Exercise and dietary interventions ameliorate MASLD via the hepatic PPARγ-miR-802-Psmd2 axis

Abstract

Background

MicroRNAs critically regulate metabolic dysfunction-associated steatotic liver disease (MASLD) progression and intervention. Lifestyle modifications, particularly exercise and time-restricted feeding (TRF), emerge as effective non-pharmacological strategies.

Methods

MiR-802 expression was quantified in plasma exosomes and hepatic tissues of MASLD mouse models using quantitative PCR and sequencing. In vivo and in vitro gain-of-function and loss-of-function experiments assessed its regulatory effects on hepatic lipotoxicity, complemented by bioinformatic predictions and luciferase-based validation of upstream transcription factors and downstream targets. Functional crosstalk between miR-802 and its regulatory network was examined under lipotoxic conditions. Therapeutic interventions, including exercise and TRF, were implemented in high-fat diet fed mice, with subsequent evaluation of miR-802 expression dynamics.

Results

Here, we show elevated miR-802 levels in plasma exosomes and livers of MASLD mice, with its inhibition reducing hepatocyte lipotoxicity. Mechanistically, peroxisome proliferator-activated receptor gamma (PPARγ) promotes miR-802 transcription, whereas miR-802 suppresses the proteasome 26S subunit ubiquitin receptor, non-ATPase 2 (Psmd2). Genetic and pharmacological modulation of PPARγ downregulates miR-802 and ameliorates MASLD. Notably, both exercise and TRF reduce hepatic PPARγ and miR-802 expression and improve MASLD via the PPARγ-miR-802-Psmd2 axis, revealing lifestyle modification-mediated epigenetic regulation in MASLD pathophysiology.

Conclusion

This study reveals the PPARγ-miR-802-Psmd2 axis as a central regulatory mechanism in MASLD pathogenesis and innovatively links lifestyle interventions (exercise and TRF) to therapeutic effects via this pathway.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Exercise and dietary interventions have been proven to ameliorate metabolic dysfunction-associated steatotic liver disease (MASLD), yet their underlying mechanisms remain incompletely understood.

• Our study elucidates these molecular mechanisms, offering novel insights for MASLD treatment.

WHAT THIS STUDY ADDS

• Our study revealed that exosomal miR-802 in plasma plays a critical role in the liver during MASLD progression and elucidates its molecular mechanisms underlying the beneficial effects of exercise and time-restricted feeding (TRF) in ameliorating MASLD.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• Our study revealed that the PPARγ-miR-802-Psmd2 axis mediates the therapeutic benefits of exercise and TRF in MASLD, offering innovative mechanistic foundations for targeted interventions.

Introduction

Metabolic dysfunction-associated steatotic liver disease (MASLD) represents a global health challenge, progressing from hepatic steatosis to metabolic dysfunction-associated steatohepatitis marked by inflammatory infiltration and hepatocyte ballooning, ultimately leading to fibrosis.¹ Current epidemiological data reveal a striking 30% global prevalence,² with an annual incidence rate of 4613 cases per 100 000 person-years.³ Projections indicate MASLD will dominate liver transplantation demands by 2030.⁴ Despite this growing burden, clinical management remains constrained by limited therapeutic options, with only two Food and Drug Administration (FDA)-approved therapeutic agents demonstrating unverified efficacy. These clinical realities underscore the urgent need for mechanistic elucidation of MASLD pathogenesis and the development of innovative therapeutic interventions.

MicroRNAs (miRNAs) are key post-transcriptional regulators in metabolic diseases, including MASLD. Among them, miR-802 has been identified as a contributor to systemic insulin resistance, as obesity-induced overexpression impairs both insulin synthesis in pancreatic β-cells and insulin signalling in hepatocytes.^{5 6} Li et al clarified the mechanism by which lignan schisanhenol counteracts MASLD progression by targeting miR-802, which functions as a critical negative regulator of AMP-activated protein kinase (AMPK) signalling through its direct binding to the 3’ untranslated region (3'UTR) of protein kinase AMP-activated catalytic subunit A1/B1 (PRKAA1/PRKAB1).⁷ Exosomes, nanoscale extracellular vesicles, serve as important carriers of miRNAs, facilitating interorgan communication in metabolic homoeostasis and disease.^{8 9}

While non-pharmacological interventions such as exercise and time-restricted feeding (TRF) show efficacy in ameliorating MASLD, the molecular mechanisms underlying their systemic benefits remain poorly understood. In particular, the role of plasma exosomal miRNAs in mediating these effects is largely unexplored. This study aims to address this gap by investigating whether exosome-mediated transfer of miR-802 contributes to the metabolic improvements induced by exercise and TRF in MASLD.

This study aimed to test the hypothesis that the peroxisome proliferator-activated receptor gamma (PPARγ)-miR-802-Psmd2 axis drives the progression of MASLD and mediates the metabolic benefits of exercise and TRF. Plasma exosomal sequencing revealed that miR-802 levels were positively correlated with MASLD severity, a finding validated by gain-of-function and loss-of-function experiments demonstrating its pro-steatotic role. Mechanistically, PPARγ transcriptionally activates hepatic miR-802, which in turn suppresses the proteasomal gene Psmd2, exacerbating lipotoxic stress. Notably, both exercise and TRF ameliorated MASLD phenotypes and coordinately downregulated the PPARγ-miR-802-Psmd2 axis. These results establish this axis not only as a key pathogenic driver in MASLD but also as a targetable pathway underlying the therapeutic effects of lifestyle interventions.

