FOXM1 drives cisplatin resistance in ovarian cancer through STIM1-mediated SOCE and STAT1 activation

Abstract

Cisplatin resistance remains a major challenge in ovarian cancer, limiting treatment efficacy and reducing patient survival. While FOXM1 and calcium (Ca²⁺ signaling have been individually implicated in chemoresistance, how transcriptional regulation is linked to Ca²⁺-dependent survival signaling remains poorly understood. To address this gap, we integrated bioinformatic analyses with functional and mechanistic studies in cisplatin-resistant (CR) ovarian cancer models. We found that FOXM1 transcriptionally activate STIM1, thereby enhancing store-operated Ca²⁺ entry (SOCE) and sustaining cytosolic Ca²⁺ levels. Elevated Ca²⁺ influx-activated downstream interferon (IFN)/STAT1 signaling and increased expression of IFN-stimulated genes associated with DNA repair and anti-apoptotic programs. FOXM1 or STIM1 depletion and pharmacological SOCE inhibition attenuated STAT1 activation and restored cisplatin sensitivity, whereas STIM1 re-expression rescued the resistance phenotype following FOXM1 depletion. Using A2780/ACR and IGROV1/ICR ovarian cancer models, together with analyses of TCGA-OV and multiple GEO cohorts, we demonstrate that FOXM1 transcriptionally activates STIM1, thereby enhancing store-operated Ca²⁺ entry and promoting STAT1-dependent IFN signaling. Genetic or pharmacological disruption of this pathway restored cisplatin sensitivity, identifying the FOXM1-STIM1/SOCE-STAT1 axis as a potential therapeutic vulnerability in platinum-resistant ovarian cancer. This study provides mechanistic insight into chemoresistance and highlights avenues for therapeutic intervention in patients with high FOXM1/STIM1 expression.

Graphical abstract

FOXM1 upregulates STIM1 to enhance store-operated Ca²⁺ entry and activate STAT1-dependent interferon programs in cisplatin-resistant ovarian cancer cells. Disrupting SOCE or JAK/STAT signaling restores cisplatin sensitivity, identifying the FOXM1-STIM1/SOCE-STAT1 axis as a therapeutic vulnerability in platinum-resistant disease.

Introduction

Ovarian cancer remains one of the most lethal gynecological malignancies, characterized by frequent recurrence and poor long-term prognosis.^1,2 Because early symptoms are often nonspecific, most patients are diagnosed at advanced stages, which complicates treatment and contributes to disease relapse.³ Platinum-based chemotherapy remains the cornerstone of first-line treatment. However, platinum resistance is a major clinical challenge that limits therapeutic efficacy and reduces progression-free and overall survival.^4,5 Elucidating the molecular mechanisms underlying chemoresistance is essential for developing more effective treatment strategies.

FOXM1 is a key transcription factor that regulates cell-cycle progression, DNA replication, and DNA damage repair.^6,7 FOXM1 also maintains mitochondrial function and redox homeostasis, thereby supporting cell survival under stress.^8,9 Its overexpression has been widely implicated in chemoresistance across multiple cancer types by enhancing DNA repair, suppressing apoptosis, and maintaining stem-like properties.^10,11 In ovarian cancer, elevated FOXM1 expression correlates with poor prognosis and reduced sensitivity to platinum-based chemotherapy.¹² However, direct therapeutic targeting of FOXM1 remains challenging due to its complex structure, limited druggable pockets, and potential compensatory resistance mechanisms.¹³ Because FOXM1 controls broad oncogenic networks, investigating its downstream effectors may provide alternative strategies for overcoming resistance.¹⁰

Intracellular Ca²⁺ signaling has emerged as a critical regulator of cancer progression, influencing proliferation, invasion, migration, and cell survival.¹⁴ In cancer cells, Ca²⁺ signaling is frequently remodeled through altered expression or activity of channels, pumps, and exchangers, enabling crosstalk with oncogenic pathways and the tumor microenvironment.^15,16,17 In non-excitable cells, store-operated Ca²⁺ entry (SOCE) is a major pathway for maintaining Ca²⁺ homeostasis. Upon endoplasmic reticulum (ER) Ca²⁺ depletion, STIM1 senses the decrease in luminal Ca²⁺ and activates ORAI1 channels at the plasma membrane to mediate Ca²⁺ influx.¹⁸ Enhanced SOCE activity and upregulation of its core components have been reported in chemoresistant cancers and are associated with cell survival, DNA repair, and apoptosis resistance.^19,20 However, the upstream regulators responsible for elevated SOCE and the mechanisms by which SOCE supports platinum tolerance remain unclear.

