This retrospective cohort study was conducted at the First Affiliated Hospital of Chongqing Medical University. The specimens were derived from 143 gynecological patients who underwent oophorectomy at the First Affiliated Hospital of Chongqing Medical University between August 2023 and January 2025. Clinical information and baseline serum HE4 levels were collected from the medical record system. The inclusion criteria were as follows: (1) Diagnosed with gynecological diseases requiring surgical intervention; (2) preoperative blood sampling for serum HE4 levels; (3) signing an informed consent form for postoperative tissue stiffness measurements, pathological examination and consent to use the relevant information in this study. Exclusion criteria were applied as follows (n = 17): (1) ovarian germ cell tumors, sex cord-stromal tumor, carcinosarcomas and metastatic tumors of the ovary; (2) patients with renal dysfunction causing elevated levels of HE4; (3) specimens that did not meet the requirements for hardness measurements (radius r < 30 mm, thickness h < 12 mm). All patients provided their written informed consent. A total of 126 patients were included and stratified by histopathological diagnosis into three groups: normal ovarian controls (n = 26; excised during surgery for uterine fibroids or cervical carcinoma), benign ovarian disease (n = 30), and EOC cases (n = 70). The EOC group included 46 cases of HGSOC and 24 cases of other types of EOC. Based on the average tissue stiffness values, the HGSOC cohort was divided into low- (n = 24) and high-stiffness (n = 22) groups.

Fagotti scores were assessed during diagnostic laparoscopy to evaluate intra-abdominal tumor dissemination. The scoring system evaluates seven parameters, including omental cake, diaphragmatic carcinomatosis, mesenteric retraction, bowel and stomach infiltration, liver surface metastases, and peritoneal carcinomatosis, with a total score ranging from 0 to 14, as described by Fagotti et al. [20].

The mechanical properties of ex vivo specimens were quantified using a digital Shore durometer (Model LX-C; San liang Measuring Tools Co., Ltd., Dongguan, China). This hardness tester employs a 0–100 HC measurement range with 0.5 HC resolution, specifically calibrated for evaluating viscoelastic materials including biological soft tissues. Specimens were subjected to hardness testing within 30 min post-excision in operation room. Using a 2.5 mm diameter spherical indenter, five equidistant measurement points were selected on the intact external surface. Hardness counts were recorded at five different locations and the mean value was calculated as the representative hardness parameter for each specimen.

Normal omental fibroblasts (NOFs) were isolated from female patients (n = 10) undergoing elective laparoscopic surgery for non-cancerous gynecological conditions. After removal of blood vessels and adipose tissue with ophthalmic scissors and forceps, tissue fragments (1 mm³) were placed on 10 cm cell culture dishes and incubated at 37°C in a humidified atmosphere (5%, CO₂). Characterization of NOFs: NOFs were distinguished by high expression of vimentin and lack of the epithelial marker cytokeratin 7 (CK7).

All procedures were conducted in accordance with the World Medical Association Declaration of Helsinki (2023 edition) for ethical conduct in human research. The research protocol was approved by the Ethics Committee of the First Affiliated Hospital of Chongqing Medical University (Approval No.: k2023-435).

Human high-grade serous adenocarcinoma OVCAR3 and OVCAR8 cells were cultured in Roswell Park Memorial Institute (RPMI) 1640 (Sigma, R6504, Saint Louis, MI, USA) medium containing 10% fetal bovine serum (FBS) (Pricella, 164210-50, Wuhan, China) and 1% penicillin and streptomycin (Pricella, C0222). 293 T cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) (Pricella, PM150210) containing 10% FBS and 1% penicillin and streptomycin. NOFs were cultured in Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12 (DMEM-F12) (Pricella, PM150312) containing 12% FBS and 1% penicillin and streptomycin. In some experiments, NOFs were cultured with rHE4 (MCE, HY-P71432, Monmouth Junction, NJ, USA)-DMEM-F12 mixture (3 μg/mL) or conditioned medium from HE4-knockdown OVCAR3/OVCAR8 cells. Induced CAFs were generated by coculture of NOFs with ovarian cancer cells. Cells were cultured in a 37°C and 5% CO₂ incubator. All cell lines were authenticated by tandem repeat sequence analysis and routinely tested Mycoplasma-negative.

