The transcript expression RNAseq data in the GTEx (https://gtexportal.org/home/) and the TCGA (https://portal.gdc.cancer.gov/) were retrieved from the “TCGA TARGET GTEx” study of UCSC Xena Toil RNA-Seq Recompute Compendium (https://xenabrowser.net/) (accessed on March 20, 2025) [14]. The transcript expression RSEM TPM dataset and OS status and time of patients with cancer were downloaded from UCSC Xena, containing 10,535 samples from TCGA. In addition, the phenotype data of OV were also downloaded from UCSC Xena. After the screen criteria, 352 OV samples were obtained to further analyses. The STAT1 and BRCA1/2 mutation profiles of pan-cancer samples interrogated via the cBioPortal platform (https://www.cbioportal.org/).
This study collected the ovarian tissues of 46 OV patients and 43 benign gynecological diseases who underwent ovariectomy in the First Affiliated Hospital of Xi’an Jiaotong University from 2021 to 2022 (the detailed clinicopathological features of the OV patients are shown in Supplementary Table S1, The cohort contain 76.7% Stage III/IV, 93.0% high-grade serous carcinoma subtype.). This study was approved by the ethics committee of the First Affiliated Hospital of Xi’an Jiaotong University (approval number: 2020-G143). All selected patients signed informed consent, and the study was conducted in accordance with the declaration of Helsinki.
Cell cultureThe human OV cell lines (including A2780, ES2, OVCAR8, OVCAR3, HEY, HO8910, SKOV3, CAOV3) and the normal ovarian epithelial cell line IOSE80 were provided by the Precision Medicine Center at the First Affiliated Hospital of Xi’an Jiaotong University. The cell lines were cultured in Dulbecco’s modified Eagle’s medium (DMEM) or RPMI-1640 medium containing 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. All of the cell lines were grown in an incubator at 37 °C with 5% CO2.
Western-blotWestern-blot analysis was conducted by extracting total protein from cells using a Radio Immunoprecipitation Assay (RIPA) lysis buffer (AccuRef Scientific, Xi’an, China), supplemented with protease inhibitors. Following protein quantification via nanodrop, an equal amount of protein was subjected to SDS-PAGE for electrophoretic separation. Subsequently, the resolved protein bands were transferred onto a polyvinylidene fluoride (PVDF) membrane (Millipore, Massachusetts, USA) employing a semi-dry transfer technique. The membrane was then blocked with 5% skimmed milk (Yili Industrial Group Co., Ltd., China) before incubation with primary antibodies, followed by secondary antibodies conjugated to horseradish peroxidase. The visualization of protein bands was achieved using ECL chemiluminescent reagents (Biosharp, Hefei China). The primary antibodies utilized in this experiment included anti-STAT1 (1:1000, Proteintech, Wuhan, China), anti-phospho-STAT1-Y701 (1:1000, ABclonal, Wuhan, China), anti-phospho-STAT1-S727 (1:1000, Abways, Shanghai, China) and anti-GAPDH (1:5000; Servicebio, Wuhan, China). The STAT1 antibody recognizes an immunogen domain spanning amino acids 2-230 and is capable of detecting both α and β isoforms.
Immunohistochemistry (IHC) and multiplex immunofluorescence stainingFor immunohistochemistry (IHC) staining, paraffin-embedded 4-µm-thick sections were deparaffinized, rehydrated in graded alcohol, blocked by endogenous peroxidase blocking solution and subjected to antigen retrieval using 10 mM citrate buffer. TMA slides were then blocked with goat serum for 2 h at 37°C and incubated overnight with primary antibody: STAT1 (1:200, Proteintech, Wuhan, China) and CD8a (1:150, Servicebio, Wuhan, China) at 4°C. Subsequently, slides were incubated with a horseradish peroxidase (HRP)-conjugated secondary antibody, followed by stain using 3,3’-diaminobenzidine solution. Finally, slides were counterstained with hematoxylin and mounted with neutral balsam. The IHC staining score of each sample was calculated as multiplying the staining intensity by the proportion of the positive staining cells [15, 16]. The staining intensity was graded as follows: 0 (negative), 1 (weak), 2 (moderate) and 3 (strong), and the proportion of the positive staining cancer cells was scored as follows: 1 (1–25%), 2 (26–50%), 3 (51–75%) and 4 (76–100%). The scores of each tumor sample were multiplied to give a final score of 0–12 [16], and the tumors were finally determined according to their STAT1 expression as low expression, score < 6, and positive expression, score ≥ 6. Two pathologists, without prior knowledge of the clinical data, independently graded the staining intensity in all cases. For each tumor specimen, four representative 20× microscopic fields were analyzed, systematically covering both the tumor center and invasive front.
