Endometrial cancer tissues were obtained from patients (n=62, aged 26-82 years old) who had undergone surgical treatment at the Department of Surgery, Kaohsiung Medical University Hospital [Kaohsiung, Taiwan]. All participants in this study were recruited between March 2025 and March 2017. Ethical approval [IRB No.: KMUHIRB-E (1)-20150026] was obtained from the Ethics Committee of Kaohsiung Medical University Hospital. Informed patient consent was waived by the Institutional Review Board due to the retrospective nature of the study.

All tissues were procured from formalin-fixed and paraffin-embedded endometrial tissue blocks. Immunohistochemical (IHC) staining for FH in endometrial tissues was performed using the Bond-Max system (Leica Microsystems GmbH). Sections were deparaffinized using Bond Dewax Solution (Leica Microsystems GmbH) and rehydrated using graded alcohol. Heat-induced antigen retrieval was achieved using Bond Epitope Retrieval Solution 1 (Leica Microsystems GmbH) for 20 min at 98°C. After washing steps, peroxidase blocking was carried out for 10 min using Bond Polymer (Leica Microsystems GmbH). Tissues were again washed and then incubated with the primary antibody, FH (cat. no. GTX110128; 1:100), for 30 min at room temperature. Post-primary IgG linker reagent was applied for 8 min, and the slides were incubated with polymeric horseradish peroxidase IgG reagent for 8 min to localize the primary antibodies. Diaminobenzidine (DAB) was used as the substrate to detect antigen-antibody binding. Then, hematoxylin was used to counterstain nuclei for 5 min at room temperature. Images of immunohistochemically stained sections were captured using Nikon Eclipse E600 fluorescence microscope (Nikon Corporation). Relative expression of FH in the endometrial cancer specimens was quantified by two pathologists independently. For the endometrial cancer samples, each specimen was assigned to one of four groups based on the percentage of positively stained normal and tumor cells: 0 (0–4%), 1 (5–24%), 2 (25–49%), 3 (50–74%) or 4 (75–100%). In addition, the immunostaining intensity was graded as: 0 (negative), 1 (weak), 2 (moderate) or 3 (strong), with the total score calculated by multiplying the percentage of positively stained cells by the graded intensity of staining for every sample. Patients with a score <4.50 were categorized as the low FH expression group and those with a score ≥4.50 were categorized as the high FH expression group.

Ishikawa, RL95-2, HEC1A, AN3CA and KLE human endometrial cancer cell lines were obtained from the Bioresource Collection and Research Center (Hsinchu, Taiwan) and cultured in RPMI 1640 (Gibco; Thermo Fisher Scientific, Inc.) and Dulbecco's modified Eagle's medium (Gibco; Thermo Fisher Scientific, Inc.). All cell lines were incubated in a humidified incubator at 37°C and 5% CO₂. All culture media contained 10% fetal bovine serum (FBS; Biological Industries; Sartorius AG), 1% penicillin G and streptomycin.

To knockdown FH expression in endometrial cancer cells, lentivirus carrying a pLKO.1_puro lentiviral vector (from National RNAi Core Facility, Academia Sinica, Taipei, Taiwan) that expressed double-stranded short hairpin (sh)RNA oligonucleotides targeting the sequences of human FH (three clones) were used: i) Clone 2: ID TRCN0000310398, target sequence: CAACGATCATGTTAATAAA, shRNA sequence: GCCCAACGATCATGTTAATAAACTCGAGTTTATTAACATGATCGTTGGGTTTTTG; ii) clone 5: ID TRCN0000052465, target sequence: CCCAACGATCATGTTAATAAA, shRNA sequence: CCGGCCCAACGATCATGTTAATAAACTCGAGTTTATTAACATGATCGTTGGGTTTTTG; and iii) clone 6 ID: TRCN0000299140, target sequence: GTGGTTATGTTCAACAAGTAA, shRNA sequence: CCGGGTGGTTATGTTCAACAAGTAACTCGAGTTACTTGTTGAACATAACCACTTTTTG (National RNAi Core Facility, Academia Sinica, Taipei, Taiwan).

A pLKO.1_puro lentiviral vector expressing shRNA targeting firefly luciferase, unrelated to the human genome sequence, was used as a negative control clone ID: TRCN0000052466, target sequence: GTGGTTATGTTCAACAAGTAA, shRNA sequence: GGGTGGTTATGTTCAACAAGTAACTCGAGTTACTTGTTGAACATAACCACTTTTTG (from National RNAi Core Facility, Academia Sinica, Taipei, Taiwan).

A ready-to-use lentivirus particle containing the pReceiver Lv105 lentiviral vector, which expressed the human FH gene, was purchased from GeneCopoeia, Inc. for overexpression of FH in endometrial cancer cells. Lentivirus particles containing an empty pReceiver Lv105 lentiviral vector (GeneCopoeia, Inc.) were used as a negative control.