Methods

Animals

Male C57BL/6 mice (8 weeks of age) were obtained from the Beijing Vital River Laboratory Animal Technology Co., Ltd (Beijing, China). All animals were housed in a specific pathogen-free facility under strictly controlled environmental conditions: a constant temperature of 22±1°C, relative humidity of 50±10% and a 12-hour light and dark cycle (lights on at 7:00 hours). Standard laboratory chow and water were provided ad libitum. Mice were allowed to acclimate to these housing conditions for at least 1 week prior to any experimental procedures. Mice were housed in individually ventilated cages with corncob bedding, with a maximum of four to five animals per cage to minimise stress from overcrowding. All surgical procedures and animal care protocols were performed in accordance with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (Publication No. 85–23, revised 1996) with approval from the Fourth Military Medical University Animal Care and Use Committee (approval number 20253406–1).

Exercise and TRF protocol

The swimming exercise protocol was adapted from a previously established method.¹⁰ Eight-week-old male C57BL/6 mice were fed a high-fat diet (HFD, 60% kcal from fat, Research Diets, Inc. (RDI; New Brunswick, NJ, USA), D12492) for 6 weeks to induce initial metabolic disturbances. Subsequently, while maintaining the HFD, the mice underwent a 12-week swimming training regimen. Training consisted of one daily session (60 min per session), 5 days per week, conducted between 9:00 hours and 11:00 hours in water maintained at 34±1°C. Sedentary control mice were briefly placed in water to mimic the stress of handling and drying without performing exercise. All mice were sacrificed at the end of the 18-week total study period (6-week HFD and 12-week intervention) for tissue collection.

For the TRF intervention, 8-week-old male C57BL/6 mice were first fed the HFD (RDI, D12492) ad libitum for 24 weeks to establish MASLD. Subsequently, they were divided into two groups for an 8-week intervention: the TRF group was allowed access to the same HFD only during an 8-hour window (22:00 hours to 6:00 hours), while the control group continued with ad libitum feeding (ALF) of the HFD. All animals were sacrificed at the end of the 32-week total duration (24-week induction and 8-week intervention) for analysis.

Plasma exosome isolation

Blood samples were taken from mice at the endpoint of these models. Human blood samples were obtained from nine patients with MASLD and nine age-matched and sex-matched healthy controls recruited from Xijing Hospital. The diagnosis of MASLD was confirmed in accordance with the latest American Association for the Study of Liver Diseases (AASLD) practice guidance, while controls were excluded if they had any history of metabolic or chronic liver diseases. Peripheral blood was drawn after an overnight fast, allowed to clot at room temperature for 30 min and then centrifuged at 3000 × g for 15 min. The resulting plasma was stored at −80°C until analysis. Exosomes were extracted from the plasma using the Plasma Exosome Extraction Kit according to the manufacturer’s instructions (MELE BIOTECH, Wuhan, China; 0001).

MiRNA library construction and sequencing

MiRNA library preparation and sequencing were conducted by a commercial service (Annoroad Gene Technology Co., Ltd., Beijing, China). Total RNA was extracted from exosomes purified from mouse plasma. Then, reverse transcription and PCR amplification were performed. The PCR products derived from the 18–30 nucleotide RNA molecules were purified by electrophoresis and sequenced using the Illumina HiSeq 2500 platform.

Cell culture, transfection and drug treatment

Alpha mouse liver 12 (AML12) cells (Procell Life Science & Technology Co., Ltd., Wuhan, China, CL-0602) and HepG2 cells (Procell, CL-0103) were seeded in 12-well plates and maintained in a defined medium (Procell, CM-0602) at 37°C in humid air with 5% CO₂ for 48 hours before other treatments. After 12 hours of culture, transfection was performed using Lipofectamine 2000 (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA, 11 668–027) according to the manufacturer’s protocol. Specifically, miR-802 mimics (RiboBio, Guangzhou, China), miR-802 inhibitors (RiboBio), si-Psmd2 (PaiviBio, Wuhan, China) and their respective negative controls were diluted in Gibco Opti-MEM (Thermo Fisher Scientific), mixed with the transfection reagent and added to the cells. The medium was replaced 6 hours post-transfection. The final working concentration was 50 nM for all transfection reagents, with the exception of the miR-802 inhibitor and its corresponding negative control, which were used at 100 nM. For pharmacological treatments, the PPARγ agonist rosiglitazone (10 nM), the PPARγ inhibitor oleuropein (Ole; 50 nM), the steatosis inducer palmitic acid (PA, 300 μM) and the proteasome inhibitor PS-341 (20 nM) were prepared as stock solutions in dimethyl sulfoxide (DMSO). The final concentration of DMSO in the culture medium did not exceed 0.1%. Cells were treated with these compounds for 24 hours prior to subsequent analyses. The sequences of miRNA mimic, inhibitor and small interfering RNA (siRNA) are shown in online supplemental table S1.

Bioinformatic analysis

To identify upstream regulators of miR-802, its murine genomic sequence with 1000 bp flanking regions was obtained from the National Center for Biotechnology Information (NCBI) and analysed using University of California Santa Cruz (UCSC) Genome Browser for transcription factor binding site prediction. High-scoring candidates were further evaluated with Joint Accessible Service for PARser databases (JASPAR), yielding six factors with scores >10. PPARγ was selected based on its elevated expression in the livers of methionine-choline-deficient diet (MCD)-fed and HFD-fed mice and its established role in obesity and MASLD. A potential PPARγ binding site in the miR-802 promoter was supported by chromatin immunoprecipitation (ChIP)-seq data from the Cistrome database.

