovl3 downregulation

TL;DR

This study shows that USP25 deficiency reduces obesity and improves insulin resistance by affecting fat cell development and lipid metabolism.

Contribution

The study reveals a novel role of USP25 in adipocyte differentiation and obesity through its regulation of PARP1 and Elovl3.

Findings

USP25-deficient mice showed reduced adipose tissue growth and better insulin resistance.

USP25 stabilizes PARP1, which promotes preadipocyte differentiation via Elovl3 regulation.

RNA sequencing revealed downregulation of lipid synthesis pathways in USP25-deficient mice.

Abstract

Ubiquitin-specific protease 25 (USP25) is a key regulator of lipid metabolism and insulin-stimulated glucose transport. Nonetheless, its involvement in adipocyte maturation remains uncertain. In this study, we aimed to explore how the expression of USP25 contributes to obesity induced by a high-fat diet (HFD). Usp25-KO and WT mice, maintained on either a normal diet or an HFD, were evaluated for weight gain, insulin resistance status, adipose tissue development, energy metabolism, and systemic inflammation status. In vitro, 3T3-L1 cells were induced to mature adipocytes in a specific culture medium. We found that USP25 expression decreased in obese mice subjected to long-term HFD feeding and in mature adipocytes. Usp25-KO mice showed restrained adipose tissue development and improved insulin resistance, whereas USP25-deficient preadipocytes failed to differentiate. RNA sequencing analysis showed the downregulation of lipid synthesis-related pathways in Usp25-KO mice. Mechanistically, USP25 binds to and stabilizes poly (ADP-ribose) polymerase 1 through deubiquitination, whereas poly (ADP-ribose) polymerase 1 facilitates preadipocyte differentiation and maturation by regulating the elongation of very long-chain fatty acid protein 3. These findings show the essential role of USP25 in adipocyte differentiation and lipid metabolism, suggesting that targeting USP25 ablation in adipocytes could be a promising therapeutic strategy for obesity treatment.

Figures

USP25 is downregulated in the adipose tissue of HFD-fed mice and mature 3T3-L1 adipocytes. A: Expression of USPs in people with mild and moderate/severe obesity. B: Fasting body weights of WT mice fed an ND (n = 4) or an HFD (n = 8) for 10 weeks. C: Representative images of H&E and Oil Red O stainings of liver tissue from WT mice prior to HFD intervention (control) and after 10 weeks of HFD feeding. D: TG content in the livers of the indicated groups (n = 6). E, Left panel: Representative Western blot showing the expression level of USP25 in the inguinal white adipose tissue of mice fed an HFD

USP25 deficiency alleviated the progression of obesity and insulin resistance in HFD-fed mice. A: Body weights of WT and *Usp25*-KO mice fed an HFD for 10 weeks were recorded weekly (n = 5 [WT + ND; KO + ND], n = 7 [KO + HFD], and n = 8 [WT + HFD]). B: Images of WT and *Usp25*-KO mice fed an HFD for 10 weeks; the scale bar represents 1 cm. C: Weights of adipose tissue from WT and *Usp25*-KO mice fed an HFD for 10 weeks (n = 5 [WT + ND; KO + ND] and n = 8 [WT + HFD; KO + HFD]). D: Images of brown adipose tissue (BAT), inguinal white adipose tissue (iWAT), and epididymal white adipose tissue (eW

Assessment of food intake, energy expenditure, and immune status in HFD-fed *Usp25-*KO mice. WT and *Usp25*-KO mice fed an HFD for 10 weeks were placed into metabolic cages. Food intake (A: WT, n = 8; KO, n = 7), heat production (B), and respiratory exchange ratio (RER) ( $[eqn]$ CO_2_/ $[eqn]$ O_2_) (C) were monitored within 48 h. D: Plasma insulin levels of WT and *Usp25*-KO mice fed an HFD for 10 weeks (n = 6). E: Levels of eight cytokines in plasma from WT and *Usp25*-KO mice fed an HFD for 10 weeks (n = 3). Immune cells were isolated from different tissues as indicated, and the relative a

Depletion of USP25 retards 3T3-L1 differentiation into adipocytes. Knockdown of USP25 was confirmed using Western blot (A, n = 4) and RT-PCR analyses (B, n = 3). C: Representative images of 3T3-L1 cells cultivated in blank medium for 7 days (control), scrambled siRNA negative control (si-NC), and si-USP25 groups differentiated for 7 days, recorded via an OM; the scale bar represents 100 μm. D, Left panel: Oil Red O staining of control, si-NC, and si-USP25 cultures subjected to 7 days of adipocyte differentiation; the scale bar represents 100 μm. Right panel: Statistical analysis of positive Oi

USP25 deletion decreased *Elovl3* expression and increased PARP1 ubiquitination, leading to impaired adipose differentiation. A: Downregulated and upregulated genes in epididymal white adipose tissue (eWAT) from WT and *Usp25*-KO mice fed an HFD for 10 weeks. B: GO term enrichment analysis. C: KEGG pathway enrichment analysis. D: RT-PCR analyses showing relative mRNA expression of *Elovl3* in eWAT from WT and *Usp25*-KO mice fed an HFD for 10 weeks (n = 6). E: RT-PCR analyses showing relative mRNA expression of *Elovl3* in 3T3-L1, si-NC, and si-USP25 cultures after 7 days of adipocyte differen

The molecular mechanism by which USP25 deficiency mitigates HFD-induced obesity progression involves proteasomal degradation of PARP1. Under HFD conditions, USP25 prevents the proteasomal degradation of PARP1 by eliminating ubiquitin chains. PARP1 ensures fatty acid synthesis during adipocyte differentiation by sustaining ELOVL3 expression, which results in the development of adipose tissue. However, this situation is reversed when USP25 is absent.

Full text

Obesity is a chronic disorder of multiple components, characterized by excessive or irregular build-up of lipid accumulation (1). This condition significantly increases the risk of severe comorbidities, including type II diabetes, atherosclerosis, hepatic steatosis, and other malignancies (2, 3, 4). Over the past two decades, the global prevalence of obesity has substantially risen owing to decreased overall physical activity and overconsumption of refined carbohydrates (5, 6). In an analysis of the Behavioral Risk Factor Surveillance System between 1990 and 2008, it was predicted that, by 2030, 51% of the population will be obese (7). Adipocytes are crucial in maintaining energy homeostasis and lipid metabolism by storing lipids as a primary source of energy (8). However, aberrant adipocyte hypertrophy and excessive intracellular lipid accumulation, rather than healthy adipocyte proliferation, are major contributors to obesity and related metabolic dysfunction. Consequently, it is essential to understand the factors that regulate adipocyte differentiation to uncover the mechanisms behind obesity and develop potential therapeutic strategies to normalize adipogenesis and obesity progression.