Although FOXM1 and Ca²⁺ signaling are both important contributors to cancer progression, their functional relationship has rarely been explored. Emerging evidence suggests that FOXM1 localization and activity may be influenced by intracellular Ca²⁺,²¹ while FOXM1 also supports mitochondrial fitness and stem-like traits linked to Ca²⁺ homeostasis.^8,9 Moreover, FOXM1 and key SOCE components have each been reported to be upregulated in chemoresistant ovarian cancer cells,^12,20,22 raising the possibility of a functional interaction. The convergence of these observations prompted us to investigate whether FOXM1 and SOCE are functionally connected in the context of ovarian cancer chemoresistance. Despite accumulating evidence implicating both FOXM1 and SOCE in cancer progression and drug resistance, two major mechanistic gaps remain. First, previous studies have largely examined FOXM1 and SOCE independently, and whether FOXM1 directly transcriptionally regulates a core SOCE component has not been established. Second, although enhanced SOCE has been associated with chemoresistant phenotypes, the downstream signaling pathway through which SOCE promotes platinum tolerance remains incompletely defined. The present study addresses these gaps by investigating whether FOXM1 directly regulates STIM1 transcription and whether STIM1-mediated SOCE activates a Ca²⁺-dependent survival program that supports cisplatin resistance.

Here, we investigated the mechanistic relationship between FOXM1 and SOCE in ovarian cancer. By combining bioinformatic analyses of patient datasets with functional studies in CR models, we evaluated the regulation of STIM1 by FOXM1, the contribution of STIM1-mediated SOCE, and the involvement of downstream IFN/STAT1 signaling in cell survival under chemotherapeutic stress. Our findings identify a novel FOXM1-STIM1/SOCE-STAT1 signaling axis that contributes to cisplatin resistance in ovarian cancer and may provide potential targets for therapeutic intervention.

Results

Identification of FOXM1 and STIM1 as key correlates of chemoresistance in ovarian cancer via integrated bioinformatic analysis

Although elevated FOXM1 expression and enhanced SOCE activity have each been implicated in chemoresistance, their potential functional relationship remains unclear. Recent studies have begun to define SOCE-related gene signatures based on the core components including STIM1/2 and ORAI1/2/3.²³ Building upon this, we first performed correlation analysis in the TCGA ovarian cancer (TCGA-OV) cohort (n = 427) using the GEPIA3 platform to investigate the relationship between FOXM1 and SOCE in ovarian cancer.²⁴ A significant positive correlation was observed between FOXM1 expression and the SOCE signature (Pearson r = 0.458, p = 1.47 × 10⁻²³; Figure 1A). This unprecedented clinical correlation suggests a functional linkage between the transcription factor FOXM1 and the SOCE machinery in ovarian cancer pathogenesis. To investigate potential causality and dissect how FOXM1 modulates individual components within the SOCE signature, we examined a public RNA sequencing (RNA-seq) dataset from an ovarian cancer cell line following FOXM1 depletion (GSE142567). Among the SOCE-signature genes, STIM1 expression was markedly decreased following FOXM1 depletion, whereas ORAI1/2/3 expressions showed no significant changes (Figures 1B and S1A). In contrast, STIM2 expression was moderately increased, possibly reflecting a compensatory response. These findings suggest that STIM1 may represent a major downstream target of SOCE associated with FOXM1.

Figure 1.

Integrative public transcriptome analysis reveals a positive association between FOXM1 and STIM1/SOCE in CR ovarian cancer

(A) Correlation between FOXM1 expression and SOCE signature (STIM1, STIM2, ORAI1, ORAI2, and ORAI3) in TCGA ovarian cancer tumors using GEPIA3. Pearson correlation coefficient (r) and p value are shown. TPM, transcript per million. (B) Effect of FOXM1 knockdown on STIM1 and ORAI1 expression in ovarian cancer cells (GSE142567, analyzed by GEO2R). (C) STIM1 expression in cisplatin-sensitive versus resistant cell lines (analyzed by GEO2R). Left: ovarian cancer cells (A2780 sensitive vs. A2780cis resistant, GSE33482); Right: triple-negative breast cancer cells (BT549 sensitive vs. MDA-MB-157 resistant, GSE103115). (D) Comparison of STIM1 mRNA expression between chemo-sensitive and chemo-resistant clinical ovarian cancer patient tumors from the GSE51373 dataset. For GSE51373, only stage III–IV ovarian cancer tumors with annotated platinum-response status were included. (E) Kaplan-Meier survival analysis showing the correlation between STIM1 expression levels and progression-free survival (PFS) in ovarian cancer patients (GSE26712 and GSE9891). Statistical significance was determined by the log-rank test. Data are presented as mean ± SE. Statistical significance was determined by Welch’s t test (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; n.s., not significant).