For downstream target prediction, miR-802 was submitted to TargetScan V.7.1 and miRDB. Genes with scores >90 in both databases were intersected, resulting in six candidates. Their expression was examined in MCD and HFD-fed mouse livers and in AML12 cells with miR-802 overexpression or knockdown. Only Psmd2 showed a consistent inverse expression pattern: downregulation in conditions of high miR-802 expression and upregulation on miR-802 inhibition. Psmd2 was therefore selected for further validation.

Animal dosing protocol

Adeno-associated virus injection

To determine whether miR-802 contributes to MASLD or exercise-induced effects, TRF mice were intravenously injected with serotype 8 adeno-associated virus (AAV8) carrying a specific sequence inhibiting or overexpressing miR-802 and the corresponding control (Hanbio Biotechnology, Shanghai, China) at 1×10¹¹ vector genomes per mouse.

GW9662 injection

To investigate the regulatory role of PPARγ on miR-802, the PPARγ inhibitor GW9662 was administered to mice via intraperitoneal injection. The dosing regimen was as follows: GW9662 was dissolved in DMSO and administered intraperitoneally at a dose of 1 mg/kg (body weight). Prior to injection, the solution was diluted with corn oil to a final volume of 100 μL. Control mice received only 100 μL of corn oil.

Liver TC and TG measurement

Liver total cholesterol (TC) or triglyceride (TG) concentrations were measured using TC Content Assay Kit or TG Content Assay Kit (Solarbio, Beijing, China) according to the manufacturer’s protocol.

Liver histology

Liver tissues fixed with 4% paraformaldehyde were either embedded in paraffin and sectioned for paraffin staining or dehydrated in 30% sucrose solution and embedded in an optimal cutting temperature compound for frozen sectioning. Oil Red O and haematoxylin and eosin (H&E) staining were performed according to standard procedures. For quantification of the positively stained area, images were analysed with Image-Pro Plus (Version 6.0).

Western blotting

The total proteins from the indicated tissues and cells were extracted using radioimmuno precipitation assay (RIPA) lysis buffer (Beyotime Biotechnology, Shanghai, China) containing 1mM phenylmethylsulfonyl fluoride (PMSF) (Sigma-Aldrich, St. Louis, MO, USA). Western blot analyses were conducted according to standard procedures using specific antibodies (listed in online supplemental table S2). Densitometric calculations were expressed as fold change in proteins relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression levels using ImageJ (Version 1.54f) software.

Quantitative real-time PCR

Total RNA was extracted from livers and AML12 cells using TRIzol reagent (Invitrogen, Thermo Fisher Scientific). The cDNA was generated using Thermo S1000 Thermal Cycler for quantitative PCR (qPCR) and real-time PCR assays were conducted with a CFX Opus 96 (Bio-Rad Laboratories, Hercules, CA, USA) using the primer sequences listed in online supplemental table S3. The relative expression of genes was determined using the comparative method (2^-△CT). U6 and ACTIN were used as internal standards for miRNAs and mRNAs, respectively.

Luciferase assays

The promoter of murine miR-802 containing either the wild type (WT) or mutated (MUT) PPARγ binding site was cloned upstream of the firefly luciferase open reading frame using specific primers. 293T cells (10 000 cells/well; Procell, CL-0005) were plated in a 96-well plate and were transfected with 190 ng of DNA (90 ng construct promoter and 90 ng transcription factors; 10 ng constitutive renilla expression plasmid as a control for transfection efficiency) using the Lipofectamine 2000 transfection reagent (Invitrogen, Thermo Fisher Scientific), according to the manufacturer’s instructions. Dual luciferase reporter assays were performed 24 hours after transfection using a Dual-Luciferase Reporter Assay System (Beyotime).

Statistical analysis

All data were analysed by GraphPad Prism (Version 9.0) and presented as mean±SD. A two-tailed unpaired Student’s t-test was used to determine the significant differences between the two groups, and ANOVA was used for multigroup difference analysis. The level of significance was set at p<0.05; *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001; ns, no significance.

Results

Upregulation of plasma exosomal miR-802 in chronic liver injury

Systematic miRNome profiling of plasma exosomes from two complementary models—MCD diet-induced steatohepatitis and carbon tetrachloride (CCl₄)-induced hepatic fibrosis—identified miR-802 as a conserved elevation signature (figure 1A). Vesicle characterisation validated exosomal properties: nanoparticle tracking analysis showed predominant 100–200 nm diameter distribution (figure 1B), while transmission electron microscopy revealed characteristic cup-shaped morphology (figure 1C).

Figure 1. MiRNA expression in plasma exosomes of chronic liver injury mice. (A) Flow chart of plasma exosomal miRNA sequencing (Created with FigDraw.com). (B) Volume distribution as a function of particle diameter (nm), as measured by nanoparticle tracking analysis. (C) Transmission electron microscope image showing isolated exosomes (highlighted in red circles, scale bar, 100 nm). (D) Heatmap of miRNAs upregulated in plasma exosomes from MCD-treated and CCl₄-treated mice compared with their respective controls. (E) Criteria for selecting candidate miRNAs: upregulated in both MCD and CCl₄ models, with log₂(fold change) >1 and p<0.05. (F) Schematic of the experimental timeline for MCD-fed mice (n=3) and HFD-fed mice (n=4) (Created with FigDraw.com). (G) MiR-802 expression levels in different organs of MCD-fed mice and chow-fed controls (heart, liver, lung, spleen, kidney, n=3). (H) MiR-802 expression levels in different organs of HFD-fed mice and chow-fed controls (heart, liver, lung, spleen, kidney, n=4). (I) MiR-802 expression in plasma exosomes from MCD-fed mice (n=3) versus chow-fed controls (CT, n=5), and from HFD-fed mice versus chow-fed controls (CT, n=6). (J) MiR-802 expression in plasma exosomes (n=9) and liver tissues (healthy, n=4; MASLD, n=5) from patients with MASLD compared with healthy controls. *p<0.05; **p<0.01. CCl₄, carbon tetrachloride; CT, control; HFD, high-fat diet; MASLD, metabolic dysfunction-associated steatotic liver disease; MCD, methionine-choline-deficient diet; miRNA, micro RNA.

