Cell culture and transfection

OC cell lines SKOV3, OVCAR8 and mouse-derived ID8 were cultured in DMEM high-glucose medium (HyClone, SH30243) with 10% FBS and 1% antibiotics, and authenticated by STR profiling. SKOV3-Luc cells were generated by transfecting SKOV3 cells with a luciferase reporter plasmid containing puromycin resistance. For stable transfection, shRNAs targeting ITGA5, ITGB1, ITGA5/ITGB1, and an ITGA5-eGFP overexpression plasmid, along with control plasmids, were designed by Genomeditech (Shanghai, China). Positive cells were selected using blasticidin (5 μg/mL) and puromycin (3 μg/mL).

For PPP2CA knockout, sgRNA (5'-3' acgtgcaagaggttcgatgt) was cloned into the LentiCRISPRv2 vector (Addgene, 52,961). Lentiviral particles were produced by transfecting HEK293FT cells with psPAX2 (Addgene, 12,260) and pMDG.2 (Addgene, 12259) using EZ trans (Life-iLab, AC04L091). SKOV3 cells were infected for 24 h and selected with puromycin (3 μg/mL) for stable knockout.

Colony formation

Cells were plated at a density of 500 cells per well in 12-well plates. Sodium oxamate (10 mM, Selleck, S6871) or Stiripentol (100 µM, Selleck, S5266) was added the following day. Colony formation was monitored daily for approximately 10 days. Once colonies exceeded 50 cells, the culture was halted. Cells were washed with PBS, fixed in 4% paraformaldehyde, stained with crystal violet, and rewashed before imaging and colony counting for analysis.

Wound healing assay

Cells were seeded in 12-well plates and cultured until they reached 80–90% confluence. A scratch was then made using a 200 µL pipette tip to create a consistent wound across the cell monolayer. Images were taken immediately at 0 h (10 × magnification) to capture the initial wound. Drug treatments were applied immediately after the scratch was made. After 24 h, images of the wound healing were taken. The migration percentage for each group was calculated by comparing the width of the scratch at 24 h to the initial width at 0 h within the same well, with 0 h defined as 0% healing. The residual wound width was normalized to the initial width at 0 h for each group, allowing for accurate comparison of migration rates between groups while accounting for initial wound size variability.

Transwell assay

Different cells and their control groups were resuspended in serum-free medium and counted (200 µL, 3 × 104 cells). The suspension was added to 8.0 µm pore membrane insert (Corning, 3422), with the lower chamber filled with complete medium containing 10% FBS. The cells were fixed with 4% paraformaldehyde after 24 h. Crystal violet staining was applied to both sides of the insert, and non-migrated cells on the upper membrane were wiped off. After drying, 10 × magnification images of the migrated cells were captured, followed by counting and statistical analysis.

Isolation of HPMCs

HPMCs were isolated from omental tissue obtained from healthy donors undergoing laparoscopic surgery, following the protocol by Cai et al. [29] In brief, approximately 10 cm2 of omental tissue was washed twice with PBS and cut into smaller Sects. (2–3 cm2). The sections were incubated in 0.25% trypsin at 37 °C for 30 min. The trypsinized solution containing HPMCs was centrifuged at 500 × g for 5 min, and the pellet was plated in DMEM with 10% FBS for culture.

Generation of OC organoids

Tumor tissue from OC patients was obtained during surgery, rinsed with DPBS, and trimmed to remove non-tumor components. The tissue was minced into fragments and digested in Tumor Tissue Digestion Solution (bioGenous, K601003) at 37 °C with vigorous shaking. When digestion was complete, the mixture was passed through a 100 µm strainer, and the cells were pelleted by centrifugation at 300 × g for 3 min. After red blood cell lysis, cell pellet was mixed with Matrigel (Mogengel, 082755) and plated into 24-well plates (50 µL per well). After 30 min of incubation, OC organoid medium (bioGenous, K2168) was added, the culture medium was refreshed every other day, and the organoid growth was monitored and recorded.

Organoid growth assay

Organoids were isolated and seeded into 96-well clear-bottom black plates (Greiner, 655090) at 100 organoids per well. After several days of growth, when organoids were nearing maturation, different concentrations of Triptolide (Selleck, S3604) were added. Organoid growth was monitored and recorded with 10 × images captured. At the experiment’s endpoint, Calcein-AM was added to stain viable organoids, and fluorescence images were taken after 30 min of staining. Organoid viability was further assessed using an ATP activity assay kit (AiMingMED, 100–347), with luminescence readings reflecting the total viability of all organoids in each well.