While FOXM1 overexpression in chemoresistant ovarian cancers has been extensively reported, the contribution of STIM1 remains poorly understood. To address this question, we first analyzed STIM1 expression in cisplatin-sensitive and CR ovarian cancer cell models and found that STIM1 and FOXM1 expressions were significantly elevated in resistant cells, respectively (GSE33482; Figures 1C [left] and S1B [left]). This pattern was further validated in a triple-negative breast cancer CR model (BT549 sensitive vs. MDA-MB-157 resistant (GSE103115; Figures 1C [right] and S1B [right]), suggesting that STIM1 upregulation in resistant cells may represent a broader, cross-cancer phenomenon.

To extend these observations to clinical specimens, we analyzed GSE51373, focusing on advanced-stage ovarian cancer patients. STIM1 mRNA levels were significantly elevated in platinum-resistant tumors compared with sensitive tumors (Figure 1D). Finally, Kaplan-Meier survival analysis across two independent ovarian cancer cohorts (GSE26712 and GSE9891) demonstrated that high STIM1 expression was significantly associated with poorer progression-free survival (Figure 1E). Collectively, these integrated bioinformatic analyses identify a previously unreported regulatory link between FOXM1 and STIM1 in the context of SOCE and chemoresistance. Our data suggest that FOXM1 may promote cisplatin resistance, at least in part, by upregulating or cooperating with STIM1 and enhancing SOCE activity. The relevance between FOXM1 and STIM1 represents a novel molecular mechanism warranting further functional investigation.

Establishment and characterization of CR ovarian cancer models with elevated FOXM1, STIM1, and SOCE activity

To validate the bioinformatic findings and further investigate the role of the FOXM1 and STIM1, we established two CR ovarian cancer sublines, ACR and ICR, from parental A2780 and IGROV1 cells, respectively. Resistant cells were generated by continuous stepwise exposure to increasing concentrations of cisplatin, starting from approximately the IC₂₅ value, until the IC₅₀ increased by approximately 8–10-fold compared with their parental counterparts (Figure S2A). Successful establishment of resistance was confirmed by CCK-8 assays, which showed that the IC₅₀ of cisplatin in ACR cells was 8.36-fold higher than in parental A2780 cells, and 9.22-fold higher in ICR cells compared with parental IGROV1 cells (Figures 2A and 2B). In addition, resistant cells displayed distinct elongated morphological alterations relative to parental cells (Figure S2B).

Figure 2.

Establishment of CR ovarian cancer cell lines and characterization of elevated FOXM1, STIM1 expression, and enhanced SOCE activity

(A and B) Cisplatin sensitivity of parental A2780 (A) and IGROV1 (I) cells, along with their CR derivatives ACR and ICR, was evaluated using the CCK-8 cell viability assay following 48 h cisplatin treatment. IC50 values were subsequently determined. Data are presented as mean ± SD from three independent experiments. (C) Protein expression levels of FOXM1, STIM1, and ORAI1 in parental and CR ovarian cancer cells were analyzed by immunoblotting. (D and E) Immunofluorescence staining of FOXM1 and STIM1 in parental and CR cells. Representative confocal images are shown, and quantitative analysis of fluorescence intensity is presented in the images at the right. (F and G) Representative single-cell cytosolic Ca²⁺ traces showing SOCE in parental and CR cells. a.u., arbitrary units. Data are presented as mean ± SE from at least three independent experiments. At least 50 cells were analyzed per group in each independent experiment. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 versus parental cells.

To determine whether FOXM1 and STIM1 signaling were altered in resistant cells, protein expression levels of FOXM1, STIM1, and ORAI1 were analyzed by immunoblotting. FOXM1 and STIM1 expression levels were markedly elevated in both ACR and ICR cells, whereas ORAI1 expression remained largely unchanged (Figure 2C). Immunofluorescence analyses further confirmed increased fluorescence intensity of FOXM1 and STIM1 in resistant cells and revealed enhanced nuclear localization of FOXM1 (Figures 2D, 2E, and S2C).

Because STIM1 is a critical regulator of SOCE, we next examined SOCE activity using Fura-2-based live-cell Ca²⁺ imaging. Cytosolic Ca²⁺ levels were monitored using the ratiometric Ca²⁺ indicator Fura-2. The fluorescence ratio obtained after excitation at 340 and 380 nm, expressed as F340/F380, was used as a relative measure of cytosolic Ca²⁺. SOCE was induced by re-addition of 2 mM extracellular Ca²⁺ after thapsigargin (TG)-mediated depletion of ER Ca²⁺ stores in Ca²⁺-free buffer. Total SOCE-mediated Ca²⁺ influx was quantified as the area under the F340/F380 curve during the extracellular Ca²⁺ re-addition period. Using this approach, both ACR and ICR cells exhibited significantly greater SOCE-mediated Ca²⁺ influx than their respective parental cells (Figures 2F and 2G).