Comparative analysis of plasma exosomal miRNAs revealed 15 (MCD model) and 16 (CCl₄ model) significantly upregulated species (fold change >2, p<0.05) relative to controls (figure 1D). Venn analysis of these dysregulated miRNAs identified miR-802 and miR-26a-2–3p as the conserved elevation signatures across both injury models (figure 1E). The evolutionary conservation of miR-802 (miR-802–5p), coupled with its previously reported liver-enriched expression pattern, prompted focused investigation into its functional relevance.⁵

Systematic integration of public MASLD hepatic miRNomes identified miR-802 as a top-ranked conserved signature across murine metabolic disease models (db/db and high-fat, high-cholesterol diet (HFHC) mice) (online supplemental figure S1). Subsequent validation in two diet-induced MASLD models (MCD and HFD) revealed a concordant upregulation of miR-802 in the hepatic parenchyma (figure 1F–H). In plasma exosomes, while an upward trend was observed, the increases did not reach statistical significance in MCD (p=0.061) (figure 1I). This suggestive pattern was corroborated by a significant upregulation in the liver of human MASLD cohorts versus healthy controls, but not in plasma exosomes (p=0.050) (figure 1J). The observed tissue-exosome expression divergence of miR-26a-2–3p across models (online supplemental figure S2) underscored miR-802’s superior candidacy as a core MASLD regulator, guiding subsequent mechanistic investigations.

Functional validation of miR-802 in hepatic lipotoxicity

Gain-of-function and loss-of-function approaches in PA-treated AML12 hepatocytes (0.3 mM, 24 hours) established miR-802’s pathogenic role in lipid dysregulation (figure 2A). MiR-802 mimic transfection induced 5×10⁴-fold overexpression (figure 2B), exacerbating PA-induced lipid accumulation through increased TC (fourfold), TGs (threefold) and Oil Red O-positive areas (figure 2C,D). Conversely, miR-802 suppression (>90% knockdown efficiency, figure 2B) significantly attenuated steatotic parameters including TC and TG levels and Oil Red O-positive areas (figure 2E,F). In PA-treated HepG2 cells, overexpression of miR-802 aggravated lipid accumulation (online supplemental figure S3A,B), as evidenced by a significant increase in Oil Red O staining area and an increasing trend in intracellular TG content (online supplemental figure S3C,D). Conversely, inhibition of miR-802 significantly alleviated lipid deposition (online supplemental figure S3F) and tended to reduce TG levels (online supplemental figure S3E), confirming the lipogenic role of miR-802 in vitro. These dose-responsive effects confirm miR-802 as a critical regulator of hepatocyte lipotoxicity.

Figure 2. Hepatocyte lipotoxicity was regulated by miR-802. (A) Schematic of AML12 cell transfection with miR-802 mimics or inhibitors followed by PA-induced steatosis modelling (Created with FigDraw.com). (B) MiR-802 expression in AML12 cells transfected with mimics or inhibitors (n=4) compared with negative control (NC: n=5). (C) TC and TG levels in NC-transfected and miR-802 mimic-transfected AML12 cells after PA induction (n=5). (D) Representative Oil Red O staining of NC and miR-802 mimic-transfected cells after PA treatment (n=6, scale bar, 100 µm). (E) TC (n=4) and TG (n=5) levels in NC-transfected and miR-802 inhibitor-transfected AML12 cells after PA induction. (F) Representative Oil Red O staining of NC and miR-802 inhibitor-transfected cells after PA treatment (n=5, scale bar, 100 µm). (G) Experimental timeline of MCD-fed mice treated with miR-802 sponge or vehicle control, and hepatic miR-802 expression by qRT-PCR (vehicle: n=7, sponge: n=8) (Created with FigDraw.com). (H) Hepatic TC and TG levels in MCD-fed mice treated with vehicle or miR-802 sponge (vehicle: n=7, sponge: n=8). (I) Representative Oil Red O and H&E staining of liver sections from vehicle-treated or miR-802 sponge-treated MCD-fed mice (vehicle: n=7, sponge: n=8, scale bar, 200 µm). (J) Experimental timeline of HFD-fed mice treated with vehicle or miR-802 sponge, and hepatic miR-802 expression (vehicle: n=5, sponge: n=6) (Created with FigDraw.com). (K) Hepatic TC and TG levels in HFD-fed mice treated with vehicle or miR-802 sponge (vehicle: n=5, sponge: n=6). (L) Representative Oil Red O and H&E staining of liver sections from vehicle-treated or miR-802 sponge-treated HFD-fed mice (vehicle: n=5, sponge: n=6, scale bar, 200 µm). *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. AAV, adeno-associated virus; AML12, alpha mouse liver 12; H&E, haematoxylin and eosin; HFD, high-fat diet; MCD, methionine-choline-deficient diet; NC, negative control; PA, palmitic acid; TC, Total cholesterol; TG, triglyceride.