To date, numerous factors have been identified as pivotal regulators of adipocyte proliferation and differentiation, including hormones, cytokines, transcription factors, and molecular chaperones (9, 10, 11, 12). Adipocyte differentiation is a complex, multistage process involving a series of interdependent events orchestrated by transcription factors and cyclins. This process ultimately facilitates mesenchymal stem cells’ transition into preadipocytes and, subsequently, mature adipocytes. Emerging evidence shows that enzymes governing post-transcriptional regulation, including phosphorylation, oxidation, and ubiquitination, play a vital role in obesity and its associated metabolic disorders (13).

Ubiquitination is a pivotal post-translational modification, widely present in prokaryotic and eukaryotic cells. This process involves the sequential enzymatic activity of E1–E2–E3 ligases, enabling the covalent conjugation of ubiquitin, a 76-amino acid polypeptide, to specific proteins through an isopeptide bond. This modification signals proteins for proteasomal degradation (14). Aberrations in ubiquitination have been implicated in various clinical diseases, including Alzheimer’s disease, cardiovascular disorders, metabolic dysfunction, and oncogenesis (15, 16, 17, 18). Conversely, deubiquitination is the process by which deubiquitinases (DUBs) remove polyubiquitin chains from target proteins, thereby modulating their stability, subcellular localization, and downstream signaling pathways. Over 100 DUBs have been identified and grouped into eight protein families (19, 20, 21): ubiquitin C-terminal hydrolases; ubiquitin-specific proteases (USPs); Machado-Joseph domain proteases; ovarian tumor proteases; JAB1, MPN, and MOV34 family proteases; motifs interacting with ubiquitin-containing novel DUB family; monocyte chemotactic protein-induced protein; and zinc finger ubiquitin peptidase 1. Among these, USPs are well studied and are crucial in repairing DNA damage and cell cycle stabilization (22, 23), contributing to cardiovascular diseases and metabolic-associated fatty liver disease (MAFLD) (24, 25).

In this research, we delved into the regulatory processes of ubiquitin-specific protease 25 (USP25) in obesity progression. USP25 was initially identified by Valero et al. (26) as a DUB, and its encoding gene is located on chromosome 21q11.2, containing 25 exons. A significant decline in USP25 expression in the adipose tissue of mice with high-fat diet (HFD)-induced obesity was observed, suggesting a potential role for USP25 in obesity progression. Reportedly, studies have identified an isoform, USP25, which is specifically expressed in muscle and fat cells. This isoform has been implicated in specific cellular dysfunctions, including muscle abnormalities in patients with Down syndrome and impaired glucose uptake in adipocytes (27, 28). However, the exact role of USP25 in adipocyte differentiation and lipid homeostasis remains unclear. We aimed to elucidate the molecular mechanisms by which USP25 mediates obesity progression, specifically focusing on its role in regulating lipogenesis and adipocyte differentiation. Our findings may provide a rationale for USP25-targeted adipocyte-specific elimination as a potential therapeutic strategy to prevent clinical obesity and associated metabolic disorders.

Materials and methods

Animals

The mice were kept in high-efficiency particulate air-filtered cages under a 12-h light/dark cycle (lights on at 07:00), at a controlled ambient temperature of 22 ± 2°C and relative humidity of 50–60%, with free access to food and water. WT mice and Usp25-KO mice were obtained from a previous study (29). Male mice were randomly sorted into four groups: 1) WT mice fed a normal diet (ND) (n = 5); 2) Usp25-KO mice fed an ND (n = 5); 3) WT mice fed an HFD (n = 8); and 4) Usp25-KO mice fed an HFD (n = 7). To induce obesity, the HFD started at 8 weeks of age. The HFD (MD12033; Medicience Biopharmaceutical Co, Ltd, Jiangsu, China) contained the following macronutrient composition (gm%/kcal%): protein, 26.2/20%; fat, 34.9/60%; carbohydrate, 26.3/20%; and micronutrients, 12.6/0. Fasting weights were measured after a 4-h fasting period, conducted in the morning to minimize circadian effects. The mice were humanely euthanized at the conclusion of the study via carbon dioxide (CO_2_). The adipose tissue was collected from the interscapular area, inguinal area, and epididymal area, subsequently being weighed and visually inspected. Serum, liver, and adipose tissues were collected for analysis. This research was approved by the Institutional Review Board of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, and was carried out in accordance with the guidelines of the Helsinki and Istanbul Declarations.

Cell culture, transfection, and adipocyte differentiation

We purchased 3T3-L1 cells from Procell Life Science & Technology Co, Ltd (CL-0006). The cells were cultured in DMEM with 4.5 g/l of glucose, 1% penicillin/streptomycin, and 10% fetal bovine serum (Bio-Channel; BC-SE-FBS01) at 37°C with 5% CO_2_. For siRNA transfection, the cells were transfected with siRNA (for sequences of targets, refer to supplemental Table S1) via Lipofectamine 3000 (OBiO Technology; OGTR(M)20221002) upon reaching 70% confluence. The cells were harvested 48 h post-transfection. For adipocyte differentiation, when the cells reached confluence, 3T3-L1 cells were cultured for 48 h to achieve contact inhibition. Differentiation was then induced via specific agents, including 10 mM rosiglitazone (Sigma-Aldrich; I8280), 0.25 mM 3-isobutyl-1-methylxanthine (Sigma-Aldrich; 15879), 1 mM dexamethasone (Sigma-Aldrich; D4902), and 10 μg/ml insulin (Bevatime; P3376). The cells were treated with an induction culture medium for 48 h and then cultured in a complete medium supplemented with 10 μg/ml insulin until they were collected at specific time points.

