PPARγ suppresses the tumorigenesis of retroperitoneal liposarcoma via inactivation of the PI3K/AKT signaling pathway

Abstract

Background

Retroperitoneal liposarcoma (RLPS) is characterized by high heterogeneity, frequent recurrence and dismal prognosis. Surgery remains the main therapeutic strategy, while no effective targeted therapy is available for advanced RLPS. Peroxisome proliferator-activated receptor γ (PPARγ) is a core regulator of adipocyte differentiation, yet its function and regulatory mechanisms in RLPS remain unclear.

Methods

Eighty clinical specimens and information from RLPS patients were collected and analyzed to examine PPARγ expression and assess its prognostic value, integrated with results from multi-database. The biological functions of PPARγ were investigated by constructing viral vectors for its overexpression and knockdown in both in vivo and in vitro. Bioinformatic analysis, multiplex immunofluorescence, co-immunoprecipitation and multiple approaches were applied to elucidate PPARγ-mediated mechanisms and its crosstalk with the downstream PI3K/AKT cascade. The anti‑tumor activity and biosafety of the PPARγ agonist troglitazone (TGZ) were assessed in vitro and in RLPS Cell Line-Derived/Patient-Derived Xenograft models.

Results

PPARγ was significantly downregulated in RLPS tissues and acted as an independent prognostic predictor. Overexpression of PPARγ inhibited malignant phenotypes and promoted apoptosis of RLPS cells, whereas PPARγ knockdown exerted the opposite effects. Mechanistically, PPARγ directly bound to PI3K and exerted tumor‑suppressive functions by suppressing the phosphorylation and activation of the PI3K/AKT pathway. TGZ upregulated PPARγ expression, regulated the PI3K/AKT pathway in vitro, and exhibited robust antitumor activity alongside favorable biosafety in vivo.

Conclusions

PPARγ exerts antitumor effects via inactivation of the PI3K/AKT signaling pathway and serves as an independent prognostic biomarker in RLPS. The PPARγ agonist TGZ holds great potential as an antitumor drug candidate for RLPS.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12967-026-08736-4.

Introduction

Retroperitoneal liposarcoma (RLPS) is a rare malignant tumor originating from retroperitoneal mesenchymal tissues [1], with well-differentiated liposarcoma (WDLPS) and dedifferentiated liposarcoma (DDLPS) as the predominant subtypes [2, 3]. Owing to the complex retroperitoneal anatomy and insidious early clinical manifestations, most patients present with large tumor masses at diagnosis, which markedly increases surgical difficulty [4, 5]. Radical resection remains the main therapeutic strategy [6, 7], yet high rates of local recurrence constitute the primary barrier to long-term survival [8, 9]. For patients with advanced, recurrent or unresectable disease, radiotherapy and chemotherapy yield limited efficacy and substantial toxicities [10, 11], resulting in a dismal overall prognosis [12, 13]. Therefore, elucidating the molecular mechanisms underlying RLPS malignant progression and identifying novel therapeutic targets represent urgent priorities in current research.

Peroxisome proliferator-activated receptor γ (PPARγ) is a pivotal transcription factor governing adipocyte differentiation and lipid metabolism [14–16], and exerts tumor-suppressive functions in multiple solid tumors by suppressing NF-κB-mediated inflammation and PI3K/AKT activation [17–19]. Although accumulating evidence suggests that PPARγ may act as a tumor suppressor in RLPS [3, 20], its precise biological functions, downstream regulatory networks, and mechanisms by which dysregulated expression promotes tumor progression remain poorly defined, severely hindering the development and clinical translation of PPARγ-targeted therapies [21].

The PI3K/AKT signaling axis constitutes a core pathway governing cell survival, proliferation, metabolism and growth [22–24], and is aberrantly activated in numerous human malignancies [25, 26]. In liposarcoma, hyperactivation of this pathway stimulates PI3K-mediated production of the second messenger PIP3, thereby promoting AKT phosphorylation and activation [27, 28]; activated AKT further modulates a panel of downstream effectors, including mTOR, GSK-3β and FoxO, markedly enhancing tumor cell survival, proliferation and invasion [29, 30]. Previous studies have established that PPARγ can modulate the activation status of the PI3K/AKT pathway by regulating the transcription of its downstream target genes [31].

Against this research backdrop, we focus on the PPARγ-PI3K/AKT signaling axis to systematically investigate the expression profiles, clinical prognostic value and biological functions of PPARγ in RLPS, elucidate the molecular mechanisms by which it modulates the malignant progression of RLPS, and comprehensively validate the anti-tumor activity of PPARγ agonists in both in vitro and in vivo models.

Materials and methods

Database analysis

Transcriptome data and prognostic information of liposarcoma and normal adipose tissue were retrieved from public databases, including GEO (GSE30929, GSE159659, https://www.ncbi.nlm.nih.gov/geo/), TCGA (https://portal.gdc.cancer.gov/), and GTEx (https://gtexportal.org/) Differential gene expression analysis and pathway enrichment analysis were performed to explore the expression pattern of PPARγ in RLPS.

Clinical specimens and patient data

This was a single-center retrospective clinical study that consecutively enrolled patients diagnosed with RLPS and surgically treated at the First Affiliated Hospital of Air Force Medical University between July 2014 and June 2024. Inclusion criteria: Age ≥ 18 years; Histopathologically confirmed primary retroperitoneal liposarcoma, with complete preoperative imaging data (CT or MRI), operative records, pathological diagnosis, and postoperative follow-up information available, and a follow-up duration of ≥ 1 year after surgery; No prior neoadjuvant radiotherapy, chemotherapy, targeted therapy, or any other form of anti-neoplastic intervention before surgery; Complete and traceable clinical medical records and follow-up data without loss of critical clinical information. Exclusion criteria: Patients with synchronous primary malignant tumors at other anatomical sites (e.g., gastric cancer, colorectal cancer); Patients lost to follow-up or with an actual postoperative follow-up duration of less than 1 year.

Cell culture

Human RLPS cell line SW872 (Procell system Biotechnology, Wuhan, China) and well-differentiated liposarcoma cell line 93T449 (Immocell Biotechnology, Xiamen, China) were used in this study. Primary adipocytes (PAC) were isolated from fresh normal retroperitoneal adipose tissue. After isolation, primary adipocytes were incubated for 2 h before immediate experiments, with no long-term culture, expansion or supplementation with differentiation/growth factors. All cells were cultured in high-glucose DMEM (Mishu Biotechnology, Xi’an, China) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (Thermo Fisher, Hangzhou, China) at 37 °C in a humidified atmosphere containing 5% CO₂. Medium was refreshed every 48–72 h, and cells were passaged when reaching 70–80% confluence. For cryopreservation, cells were resuspended in non-programmed cryopreservation solution (Servicebio, Wuhan, China) and stored at -80 °C for short-term or liquid nitrogen for long-term preservation.

Isolation of primary adipocytes

Fresh sterile retroperitoneal adipose tissues were briefly preserved in 4 °C DMEM, rinsed with PBS containing 10% antibiotics in a biosafety cabinet, and minced into ~ 1 mm³ fragments. Tissues were digested with 3% collagenase I (400 U/mL, 50× volume of tissues) in serum-free DMEM at 37 °C for 2–4 h. The single-cell suspension was obtained via a 100-mesh cell strainer, centrifuged at 1000 rpm for 5 min, rinsed with PBS, resuspended in complete medium, and seeded into T25 flasks at 1–2 × 10⁶ cells/flask (6 mL total volume) for culture at 37 °C with 5% CO₂.