Lactate assay

Cell were collected and resuspended in lactate assay buffer. The suspension was quickly pipetted up and down, then centrifuged at maximum speed at 4 °C for 2–5 min to remove insoluble material. The supernatant was collected and kept on ice for further analysis. The remaining steps were performed following the manufacturer’s protocol (Abcam, ab65331). Absorbance at 450 nm was recorded, and lactate concentrations were calculated based on the standard curve.

Exosome isolation and characterization

SKOV3 cells were cultured in complete medium supplemented with exosome-depleted FBS (Umibio, UR50202) for 24 h. The collected supernatant was centrifuged at 2000 × g for 30 min to remove cells and debris. The supernatant was then mixed with reagents from the Total Exosome Isolation Kit (Invitrogen, 4478359) and incubated overnight. Exosomes were collected the next day after centrifugation.

For electron microscopy, extracted exosomes were applied onto copper grids, followed by staining with uranyl acetate for 1 min, and the grid was air-dried at RT for a few minutes. Imaging was performed using a transmission electron microscope at 100 kV. For Nanoparticle Tracking Analysis (NTA), exosomes were analyzed for particle size and concentration using a NanoFCM system. Western blot was performed using antibodies against exosome markers CD9 (Proteintech, 20597-1-AP), CD81 (Proteintech, 66866-1-Ig), and TSG101 (Proteintech, 28283-1-AP) to confirm the presence of exosomes.

Nuclear and cytoplasmic protein extraction and western blot

Cells were lysed and centrifuged at 10,000 × g for 10 min at 4 °C, and the supernatant was collected. Nuclear and cytoplasmic proteins were extracted using a commercial kit (Beyotime, P0028) to ensure efficient separation. Equal amounts of protein were separated by SDS-PAGE and transferred onto PVDF membranes. Membranes were blocked with quick-blocking solution (Beyotime, P0252) and incubated with primary antibodies including ITGA5 (Proteintech, 10569-1-AP), ITGB1 (Proteintech, 26918-1-AP), β-actin (Proteintech, 66009-1-Ig), Slug (Proteintech, 12129-1-AP), N-cadherin (Proteintech, 22018-1-AP), LDHA (Proteintech, 21799-1-AP), and YAP1 (Proteintech, 13584-1-AP). Membranes were incubated with HRP-secondary antibodies the following day and visualized. For organoid total protein extraction, the procedure was similar, with lysis performed for 30 min and pipetting every 5 min to ensure thorough disruption.

Animal models

OC tissues obtained from surgical patients were cut into small pieces (2–3 mm3), with necrotic and non-tumor parts removed. The prepared tumor tissues were subcutaneously implanted into the right posterior axilla of 4–6 week-old female NSG mice (n = 3). Tumor growth was regularly monitored, and after two stable passages in NSG mice, the tumors were harvested and re-implanted into new mice for treatment with sodium oxamate and stiripentol. Tumor size was measured, and mice were randomly assigned to treatment groups and a control group. The treatment group received sodium oxamate (500 mg/kg) and stiripentol (200 mg/kg). Tumor growth was monitored and recorded, with volume calculated using the formula V = (L × W2) / 2. After two weeks, and tumors were extracted and weighed.

To establish the SKOV3 cell-derived subcutaneous xenograft model, SKOV3 sgCtrl and sgPPP2CA cells were implanted subcutaneously into 4–6 week-old female athymic nude mice (n = 5). Each mouse received 5 × 10⁶ cells mixed with matrigel in a 1:1 ratio. Tumor growth was recorded as previously described. The mice were closely observed for any health changes, and at the experimental endpoint, tumors were excised and weighed. Tumor samples were harvested for immunofluorescence, immunohistochemistry, and HE staining. A similar procedure was followed for shCtrl and shITGA5/ITGB1 cell groups.

To further investigate the effect of Triptolide on PPP2CA-knock out tumors, 20 nude mice were randomly divided into two groups (n = 10) and subcutaneously implanted with sgCtrl and sgPPP2CA cells, Following implantation, each group was further divided into a control group and a Triptolide treatment group (0.3 mg/kg), Triptolide was administered via intraperitoneal injection every other day for two weeks. At the experimental endpoint, mice were euthanized and tumors were excised.