Multiplex immunofluorescence staining was performed using Tyramide Signal Amplification (TSA) on formalin-fixed paraffin-embedded tissues. After deparaffinization, antigen retrieval was conducted under microwave heating in EDTA buffer (pH 8.0), followed by endogenous peroxidase blocking with 3% H2O2. Four iterative staining cycles were performed, each consisting of serum blocking (3% BSA or 10% rabbit serum depending on primary antibody species), overnight primary antibody incubation, HRP-conjugated secondary antibody treatment, TSA-based fluorescent tyramide labeling (utilizing different fluorophores: iF440, iF488, iF555, and iF647; 10 min/dark), and antibody stripping to remove bound antibodies. After the cycles, nuclei were stained with DAPI, autofluorescence was quenched, and slides were mounted for imaging. The slides were imaged using a scanner, and the acquired image data were analyzed using CaseViewer (v2.4, 3DHISTECH) software.
Survival analysis and cox regression analysisThe “surv_cutpoint” function from the “survminer” R software package was utilized to ascertain the optimal cut-off values for the variables of interest. Based on these values, patients were stratified into high and low expression groups, and a Kaplan-Meier curve was plotted to compare survival rates between the two expression levels. Univariate and multivariate Cox proportional hazards regression models were conducted using the “coxph” function from the “survival” package to assess the potential of different STAT1 transcript expression levels as independent prognostic biomarkers in OV patients. The multivariate analysis incorporated factors such as age, stage, tumor grade, and transcriptional expression levels. Subsequently, a forest plot generated by the “forestplot” R software package clearly depicted the p-values, hazard ratios (HRs), and 95% confidence intervals (CIs) for each variable considered.
Exploration of TIDE signaturesBy utilizing the Tumor Immune Dysfunction and Exclusion (TIDE, http://tide.dfci.harvard.edu/) database [17], we evaluated the TIDE score, dysfunctional cytotoxic T cells score, exclusion cytotoxic T cells score, interferon gamma (IFNG) score, M2 subtype of tumor-associated macrophages (TAMs) score, cancer-associated fibroblasts (CAFs) score, and myeloid-derived suppressor cell (MDSC) infiltrations score, CD274 (PD-L1), CD8, and microsatellite instability (MSI) score. These indicators collectively reflect the anti-tumor and tumor immune escape capabilities of each OV sample.
Correlation analysis and functional enrichment analysisSpearman correlation analysis was used to determine the correlation between the expression levels of different transcripts and various immune prediction scores of TIDE. The top 100 genes most significantly correlated with STAT1 (ranked by descending order of absolute Pearson’s correlation coefficient, requiring p < 0.05 and |r| >0.55) were selected for functional enrichment analysis. The “clusteprofiler” package in R was used to check the cell component (CC), molecular function (MF) and biological process (BP) categories in the gene ontology (go). The “simplification” function in the R package is used to reduce redundancy in the output of rich go terms.
Statistical testsData analysis in the study was conducted primarily using R language (version 4.2.2). The “ggplot2” package facilitated figure creation. The Wilcoxon test estimated statistical significance for quantitative data comparisons between two types. Kaplan-Meier survival analysis employed the log-rank method to determine statistical significance. Values of p < 0.05 were deemed statistically significant, with significance levels denoted as *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.
To investigate whether various immune scores could serve as mediators explaining the relationship between independent variables (transcript expression levels of STAT1 α/β subtypes) and dependent variables (patient survival outcomes), we conducted causal mediation analysis using the “mediation” package in R. Specifically, under the counterfactual framework of causal mediation analysis, the total effect of exposure variables (different transcript expression levels) on the outcome variable (overall survival rate, OS) was decomposed into direct effects and indirect effects mediated through different scores. All mediation analyses employed a quasi-Bayesian Monte Carlo method with 1,000 simulations based on normal approximations. This method estimates uncertainty in mediation effects through repeated simulations, accounting for potential confounding. Results present the effect magnitude and p-values for average direct effects (ADE) and average causal mediation effects (ACME), along with the proportion of mediated effects.
Comments (0)