Cell climbing slice preparation

Sterilized 24 mm×24 mm coverslips were placed in 6-well plates. ~500 µL cell suspension (1 × 10⁵ cells) was added to each coverslip, incubated at 37 °C with 5% CO₂ for 5 h, then supplemented with 1 mL complete medium and cultured overnight. After cell adherence, the medium was removed, cells were gently rinsed twice with PBS, and fixed with 1 mL paraformaldehyde at room temperature for 10–15 min.

RNA extraction and quantitative real-time PCR (qPCR)

Total RNA was extracted from tissues and cells using Trizol reagent (Jingzhun Biotechnology, Hangzhou, China) according to the manufacturer’s instructions. RNA concentration and purity were determined by a UV spectrophotometer (Bio-Rad, Hercules, USA) by measuring the absorbance ratio at 260/280 nm. Complementary DNA (cDNA) was synthesized from 500 ng of total RNA using a reverse transcription kit (Accurate Biology, Changsha, China). qPCR was performed on a real-time PCR system (Bio-Rad, Hercules, USA) with SYBR Green PCR Master Mix (Accurate Biology, Changsha, China). The reaction conditions were: 95 °C for 3 min, followed by 40 cycles of 95 °C for 10 s and 58 °C for 30 s. GAPDH was used as an internal reference gene, and the relative expression levels of target genes were calculated using the 2⁻ΔΔCt method. The primer sequences used are as follows:

PPARγ-Forward: 5’-GGTGACCAGAAGCCTGCATT-3’,

PPARγ-Reverse: 5’-TGTCAACCATGGTCATTTCGTT-3’;

GAPDH-Forward: 5’-AGAAGGCTGGGGCTCATTTG-3’,

GAPDH-Reverse: 5’-AGGGGCCATCCACAGTCTTC-3’.

Western blot

Total proteins were extracted from tissues and cells using RIPA lysis buffer (Yeasen Biotechnology, Shanghai, China) supplemented with protease and phosphatase inhibitors (Biocytocare Biotechnology, Guangzhou, China; Servicebio, Wuhan, China). Protein concentration was determined by the BCA protein assay kit (Boxbio Biotechnology, Beijing, China). Equal amounts of protein (20–40 µg) were separated by 8–12% SDS-PAGE and transferred onto PVDF membranes (Thermo Fisher, Hangzhou, China). Membranes were blocked with 5% non-fat milk in TBST for 1–2 h at room temperature, then incubated with primary antibodies overnight at 4 °C. Primary antibodies included PPARγ, GAPDH, PTEN, E-cadherin, Cyclin B1, Cyclin D1, Bax, Bcl-2, PI3K, AKT, p-PI3K, and p-AKT (CST, Boston, USA; Proteintech, Wuhan, China; Thermo Fisher, Hangzhou, China). After washing three times with TBST, membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies (Thermo Fisher, Hangzhou, China) for 1–2 h at room temperature. Protein bands were visualized using an enhanced chemiluminescence (ECL) kit (Beyotime, Shanghai, China) and quantified with Image Lab software (Bio-Rad, Hercules, USA).

Oil Red O staining

Frozen tissue sections or cells were fixed with 4% paraformaldehyde, then stained with Oil Red O working solution (6 parts saturated Oil Red O solution + 4 parts deionized water, filtered twice) for 8–10 min in the dark. Sections were differentiated with 60% isopropanol for 3–5 s, rinsed with water, counterstained with hematoxylin, and mounted with glycerol gelatin.

Bodipy staining

Cells were incubated with Bodipy staining solution for 20–30 min at 37 °C, rinsed with PBS, and observed under a fluorescence microscope. The positive area ratio or positive cell ratio was quantified using ImageJ software.

Immunohistochemistry (IHC) staining

Paraffin-embedded tissue Sect.  (4 μm thick) were deparaffinized in xylene and rehydrated through a graded ethanol series. Antigen retrieval was performed by boiling sections in citrate buffer (pH 6.0) for 8 min followed by 7 min at 80 °C. Endogenous peroxidase activity was blocked with 3% H₂O₂ for 10 min. Sections were blocked with 5% BSA for 1 h at room temperature, then incubated with the PPARγ primary antibody overnight at 4 °C. After washing with PBS, sections were incubated with secondary antibody for 1 h at room temperature, followed by DAB staining and hematoxylin counterstaining. Sections were dehydrated, cleared, and mounted with neutral gum. Images were captured under a light microscope, and the positive cell ratio was quantified using ImageJ software.

Immunofluorescence (IF) staining

Cells grown on coverslips were fixed with 4% paraformaldehyde for 15 min at room temperature, permeabilized with 0.1% Triton X-100 for 10 min, and blocked with 5% BSA for 30 min. Cells were incubated with primary antibodies against PPARγ, Ki67, or other target proteins overnight at 4 °C, then with fluorescently labeled secondary antibodies for 1 h at room temperature in the dark. Nuclei were stained with DAPI for 10 min. Coverslips were mounted with anti-fluorescence quenching mounting medium and observed under a laser confocal microscope (Olympus, Tokyo, Japan). The average fluorescence intensity was quantified using ImageJ software.

Lentiviral transfection and stable cell line establishment

PPARγ overexpression lentivirus (Hanheng Biotechnology, Shanghai, China) and PPARγ knockdown lentivirus (Shaanxi Qingke Biotechnology, Xi’an, China) were used to construct stable cell lines. Cells were seeded in 24-well plates at a density of 1 × 10⁵ cells/well and transfected with lentivirus at the optimal multiplicity of infection (MOI = 30) determined by preliminary experiments. At 48–72 h post-transfection, cells were selected with 4 µg/mL puromycin (CAS9 Biotechnology, Xi’an, China) for 2 weeks to obtain stable cell lines. The transfection efficiency and PPARγ expression level were verified by qPCR, Western blot, and immunofluorescence (IF) staining.

CCK-8 assay

Cells were seeded in 96-well plates at a density of 5 × 10³ cells/well and cultured for 24 h. CCK-8 solution (10 µL/well) was added, and the plates were incubated for another 2 h at 37 °C. The absorbance at 450 nm was measured using a microplate reader (Bio-Rad, Hercules, USA).

Colony formation assay

Cells were seeded in 6-well plates at a density of 400–1000 cells/well and cultured for 14 days. Colonies were fixed with 4% paraformaldehyde for 30 min, stained with 0.1% crystal violet for 30 min, rinsed with PBS, and counted under a light microscope.

Wound-healing assay

Cells were seeded in 6-well plates and cultured until confluence > 90%. A scratch was made using a 200 µL pipette tip, and cells were rinsed with PBS to remove detached cells. Images were captured at 0 h, 24 h, and 48 h under a microscope, and the wound closure rate was calculated.