The shRNA target sequences, sh-HE4-1: CCGGGGCTGACCAGAACTGCACGCAACTCGAGTTGC- GTGCAGTTCTGGGTCAGCTTTTTT, sh-HE4-2: CCGGTTCGGGCTTCACCCTAGTCTCACTCG- AGTGAGACTAGGTGTGAAGCCGAGAAGTTTTT and sh-HE4-Negative Control (NC): CCGG-GGTTCTCCGAACGTGTCACGT-CTCGAG-ACGTGACACGTTCGGAGAACC-TTTTTG, which were designed by Tsingke Biotechnology Co., Ltd. (Xi’an, China), and made to clone them into the PLKO.1 backbone. 293 T cells were inoculated at 1 × 105 per well in 6-well plates in DMEM medium containing 10% FBS for lentiviral preparation. After 12 h, 2 μg of expression vector, 1.25 μg of pSPAX2, and 0.75 μg of pLP/VSV-G packaging vector were used to transfect the cells. Viral supernatants were collected at 48 and 72 h post-transfection, filtered through a 0.22 μm filter (Nest, 331011, Wuxi, China) and added to OVCAR3 and OVCAR8 cell culture dishes for 72 h, and then screened with puromycin (5 μg/mL). As a control, lentivirus containing a random sequence shRNA was used. The transfection efficiency was verified by western blot analysis, and the shRNA knockdown cells obtained were screened for subsequent experiments.

Total Protein Extraction: Total protein was extracted from tumor tissues and cultured cells (OVCAR3/8 and primary fibroblasts). Approximately 30 mg of frozen tumor tissue was homogenized in RIPA lysis buffer (Beyotime, P0013B, Shanghai, China) containing 1% PMSF (Beyotime, ST506) at a 1:10 (w/v) tissue-to-buffer ratio, with three 2 mm zirconia beads, using a bead mill (JINGXIN, JXFSTPRP-CL, Shanghai, China) at 80 Hz for 10 cycles (5 s on/5 s off, 4°C). Cells at 80%–90% confluence were washed with 4–5 mL PBS (0.01 M, pH: 7.4) and lysed in the same buffer. All lysates were incubated on ice for 15 min, ultrasonicated (SCIENTZ, Scientz-IID, Zhejiang, China; 80 W, 12 cycles, 3 s on/3 s off), and incubated for another 15 min on ice. Samples were then centrifuged at 12,000× g for 15 min at 4°C (Thermo Fisher, ST 16R, Bremen, Germany). Supernatants were collected for Bicinchoninic Acid Assay (BCA) protein quantification. The remaining supernatants were mixed with 5× loading buffer (Beyotime, P0010) at 1:4 ratio, denatured at 100°C for 5 min, and stored at −20°C.

Protein Quantification: Protein concentrations were determined using the BCA assay kit (Beyotime, P0010). Samples were incubated with working reagent (37°C, 30 min) and measured at 562 nm (Tecan, Infinite M PLEX, Männedorf, Swiss).

Total protein lysates underwent separation via sodium dodecyl sulfate polyacrylamide gel electrophoresis. In brief, protein samples were electrophoresed at 80–120 V under a constant electrical current for 60–120 min. Subsequently, the proteins were transferred to a polyvinylidene fluoride membrane (Millipore, Burlington, MA, USA) and blocked with 5% skimmed milk at 37°C for 2 h. The membrane was then incubated overnight at 4°C with primary antibodies HE4 (1:1000; Proteintech, 66557-1-Ig, Wuhan, China), α-SMA (1:1000; Proteintech, 14395-1-Ig), FAP (1:800; HUABIO, ET1704-23, Hangzhou, China), Collagen I (1:800; Proteintech, 67288-1-Ig), Vimentin (1:2000; HUABIO, ET1610-39) and GAPDH (1:10,000; Proteintech, 60004-1-Ig). Membranes were washed with TBST(20 mM Tris, 137 mM NaCl, 0.1% Tween-20, pH 7.4) buffer, prepared by adding 0.1% (v/v) Tween-20 (BioFroxx, 1247ML100, Hesse, Germany) to 1× TBS solution reconstituted from powder (Servicebio, G0001-15, Wuhan, China) and incubated with HRP-conjugated secondary antibodies (1:10,000; Boster, Wuhan, China) at 37°C for 1 h. Signals were detected using ECL luminescence (KeyGEN BioTECH, KGC4602-100, Nanjing, China) and exposed in a Bio-Rad imaging system (BIO-RAD, V3 Western Workflow, Hercules, CA, USA).