To establish causal involvement, we engineered a liver-directed AAV8-thyroxine-binding globulin promoter-sponge vector for miR-802 knockdown (50% suppression efficiency, figure 2G) through systemic administration. In MCD diet-induced MASLD mice (C57BL/6, 8-week), this targeted inhibition attenuated hepatic steatosis, evidenced by a trend toward TG reduction (25% decrease, p=0.054), no significant change in TC, a significant decrease in lipid droplet coverage (Oil Red O quantification) and marked histological resolution of macrovesicular fat (figure 2H,I). The therapeutic concordance between in vitro and in vivo models further strengthens miR-802’s pathological role in steatosis progression.

To validate the therapeutic universality of miR-802 inhibition, HFD-fed C57BL/6 mice (12-week pre-established steatosis) received hepatic miR-802 sponge delivery followed by an 8-week continued HFD challenge. This intervention achieved robust hepatic miR-802 suppression (60%, p<0.01 vs controls, figure 2J), concomitant with significant steatosis attenuation: 30% TG reduction (figure 2K), 37% lipid droplet area decrease (Oil Red O quantification) and histological resolution of macrovesicular fat (figure 2L). The therapeutic congruence across MCD and HFD models establishes miR-802 suppression as a cross-model disease-modifying strategy.

Exercise and TRF attenuate MASLD via miR-802 suppression

To interrogate miR-802’s involvement in exercise benefits, an HFD-fed C57BL/6 murine model underwent a 12-week swimming training (figure 3A). Exercise significantly reduced hepatic and circulating exosomal miR-802 levels (p<0.05), effects that were reversed by AAV8-mediated hepatic miR-802 overexpression (figure 3B). The antisteatotic effects of exercise were evidenced by 52% lower hepatic TG (p<0.001) and more than 70% reduced lipid droplet area (Oil Red O and H&E positive quantification) versus sedentary counterparts. Crucially, miR-802 overexpression abolished exercise benefits, increasing TG level and Oil Red O positive area, partially reversing the reduction in TG and restoring the Oil Red O positive area to non-exercised levels, respectively (figure 3C,D). A concurrent increase in H&E vacuolation area was observed, though not statistically significant (p=0.196). Moreover, exercise intervention significantly reduced body weight and aspartate aminotransferase (AST) levels compared with the sedentary group; however, these beneficial effects were reversed by AAV8-mediated overexpression of miR-802 (online supplemental figure S4D). This establishes miR-802 as a mandatory mediator of exercise-induced MASLD attenuation.

Figure 3. Exercise and TRF suppress miR-802 to ameliorate MASLD in mice. (A) Schematic of the exercise intervention protocol in HFD-fed mice (Created with FigDraw.com). (B) Hepatic and plasma exosomal miR-802 expression in HFD-fed mice under sedentary (Sed, n=5), exercise (Exe, n=7) or exercise combined with AAV8-miR-802 overexpression (Exe+AAV8-miR-802, n=8) conditions. Plasma exosome sample size: n=4 per group. (C) Hepatic TC and TG levels in Sed (n=5), Exe (n=7), and Exe+AAV8-miR-802 groups (n=8). (D) Representative Oil Red O and H&E staining of liver sections from Sed (n=5), Exe (n=7) and Exe+AAV8-miR-802 mice (n=8, scale bar, 200 µm). (E) Schematic of the TRF intervention protocol in HFD-fed mice (Created with FigDraw.com). (F) Hepatic and plasma exosomal miR-802 expression in HFD-fed mice under ALF (n=5), TRF (n=5) or TRF combined with AAV8-miR-802 overexpression (TRF+AAV8-miR-802, n=4) conditions. Plasma exosome sample size: ALF n=3, TRF and TRF+AAV8-miR-802 n=4 each. (G) Hepatic TC and TG levels in ALF (n=5), TRF (n=5) and TRF+AAV8-miR-802 groups (n=4). (H) Representative Oil Red O and H&E staining of liver sections from ALF (n=5), TRF (n=5) and TRF+AAV8-miR-802 mice (n=4, scale bar, 200 µm). *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. AAV8, serotype 8 adeno-associated virus; ALF, ad libitum feeding; Exe, exercise; H&E, haematoxylin and eosin; HFD, high-fat diet; MASLD, metabolic dysfunction-associated steatotic liver disease; miRNA, microRNA; ns, no significance; Sed, sedentary; TC, total cholesterol; TG, triglyceride; TRF, time-restricted feeding.

TRF is also a beneficial intervention for MASLD, yet whether it acts through miR-802 regulation, similar to exercise, remains unknown. We therefore subjected HFD-fed mice to TRF or ALF for 8 weeks after 24 weeks of HFD feeding. Prior to TRF, AAV8-miR-802 was administered via the tail vein to achieve hepatocyte-specific miR-802 overexpression (figure 3E). Plasma and liver samples were collected from the ALF, TRF and TRF and miR-802 groups for analysis. TRF tended to reduce miR-802 in plasma exosomes (p=0.063) and, to a lesser extent, in the liver (p=0.078), while effectively alleviating hepatic steatosis, as evidenced by decreased Oil Red O staining, lipid vacuolation and hepatic TG. Conversely, miR-802 overexpression blunted TRF’s benefits and exacerbated steatosis, with increased Oil Red O-positive area and H&E vacuolation area, the hepatic TG also showed a numerical increase but did not reach statistical significance (p=0.165) (figure 3F–H). Similarly, while TRF treatment markedly decreased body weight and ALT levels, miR-802 overexpression significantly attenuated the reduction in ALT, although it did not further alter body weight or AST levels in the TRF cohort (online supplemental figure S4E).