Transwell assay

For the migration assay, cells (1–10 × 10⁵ cells/mL) in serum-free medium were added to the upper chamber of Transwell inserts (8 μm pore size). Medium containing 10% FBS was added to the lower chamber. After incubation for 12–48 h, cells that migrated to the lower surface of the membrane were fixed with 4% paraformaldehyde, stained with crystal violet, and counted. For invasion assay, the upper chamber was precoated with Matrigel (1:8 dilution) for 1–3 h at 37 °C before seeding cells, and the remaining steps were the same as the migration assay.

Cell cycle analysis

Cells were harvested, fixed with 70% ethanol at 4 °C for 30 min, rinsed with PBS, and stained with PI/RNase staining solution for 30 min at 37 °C in the dark. Cell cycle distribution was analyzed by flow cytometry (Thermo Fisher, Hangzhou, China).

Apoptosis analysis

Cells were collected (including supernatant and adherent cells), centrifuged at 500 g for 5 min, rinsed with PBS twice, and resuspended at 1 × 10⁶ cells/mL. Annexin V-FITC and PI staining solutions were added, and cells were incubated for 8–10 min in the dark. Apoptosis rate was analyzed by flow cytometry within 1 h.

Cell-derived xenograft (CDX) model

Male nude mice (5 weeks old, 18–21 g) were purchased from Huachuang Sino (Jiangsu, China) and housed in a specific pathogen-free (SPF) environment for 1 week of adaptation. Stable PPARγ-overexpressing or control SW872 cells (5 × 10⁷ cells/mL) were subcutaneously injected into the right axilla of nude mice (100 µL/mouse). Tumor volume was measured every 7 days using a vernier caliper, and volume was calculated as V = 1/2 × L × W² (L: long diameter, W: short diameter). At 42 days post-inoculation, mice were anesthetized with isoflurane, and in vivo fluorescence imaging was performed using a small animal imaging system (Perkinelmer, Shanghai, China) after intraperitoneal injection of D-luciferin potassium salt (10 µL/g). Mice were then euthanized, and tumors were dissected, photographed, weighed, and measured. Tumor tissues were stored at -80 °C or fixed in 4% paraformaldehyde for subsequent experiments. Animal experiments were approved by the Animal Ethics Committee of Air Force Medical University (Approval No. 20250142).

Patient-derived xenograft (PDX) model

Fresh RLPS tissues from patients were collected during surgery, immediately placed in pre-cooled tissue preservation solution (Miltenyi Biotec, Bergisch Gladbach, Germany), and transferred to the ultra-clean bench on ice. Tissues were washed with PBS containing 10% penicillin-streptomycin, trimmed into 1.5 mm³ fragments, and subcutaneously implanted into the right axilla of nude mice under anesthesia. Tumor growth was monitored every 7 days by measuring volume and weight. When the tumor volume of the F1 generation reached 1000–1500 mm³, tumors were dissected and passaged to the F3 generation. The validity of the PDX model was verified by pathological examination and molecular marker detection. For drug treatment (TargetMol Chemicals, Boston, USA), mice were randomly divided into the methyl cellulose (MC) group and the TGZ group. TGZ (200 mg/kg) or MC solvent (100 µL/mouse) was administered intragastrically daily for 42 days. Tumor growth, body weight changes, and molecular alterations were evaluated.

Molecular docking and molecular dynamics simulation

The crystal structures of PPARγ and PI3K were retrieved from the AlphaFold and RCSB PDB databases. Water molecules and phosphate ions were removed using PyMOL 2.6.0 software. Molecular docking was performed using HDOCK software with the docking region set to full-surface mode, and 100 conformations were generated. The binding affinity was evaluated based on the docking score, and the results were visualized using PyMOL 2.6.0 and Ligplot. A 100-ns molecular dynamics simulation was performed to verify the stability of the PPARγ-PI3K complex, and parameters including RMSD, RMSF, Rg, and hydrogen bond number were analyzed.

Co-immunoprecipitation (Co-IP) assay

Cells were lysed in IP lysis buffer, and protein concentration was determined. Equal amounts of protein were incubated with PPARγ or PI3K primary antibody overnight at 4 °C, followed by incubation with protein A/G agarose beads for 4 h at 4 °C. Beads were washed with IP lysis buffer three times, and the immunoprecipitated proteins were analyzed by Western blot.

Statistical analysis

All statistical analyses were performed using SPSS 25.0 software, and graphs were plotted using GraphPad Prism 8.0 software. Continuous data were presented as the mean ± standard deviation (SD). Comparisons between two groups were performed using an unpaired Student’s t-test or a paired t-test. Comparisons among multiple groups were performed using one-way analysis of variance (ANOVA) followed by post-hoc tests. The correlation between PPARγ expression and clinicopathological characteristics was analyzed using Pearson’s χ² test. Survival analysis was performed using the Log-rank test. P < 0.05 was considered statistically significant (*P < 0.05, **P < 0.01, ***P < 0.001; ns, no statistical significance).

Results

Integrative database and transcriptome sequencing analyses reveal PPARγ downregulation in RLPS

To elucidate molecular expression disparities between RLPS and WAT, we integrated multi-public database analyses with transcriptome sequencing validation. Comparative analysis of TCGA-SARC DDLPS and WAT datasets revealed marked PPARγ downregulation in DDLPS (Fig. 1A); KEGG enrichment analysis showed their differentially expressed genes were significantly enriched in the PPAR signaling pathway (Fig. 1C), and GSEA confirmed the statistical significance of this pathway’s expression divergence (p < 0.001) (Fig. 1B). Analysis of 15 DDLPS, WDLPS and WAT samples each from the GEO dataset identified a graded PPARγ decrease (highest in WAT, intermediate in WDLPS, and lowest in DDLPS) (Fig. 1D), consistent with TCGA results and indicating progressive PPARγ downregulation with increasing RLPS malignant grade. GTEx analysis demonstrated the highest PPARγ basal expression in WAT (Fig. 1E), suggesting its downregulation in RLPS may correlate with the early stage of WAT-to-RLPS malignant transformation. Transcriptome sequencing of five WAT, WDLPS and DDLPS samples each showed GSEA-validated differentially expressed genes from WAT vs. DDLPS were significantly enriched in the PPAR pathway (Fig. 1F, H), further corroborating multi-database conclusions. We identified 4028 upregulated and 2791 downregulated genes in this comparison (Fig. 1G), with PPARγ and its associated ligands/receptors showing prominent downregulation in RLPS (Fig. 1I). In aggregate, PPARγ expression gradually decreases with RLPS grade and may play a key role in the malignant transformation of adipose tissue through the PPAR signaling pathway.

Fig. 1.