Cells (primary fibroblasts) adhered to slides were fixed with 4% paraformaldehyde (Biosharp, BL539A, Hefei, China), permeabilised with an immunostaining potent permeabilising solution (Beyotime, P0097). After blocking with goat serum albumin (Boster, AR0009) for 30 min at room temperature, then 1× PBS (pH: 7.4) diluted anti-α-SMA (1:500; Proteintech, 14395-1-AP), anti-cytokeratin7 (1:400; HUABIO, R1309-4), anti-Vimentin (1:800; HUABIO, EM0401) primary antibody mixture was added, and the cells were incubated overnight at 4°C. Cells were washed three times with PBS. Cells were incubated with DyLight 594 (1:400; Abbkine, A23420, Wuhan, China) and DyLight 488 (1:400; Abbkine, A23210) for 1 h at room temperature, and the nuclei were stained with DAPI (Boster, AR1176) for 5 min and washed three times with PBS. The slices were sealed with a droplet of anti-fluorescence quenching sealer (Beyotime, P0126), and images were acquired using a confocal microscope (Zeiss, LSM 800, Oberkochen, Germany).

5 μm Formalin Fixed Paraffin Embedded (FFPE) sections were deparaffinized in xylene (undiluted, 100%) and rehydrated through graded ethanol. Antigen retrieval was performed using EDTA antigen repair buffer (pH 9.0, Servicebio, G1203) at 95°C for 20 min. Sections were treated with 3% H₂O₂ for 20 min, blocked with goat serum for 30 min, and then incubated with anti-HE4 (1:800; Proteintech, 66557-1-Ig), anti-α-SMA (1:800; Proteintech, 14395-1-AP), anti-FAP (1:200; HUABIO, ET1704-23), anti-Collagen I (1:1000; Proteintech, 67288-1-Ig) antibodies at 4°C overnight. The following day, the slices were incubated with ready-to-use horseradish peroxidase (HRP)-conjugated goat anti-rabbit/mouse secondary antibodies (undiluted, as per manufacturer’s protocol; Zhongshan Goldenbridge, PV9000, Beijing, China) for 30 min at room temperature, and stained using DAB (Zhongshan Goldenbridge, ZLI-9017). Nuclei were stained blue using haematoxylin (ready-to-use, Zhongshan Goldenbridge, ZLI-9610). The sections were then dehydrated, cleared with xylene (undiluted, 100%), and mounted. Images were acquired using a digital histopathology system (Zeiss, Axio Imager A2, Oberkochen, Germany).

96-well plates were coated with 2% BSA solution prepared from BSA powder (Solarbio, A8010, Beijing, China) dissolved in PBS (0.01 M, pH: 7.4, Servicebio, G4202-500 mL) and filter-sterilized through a 0.22 μm filter (Nest, 331011). Cells (primary fibroblasts) were dissociated with 0.25% trypsin (Servicebio, G4001-100 mL), diluted to 1 × 10⁶ cells/mL in growth medium. 100 μL of the diluted cells were combined with 100 μL of rat tail collagen (Corning, 354236, Corning, NY, USA) working solution. This mixture was then transferred to the prepared 96-well plates. Following a 30-min incubation period, the collagen gel was detached from the well edges using a pipette tip, and 200 μL of cell culture medium was added. Gel contraction was assessed via photography at 48 h, and collagen area was quantified using ImageJ software (National Institutes of Health; version 1.53t, Bethesda, MD, USA).