Western blot

After being mixed in RIPA buffer containing 1% PMSF and phosphatase inhibitors, the tissues and cells were broken down. The resulting lysates were then spun at 14,000 rpm for 20 min at 4°C to remove any solid particles. The adipose tissue homogenates underwent serial centrifugation steps at 7,000 rpm for 25 min to eliminate the floating layer. Protein was quantified using a BCA assay kit (New Cell & Molecular Biotech; WB6501) per the manufacturer’s instructions. Proteins were separated using 10% SDS-PAGE and transferred onto 0.45 μm nitrocellulose membranes. The membranes were blocked with 5% bovine serum albumin for 1 h, then incubated with specific primary antibodies (refer to supplemental Table S2) at 4°C for 14 h, followed by washing to eliminate any unattached primary antibodies. The blots were then exposed to the correct secondary antibody (refer to supplemental Table S2) for 2 h. Images were acquired and analyzed using an enhanced chemiluminescence detection system, and relative protein expression was determined with ImageJ 1.54g (Wayne Rasband and contributors National Institutes of Health, Bethesda, MD).

RT-PCR

Total RNA from cells or tissues was isolated with a kit (Fastegen Shanghai; 220011). Equal amounts of RNA were reverse transcribed into complementary DNA (cDNA) using a Prime Script kit (Takara; RR037A). Finally, RT-PCR was conducted on the thermal cycler (StepOne, Thermo Fisher Scientific) using TB-Green fluorescent dye (Takara; RR820A/B). The relative expression of genes was determined using the ΔΔCt technique. supplemental Table S3 shows a list of primer pairs.

Histological analysis and Oil Red O staining

For H&E staining, liver and adipose tissues were preserved in 4% formalin, encased in paraffin, and sliced into 6 μm sections. These sections were dyed with H&E. Oil Red O provided by Hubei Biossci Biotechnology Co, Ltd. Specifically, Oil Red O staining of liver tissue was performed on frozen sections (8 μm) prepared using a cryostat (Thermo Fisher Scientific; HM525 NX U). Sections were fixed in 4% paraformaldehyde for 15 min, rinsed in 60% isopropanol, and stained with freshly prepared Oil Red O working solution for 10–15 min. Sections were then washed in 60% isopropanol, counterstained with hematoxylin, mounted with aqueous mounting medium, and imaged under an optical microscope (OM; Olympus Corporation; CX-31). For Oil Red O staining in vitro, cells were washed with PBS thrice and then fixed in 4% paraformaldehyde for 15 min. Then, the cells were exposed to 60% (v/v) isopropanol solution for 5 min before being stained with Oil Red O. Following the removal of the Oil Red O, the cells were rinsed with distilled water five times. observed using an OM (Nikon Corporation; ECLIPSE Ts2R). The Oil Red O-positive area from the cells was then quantified with ImageJ software to provide a relative measure of lipid accumulation.

Total triglyceride content assays

Liver tissues were weighed and homogenized in a lysis buffer suitable for lipid extraction. The homogenates were centrifuged to remove cellular debris, and the resulting supernatants were collected and diluted with normal saline for analysis. Triglyceride (TG) levels were determined using the Automated Biochemical Analyzer (Chemray 240; Rayto Life and Analytical Sciences, Shenzhen, China). A dedicated dual-reagent enzymatic colorimetric kit (S03027; Rayto Life and Analytical Sciences, Shenzhen, China) based on the Glycerol-3-phosphate oxidase–peroxidase method was employed for the assay, strictly adhering to the manufacturer's standard operating procedures. Briefly, the analyzer was programmed with optimized parameters for the sequential addition of working solution R1 (buffer and enzymatic cofactors) and working solution R2 (chromogenic substrate). Following sample loading and incubation at 37°C, the instrument automatically performed reagent dispensing, and absorbance was measured photometrically. Absolute TG concentrations were calculated against a multipoint calibration curve generated from standard reference materials, and data were automatically exported for analysis. Liver TG content was expressed as micrograms of TG per milligram of liver wet weight (μg/mg liver).

Coimmunoprecipitation and in vitro ubiquitination assays

Proteins were extracted from 3T3-L1 cells using an immunoprecipitation (IP) lysis buffer. A fraction of the lysate was retained as input. Then, the lysates were gently agitated and incubated at 4°C overnight with the primary antibody (refer to supplemental Table S2). Next, we added protein agarose beads (Santa Cruz Biotechnology; sc-2003) to the lysate, which was subsequently incubated at 4°C for 7 h. The beads were separated by centrifugation and washed thrice with PBS. The bead-bound proteins were eluted by heating at 100°C for 15 min. Western blot analysis was then conducted on the samples. For in vitro ubiquitination assays, the samples were pretreated with 5 μM MG132 for 5 h before collection. Co-IP experiments were conducted using primary antibodies (refer to supplemental Table S2), and poly (ADP-ribose) polymerase 1 (PARP1) ubiquitination levels were assessed using Western blot.

Sample preparation and RNA sequencing

Total RNA was prepared as described previously (30). Eukaryotic mRNA was used to construct specific cDNA libraries. cDNA libraries were sequenced using combination probe-anchor synthesis on DNBSEQ-T7 (Bioyi Biotechnology Co Ltd, Wuhan, China). The paired-end clean reads were aligned to the reference genome using HISAT2 v2.1.0 (Johns Hopkins University, Baltimore, MD). The raw gene expression levels were quantified with StringTie v2.1.5 (Johns Hopkins University, Baltimore, MD) and normalized to the number of fragments per kilobases per million fragments. Differential gene expressions were compared using DESeq2 R package v1.30.1 (Open Source in Bioconductor). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses of the indicated groups were conducted using the clusterProfiler R package v3.8.1 (Open Source in Bioconductor).

Glucose metabolism analysis

Glucose tolerance test (GTT) and insulin tolerance test (ITT) were used to evaluate glucose metabolism and insulin sensitivity. For the GTT, mice were injected intraperitoneally with glucose at a dose of 2 mg/g of body weight. For the ITT, crystalline zinc insulin was injected at a dose of 1 U/kg of body weight. We obtained blood samples at specified intervals following the injection of either glucose or insulin. Circulatory glucose levels were detected, with the 0 min reading serving as the baseline. Insulin levels were measured using an ELISA kit (Elabscience; E-EL-M1382) per the manufacturer’s instructions.