Integrated database analysis and transcriptome sequencing reveal downregulated PPARγ expression in retroperitoneal liposarcoma. A Volcano plot of differentially expressed genes between DDLPS and normal adipose tissue in the TCGA dataset. B GSEA enrichment analysis of the PPAR signaling pathway based on the TCGA dataset. C KEGG pathway enrichment analysis of differentially expressed genes in the TCGA dataset. D PPARγ expression distribution in DDLPS, WDLPS and normal adipose tissue in the GEO database. E PPARγ RNA expression levels in abdominal visceral tissues in the GTEx database. F Expression heatmap of PPAR signaling pathway-related genes in transcriptome sequencing (n = 5). G Volcano plot of differentially expressed genes in transcriptome sequencing. H GSEA enrichment analysis of the PPARγ signaling pathway in transcriptome sequencing. I Expression levels of PPARγ and its associated ligands/receptors (RARG, PPARGC1B, PPARGC1A) in transcriptome sequencing (n = 5). * p < 0.05, ** p < 0.01

PPARγ expression and lipid accumulation are reduced with elevated malignant grade of liposarcoma

To investigate PPARγ protein expression characteristics and its association with lipid metabolism, three samples each of WAT, WDLPS and DDLPS were selected, and the correlation between lipid accumulation and PPARγ expression in RLPS of varying malignant grades was analyzed via Oil Red O staining and IHC assays. Compared with WAT, RLPS exhibited progressive reduction in lipid accumulation with increasing malignant grade (Fig. 2A, B); PPARγ protein was predominantly nuclear-localized with minimal cytoplasmic expression, and its levels decreased markedly with elevated RLPS malignancy. Western blot validation confirmed high PPARγ expression in WAT, with its protein levels gradually downregulated as RLPS malignant grade increased (Fig. 2D, E). qPCR assays of 26 WAT, 28 WDLPS and 52 DDLPS clinical samples revealed that PPARγ mRNA expression was also significantly reduced with increasing RLPS malignancy (Fig. 2F). To validate these clinical findings at the cellular level, we performed in vitro assays with primary adipocytes (PAC) as controls, alongside the well-differentiated liposarcoma cell line 93T449 and the dedifferentiated SW872. Western blot and qPCR confirmed high PPARγ expression in PAC, with its protein and RNA levels decreasing significantly with increasing cellular malignant grade (Fig. 2G-I). Bodipy lipid staining showed a gradual reduction in intracellular lipid accumulation with elevated malignancy (Fig. 2J, K), while IF staining revealed diminished PPARγ expression in more malignant cells (Fig. 2J, L). These findings indicate that lipid accumulation and PPARγ expression decrease progressively with increasing malignancy of RLPS.

Fig. 2.

PPARγ expression and lipid accumulation are negatively correlated with the malignant grade of liposarcoma in clinical specimens and cellular models. A Representative images of Oil Red O staining and PPARγ IHC staining in tissues. B Quantification of the positive area ratio of Oil Red O staining. C Quantification of the positive cell ratio of PPARγ IHC staining. D Detection of PPARγ protein expression in tissues by Western blot. E Quantitative analysis histogram of PPARγ protein expression in tissues(n = 8). F Detection of PPARγ mRNA expression in tissues by qPCR (n WAT: WDL: DDL = 26:28:52). G Detection of PPARγ protein expression in cells by Western blot. H Quantitative analysis histogram of PPARγ protein expression in cells. I Detection of PPARγ mRNA expression in cells by qPCR. J Representative images of cellular Bodipy lipid staining and PPARγ IF staining. K Quantification of the positive cell ratio of Bodipy lipid staining. L Quantification of the positive cell ratio of PPARγ IF staining. All experiments were performed in triplicate (n = 3) unless otherwise specified, ** p < 0.01, *** p < 0.001

Low PPARγ expression is an independent predictor of poor prognosis in patients with RLPS

To investigate the association between PPARγ expression and clinicopathological characteristics in patients with RLPS, this study retrospectively analyzed data from RLPS patients admitted to Xijing Hospital between July 2014 and June 2024. A total of 181 RLPS patients were initially enrolled, including 116 who underwent primary surgery at our hospital and 65 treated at external institutions. After strict application of inclusion and exclusion criteria, 12 patients with other primary malignancies, 4 with incomplete clinical or follow-up data, 8 lost to follow-up, and 12 with preoperative radiotherapy or chemotherapy were excluded, leaving 80 eligible patients (28 WDLPS and 52 DDLPS). Pathological diagnoses and clinical staging were independently reviewed in a blinded manner by two experienced pathologists and confirmed according to the 7th edition of the AJCC staging system. The follow-up deadline was June 30, 2025.

PPARγ mRNA expression in 80 tissue specimens was measured by qPCR, and patients were stratified into high- and low-PPARγ expression groups based on the median value. Continuous clinical variables were similarly dichotomized at the median for correlation analysis. PPARγ expression was significantly associated with operative duration and MDM2 gene amplification percentage, but showed no significant correlation with age, sex, hospital stay, comorbidities, unrelated surgical history, BMI, number of primary tumors, surgical margin status, intraoperative blood loss, severe postoperative complications, TNM stage, MDM2 protein expression, CDK4 expression, or administration of adjuvant radiotherapy/chemotherapy (Table 1).

Table 1.

Relationship between PPARγ expression and clinicopathological characteristics in DDLPS patients (n=52)

Characteristics	PPARγ expression		Total(n = 52)	χ2	p Value
	High(n=26)	Low(n=26)
Age(year)
<60	13	12	25	0.077	0.7812
≥60	13	14	27
Gender
Male	18	17	35	0.087	0.7678
Female	8	9	17
Inpatient days(day)
≤13	15	13	28	0.309	0.5776
>13	11	13	24
Other complications
0	17	18	35	0.087	0.7678
1	9	8	17
History of surgeries
0	17	18	35	0.087	0.7678
1	9	8	17
BMI
<25	21	22	43	0.134	0.7142
≥25	5	4	9
Number of primary tumors
Solitary	19	21	40	0.433	0.5104
Multiple	7	5	12
Surgical method (R0/R1)
R0	20	23	43	1.209	0.2715
R1	6	3	9
Surgery duration (min)
≤180	17	9	26	4.923	0.0265
>180	9	17	26
Peroperative bleeding(mL)
≤600	14	12	26	0.307	0.579
>600	12	14	26
Postoperative complications
0	24	23	47	0.221	0.6383
1	2	3	5
Tumor size(T)
T1/T2	2	4	6	0.753	0.3851
T3/T4	24	22	46
Lymph node metastasis(N)
0	25	25	50	0	1
1	1	1	2
Distant metastasis(M)
0	19	20	39	0.102	0.7482
1	7	6	13
MDM2 expression
Negative	3	10	13	1.8	0.1795
Positive	13	16	29
CDK4 expression
Negative	3	4	7	0.165	0.6846
Positive	23	22	45
MDM2 gene amplification(%)
≤80	6	16	22	7.878	0.005
>80	20	10	30
Chemoradiotherapy
0	23	23	46	0	1
1	3	3	6

Given the limited sample size, the correlation between PPARγ expression and clinicopathological features in WDLPS patients was not representative and is therefore not presented here; detailed results are shown in Supplementary Table S1.

To determine the association between PPARγ expression and prognosis in RLPS patients, TCGA database analysis showed significantly shorter OS in the low-PPARγ group (HR = 0.13, 95% CI 0.2–0.98, P = 0.02; Fig. 3A), indicating that low PPARγ expression predicts poor prognosis in DDLPS. WDLPS also exhibited consistent results, and the relevant survival curves are shown in Supplementary Fig. S1. Clinical cohort validation confirmed markedly higher survival in DDLPS patients with high PPARγ expression (HR = 0.13, 95% CI 0.05–0.34, P = 1.9 × 10⁻⁶; Fig. 3B). PPARγ expression showed favorable predictive efficacy for 1-, 3-, and 5-year survival with AUC values of 0.85, 0.89, and 0.92 (Fig. 3C).

Fig. 3.