PPARγ transcriptionally activates miR-802 to drive MASLD pathogenesis

To elucidate upstream regulators of miR-802 in MASLD, bioinformatic screening (JASPAR and UCSC Genome Browser) identified PPARγ as a candidate transcriptional regulator. Consistently, the hepatic mRNA and protein levels were significantly upregulated in HFD mice and showed a concordant, although non-significant, increasing trend in the MCD model (p=0.088) (figure 4A–D, online supplemental figure S4A). Pharmacological modulation in PA-treated AML12 cells showed rosiglitazone (agonist) increased miR-802 expression, whereas Ole (antagonist) suppressed it, correlating with lipid accumulation changes (TG and Oil Red O; figure 4E–G). We also treated PA-induced HepG2 cells with Ole or rosiglitazone (Rosi) (online supplemental figure S5A). Ole treatment significantly suppressed PA-induced miR-802 expression (online supplemental figure S5B) and concurrently alleviated lipid accumulation, as shown by reduced Oil Red O staining and tended to reduce TG content (online supplemental figure S5C,D). In stark contrast, Rosi treatment further tended to elevate miR-802 levels and exacerbated hepatic steatosis compared with the PA group. Luciferase reporter assays confirmed PPARγ binding to the miR-802 promoter’s conserved motif, with binding-site mutation abolishing transcriptional activation (figure 4H). These mechanistic insights establish PPARγ-miR-802 signalling as a key axis in steatotic progression.

Figure 4. MiR-802 is positively regulated by PPARγ. (A) Prediction of transcription factors binding to the miR-802 promoter using JASPAR database and UCSC Genome Browser. (B) Publicly available ChIP-seq data from the Cistrome database showing predicted PPARγ binding sites on the miR-802 promoter region. (C–D) Hepatic PPARγ mRNA (C) and protein (D) expression in HFD-fed or MCD-fed mice compared with their respective chow-fed CT. Sample sizes: HFD and CT, n=5; MCD, n=5; MCD, n=5, CT, n=4 for RNA; n=3 per group for protein. (E) MiR-802 expression in AML12 cells treated with PA alone or in combination with the PPARγ agonist Rosi or inhibitor Ole. (CT, n=6; PA, n=6; PA+Rosi, n=7; PA+Ole, n=8). (F–G) Representative Oil Red O staining (CT, PA, PA+Rosi: n=4, PA+Ole: n=3,F) and quantification of TG levels (G) in AML12 cells under the indicated treatments. Sample sizes for TG: Ole experiment (n=4 per group); Rosi experiment (n=7). (H) Luciferase reporter assay in 293T cells co-transfected with an empty vector or miR-802 promoter construct, together with a PPARγ overexpression plasmid or its control (CT: n=4, 802: n=6, 802 and Mut+PPARγ:n=3). (I–J) Hepatic PPARγ mRNA (I) and protein (J) levels in HFD-fed mice subjected to Sed, Exe, ALF or TRF regimens. Sample sizes for mRNA: Sed, n=6; Exe, n=5; ALF, n=8; TRF, n=7. Protein analysis: n=3 per group. (K) Gene Ontology enrichment analysis of differentially expressed genes between livers of exercised and sedentary HFD-fed mice. (L) Schematic of the experimental design: HFD-fed mice were treated with the PPARγ antagonist GW9662 and/or AAV8-miR-802 overexpression virus (Created with FigDraw.com). (M–N) Hepatic TG levels (M) and representative Oil Red O and H&E staining (N) of liver sections from HFD-fed mice treated as indicated. Sample sizes: CT, n=5; GW9662, n=7; GW9662+AAV8-miR-802, n=8. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. AAV8, serotype 8 adeno-associated virus; ALF, ad libitum feeding; AML12, alpha mouse liver 12; CT, control; Exe, exercise; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; H&E, haematoxylin and eosin; HFD, high-fat diet; ip qod, intraperitoneal every other day; JASPAR, Joint Accessible Service for PARser databases; MCD, methionine-choline-deficient diet; Mut, mutated; ns, no significance; Ole, oleuropein; PA, palmitic acid; PPARγ, peroxisome proliferator-activated receptor gamma; Rosi, rosiglitazone; Sed, sedentary; TG, triglyceride; TRF, time-restricted feeding; UCAC, University of California Santa Cruz.

Exercise and TRF interventions attenuated hepatic PPARγ expression in both models, as quantified by quantitative reverse transcription PCR and immunoblotting, with statistically significant reductions observed in most comparisons (figure 4I,J, online supplemental figure S4B). Transcriptomic profiling revealed concomitant downregulation of PPARγ signalling in exercised liver tissue, with Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis showing pathway suppression (figure 4K).

To mechanistically interrogate PPARγ-miR-802 cross-regulation, GW9662-mediated PPARγ inhibition was implemented in HFD-fed mice with hepatic miR-802 overexpression (AAV8 delivery post 24-week HFD, figure 4L). Pharmacological PPARγ blockade (weeks 30–32) ameliorated steatosis biomarkers (lipid droplets, vacuolar area) but not in TG levels (p=0.206), whereas miR-802 overexpression abrogated these protective effects (figure 4M,N).