Briefly, the cells were seeded at 5×10⁵ cells/well in 6 cm plates (Corning, Inc.) and incubated overnight at 37°C in 5% CO₂ atmosphere. Lentiviral infection was achieved by adding virus solution to cells in culture media containing 8 g/ml polybrene (TR-1003; Sigma-Aldrich; Merck KGaA). The number of viruses was added according to the recommended infection MOI for Ishikawa, RL95-2, KLE, and AN3CA cells (MOI=5). Following a 24 h incubation at 37°C in 5% CO₂ atmosphere, 2 g/ml puromycin (cat. no. A11138-03; Gibco; Thermo Fisher Scientific, Inc.) was added for selection. Selected cells were cultured in 2 g/ml puromycin for 2 weeks to establish cells with stable overexpression or knockdown of FH.

The cell proliferation rate was determined using a XTT colorimetric assay (Roche Diagnostics GmbH) following the detailed procedure in accordance with a previous report (13).

A human phosphokinase antibody array (ARY003C; R&D systems, Inc.) was applied to explore the kinases that are affected by FH in endometrial cancer cells. The site-specific phosphorylation of 43 kinases was determined in a single sample. Total cell lysates of RL95-2 cells with or without FH knockdown were harvested for phosphokinase array analysis according to the manufacturer's instruction.

FH and twist RNA Seq datasets of uterine corpus endometrial carcinoma (UCEC) in The Cancer Genome Atlas (TCGA) database were retrieved from TCGA website (Project ID: TCGA-UCEC; http://xena.ucsc.edu). The Pearson's correlation between interested proteins was analyzed using the data of patients from TCGA-UCEC.

FH, VIM and CLDN1 datasets of uterus in the Genotype-Tissue Expression (GTex) database were retrieved from the GTex website (https://gtexportal.org/home/). The correlation between FH and VIM and FH and CLDN1 was analyzed using the data of patients from GTex-Uterus.

Cell migration assays were performed in 24-well plates with Transwell (Corning Inc.) membrane filter inserts (6.5 mm diameter, 8 µm pore size). Endometrial cancer cells, after FH overexpression or knockdown, were trypsinized, suspended in serum-free medium and seeded (1×10⁵ cells) in the upper chamber of the Transwell filters. Medium containing 10% FBS was added to the lower chamber and the plates were incubated for 24 h at 37°C. Following incubation, cells were stained with crystal violet for 2 h at room temperature. Non-migrating cells were removed by wiping the upper surface of the filter. Migrated cells were imaged using an Olympus SZX10 stereo light microscope (Olympus Corporation) and analyzed using ImageJ software (ij153-win-java8; National Institutes of Health).

For invasion assays, BioCoat Matrigel (BD Biosciences) invasion chambers were rehydrated according to the manufacturer's instructions and subsequent steps were identical to the migration assay.

To study the effect of EGFR phosphorylation on endometrial cancer cell metastasis, Gefitinib, an EGFR phosphorylation inhibitor, was purchased from Sigma-Aldrich (Merck KGaA; cat. no. SML1657). RL95-2 endometrial cancer cells were treated with Gefitinib for 24 h before performing the migration and invasion assays.

Western blotting was performed to assess the knockdown efficiency following lentivirus infection and to assess the protein expression of other proteins. The detailed procedure according to a previous study was followed (13). In brief, the cells were lysed with RIPA buffer (20 mM Tris-HCL pH 7. 4, 150 mM NaCl, 1 mM EDTA, 1% Triton-X100, 1% sodium deoxycholate, 0.1% SDS) and cell lysates were collected. The BCA protein assay (cat. no. 23225; Sigma-Aldrich; Merck KGaA) was used to quantify total protein. The samples were electrophoresed on a SDS-PAGE gel (8–12%; 20 µg/lane). After protein transfer, the polyvinylidene fluoride (PVDF) membrane was blocked with 2% BSA in 1X TBST solution for 1 h at room temperature. The membranes were incubated overnight with primary antibodies at 4°C. Antibodies against FH (cat. no. GTX110128; 1:2,000), vimentin (cat. no. GTX100619; 1:1,000), twist (cat. no. GTX127310; 1:1,000), EGFR (cat. no. GTX100448; 1:1,000) and phosphorylated (p-)EGFR (cat. no. GTX133599; 1:1,000) were purchased from GeneTex Inc. Antibodies against JNK1/2/3 (cat. no. ab179461; 1:1,000) and p-JNK1/2/3 (cat. no. ab124956; 1:1,000) were purchased from Abcam. An antibody against β-actin (cat. no. A5441; 1:5,000; Sigma-Aldrich; Merck KGaA) was used as the internal control. After the incubation with primary antibodies and subsequent washing, PVDF membranes were incubated with rabbit (HRP conjugate; cat. no. GTX2131101; 1:5,000; GeneTex, Inc.) or mouse (HRP conjugate; cat. no. GTX213111; 1:5,000; GeneTex, Inc.) secondary antibodies for 1 h at room temperature. The protein bands on the PVDF membrane were visualized using enhanced chemiluminescence reagent (PerkinElmer, Inc.) and Image Lab software 6.0.1 (Bio-Rad Laboratories, Inc.).