Low PPARγ expression is an independent predictor of poor prognosis in patients with RLPS. A Survival analysis of the association between PPARγ expression and OS in the TCGA database. B Survival analysis of PPARγ expression and OS in the clinical cohort from Xijing Hospital. C AUC curve. D Univariate survival analysis of the relationship between PPARγ and prognosis in DDLPS patients. E Multivariate survival analysis results. F Nomogram constructed from multivariate survival analysis. G ROC curves for predicting 12-, 36-, and 60-month survival probabilities

Univariate Cox analysis revealed that age, hospital stay, surgical margin, intraoperative blood loss, MDM2 amplification, and PPARγ expression were significantly associated with DDLPS survival (Fig. 3D). After multivariate adjustment, hospital stay, MDM2 amplification, age, intraoperative blood loss, and PPARγ expression were identified as independent prognostic factors for RLPS (Fig. 3E). A nomogram incorporating these five factors showed that higher total scores predicted poorer survival (Fig. 3F), and calibration curves confirmed its reliability (Fig. 3G).

In conclusion, low PPARγ expression serves as an independent predictor of poor prognosis in RLPS patients.

PPARγ overexpression inhibits proliferation, migration, and invasion and promotes apoptosis in RLPS

To investigate the role of PPARγ in the malignant progression of liposarcoma, we established PPARγ-overexpressing and knockdown stable cell lines in 93T449 and SW872 cells using a lentiviral system. Transfection efficiency exceeded 90% at MOI = 30, and stable lines were obtained after selection with 4 µg/mL puromycin. qPCR, Western blot and IF confirmed markedly elevated PPARγ expression in overexpressing cells (Fig. 4A-C). CCK-8 and colony formation assays demonstrated that PPARγ overexpression significantly suppressed proliferation in SW872 cells (Fig. 4D, E). Wound-healing and Transwell assays revealed that PPARγ overexpression markedly reduced cell migration and invasion (Fig. 4F–I). Flow cytometry showed that PPARγ overexpression induced G0/G1 phase arrest and promoted apoptosis in SW872 cells (Fig. 4J, K). Western blot analysis revealed an increased Bax/Bcl-2 ratio and PTEN, downregulated Cyclin B1, Cyclin D1 and N-cadherin upon PPARγ overexpression (Fig. 4L).

Fig. 4.

PPARγ overexpression inhibits proliferation, migration and invasion and promotes apoptosis in RLPS cells. A qPCR analysis to measure the overexpression level of PPARγ mRNA. B Western blot analysis to determine the overexpression efficiency of PPARγ protein. C IF assay to detect the expression and localization of PPARγ protein (n = 4). D CCK-8 assay to evaluate the effect of PPARγ overexpression on the proliferative capacity of RLPS cells. E Colony formation assay to assess the impact of PPARγ overexpression on cell growth. F Wound-healing assay to determine the effect of PPARγ overexpression on cell migratory ability. G Transwell migration and invasion assays to evaluate the effects of PPARγ overexpression on cell migration and invasion. H Quantitative analysis of Transwell migration assay. I Quantitative analysis of Transwell invasion assay. J Flow cytometry analysis to examine the effect of PPARγ overexpression on cell cycle distribution. K Flow cytometry analysis to investigate the impact of PPARγ overexpression on cell apoptosis L Western blot analysis to detect the expression of proliferation-, metastasis-, cell cycle- and apoptosis-related proteins (PTEN, N-cadherin, CyclinB1, CyclinD1, and Bax/Bcl-2). All experiments were performed in triplicate (n = 3) unless otherwise specified, * p < 0.05, ** p < 0.01, *** p < 0.001

In vivo assays showed that tumor fluorescence intensity, volume and weight were significantly lower in the PPARγ overexpression group than in the control group, with markedly suppressed tumor growth (Fig. 5A–E). Histological staining revealed increased lipid droplets and reduced Ki67 positivity in PPARγ-overexpressing tumors (Fig. 5F–I). Since 93T449 and SW872 cells yielded similar results, only data from SW872 cells are shown in the main text, with those for 93T449 provided in Supplementary Fig. S2.

Fig. 5.

PPARγ overexpression inhibits RLPS proliferation in vivo A Subcutaneous xenograft model of SW872 cells in nude mice (n = 6). B Tumor fluorescence intensity measured by in vivo small animal imaging (n = 6). C Quantitative analysis of xenograft weight (n = 6). D Growth curve of xenograft volume over time (n = 6). E Representative images of dissected solid tumors from the axilla of nude mice (n = 6). F Representative images of Oil Red O staining in tumor tissues (n = 3). G Representative images of IF staining in tumor tissues. H Quantitative analysis of the average fluorescence intensity of PPARγ (n = 3). I Quantitative analysis of the average fluorescence intensity of Ki67 (n = 3). ** p < 0.01, *** p < 0.001

Altogether, PPARγ overexpression markedly inhibits proliferation, migration and invasion of RLPS cells in vitro and in vivo, and promotes tumor cell apoptosis.

PPARγ knockdown promotes proliferation, migration, and invasion and inhibits apoptosis in RLPS cells

Having previously demonstrated that PPARγ overexpression suppresses the malignant biological behaviors of liposarcoma cells, we further verified whether PPARγ knockdown exerts the opposite effect by silencing PPARγ in liposarcoma cells using the CRISPR/Cas9 lentiviral system. Western blot and qPCR confirmed successful knockdown (Fig. 6A, B). Using SW872 cells, CCK-8 and colony formation assays showed PPARγ knockdown significantly enhanced proliferation (Fig. 6C, D); wound-healing and transwell assays demonstrated increased migration and invasion (Fig. 6E-H); flow cytometry revealed reduced G0/G1 phase, increased S phase, and decreased apoptosis (Fig. 6I, J). Western blot showed upregulated N-cadherin, Cyclin B1, and Cyclin D1, and a reduced Bax/Bcl-2 ratio, PTEN after knockdown (Fig. 6K). Consistent results were observed in WDLPS cells (Supplementary Fig. S2). Since 93T449 and SW872 cells yielded similar results, only data from SW872 cells are shown in the main text, with those for 93T449 provided in Supplementary Fig. S3. Taken together, PPARγ acts as a tumor suppressor in liposarcoma, with high expression inhibiting and low expression promoting malignancy.

Fig. 6.

PPARγ knockdown promotes proliferation, migration and invasion and inhibits apoptosis in RLPS cells. A qPCR analysis of PPARγ mRNA knockdown efficiency. B Western blot analysis of PPARγ protein knockdown efficiency (n = 4). C CCK-8 assay for the effect of PPARγ knockdown on cell proliferation. D Colony formation assay for the effect of PPARγ knockdown on cell growth. E Wound-healing assay for the effect of PPARγ knockdown on cell migration. F Transwell migration and invasion assays for the effects of PPARγ knockdown on cell migration and invasion. G Quantitative analysis of Transwell migration. H Quantitative analysis of Transwell invasion. I Flow cytometric analysis of the effect of PPARγ knockdown on cell cycle distribution. J Flow cytometric analysis of the effect of PPARγ knockdown on cell apoptosis. K Western blot analysis of the expression of proliferation-, metastasis-, cell cycle- and apoptosis-related proteins (PTEN, N-cadherin, Cyclin B1, Cyclin D1, Bax/Bcl-2). All experiments were performed in triplicate (n = 3) unless otherwise specified, * p < 0.05, ** p < 0.01, *** p < 0.001; ns indicates no statistical significance

PPARγ inhibits the PI3K/AKT pathway by interacting with PI3K in RLPS

Our previous results demonstrated that PPARγ knockdown markedly enhances the malignant biological behaviors of retroperitoneal liposarcoma cells. To further elucidate the molecular mechanism by which PPARγ modulates the initiation and progression of RLPS, we performed RNA sequencing in PPARγ-knockdown and control SW872 cells. 104 upregulated genes and 466 downregulated genes were identified (Fig. 7A). GSEA revealed significant enrichment of the PPAR signaling pathway (Fig. 7B), and KEGG pathway analysis indicated that the PI3K/AKT pathway was among the most significantly enriched terms (Fig. 7C).