MiR-802 directly suppresses Psmd2 to attenuate lipotoxicity via proteasome-dependent pathways

Multiplatform bioinformatic screening (TargetScan and miRDB) prioritised Psmd2 as a conserved miR-802 target (figure 5A). Hepatic miR-802 overexpression significantly downregulated Psmd2 expression in vitro, and Psmd2 expression is also decreased in HFD livers, with quantitative assessment showing significant mRNA and protein reduction compared with controls (p<0.05, figure 5B,C, online supplemental figure S4C).

Figure 5. Validation of target genes of miR-802. (A) Schematic workflow for identifying miR-802 target genes using miRDB and TargetScan databases, followed by intersection analysis. (B) Psmd2 mRNA levels in AML12 cells transfected with miR-802 mimics or negative control (NC) (mimics: n=4, NC: n=6), and in livers of HFD-fed versus chow-fed mice (n=5 per group). (C) Psmd2 protein levels in AML12 cells transfected with miR-802 mimics or NC (n=5), and in livers of HFD-fed versus chow-fed mice (n=5 per group). (D) Experimental scheme: AML12 cells were treated with si-Psmd2 or the proteasome inhibitor PS-341, followed by PA-induced steatosis (Created with FigDraw.com). (E) Psmd2 mRNA levels in AML12 cells transfected with si-Psmd2 or NC (n=5). (F–G) Representative Oil Red O staining (F) and quantification of cellular TG levels (G) in AML12 cells treated without or with PA alone or in combination with si-Psmd2 or PS-341 (n=4 per group). Scale bar, 100 µm. (H–I) Rescue experiments: Representative Oil Red O staining (H) and cellular TG levels (I) in PA-treated AML12 cells co-transfected with miR-802 inhibitors and either si-Psmd2 or its NC (n=4 per group). Scale bar, 100 µm. (J) Luciferase reporter assay in 293T cells co-transfected with miR-802 mimics or NC, together with a reporter plasmid containing either the wild type (WT, n=6) or mutated (Mut, n=6) 3’UTR of Psmd2. Control (CT, n=3) group received an empty vector. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. AML12, alpha mouse liver 12; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HFD, high-fat diet; NC, negative control; PA, palmitic acid; TG, triglyceride.

In PA-treated AML12 hepatocytes, siRNA-mediated Psmd2 knockdown (40% efficiency, p<0.001; figure 5D,E) and proteasome inhibition (PS-341) synergistically exacerbated lipid accumulation (1.2-fold and 2-fold Oil Red O postive area; 1.3-fold and 3-fold TG elevation vs controls; figure 5F,G), paralleling miR-802 overexpression effects. We then knocked down PSMD2 in HepG2 cells prior to PA treatment (online supplemental figure S6A), with knockdown efficiency confirmed by qPCR (online supplemental figure S6B). Silencing PSMD2 significantly aggravated PA-induced intracellular TG accumulation (online supplemental figure S6C). Consistent with this, Oil Red O staining revealed more extensive lipid droplets in the si-PSMD2 group, showing an increasing trend in lipid content compared with the PA group (p=0.088, online supplemental figure S6D). These findings position Psmd2 as a functional miR-802 target modulating steatosis through proteasome-dependent degradation.

Dual modulation in PA-challenged AML12 hepatocytes (anti-miR-802 and Psmd2 siRNA) dissected this regulatory axis. Building on established antisteatotic effects of miR-802 blockade, Psmd2 silencing paradoxically tended to elevate lipid accumulation (20% increase Oil Red O positive area (p=0.067); 1.5-fold TG increase vs miR-802 inhibition alone, figure 5H–I), establishing a causal miR-802-Psmd2 interaction in steatotic regulation. Furthermore, Ole treatment significantly upregulated PSMD2 mRNA expression compared with the control group. We transfected Ole-treated HepG2 cells with miR-802 mimics and found that this induction was effectively abrogated by transfection with miR-802 mimics (online supplemental figure S7A,B). Consistent with the transcript levels, Western blot analysis confirmed that miR-802 overexpression significantly suppressed PSMD2 protein levels in Ole-treated HepG2 cells (online supplemental figure S7C).

To validate binding specificity, we engineered luciferase reporters containing WT or binding-site MUT Psmd2 3′UTRs. Co-transfection with miR-802 mimics suppressed WT reporter activity by 50% (p<0.05) while MUT constructs remained unaffected (figure 5J), conclusively demonstrating miRNA-802’s direct interaction with Psmd2 3′UTR to repress gene expression.

Discussion

Hepatic and circulating exosomal miR-802 were significantly elevated in MASLD murine models, with gain-of-function and loss-of-function studies confirming its pathogenic role in steatosis progression. Extending previous findings by Kornfeld et al⁵ on miR-802-mediated insulin signalling disruption, we demonstrated its direct contribution to lipotoxicity through proteasomal dysregulation. Multitissue qPCR analysis localised excess exosomal miR-802 primarily to hepatocyte overproduction rather than extrahepatic sources, establishing hepatic-paracrine signalling in MASLD pathophysiology.