Metabolic cage analysis

The mice were subjected to 48 h of data collection following a week of acclimation to metabolic cages (Columbus Instruments, Columbus, OH). The oxygen consumption ( $[eqn]$ O_2_), CO_2_ production ( $[eqn]$ CO_2_), heat production, and average food intake were assessed. The respiratory exchange ratio was calculated as the ratio of $[eqn]$ O_2_ to $[eqn]$ CO_2_.

Flow cytometry

Liver and white adipose tissue were minced and enzymatically digested to isolate immune cells. The Fc receptors were blocked using an anti-CD16/32 antibody (BioLegend; 101303). Next, the cells were exposed to the antibody solutions at 4°C for 15 min without light. The following antibodies (BioLegend) were used for surface staining: anti-CD11b-PE-Cy7 (101279), anti-CD45-Percp (103129), anti-F4/80-PE (111604), anti-CD86-FITC (159219), anti-CD19-APC (152409), anti-CD3-BV510 (100234), anti-CD4-FITC (100405), and anti-CD8-APC-Cy7 (100713). Following the immunofluorescence staining process, the cells were rinsed and assessed using a FACS flow cytometer (BD Biosciences). The data were analyzed with FlowJo software (version 10; TreeStar).

Cytokine assays

Serum samples from WT and Usp25-KO mice fed an HFD (n = 3 for each group) were analyzed for cytokine profiles as described previously (31). The median fluorescence intensity was determined using FCAP Array™ v3.0 software (BD Bioscience). All samples were analyzed in triplicate to ensure data reliability for statistical analyses. Data points with values deviating over 20% from the mean were excluded, and the mean was recalculated.

Statistical analysis

The data are expressed as the means ± standard deviations. Analyses and visual representations were conducted using GraphPad Prism (version 10.0, GraphPad Software, San Diego, CA). The statistical tests used are described in the figure captions and were selected based on the assumptions of a normal distribution and equal variances. An unpaired, two-tailed Student’s t-test was used when two groups were compared. For analyses involving multiple groups, ANOVA followed by Sidak’s multiple comparisons test was used. Two-way ANOVA was conducted to evaluate the interaction between two independent variables. Area under the curve (AUC) values were calculated using the “Area Under Curve” analysis in GraphPad Prism 10. To control baseline (0-min) differences, baseline-adjusted AUCs were additionally analyzed by ANCOVA with fasting glucose as a covariate. Statistical significance was determined at P < 0.05, and the P values are provided in the figure captions.

Results

USP25 is downregulated in adipose tissue following long-term HFD exposure

We analyzed data from a previous study on the human subcutaneous and visceral adipocyte atlas to investigate the relationship between obesity and USP25 (32). The analysis showed differential expression of the USP family in adipocytes across various obesity categories. Notably, USP25 expression was reduced in individuals with moderate/severe obesity compared with those with mild obesity (Fig. 1A). In vivo, we established a murine obesity model with C57BL/6 mice using HFD feeding to elucidate USP25 role in obesity development. The model was validated by comparing fasting body weight (Fig. 1B). For Oil Red O staining, frozen liver sections were prepared from HFD-fed mice after 10 weeks and compared with those from baseline control mice at 6–8 weeks of age prior to dietary intervention (Fig. 1C). As expected, HFD-fed mice showed markedly increased hepatic total lipid accumulation compared with the control group. Oil Red O staining is a qualitative indicator to visualize the intrahepatic distribution and morphology of lipid droplets. Recognizing the semiquantitative nature of histological staining, we subsequently performed a rigorous enzymatic quantification of total hepatic TG content to validate our findings. These biochemical assays corroborated the histological evidence, demonstrating that hepatic TG levels were significantly elevated in the HFD-fed cohort compared with the standard diet controls (Fig. 1D), thereby confirming the establishment of the hepatic steatosis phenotype. Notably, the investigation showed significantly reduced USP25 protein levels in the inguinal white adipose tissue of obese mice after 10 weeks of HFD feeding, compared with those at baseline (0 weeks) and after 5 weeks of HFD (Fig. 1E). The molecular mass of the USP25 protein in white adipose tissue was found to be approximately 135 kDa, consistent with the predicted 130 kDa size of USP25m, suggesting that USP25m is the major splice form expressed in adipose tissue (28). Due to the lengthy model construction period, to eliminate the influence of age on USP25 expression in mouse adipose tissue, we examined USP25 expression in adipose tissue of mice at different weeks of age under ND feeding conditions. The results showed no significant differences in USP25 expression in adipose tissue among mice of different ages (supplemental Fig. S1A, B). In vitro experiments were conducted by differentiating 3T3-L1 cells into mature adipocytes in DMEM supplemented with rosiglitazone, methyl isobutyl xanthine, dexamethasone, and insulin. The induction after 7 days was confirmed using OM observing and Oil Red O staining (Fig. 1F). Similarly, the in vitro results showed significantly reduced USP25 protein expression in induced adipocytes (Fig. 1G), which was further confirmed at the transcriptional level using RT-PCR (Fig. 1H). Collectively, these findings show that USP25 expression was associated with the severity of obesity.Fig. 1USP25 is downregulated in the adipose tissue of HFD-fed mice and mature 3T3-L1 adipocytes. A: Expression of USPs in people with mild and moderate/severe obesity. B: Fasting body weights of WT mice fed an ND (n = 4) or an HFD (n = 8) for 10 weeks. C: Representative images of H&E and Oil Red O stainings of liver tissue from WT mice prior to HFD intervention (control) and after 10 weeks of HFD feeding. D: TG content in the livers of the indicated groups (n = 6). E, Left panel: Representative Western blot showing the expression level of USP25 in the inguinal white adipose tissue of mice fed an HFD for 0, 5, and 10 weeks. Right panel: density analysis of USP25 protein level normalized to that of GAPDH (n = 6). F: Left panel: OM observation and Oil Red O staining of 3T3-L1 cells differentiated for 7 days. Right panel: statistical analysis of positive Oil Red O-stained areas in the indicated groups (n = 3). G: Left panel: Representative Western blot showing the expression level of USP25 in 3T3-L1 cells and induced 3T3-L1 adipocytes. Right panel: Density analysis of USP25 protein level normalized to that of GAPDH (n = 4). H: RT-PCR analyses of Usp25 expressed by 3T3-L1 cells and mature 3T3-L1 adipocytes (n = 6). Data are expressed as means ± SD. Significant differences were indicated with P values.