Fig. 7.

PPARγ interacts with PI3K and regulates the PI3K-Akt signaling pathway A Volcano plot of differentially expressed genes. B GSEA analysis of PPAR signaling pathway enrichment. C KEGG pathway enrichment analysis. D Hydrogen bond interactions between PPARγ and PI3K residues. E RMSD curve of the PPARγ-PI3K complex. F RMSF curve of the PPARγ-PI3K complex. G Free energy landscape of PI3K. H Free energy landscape of PPARγ. I Radius of gyration (Rg) curve of the PPARγ-PI3K complex. J Hydrogen bond number variation of the PPARγ-PI3K complex. K Co-IP analysis of endogenous PI3K pulled down by PPARγ. L Co-IP analysis of endogenous PPARγ pulled down by PI3K. M The effects of DMSO, PI3K-IN, and PPARγ on the binding of PPARγ to PI3K. N Western blot analysis of PI3K/AKT and phosphorylated levels after PPARγ knockdown (n = 4). *** p < 0.001

Consistent with previous studies showing that PPARγ agonists suppress hypopharyngeal squamous cell carcinoma proliferation by upregulating PTEN and inactivating the PI3K/AKT pathway [32], we hypothesized that in RLPS, PPARγ may inhibit cell proliferation by suppressing PI3K/AKT pathway activation.

Molecular docking analysis indicated stable binding between PPARγ and PI3K, with a docking score of -285.10 and a confidence score of 0.9371 (Fig. 7D). A 100-ns molecular dynamics simulation showed that the PPARγ-PI3K complex reached equilibrium after 30 ns. Further analyses of RMSD, RMSF, radius of gyration, and hydrogen bonding collectively confirmed tight and stable binding between the two proteins (Fig. 7E–J). Co-IP assays verified the endogenous interaction between PPARγ and PI3K, and overexpression of PPARγ further enhanced their association (Fig. 7K–M). Western blot analysis revealed that PPARγ knockdown significantly elevated the phosphorylation levels of PI3K and AKT without altering their total protein expression (Fig. 7N), suggesting that PPARγ blocks PI3K/AKT activation by inhibiting their phosphorylation rather than affecting protein expression.

Collectively, PPARγ forms a stable complex with PI3K and suppresses the phosphorylation-mediated activation of the PI3K/AKT pathway, thereby restraining the malignant progression of RLPS.

Inhibition of the PI3K/AKT signaling pathway attenuates PPARγ knockdown-mediated malignant progression of RLPS cells

To determine the mediating role of the PI3K/AKT pathway in PPARγ‑regulated malignant progression of RLPS, we performed pathway intervention using the specific PI3Kα/δ inhibitor Pictilisib (GDC‑0941) [33, 34]. Preliminary concentration screening identified 50 nM as the optimal working concentration (Fig. 8A). Western blot analysis confirmed that 50 nM Pictilisib markedly reduced the levels of p‑PI3K and p‑AKT in sh-NC and PPARγ‑knockdown cells without altering total protein expression, indicating that Pictilisib effectively reversed the excessive activation of the PI3K/AKT pathway induced by PPARγ depletion (Fig. 8B). In vitro functional assays demonstrated that Pictilisib treatment significantly suppressed cell proliferation and colony formation (Fig. 8C, D), and attenuated cell migration and invasion (Fig. 8E–H). Flow cytometry revealed that Pictilisib induced G0/G1 phase arrest and promoted apoptosis (Fig. 8I, J). Western blot analysis further showed that Pictilisib downregulated N-cadherin, Cyclin B1 and Cyclin D1, elevated the Bax/Bcl‑2 ratio, and upregulated PTEN expression, consistent with the phenotypic changes (Fig. 8K). Since 93T449 and SW872 cells yielded similar results, only data from SW872 cells are shown in the main text, with those for 93T449 provided in Supplementary Fig. S4. On the whole, Pictilisib effectively reverses PPARγ knockdown‑mediated malignant phenotypes of RLPS by inhibiting PI3K/AKT activation, confirming that the regulatory effect of PPARγ on RLPS progression is dependent on the PI3K/AKT signaling pathway.

Fig. 8.

PI3K inhibitors mitigate PPARγ knockdown-induced malignant progression in RLPS cells. A Western blot analysis of PI3K and p-PI3K expression after 24 h treatment with different concentrations of Pictilisib. B Western blot analysis of PI3K/AKT and their phosphorylated protein expression following Pictilisib treatment. C CCK-8 assay to evaluate the effect of Pictilisib on the proliferative capacity of PPARγ-knockdown cells. D Colony formation assay to detect the regulatory effect of Pictilisib on cell colony formation ability. E Wound-healing assay to analyze the effect of Pictilisib on cell migratory capacity. F Transwell migration and invasion assays to assess the effects of PPARγ knockdown and Pictilisib treatment on cell migration and invasion. G Quantitative analysis of Transwell migration assay. H Quantitative analysis of Transwell invasion assay. I Flow cytometric analysis of cell cycle distribution after Pictilisib treatment. J Flow cytometric analysis of the effect of Pictilisib on cell apoptosis. K Western blot analysis of the expression of proliferation-, metastasis-, cell cycle- and apoptosis-related proteins (PTEN, N-cadherin, CyclinB1, CyclinD1, Bax/Bcl-2). All experiments were performed in triplicate (n = 3), * p < 0.05, ** p < 0.01, *** p < 0.001; ns indicates no statistical significance

PI3K inhibitors strengthen the tumor-suppressive function of PPARγ overexpression in RLPS cells

PPARγ overexpression alone significantly reduced the phosphorylation levels of PI3K and AKT, while additional Pictilisib treatment further suppressed PI3K/AKT activation in OE-PPARγ cells (Fig. 9A). Consistently, functional assays demonstrated that PPARγ overexpression inhibited cell proliferation, colony formation, migration, and invasion, induced G0/G1 phase arrest, and promoted apoptosis. These anti-tumor phenotypes were significantly enhanced by Pictilisib (Fig. 9B–I). Western blot analysis further confirmed that Pictilisib reinforced the upregulation of PTEN and the Bax/Bcl-2 ratio, as well as the downregulation of N-cadherin, Cyclin B1, and Cyclin D1 in PPARγ-overexpressing cells (Fig. 9J).

Fig. 9.