Ten 5-week-old female Balb/c nude mice (Vital River, Beijing, China) were randomly divided into two groups (n = 5 per group) and subcutaneously injected with either OVCAR3-HE4-MOCK or OVCAR3-HE4-knockdown cells (5 × 10⁶ cells/100 μL PBS (0.01 M, pH: 7.4)). Randomization was performed using a computer-generated random number sequence to minimize allocation bias. The sample size was calculated using G*Power software (version 3.1; Heinrich-Heine-Universität Düsseldorf, Düsseldorf, Germany) based on a medium effect size (Cohen’s d = 0.8), 80% power, and a significance level of 0.05, resulting in a minimum of 5 mice per group. Tumor growth was monitored every two days by caliper measurements, and body weight was recorded concurrently. Tumor volume was calculated as V=ab2/2, where a is the maximum and b the minimum tumor diameter. Tumors became palpable by day 7 post-injection. Formalin-fixed tumors were paraffin-embedded, sectioned at 5 μm, and subjected to immunohistochemical staining. All animal procedures were approved by the Animal Care and Use Committee of Chongqing Medical University (Approval No.: IACUC-CQMU-2025-0212).

All statistical analyses and data visualizations, including violin plots, were performed using GraphPad Prism 8.0 (GraphPad Software, San Diego, CA, USA). Categorical variables were expressed as frequencies and percentages (%), and differences between groups were assessed using Pearson’s chi-square test or Fisher’s exact test depending on the expected frequencies (Pearson’s test when all expected frequencies ≥5; Fisher’s exact test when any expected frequency <5). Data normality was evaluated using the Shapiro-Wilk test (α = 0.05), and homogeneity of variances was assessed by Levene’s test. Normally distributed data with equal variances were presented as mean ± standard deviation (SD), and comparisons between two groups were performed using Student’s t-test; for three or more groups, one-way analysis of variance (ANOVA) followed by appropriate post hoc multiple comparison tests (e.g., Bonferroni or Tukey’s test) was used. For non-normally distributed data, values were expressed as median and interquartile range (IQR), and group differences were analyzed using the Mann-Whitney U-test (two groups) or Kruskal-Wallis H-test (more than two groups). Correlations between ordered variables were assessed using Spearman’s rank correlation coefficient. All statistical tests were two-sided, and p-values < 0.05 were considered statistically significant unless adjusted for multiple comparisons. All experiments were performed independently at least three times.

Previous studies have demonstrated that tumor tissues exhibit higher stiffness than normal tissues [21]. To investigate the relationship between tissue stiffness and ovarian pathology, we collected clinical data from 126 patients and classified them into three groups based on histopathological diagnosis: normal ovary (n = 26), benign ovarian disease (n = 30), and EOC (n = 70). Tissue stiffness was measured using a durometer. Baseline characteristics of the patients are summarized in Table 1, and detailed characteristics of EOC subtypes are provided in Table S1. Tissue stiffness was significantly higher in EOC tissues than in both normal and benign groups, with no significant difference between the normal and benign groups (Fig. 1A). Notably, the stiffness difference between HGSOC and the normal/benign groups was more pronounced than that observed in the overall EOC group (Fig. 1B).

Some studies have reported an association between HE4 and fibrotic progression in various diseases [22–24]. Quantitative analysis confirmed that serum HE4 levels were positively correlated with tissue stiffness in EOC, with a notably stronger association observed in HGSOC (Fig. 1C,D). This finding is consistent with the characteristic desmoplastic microenvironment of HGSOC. However, the correlation remains in a preliminary exploratory phase, and the relatively low r values may be affected by various factors, including sample size and experimental variability. These findings suggest that HE4 may serve as a biomarker for increased tissue stiffness in ovarian cancer and may contribute to the development of the mesenchymal phenotype in HGSOC.

To further validate the association between HE4 expression and tissue stiffness in HGSOC, a cohort of 46 patients pathologically diagnosed with HGSOC was analyzed. Patients were classified into high-stiffness (n = 24) and low-stiffness (n = 22) groups based on the average tissue stiffness. Baseline clinical characteristics are presented in Table 2. It was found that patients in the high-stiffness group exhibited significantly higher Fagotti scores.