USP25 deficiency mitigates HFD-induced obesity and insulin resistance in mice

Given that USP25 is essential in modulating proinflammatory pathways and glycolysis (33, 34), with both contributing to the pathophysiology of obesity, we investigated the effect of USP25 deletion on adipose tissue development during obesity progression. The whole-body Usp25-KO mice model was generated based on our previous research (29) and subjected to a 10-week HFD regimen. Notably, compared with HFD-fed WT mice, HFD-fed Usp25-KO mice showed significantly reduced weight gain (Fig. 2A, B). Consistent with these observations, the weights of adipose tissue were assessed and found to be significantly lower in HFD-fed Usp25-KO mice than in their WT counterparts, with no significant differences observed between WT and Usp25-KO ND-fed mice (Fig. 2C). These quantitative findings were further supported by visual inspection of the adipose tissues (Fig. 2D). Histological analysis showed that adipose tissue of HFD-fed Usp25-KO mice contained smaller adipocytes and showed reduced cell cluster density (Fig. 2E). This observation suggests that these mice have impaired susceptibility to HFD-induced adipocyte maturation. Additionally, we analyzed the glucose metabolic status of the indicated four groups using GTT and ITT experiments. The GTT curve results showed that USP25 deficiency significantly decreased plasma glucose levels under HFD conditions (Fig. 2F), indicating a significant improvement in glucose tolerance in HFD-fed Usp25-KO mice. ITT curve results also showed decreased glucose levels after insulin injection, suggesting enhanced insulin tolerance (Fig. 2G). For the overall changes in glucose levels, we performed ANCOVA on the GTT and ITT data using baseline blood glucose levels as covariates. The baseline-adjusted AUC analyses for both the GTT (Fig. 2H) and ITT (Fig. 2I) demonstrated that USP25 deficiency is associated with improved glucose tolerance and enhanced insulin sensitivity, independent of fasting glucose variations. To further dissect the temporal dynamics of this metabolic improvement, we examined blood glucose levels normalized to baseline at specific time points. And the analysis revealed that the statistically significant disparity between the WT and Usp25-KO mice persisted at 60 min postchallenge in the GTT and at 15 min during the ITT (supplemental Fig. S2A, B), confirming that the metabolic protection of Usp25 KO is driven by enhanced glucose clearance kinetics and insulin responsiveness. In addition, we examined whether USP25 deficiency leads to ectopic lipid accumulation in the liver, which is a typical complication in lipodystrophy (35). Representative H&E and Oil Red O staining of liver sections from ND-fed Usp25 KO and WT mice revealed that KO livers exhibited normal histological architecture with no evidence of increased lipid droplet accumulation compared with WT controls (supplemental Fig. S2C). Under HFD conditions, hepatic lipid accumulation in Usp25-KO mice did not differ significantly from WT, indicating that USP25 deficiency does not exacerbate HFD-induced hepatic steatosis (supplemental Fig. S2D, E). These findings suggest that the reduced adiposity observed in Usp25-KO mice is not attributable to ectopic fat deposition in the liver. In summary, these findings showed that USP25 deficiency mitigates HFD-induced obesity and insulin resistance.Fig. 2USP25 deficiency alleviated the progression of obesity and insulin resistance in HFD-fed mice. A: Body weights of WT and Usp25-KO mice fed an HFD for 10 weeks were recorded weekly (n = 5 [WT + ND; KO + ND], n = 7 [KO + HFD], and n = 8 [WT + HFD]). B: Images of WT and Usp25-KO mice fed an HFD for 10 weeks; the scale bar represents 1 cm. C: Weights of adipose tissue from WT and Usp25-KO mice fed an HFD for 10 weeks (n = 5 [WT + ND; KO + ND] and n = 8 [WT + HFD; KO + HFD]). D: Images of brown adipose tissue (BAT), inguinal white adipose tissue (iWAT), and epididymal white adipose tissue (eWAT) originating from WT and Usp25-KO mice fed an HFD for 10 weeks; the scale bar represents 1 cm. E: Representative images of H&E staining of brown adipose, iWAT, and eWAT tissues; the scale bar represents 100 μm. F: GTT of WT and Usp25-KO mice fed an ND (left panel, n = 5) or an HFD (right panel, n = 6) for 10 weeks. G: ITT of WT and Usp25-KO mice fed an ND (left panel, n = 4) or an HFD (right panel, n = 7) for 10 weeks. The baseline-adjusted AUC of GTT (H) and ITT (I) of the indicated groups. Data are expressed as means ± SD. #P < 0.05 compared with the WT + ND group, ∗P < 0.05 compared with the WT + HFD group. The other significant differences were indicated with P values.

Assessment of food intake, energy expenditure, and immune status in HFD-fed Usp25-KO mice

To explore the multifaceted effects of USP25 on diet-induced obesity, WT and Usp25-KO mice were randomly selected and placed in metabolic cages, both of which were subjected to a 10-week HFD regimen to investigate whether the impaired obesity phenotype in Usp25-KO HFD-fed mice was attributable to reduced food intake, decreased activity-dependent energy expenditure, or alterations in systemic inflammatory status. The mice in metabolic cages were acclimated for 7 days, followed by 48 h of data recording. Metabolic cage data showed no significant difference in food intake normalized to body weight between HFD-fed WT and Usp25-KO mice (Fig. 3A), whereas Usp25-KO mice exhibited a trend toward lower absolute food intake (supplemental Fig. S3A). This minor reduction in caloric input over the course of the study likely contributes to the observed resistance to diet-induced obesity, consistent with the cumulative effects of energy balance. Furthermore, hourly heat production measurements showed no significant differences in energy expenditure between the groups (Fig. 3B). Similarly, $[eqn]$ O_2_, $[eqn]$ CO_2_ values, and respiratory exchange ratio were comparable between the groups (Fig. 3C, supplemental Fig. S3B, C), indicating that USP25 deficiency did not significantly affect thermogenesis or subsequent fat consumption. Serum biochemical assays showed no significant differences in circulatory insulin levels between WT and Usp25-KO mice after 10 weeks of HFD feeding (Fig. 3D). Serum-soluble cytokines were analyzed using flow cytometry as described previously (31). Usp25-KO mice presented lower levels of TNF-α, IL-5, and IL-10 than did WT mice after 10 weeks of HFD consumption; however, most proinflammatory cytokine levels remained unchanged (Fig. 3E). Additionally, based on immune cells, the relative abundances of hepatic macrophages, epididymal white adipose tissue macrophages, or epididymal white adipose tissue CD4^+^/CD8^+^ T cells in HFD-fed mice were not significantly affected by USP25 deficiency (Fig. 3F–I). In summary, Usp25 KO did not affect fat consumption or immune status in HFD-induced obese mice.Fig. 3. Assessment of food intake, energy expenditure, and immune status in HFD-fed *Usp25-*KO mice. WT and Usp25-KO mice fed an HFD for 10 weeks were placed into metabolic cages. Food intake (A: WT, n = 8; KO, n = 7), heat production (B), and respiratory exchange ratio (RER) ( $[eqn]$ CO_2_/ $[eqn]$ O_2_) (C) were monitored within 48 h. D: Plasma insulin levels of WT and Usp25-KO mice fed an HFD for 10 weeks (n = 6). E: Levels of eight cytokines in plasma from WT and Usp25-KO mice fed an HFD for 10 weeks (n = 3). Immune cells were isolated from different tissues as indicated, and the relative abundance of F4/80^+^ macrophages in the liver (F) and epididymal white adipose tissue (eWAT) (G), CD4^+^ T cells, and CD8^+^ T cells in eWAT (H) were analyzed via fluorescence-activated cell sorting (n = 3). I: Statistical analysis of immune cell abundance measured using flow cytometry (n = 3). Data are expressed as means ± SD. ∗P < 0.05, ∗∗P < 0.01.