All statistical analyses were performed using the SPSS 14.0 statistical package for PC (SPSS, Inc.). Comparisons between FH expression with various variables, including stage, tumor size, grade lymph node metastasis and myometrium invasion, were investigated by χ² test. Student's t-test (unpaired) was used to compare the difference between two groups. One-way analysis of variance with post-hoc Tukey's test was used for multiple group comparisons. P<0.05 was considered to indicate a statistically significant difference.

To evaluate FH expression in normal endometrium and endometrial cancer tissues, IHC staining of 59 normal endometrial and 62 endometrial cancer tissue samples was performed. It was demonstrated that endometrial cancer tissue had low FH expression compared with normal endometrial tissue (Fig. 1A and B). FH expression was further correlated with the clinicopathological characteristics of patients with endometrial cancer and it was found that FH had a negative correlation with tumor size (P=0.028) and lymph node metastasis (P=0.044; Table I). Moreover, the patients were classified into an FH-low group (<4.50%) and an FH-high group (≥4.50%; Table I) using receiver operating characteristic (ROC) curve.

The endogenous expression of FH was further assessed in five endometrial cancer cell lines. The result demonstrated that the HEC1A, Ishikawa and RL95-2 cell lines (which are relatively more invasive cell lines) had low FH protein expression, while the AN3CA and KLE cell lines (which are relatively less invasive cell lines) had high FH protein expressions (Fig. 2A). FH was subsequently recombinantly overexpressed in Ishikawa cells (which had low endogenous FH expression) and knocked down in RL95-2 cells (which had high endogenous FH expression), to evaluate the effect of FH expression on the proliferation, migration and invasion abilities of endometrial cancer cells. A total of three shRNAs for FH expression knockdown were assessed. It was discovered that KD6 had the best knockdown efficiency and KD6 was therefore used in further experiments.

The cell proliferation assay results revealed that the proliferation of FH-overexpressing Ishikawa cells was decreased, while the proliferation of FH-knockdown RL95-2 cells was increased (Fig. 2B and C). Furthermore, the migration and invasion abilities of FH-overexpressing Ishikawa cells were also decreased. By contrast, the migration and invasion abilities of FH-knockdown RL95-2 cells were increased. The migration ability of cells transfected with KD2 and KD5 clones was also assessed and it was found that cell migration was also increased significantly compared with the control group (Fig. S1B).

It is shown in Table I that FH expression was negatively correlated with lymph node metastasis. Therefore, the protein expression of various epithelial-mesenchymal transition (EMT) markers, which play critical role in cancer cell metastasis (14), in FH-knockdown RL95-2, KLE and AN3CA cells was further evaluated. It was demonstrated that expression of the mesenchymal markers, vimentin and twist, was upregulated in FH-knockdown cells; by contrast, expression of the epithelial marker, claudin 1, was downregulated significantly compared with the control group (Fig. 3A-C). To access whether FH mRNA expression is correlated with the mRNA expression of EMT markers, the expression levels in the TCGA (https://xena.ucsc.edu/) and GTex (https://gtexportal.org/home/datasets) datasets were analyzed. A negative correlation was observed between FH and two mesenchymal markers, vimentin (r=−0.33; P=0.003) and twist (r=−0.206; P=0.003), whereas a positive, but not statistically significant, correlation was observed between FH and the epithelial marker, claudin-1 (r=0.18, P=0.12), as shown in Fig. 3D and E.

Next, a human phosphokinase array was used to identify the possible kinases regulating FH-mediated endometrial cancer cell behavior. The results demonstrated that the levels of p-JNK1/2/3 and p-EGFR were increased in FH-knockdown RL95-2 cells compared with control cells (Figs. 4A and S2A), which was further validated by western blotting (Fig. 4B-D). The p-EGFR protein level was significantly increased in FH-knockdown endometrial cancer cell lines including RL95-2 (Fig. 4B), KLE (Fig. 4C) and AN3CA (Fig. 4D), while p-JNK1/2/3 expression did not change in FH-knockdown RL95-2 cells (Fig. S2B). In addition, a connection between FH and EGFR, mediated by TP53, was observed using the STRING online database (https://string-db.org/) (Fig. S3).

Further analysis was performed using the p-EGFR inhibitor, gefitinib (1 µM), to treat FH-knockdown RL95-2 cells. The results demonstrated that gefitinib inhibited p-EGFR protein expression (Fig. S2C), while the migration and invasion abilities of FH-knockdown cells were significantly decreased after gefitinib treatment, compared with the vehicle treatment group (Fig. 4C and D). These results suggested that FH knockdown promoted EGFR phosphorylation and hence upregulated the migration and invasion of endometrial cancer cells.