To investigate the potential mechanisms underlying increased tissue stiffness, the expression levels of HE4, fibroblast activation markers, and type I collagen were assessed. Elevated serum HE4 levels were observed in the high-stiffness group (Fig. 2A), and increased HE4 expression in tumor tissues was confirmed by Western blot and immunohistochemistry, with stronger expression localized to tumor epithelial cells (Fig. 2B,C).

Activated CAFs, particularly myofibroblasts, have been recognized as the major ECM-producing cells in the tumor microenvironment [25]. In the high-stiffness group, elevated expression of α-SMA and FAP was detected (Fig. 2D,E), predominantly distributed in the tumor stroma. Moreover, a marked increase in type I collagen deposition was identified in high-stiffness tumor tissues (Fig. 2E). Collectively, these findings indicate that tissue stiffness may reflect the activation state of fibroblasts and the level of ECM production. The activation of CAFs and the ECM deposition are likely to be key driving factors of increased stiffness in ovarian cancer tissues.

To determine whether HE4 directly promotes the transdifferentiation of normal fibroblasts into myofibroblasts, primary normal fibroblasts were isolated and characterized (Fig. 3A). Western blot analysis revealed that treatment with rhHE4 (3 μg/mL) significantly upregulated the expression of myofibroblast markers α-SMA and FAP compared to the control group treated with BSA (3 μg/mL) (Fig. 3B). This finding was further confirmed by immunofluorescence, which demonstrated that rhHE4 induced the transdifferentiation of normal fibroblasts into α-SMA–positive myofibroblasts (Fig. 3C).

Notably, the collagen gel contraction assay—a classical 3D model for evaluating fibroblast-mediated matrix remodeling—showed that fibroblasts treated with rhHE4 exhibited significantly enhanced contractile ability (Fig. 3D). In addition, expression of type I collagen, a key component of ECM, was markedly increased following rHE4 treatment (Fig. 3E).

The effects of ovarian cancer cell–derived HE4 on fibroblast function were assessed through indirect co-culture, Western blotting, immunofluorescence, and collagen gel contraction assays. The WFDC2 gene, which encodes HE4, was silenced in the ovarian cancer cell lines OVCAR3 and OVCAR8 using shRNA (Fig. 4A). Conditioned media from the HE4-knockdown group (sh-HE4) and the negative control group (NC) were collected and indirectly co-cultured with normal fibroblasts.

Western blot analysis revealed that α-SMA and FAP expression levels were decreased in fibroblasts following co-culture with sh-HE4 cells (Fig. 4B). Immunofluorescence further demonstrated a reduction in α-SMA fluorescence intensity (Fig. 4C). Moreover, the collagen gel contraction capacity of fibroblasts was significantly reduced after treatment with sh-HE4 conditioned media (Fig. 4D). In addition, type I collagen expression in fibroblasts was inhibited by the sh-HE4 conditioned media (Fig. 4E).

These findings indicate that HE4 knockdown suppresses the ability of ovarian cancer cells to induce the transdifferentiation of normal fibroblasts into myofibroblastic cancer-associated fibroblasts (myCAFs).

To further investigate the potential role of HE4 in vivo, subcutaneous xenograft tumor experiments were performed using OVCAR3 ovarian cancer cells with stable knockdown of WFDC2. Tumor volume and weight were found to be significantly reduced in the sh-HE4 group compared to controls (Fig. 5A–C), consistent with previous reports indicating that HE4 promotes ovarian cancer cell proliferation. Immunohistochemical analysis confirmed that HE4 expression was downregulated in the sh-HE4 group (Fig. 5D).

Meanwhile, the expression of fibroblast activation markers α-SMA and FAP was markedly decreased in the knockdown group, accompanied by a reduction in type I collagen deposition (Fig. 5E). These findings further support the critical role of HE4 in promoting the transdifferentiation of fibroblasts into cancer-associated fibroblasts in ovarian cancer.