The differentiation of 3T3-L1 cells into mature adipocytes is hindered by a lack of USP25

We introduced undifferentiated 3T3-L1 preadipocytes to an siRNA aimed at USP25 to further elucidate how USP25 influences the transcriptional control of adipogenesis in an adipocyte induction model. In this study, we used scrambled siRNA for comparison. Compared with those in the untreated group, the USP25 protein levels in the USP25-siRNA-transfected 3T3-L1 cells were significantly lower (Fig. 4A). RT-PCR analysis further showed the interference of the siRNA with the normal transcription of Usp25 (Fig. 4B). Using an in vitro induction protocol, preadipocytes were differentiated into mature adipocytes. Subsequent OM observations and Oil Red O staining showed impaired differentiation of USP25-depleted preadipocytes into mature adipocytes (Fig. 4C, D). Moreover, the transcript levels of adipocyte-related genes in 3T3-L1 cells transfected with USP25-siRNA were significantly lower than those in the control cells after 7 days of postinduction (Fig. 4E). These findings show that USP25 expression is essential for 3T3-L1 preadipocyte differentiation into mature adipocytes.Fig. 4. Depletion of USP25 retards 3T3-L1 differentiation into adipocytes. Knockdown of USP25 was confirmed using Western blot (A, n = 4) and RT-PCR analyses (B, n = 3). C: Representative images of 3T3-L1 cells cultivated in blank medium for 7 days (control), scrambled siRNA negative control (si-NC), and si-USP25 groups differentiated for 7 days, recorded via an OM; the scale bar represents 100 μm. D, Left panel: Oil Red O staining of control, si-NC, and si-USP25 cultures subjected to 7 days of adipocyte differentiation; the scale bar represents 100 μm. Right panel: Statistical analysis of positive Oil Red O-stained area in the indicated groups (n = 3). E: RT-PCR analyses of adipocyte-specific genes in control, si-NC, and si-USP25 cultures after 7 days of adipocyte differentiation (n = 3). Data are expressed as means ± SD. Significant differences were indicated with P values.

USP25 deficiency promotes PARP1 ubiquitination, leading to decreased Elovl3 expression and compromised adipogenesis

After HFD feeding for 10 weeks, we conducted RNA sequencing (RNA-Seq) analysis of the epididymal white adipose tissue of WT and Usp25-KO mice to investigate the mechanisms of impaired adipose tissue development in Usp25-KO mice. Principal component analysis of the differentially expressed gene clusters showed distinct gene expression profiles between the WT and Usp25-KO HFD-fed mice (supplemental Fig. S4A, B). An analysis of the statistical data showed 422 transcripts that were distinctively expressed between the two groups (false discovery rate-adjusted P value <0.05; log2-fold change greater than x > 0.5) (Fig. 5A, supplemental Table S4). Down GO and KEGG enrichment analysis showed that the transcripts significantly downregulated upon Usp25 KO were enriched primarily in genes associated with cellular mitosis and lipid metabolism (Fig. 5B, C). Among the 267 downregulated protein-coding transcripts, elongation of very long-chain fatty acid protein 3 (ELOVL3) expression was significantly reduced in the Usp25-KO HFD-fed mice compared with their WT counterparts (Tables S4–S6). RT-PCR analysis of RNA obtained from adipose tissue showed that the ELOVL3 transcript level was significantly lower in Usp25-KO HFD-fed mice than in WT mice (Fig. 5D). Similarly, the induction of ELOVL3 expression was significantly blunted in 3T3-L1 cells transfected with Usp25-siRNA following stimulation with induction culture medium (Fig. 5E). Furthermore, to elucidate the relationship between USP25 and ELOVL3, we monitored the expression of USP25 and ELOVL3 at specific time points throughout the differentiation process. Analysis of both protein (supplemental Fig. S4C, D) and mRNA (supplemental Fig. S4E) levels revealed a correlation between USP25 upregulation and ELOVL3 induction, suggesting that USP25 abundance is closely coupled to the activation of the lipogenic program during adipogenesis.Fig. 5USP25 deletion decreased Elovl3 expression and increased PARP1 ubiquitination, leading to impaired adipose differentiation. A: Downregulated and upregulated genes in epididymal white adipose tissue (eWAT) from WT and Usp25-KO mice fed an HFD for 10 weeks. B: GO term enrichment analysis. C: KEGG pathway enrichment analysis. D: RT-PCR analyses showing relative mRNA expression of Elovl3 in eWAT from WT and Usp25-KO mice fed an HFD for 10 weeks (n = 6). E: RT-PCR analyses showing relative mRNA expression of Elovl3 in 3T3-L1, si-NC, and si-USP25 cultures after 7 days of adipocyte differentiation (n = 3). F: Co-IP assays were conducted to evaluate the interactions between USP25 and ELOVL3 in 3T3-L1 cells. G: Co-IP assays were conducted to evaluate the interactions between USP25 and PARP1. H, Left panel: Representative Western blot showing the expression of PARP1 in the indicated groups. Right panel: Density analysis of PARP1 protein level normalized to that of GAPDH (n = 5). I: Western blot analysis of the level of ubiquitinated PARP1 in 3T3-L1 cells transfected with USP25 siRNA, with MG132 added for 6 h before harvesting. Knockdown of PARP1 was confirmed by Western blot (J, n = 4) and RT-PCR analyses (K, n = 3). L, Left panel: Oil Red O staining of 3T3-L1 cells cultivated in blank medium for 7 days, si-NC, and si-PARP1 cultures subjected to 7 days of adipocyte differentiation; the scale bar represents 100 μm. Right panel: Statistical analysis of positive Oil Red O-stained area in the indicated groups (n = 3). M: RT-PCR analyses showing relative Elovl3 mRNA expression in 3T3-L1, si-NC, and si-PARP1 cultures after 7 days of adipocyte differentiation (n = 3). Data are expressed as means ± SD. Significant differences were indicated with P values.