PI3K inhibitors strengthen the tumor-suppressive function of PPARγ overexpression in RLPS cells. A Western blot analysis of PI3K/AKT and their phosphorylated protein expression following Pictilisib treatment. B CCK-8 assay to evaluate the effect of Pictilisib on the proliferative capacity of PPARγ-overexpression cells. C Colony formation assay to detect the regulatory effect of Pictilisib on cell colony formation ability. D Wound-healing assay to analyze the effect of Pictilisib on cell migratory capacity. E Transwell assays to assess the effects of PPARγ overexpression and Pictilisib treatment on cell migration and invasion. F Quantitative analysis of Transwell migration assay. G Quantitative analysis of Transwell invasion assay. H Flow cytometric analysis of cell cycle distribution after Pictilisib treatment. I Flow cytometric analysis of the effect of Pictilisib on cell apoptosis. J Western blot analysis of the expression of proliferation-, metastasis-, cell cycle- and apoptosis-related proteins (PTEN, N-cadherin, CyclinB1, CyclinD1, Bax/Bcl-2). All experiments were performed in triplicate (n = 3), * p < 0.05, ** p < 0.01, *** p < 0.001; ns indicates no statistical significance

Collectively, these results, together with the rescue data in PPARγ-knockdown cells, demonstrate that the regulatory effects of PPARγ on RLPS malignant progression are mediated, at least in part, by the PI3K/AKT signaling pathway. PPARγ overexpression and PI3K inhibition exert synergistic anti-tumor effects, providing a rationale for combining PPARγ agonists with PI3K inhibitors in the treatment of RLPS.

Troglitazone inhibits malignant phenotypes of RLPS cells via activating PPARγ

Troglitazone (TGZ) is an orally active and specific agonist of PPARγ. Its classic pharmacological function is to ameliorate insulin resistance and enhance insulin sensitivity [35–37]. Moreover, TGZ has been reported to exert tumor‑suppressive activity in pancreatic cancer, lung cancer and other malignancies, providing a theoretical basis for its application in RLPS [38–40].

To determine the optimal concentration of TGZ, SW872 cells were treated with 0, 0.5, 1, 5, 10, and 100 µM TGZ for 24 h. Western blot and qPCR analyses revealed that PPARγ expression was upregulated in a concentration-dependent manner (Fig. 10A, B). CCK-8 assay showed that the inhibitory effect on proliferation plateaued at 10 µM TGZ (Fig. 10C). Treatment with 10 µM TGZ for 24 h significantly elevated PPARγ mRNA and PPARγ protein expression (Fig. 10D, E), confirming efficient activation of PPARγ by TGZ. In vitro functional assays demonstrated that TGZ treatment markedly suppressed cell proliferation and colony formation (Fig. 10F, G), and attenuated cell migration and invasion (Fig. 10H–K). Flow cytometry revealed that TGZ induced G0/G1 phase arrest and promoted apoptosis (Fig. 10L, M). Western blot analysis showed that TGZ downregulated N-cadherin, Cyclin B1 and Cyclin D1, increased the Bax/Bcl-2 ratio, and upregulated PTEN expression (Fig. 10N). Since 93T449 and SW872 cells yielded similar results, only data from SW872 cells are shown in the main text, with those for 93T449 provided in Supplementary Fig. S5. These molecular changes were consistent with phenotypic alterations, indicating that TGZ reverses the pro-tumor phenotype mediated by PPARγ knockdown.

Fig. 10.

Troglitazone inhibits malignant phenotypes of RLPS cells via activating PPARγ. A qPCR analysis of PPARγ mRNA expression treated with different concentrations of TGZ for 24 h. B Western blot analysis of PPARγ protein expression treated with different concentrations of TGZ for 24 h. C CCK‑8 assay for the effect of different concentrations of TGZ on SW872 cell proliferation. D Western blot analysis of PPARγ protein expression in PPARγ‑knockdown cells after TGZ treatment. E qPCR analysis of PPARγ mRNA expression in PPARγ‑knockdown cells after TGZ treatment. F CCK‑8 assay for the effect of TGZ on proliferation of PPARγ‑knockdown cells. G Colony formation assay for the effect of TGZ on colony formation of PPARγ‑knockdown cells. H Wound‑healing assay for the effect of TGZ on migration of PPARγ‑knockdown cells. I Transwell migration and invasion assays for the effects of TGZ on migration and invasion of PPARγ‑knockdown cells. J Quantitative analysis of Transwell migration. K Quantitative analysis of Transwell invasion. L Flow cytometric analysis of cell cycle distribution in PPARγ‑knockdown cells after TGZ treatment. M Flow cytometric analysis of the effect of TGZ on apoptosis in PPARγ‑knockdown cells. N Western blot analysis of the expression of proliferation‑, metastasis‑, cell cycle‑ and apoptosis‑related proteins (PTEN, E‑cadherin, CyclinB1, CyclinD1, Bax/Bcl‑2). All experiments were performed in triplicate (n = 3), * p < 0.05, ** p < 0.01, *** p < 0.001; ns indicates no statistical significance

TGZ suppresses the growth of RLPS in vivo

To further verify the effect of TGZ on RLPS growth in vivo, a CDX model was established using SW872 cells in immunodeficient nude mice. Six mice were used as the NC control group, and 18 mice implanted with PPARγ‑knockdown cells were randomly divided into sh‑PPARγ, Methyl cellulose (MC), and TGZ‑treated groups (200 mg/kg). After 42 days of intragastric administration, in vivo fluorescence imaging and tumor weighing revealed that tumor fluorescence intensity, volume, and weight were markedly reduced in the TGZ group compared with the MC group (Fig. 11A–E). IF staining of tumor tissues showed that TGZ upregulated PPARγ expression, promoted lipid accumulation, and decreased the Ki67‑positive rate (Fig. 11F–I). These results indicate that TGZ significantly inhibits the in vivo growth of RLPS in the CDX model by upregulating PPARγ, enhancing lipid accumulation, and suppressing cell proliferation.

Fig. 11.

TGZ inhibits RLPS proliferation in the CDX model. A In vivo small animal imaging of the nude mouse CDX model (n = 6). B Quantitative analysis of fluorescence intensity from in vivo imaging (n = 6). C Dissected solid tumor specimens (n = 6). D Tumor volume statistics (n = 6). E Tumor weight statistics (n = 6). F Representative images of Bodipy staining (n = 3). G Representative images of IF staining. H Quantitative analysis of the average fluorescence intensity of PPARγ (n = 3). I Quantitative analysis of the average fluorescence intensity of Ki67 (n = 3). * p < 0.05, ** p < 0.01, *** p < 0.001; ns indicates no statistical significance

To further validate the in vivo antitumor activity and translational potential of TGZ, a PDX model of retroperitoneal DDLPS was employed. Compared with the control group, TGZ treatment markedly suppressed tumor growth in the PDX model without obvious fluctuations in body weight, suggesting favorable in vivo safety (Fig. 12A–D). IF staining demonstrated that TGZ significantly increased PPARγ expression and reduced Ki67 levels in tumor tissues (Fig. 12E–G), consistent with the findings in the CDX model and in vitro experiments.

Fig. 12.