Therapeutic lifestyle interventions, particularly structured exercise and TRF, significantly ameliorated MASLD pathology in diet-induced obese mice, concomitant with hepatic and circulating exosomal miR-802 downregulation. Emerging evidence positions weight reduction as a primary therapeutic target, with ≥10% loss demonstrating steatohepatitis resolution and >5% yielding histological improvements in the non-alcoholic fatty liver disease activity score (NAS scores).¹¹ Structured exercise modalities (aerobic, resistance and high-intensity interval training) mitigate intrahepatic lipid accumulation through multimodal mechanisms: (1) enhanced fatty acid β-oxidation, (2) hepatocyte autophagy activation and (3) restoration of insulin sensitivity.12,16 Concurrently, exercise suppresses oxidative stress via NRF2-mediated antioxidant upregulation and dampens proinflammatory cytokine production.¹⁷ Besides exercise, different dietary strategies (ketogenic diet, high-protein diet, low-carbohydrate diet and TRF, etc) also have therapeutic effects on fatty liver through calorie restriction.¹⁸ Dietary treatments include two major intervention strategies: intermittent fasting (IF), where caloric intake is severely constrained for short periods and TRF, which limits the daily eating window with fasting for the remainder of the day.¹⁹ It has been reported that both IF and TRF have positive effects on total body weight, fat mass reduction and glucose homoeostasis in HFD mice along with reprogramming of numerous metabolic and stress resistance pathways including increased DNA repair, mitochondrial biogenesis, autophagy, expression of antioxidants and downregulation of inflammatory pathways.20,22 In addition, IF and TRF lead to a metabolic switch from glucose to fatty acids and increased production of ketone bodies, which act as potent signalling molecules triggering downstream protective pathways via PPARγ coactivator 1α and fibroblast growth factor 21.^{23 24}

Additionally, exercise and TRF can benefit multiple organs by affecting circulating miRNAs. Heianza et al demonstrated that lifestyle interventions lead to increased circulating miR-375–3 p, which was associated with visceral and intrahepatic fat reduction.²⁵ Hou et al found that long-term exercise-derived circulating exosomes protect the heart against myocardial ischaemia and reperfusion injury via exosomal miR-342-5p.²⁶ Our data suggest that the beneficial effects of exercise and TRF on steatosis hepatocytes may be mediated, at least partially, by alterations in the circulating exosome miR-802 expression pattern.

As a transcription factor, PPARγ was observed to be responsible for the induction and activation of miR-802. Studies have shown that PPARγ is upregulated in the liver and adipose tissue in obese mice. It appeared to be a core obesity gene, and obesity candidate genes were highly interconnected, suggesting a coordinately regulated molecular network in obese mice.²⁷ We hypothesised that exercise and TRF influenced miR-802 levels by downregulating PPARγ, thereby alleviating MASLD. Psmd2 encodes the 26S proteasome non-ATPase regulatory subunit 2 within the ubiquitin-proteasome system and is recognised as an oncogene in various cancers. Current research on Psmd2 primarily focuses on its role in tumours, where it promotes cancer progression and affects prognosis in oesophageal carcinoma, renal cell carcinoma and breast cancer by inhibiting autophagy, enhancing T cell exhaustion and regulating proteasomal degradation of p21 and p27.²⁸ However, the function of Psmd2 in MASLD remains unexplored. Our experiments demonstrated that under sodium palmitate-induced high-lipid conditions, knockdown of Psmd2 via small interfering RNA—mimicking the suppressive effect of miR-802—resulted in increased lipid deposition, consistent with miR-802 overexpression. Furthermore, dual-luciferase reporter assays confirmed that miR-802 negatively regulates Psmd2 by binding to a specific site within its 3'UTR. These findings identify Psmd2 as a novel target of miR-802 and elucidate its role in cellular lipid deposition.

Our study has several limitations. First, the origins of miR-802 need to be further explored, although we found that liver miR-802 expression was higher than that in the heart, lung, spleen and kidney. In addition, we discovered that miR-802 is predominantly expressed in hepatocytes rather than non-parenchymal cells in the liver (online supplemental figure S2C). Except for the liver, other organs also secrete miR-802-enriched exosomes under the conditions of obesity, and exosomes are released into the circulation. For instance, adipocyte-derived exosomal miR-802 induces cardiac insulin resistance through downregulation of HSP60, and high expression of miR-802 in the kidney of obese mice leads to kidney damage associated with diabetes through inflammatory and immune pathways.^{29 30} Second, as target organs of MASLD, we only examined the role of miR-802 in the liver affected by MASLD. Zhang et al have found that obesity-induced miR-802 overexpression in pancreatic islets impairs insulin transcription and secretion, leading to insulin resistance in target organs.⁶ The role of miR-802 secreted in other tissues in MASLD requires further investigation. Third, while our research used established models of MASLD, we did not include two widely recognised and representative dietary models—the Western diet and the choline-deficient, L-amino acid-defined HFD. Their inclusion in future work would help to further validate and generalise our findings across different metabolic contexts. Furthermore, our investigation focused on the effects of PPARγ activation; the context-dependent consequences of PPARγ blockade were not examined. This represents a significant gap, as the net outcome of PPARγ modulation may vary under different physiological and pathological conditions. Future studies specifically designed to dissect the effects of PPARγ inhibition in various stages of MASLD progression are warranted. Fourth, to evaluate the diagnostic potential of plasma miR-802, we analysed its correlation with hepatic steatosis in a cohort of nine patients with MASLD and nine healthy controls. The groups were matched for age (p=0.653) and sex (p=0.807), though body mass index was significantly higher in the MASLD group (mean NAS score: 1.78) (online supplemental table S4). We found that plasma exosomal miR-802 levels were elevated in patients with MASLD and positively correlated with steatosis severity (online supplemental figure S8). Future studies with larger sample sizes are needed to examine correlations with liver inflammation and fibrosis.

In conclusion, this study identifies the PPARγ–miR-802–Psmd2 axis that promotes hepatic lipid accumulation in MASLD. Exercise and TRF ameliorate steatosis at least partly by suppressing this pathway, highlighting miR-802 as a potential biomarker and therapeutic target linking lifestyle intervention, including exercise and TRF, to metabolic improvement.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.