Reportedly, USP25 facilitates substrate deubiquitination through direct interaction with target proteins (36, 37). Then, co-IP experiments were conducted to show the relationship between USP25 and ELOVL3. Surprisingly, there were no detectable interactions between the two proteins (Fig. 5F). However, we observed an interaction between USP25 and PARP1 (Fig. 5G). Previous studies have implicated PARP1-mediated poly-ADP ribosylation in the etiology of various lipid metabolism disorders (38). Subsequently, we transfected 3T3-L1 cells with siRNA targeting USP25, which led to a significant reduction in PARP1 expression level in the transfected group compared with that in the scrambled siRNA control group (Fig. 5H). Furthermore, in vitro ubiquitination assays showed increased ubiquitin conjugation to PARP1 in USP25-siRNA-transfected 3T3-L1 cells (Fig. 5I), indicating that the reduction in PARP1 protein levels was mediated by ubiquitination and proteasomal degradation. After confirming that USP25 modulates PARP1 deubiquitination, we investigated the effect of PARP1 on ELOVL3 expression. We transfected 3T3-L1 cells with siRNA-targeting PARP1 to assess the effects of PARP1 depletion in a murine adipogenesis model. Western blotting and RT-PCR analyses showed effective PARP1 knockdown after siRNA treatment (Fig. 5J, K). 3T3-L1 cells were subsequently induced to differentiate. The result of Oil Red O staining showed a notable hindrance in the differentiation of PARP1-deficient 3T3-L1 cells into mature adipocytes (Fig. 5L). RT-PCR analysis of 3T3-L1 cells, induced adipocytes, and PARP1-siRNA-transfected differentiated cells showed that ELOVL3 expression was significantly lower in transfected differentiated cells than in mature adipocytes, with no significant difference from that in 3T3-L1 cells (Fig. 5M).

Discussion

In this study, we identified a novel mechanism by which USP25 influences lipid synthesis in the context of obesity. We observed a significant reduction in USP25 expression in the adipose tissue of HFD-induced obese mice and in mature adipocytes differentiated in vitro. Previously, studies have shown that USP25 overexpression mitigates obesity-associated complications, including MAFLD (39) and obesity-related cardiomyopathy (40). Conversely, in this study, USP25 deficiency in murine models and preadipocytes impaired adipocyte differentiation. Research has shown that Elovl3-KO mice show decreased lipid accumulation and compromised metabolic function in brown adipose tissue (41), which is essential for adipocyte differentiation and maturation. Mechanistically, differential gene enrichment analysis showed a significant decrease in ELOVL3 expression in the adipose tissue of Usp25-KO mice despite USP25 indirectly interacting with or regulating ELOVL3 expression. Moreover, PARP1 is a pivotal regulator of adipocyte maturation (42). In this study, we discovered that USP25 directly modulated the PARP1 ubiquitination status. The reduced PARP1 protein levels significantly reduced Elovl3 transcription, thereby inhibiting preadipocyte differentiation and lipid production. The USP25-PARP1-ELOVL3 axis, which specifically targets preadipocytes, may be a promising therapeutic target for managing the progression of obesity.

Studies on USP25 have primarily focused on oncology, neurodegenerative disorders, and antiviral immunity (15, 43, 44, 45). However, recent results have implicated dysregulated USP25 expression in a range of obesity-related diseases, suggesting a role in obesity development and progression. For instance, Habtemichael et al. (28) showed that Usp25 knockdown impaired insulin-stimulated GLUT4 translocation, thereby disrupting downstream signaling. Throughout our investigation, we observed a significant reduction in USP25 expression in the adipose tissue of HFD-induced obese mice, a result that was also confirmed in vitro in 3T3-L1 adipocytes. Importantly, our results provide the first evidence that USP25 deficiency impairs adipose tissue development and mature adipocyte differentiation, underscoring its importance in obesity pathogenesis.

Prolonged HFD exposure induces a proinflammatory physiological state and reduces insulin sensitivity (46). Chronic inflammation and insulin resistance are essential contributors to lipid metabolic dysfunction. The bidirectional relationship between insulin resistance and proinflammatory states has been investigated. For instance, proinflammatory cytokines, including IL-1β and TNF-α, interfere with insulin function through the inhibition of insulin signal transduction (47), whereas insulin resistance impairs the anti-inflammatory effects of insulin and eventually exacerbates inflammation (48). With the exclusion of the USP25 effect on food intake and energy metabolism, glucose metabolism analysis showed that Usp**25-KO mice had alleviated systemic insulin resistance under HFD conditions without significantly increasing the levels of most serum proinflammatory cytokines. Additionally, USP25 deficiency showed no significant effect on proinflammatory cells in the liver or adipose tissue. Interestingly, Usp**25-KO mice displayed improved insulin sensitivity following HFD feeding; nevertheless, their circulating insulin concentrations remained elevated, with no significant differences observed from those in the HFD-fed control group. The mismatch between insulin sensitivity and circulating insulin levels is possibly due to the minimal effect of USP25 deficiency on the levels of inflammatory cytokines, which can still stimulate islet β-cells to secrete insulin (49). Collectively, the in vivo results show that USP25 deficiency has no apparent effect on the HFD-induced proinflammatory physiological state; nonetheless, it exerts a protective effect on systemic insulin sensitivity.