Co-culture experiment

For the indirect co-culture experiment, mesothelial cells were seeded at 2 × 105 cells/2.6 mL in a six-well plate. Different cells were seeded at 1 × 105 cells/1.5 mL per well into Millicell cell culture inserts (Millipore, PTHT06H48) with 0.4 μm pore size. After 48 h of co-culture, HPMC were counted and resuspended for transwell migration assay. The number of migrating cells was recorded and representative images were captured after 24 h, as previously described.

For the direct co-culture with exosomes, mesothelial cells were adjusted to 2 × 105 cells/mL in serum-free medium. Then, 3 mg of exosomes were mixed with 200 μL of the mesothelial cell suspension and added to an transwell insert. The number of migrated cells was recorded after 24 h, and representative images were captured.

Tissue and organoid immunofluorescence

Tumor tissue sections from mice were subjected to antigen retrieval to expose the epitopes. Subsequently, 3% BSA was applied to block nonspecific binding sites at RT for 30 min, after which the primary antibody YAP1 (Proteintech, 13,584–1-AP) was added. Sections were incubated with a fluorescently labeled secondary antibody at RT, followed by nuclear staining with DAPI. YAP expression and localization were assessed.

For organoid immunofluorescence, the process mirrors the above steps. After incubation with the secondary antibody, Tyramide signal amplification (TSA) staining was applied using a fluorophore-labeled tyramide reagent. Once TSA staining was complete, blocking was performed before introducing new antibodies for subsequent labeling.

Hematoxylin–eosin (HE) and immunohistochemistry staining

For HE staining, tissue sections were deparaffinized in xylene and rehydrated through a graded ethanol series. After rehydration, sections were stained in hematoxylin for 5 min to visualize nuclei. The sections were then rinsed under running water and differentiated in acidic alcohol. Subsequently, they were counterstained with eosin for approximately 30 s to stain the cytoplasm and extracellular matrix. The sections were dehydrated in ethanol, cleared with xylene, then prepared for microscopic examination.

For immunohistochemistry, staining, tissue sections were incubated with primary antibodies against ITGA5, ITGB1, and Ki67. DAB was used for color development, with staining monitored microscopically. After rinsing with water to remove excess DAB, the sections were counterstained with hematoxylin for 30 s, dehydrated, and mounted.

RNA sequencing and data analysis

RNA from tumor cells was extracted using the Rneasy Mini Kit (QIAGEN, 74104), RNA concentration and quality were assessed respectively. Library preparation was performed for paired-end multiplexed sequencing on the Illumina platform. Sequencing was carried out on the Illumina Novaseq 6000, generating 150 bp paired-end reads. Differentially expressed genes were identified based on a p-value < 0.05 using the limma package. Raw sequence data are accessible in the Genome Sequence Archive (GSA: CRA015120) via the National Genomics Data Center at https://ngdc.cncb.ac.cn/gsa. Gene set annotation was sourced from MsigDB, and pathway enrichment was performed using the fgsea package (version 1.30.0).

Survival analysis

Glycolysis scores were first calculated for each patient in the TCGA-OV dataset by aggregating the expression values of genes included in the KEGG “Glycolysis / Gluconeogenesis” pathway (hsa00010), as defined in the Molecular Signatures Database (MSigDB; https://www.gsea-msigdb.org/gsea/msigdb/collections.jsp). The full list of genes used for scoring is provided in Supplementary File 1. This was performed using the gsva function from the GSVA R package, with the parameters set to method = ”gsva” and kcdf = ”Gaussian”. Based on these glycolysis scores, the surv_cutpoint function from the survminer package was used to determine the optimal cutoff value for overall survival. Patients with scores above this cutoff were classified into the “high glycolysis” group, and those below into the “low glycolysis” group. Kaplan–Meier survival analysis and the Log-rank test were then used to compare survival differences between the two groups.

Statistical analysis and data availability

All statistical analyses were conducted using GraphPad Prism 9. For comparisons between two groups, Student’s t-test was used. A p-value of less than 0.05 was considered statistically significant. Data are presented as mean ± SEM. RNA expression data for OC cell lines are available from the Broad Institute’s DepMap portal (DepMap Cancer Cell Line Encyclopedia (CCLE) Public 23Q2; https://depmap.org/portal/download/). Public RNA sequencing datasets of metastatic OC patients (GSE218939, GSE137237 and GSE98281) are available at GEO database.