TGZ inhibits RLPS proliferation in the PDX model. A Schematic diagram of PDX model establishment. B Tumor growth of PDX models transplanted with dedifferentiated RLPS after MC and TGZ treatment. C Tumor volume changes were analyzed via two-way ANOVA (n = 6). D Statistical analysis of nude mouse body weight changes (n = 6). E IF staining showing the expression of PPARγ and Ki67. F Quantitative analysis of the average fluorescence intensity of PPARγ (n = 3). G Quantitative analysis of the average fluorescence intensity of Ki67 (n = 3). ** p < 0.01

Discussion

This study is the first to systematically elucidate the critical regulatory role of PPARγ in the pathogenesis of RLPS, clearly revealing its dual value as a prognostic biomarker and a potential therapeutic target. We provide conclusive evidence that PPARγ exerts its tumor-suppressive effect by directly inhibiting the activation of the PI3K/AKT signaling pathway, and its specific agonist TGZ exhibits significant antitumor activity in both in vitro cell experiments and in vivo animal models. These findings offer a novel strategic option for the clinical treatment of RLPS and open up a new direction for the research on targeted therapy of this disease.

We first validated through multi‑dimensional analyses that decreased PPARγ expression in RLPS patient tissues is significantly associated with poor prognosis, suggesting that PPARγ may serve as a reliable prognostic biomarker. Further mechanistic investigations focused on the pivotal PI3K/AKT pathway. A core finding of the present work is that multiple lines of structural and experimental evidence including molecular docking, molecular dynamics simulation and co-immunoprecipitation assays substantiated the direct binding between PPARγ and PI3K. PPARγ directly interacts with the PI3K/AKT cascade, thereby inhibiting RLPS cell proliferation, migration, and invasion while promoting apoptosis [41, 42]. These findings not only deepen the understanding of the molecular pathogenesis of RLPS, but also provide a new perspective for delineating the tumor‑suppressive mechanisms of PPARγ. Either endogenous activation of PPARγ or exogenous administration of its agonist TGZ effectively blocks PI3K/AKT signaling activation and consequently restrains RLPS malignant progression. These effects are fully validated in preclinical models, laying a solid foundation for subsequent drug development and clinical translation.

As a highly heterogeneous malignant tumor, RLPS has long been primarily treated with surgical resection [43, 44]; however, high recurrence rates and distant metastasis remain major clinical challenges [7, 45, 46]. The present study identifies PPARγ as a key molecular target for RLPS, which is highly consistent with the current development trend of precision medicine [47]. Although PPARγ has been shown to exert potential antitumor effects in other solid tumors, such as bladder cancer, pancreatic cancer, and lung cancer [17, 38, 39], its specific regulatory mechanisms and targeting potential in RLPS have not been fully elucidated previously. Our findings fill this research gap and provide important theoretical support and experimental basis for the transformation of the existing “surgery-dominated” diagnosis and treatment model of RLPS to a “molecular-targeted” precision therapy strategy. Furthermore, the systematic validation of the antitumor effect of TGZ in this study highlights the great potential of “old drugs for new uses” of existing medications, which may accelerate the clinical translation of novel therapeutic strategies for RLPS and offer new possibilities for improving patient prognosis.

Although this study established a relatively complete evidence chain, several limitations require objective scrutiny. First, the clinical sample size (80 cases) and cell line types are relatively limited, which may affect the generalizability of conclusions and require validation by large-scale, multi-center prospective studies [48, 49]. Second, while we clarified the inhibitory effect of PPARγ on the PI3K/AKT pathway, the specific domains mediating PPARγ-PI3K interaction and the upstream regulatory mechanisms (e.g., epigenetic modifications) of PPARγ remain unelucidated, failing to form a complete regulatory network. Furthermore, TGZ has a history of hepatotoxicity in previous clinical applications [50, 51], which warrants careful hepatological monitoring in future clinical trials.

Based on the findings and limitations of this study, future research should focus on the following aspects. First, conduct multi-center, large-sample clinical cohort studies to validate conclusions, establish a standardized PPARγ detection protocol, and develop prognostic and efficacy prediction models. Second, use structural biology techniques to resolve the details of PPARγ-PI3K interaction and explore its upstream signaling pathways and epigenetic regulatory mechanisms. In particular, screening and identifying key intermediate proteins that mediate the interaction between PPARγ and PI3K will be a core direction for subsequent mechanistic research. Third, develop targeted delivery systems for TGZ or novel PPARγ agonists to reduce toxicity, and evaluate their combined effects and synergistic potential with existing therapies [52, 53]. Fourth, considering the complex clinical characteristics of RLPS and restrictions related to animal ethics, long-term survival assessment was not conducted in this study. Moreover, we intend to optimize animal models and experimental protocols in follow-up studies to further explore the translational potential of the present findings. Establish humanized models simulating the tumor microenvironment to explore the interaction between PPARγ and tumor microenvironment cells [54, 55], as well as crosstalk with other key signaling pathways [56] (e.g., MAPK), to improve the regulatory network.

In summary, this study is the first to show that PPARγ, a key tumor suppressor in RLPS, exerts antitumor effects by inhibiting the PI3K/AKT pathway, laying a theoretical foundation for RLPS molecular research, targeted therapy and therapeutic model transformation. Future interdisciplinary research is expected to translate this into clinical practice to improve patient prognosis.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1 ^{(29KB, xls)}

Supplementary Material 2 ^{(9.9MB, doc)}

Supplementary Material 3 ^{(6.3MB, doc)}

Supplementary Material 4 ^{(28.9KB, xlsx)}

Supplementary Material 5 ^{(1.2MB, docx)}

Supplementary Material 6 ^{(23.2KB, xlsx)}

Author contributions

Niu Dai: conceptualization, experimental design, data acquisition, data analysis, data interpretation, funding acquisition and original manuscript drafting. Juzheng Yuan: experimental design, data acquisition, data analysis, data interpretation, funding acquisition, manuscript reviewing and editing. Miaojie Tian: data acquisition and data analysis. Haohao Ding: data acquisition and data analysis. Xudan Wang: data acquisition and data analysis. Fuyaun Liu: data acquisition and data analysis. Xuqiang Liu: facility management, data acquisition and data analysis. Yongxing Wang: data interpretation, scientific discussion, supervision, and funding acquisition. Xiao Li: data interpretation, scientific discussion, supervision, funding acquisition, manuscript reviewing and editing. Shuqiang Yue: conceptualization, data interpretation, scientific discussion, supervision, funding acquisition, manuscript reviewing and editing. All authors critically revised the article for important intellectual content and approved the submission for publication.

Funding

This study was supported by the Clinical Research Program of Air Force Medical University (2022LC2205), the Discipline Promotion Program of Xijing Hospital (XJZT24LY45), the Graduate Research Fund of Xijing Hospital (2025SZ010, XT2023E2043), the Key Research and Development Program of Shaanxi Province (2023-YBSF-429, 2024SF2-GJHX-08), and the Science and Technology Program of the Public Health Research Joint Fund for Public Hospitals of Inner Mongolia Academy of Medical Sciences (2025GLLH0437).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

All experimental procedures involving animals and human‑derived materials were performed in strict accordance with relevant guidelines and regulations. Animal study protocols were reviewed and approved by the Animal Ethics Committee of Air Force Medical University (Approval No. 20250142). The use of human tissue samples, clinical data, and all related experimental procedures were approved by the Ethics Committee of Xijing Hospital (Approval No. XJLL‑KY‑2025237). Written informed consent was obtained from all patients.

Consent for publication

All authors have read the manuscript and provided their consent for the submission.

Competing interests

The authors declare no competing interests.