Adipocyte differentiation is another important factor in obesity development. With our observation of impaired adipose tissue development in Usp**25-KO HFD-fed mice, we hypothesized that USP25 deficiency might affect adipocyte differentiation. Consistent with our findings in adipose tissue, in vitro studies showed that USP25 knockdown in 3T3-L1 cells significantly inhibited their differentiation into mature adipocytes. Gene expression analysis further showed impaired adipocyte functions, especially lipid synthesis. These findings highlight the importance of USP25 in adipocyte differentiation and lipid synthesis.

ELOVL3 is a key enzyme in adipose tissues, primarily responsible for synthesizing C20-C24 saturated and monounsaturated very long-chain fatty acids. Zadravec et al. (41) reported that ELOVL3-deficient mice had reduced adipose tissue expansion and resistance to diet-induced obesity. Additionally, their findings further showed that ELOVL3 deficiency impairs hepatic lipogenesis and reduces TG levels in the liver and serum, highlighting its importance in lipid metabolic homeostasis. Another study reported that liver-specific Elovl3 KO did not observably alter hepatic lipid homeostasis (50). However, researchers have identified a compensatory role for extrahepatic ELOVL3 in regulating metabolic homeostasis. In this study, RNA-Seq data showed a significant reduction in ELOVL3 transcript levels in the adipose tissue of Usp**25-KO mice fed an HFD, with down GO and KEGG enrichment analysis showing a significant downregulation of cholesterol and steroid synthesis pathways. Consistent with these findings, we observed significantly lower ELOVL3 mRNA levels in the epididymal white adipose tissue of Usp**25-KO mice than in those of control mice, which was further supported using in vitro experiments. Notably, although we did not observe downregulation of classical adipogenic regulators, such as Pparg or Fasn, we detected a reduction in Acss2, a lipid biosynthetic gene reported to participate in acetyl-CoA metabolism and lipogenesis (51, 52, 53). This suggests that the attenuated adipose expansion in Usp**25-KO mice may primarily reflect impaired lipid synthesis due to the downregulation of multiple lipid metabolic genes, rather than a block in preadipocyte differentiation per se. These findings mean that Usp25 KO ameliorates obesity through a reduction in ELOVL3 expression. Nevertheless, co-IP experiments showed that USP25 indirectly interacts with ELOVL3, ruling out the possibility of stabilizing protein expression through deubiquitination.

PARP1 is a well-characterized member of the PARP family that is essential in inflammatory responses and repairs of DNA damage. After being activated by damaged DNA, PARP1 mediates PARylation of downstream targets, influencing multiple biological processes (42). In fact, the impact of PARP1 on obesity susceptibility and lipid metabolism remains controversial. For instance, Erener et al. (54) reported that PARP1-deficient mice display increased hepatic steatosis and impaired adipocyte function. By contrast, Ai et al. (55) have reported that pharmacological inhibition of PARP1 exerts beneficial effects on metabolic diseases, including MAFLD. Consistently, Hans et al. (56) demonstrated that PARP1 inhibition ameliorates HFD-induced atherosclerosis by lowering total cholesterol and low-density lipoprotein levels. Our findings show that USP25 directly interacts with PARP1, removing its ubiquitin chains and promoting its stability. Furthermore, USP25 deficiency significantly reduced the PARP1 protein level in 3T3-L1 cells. We investigated the role of PARP1 in adipogenesis using adipocyte differentiation assays. The Oil Red O staining results showed a decrease in lipid droplet formation of PARP1-knockdown cells. Importantly, RT-PCR results showed that ELOVL3 expression was significantly reduced in the PARP1-knockdown cells than in the normally differentiated cells. These findings show that PARP1 suppresses adipocyte differentiation through ELOVL3 transcription decline.

This study had some limitations. Adipocyte-specific KO mice were not used in the experiment, and reducing the expression of USP25 in adipocytes specifically to eliminate the influence of other cell types is necessary. The exact mechanism by which PARP1 deletion alters ELOVL3 expression remains unclear, and overexpression of ELOVL3 in Usp**25-KO 3T3-L1 cells would provide definitive evidence for the mechanism. This study was conducted exclusively in male mice, which may limit the generalizability of the findings. The precise mechanism underlying the reduced body weight observed in USP25-deficient mice under HFD feeding remains unclear, and further validation through direct fecal energy measurements and complementary metabolic phenotyping will be required in future work. In addition, our transcriptomic analysis revealed enrichment of mitosis- and cell cycle-related pathways, raising the possibility that USP25 may influence proliferative processes during early adipogenesis. However, this aspect was not specifically addressed in our current experimental design and will require dedicated studies to clarify. Whole-body composition analyses, such as EchoMRI or dual-energy X-ray absorptiometry, which would provide accurate assessments of lean and fat mass, were not available to us during this study, and this represents another limitation. Finally, this is a preclinical study, and additional safety verification is required.

Together, our findings establish USP25 as a key regulator of adipose tissue expansion in diet-induced obesity. Mechanistically, USP25 deficiency attenuates obesity progression by promoting PARP1 ubiquitination and downregulating ELOVL3, thereby restricting adipocyte differentiation and maturation (Fig. 6). Notably, although Usp25 KO did not broadly reduce inflammatory cytokines, it markedly improved glucose tolerance and insulin sensitivity and limited adipose tissue expansion in HFD-fed mice. These results illustrate that curtailing adipose tissue expansion via USP25 yields metabolic benefits in obesity independent of classical inflammatory pathways.Fig. 6. The molecular mechanism by which USP25 deficiency mitigates HFD-induced obesity progression involves proteasomal degradation of PARP1. Under HFD conditions, USP25 prevents the proteasomal degradation of PARP1 by eliminating ubiquitin chains. PARP1 ensures fatty acid synthesis during adipocyte differentiation by sustaining ELOVL3 expression, which results in the development of adipose tissue. However, this situation is reversed when USP25 is absent.

Data availability

Raw bulk RNA-Seq data were deposited to the Genome Sequence Archive website (https://ngdc.cncb.ac.cn/gsub/) affiliated with the National Genomics Data Center (https://ngdc.cncb.ac.cn/) with identifier (CRA030956). All other data will be made available upon reasonable request.

Supplemental data

This article contains supplemental data.

Conflict of interest

The authors declare that they have no conflicts of interest with the contents of this article.