Collectively, these findings demonstrate that CR ovarian cancer cells exhibit elevated FOXM1 and STIM1 expressions, accompanied by enhanced SOCE activity. These observations are consistent with the bioinformatic analyses (Figure 1) and support a potential role of FOXM1, STIM1, and SOCE in the development of chemoresistance.

FOXM1 positively regulates STIM1 expression and SOCE activity

To directly investigate whether FOXM1 regulates STIM1 expression and SOCE activity, we first applied the FOXM1 inhibitor Thiostrepton (Thio), which has been reported to suppress FOXM1 expression and transcriptional activity through direct inhibition of FOXM1-dependent transcription.²⁵ Immunoblotting analyses revealed that Thio treatment markedly reduced STIM1 expression, whereas ORAI1 expression remained largely unchanged (Figures 3A and 3C). In parallel, FOXM1 inhibition significantly decreased cell viability (Figures S3A and S3B). To determine whether FOXM1 inhibition also affects SOCE activity, we next examined SOCE activity using live-cell Ca²⁺ imaging. Thio treatment significantly suppressed SOCE-mediated Ca²⁺ influx, indicating that FOXM1 activity is required for maintaining SOCE function (Figures 3B and 3D).

Figure 3.

Pharmacological inhibition and genetic knockdown of FOXM1 suppress STIM1 expression and SOCE activity in both parental and CR ovarian cancer cells

(A and C) Protein expression levels of FOXM1, STIM1, and ORAI1 in parental and CR A2780 (A) and IGROV1 (C) cells after treatment with 0.5, 1, 2 μM Thiostrepton (Thio) for 48 h. (B and D) Representative single-cell cytosolic Ca²⁺ traces showing SOCE activity in parental and CR A2780 (B) and IGROV1 (D) cells. Quantification of SOCE-mediated Ca²⁺ influx was calculated as the area under the curve and is presented in the images at the right. (E and G) Protein expression levels of FOXM1, STIM1, and ORAI1 in parental and CR A2780 (E) and IGROV1 (G) cells after stable knockdown of FOXM1 using two independent shRNAs (shF#1 and shF#2). shC served as the non-targeting control. (F and H) Representative single-cell cytosolic Ca²⁺ traces showing SOCE activity in parental and CR A2780 (F) and IGROV1 (H) cells following FOXM1 knockdown. Quantification of SOCE-mediated Ca²⁺ influx (area under the curve) is shown in the images at the right. Data are presented as mean ± SE from at least three independent experiments. At least 50 cells were analyzed per group for Ca²⁺ measurements. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

To further validate the regulatory role of FOXM1, two independent FOXM1-specific shRNAs were used to genetically suppress endogenous FOXM1 expression. Consistent with the pharmacological inhibition results, FOXM1 knockdown markedly reduced STIM1 expression without substantially altering ORAI1 levels (Figures 3E and 3G). FOXM1 knockdown also significantly attenuated SOCE activity, particularly in CR cells (Figures 3F and 3H). Furthermore, depletion of FOXM1 restored cisplatin sensitivity, further supporting the contribution of FOXM1 to chemoresistance (Figures S3C and S3D).

To further support the positive regulatory relationship between FOXM1 and STIM1, FOXM1 was ectopically overexpressed using two independent FOXM1 expression vectors (pcDNA3.1-FOXM1 and pLXSN-FOXM1). FOXM1 overexpression consistently increased STIM1 protein levels (Figure S4A). Moreover, since NFAT is a key Ca²⁺-dependent transcription factor downstream of SOCE, we examined its nuclear localization following FOXM1 depletion. Compared with parental cells, CR cells exhibited increased nuclear accumulation of NFAT, whereas FOXM1 knockdown markedly reduced the nuclear-to-cytoplasmic ratio of NFAT (Figure S4B), further supporting the role of FOXM1 in regulating Ca²⁺ signaling activity.

Taken together, both pharmacological inhibition and genetic knockdown of FOXM1 consistently suppressed STIM1 expression and attenuated SOCE activity, with more pronounced effects observed in CR cells. These findings reveal a previously underappreciated role of FOXM1 in positively regulating STIM1 expression and SOCE activity, supporting a functional link between FOXM1 and STIM1-mediated Ca²⁺ signaling, particularly in the context of chemoresistant ovarian cancer.

FOXM1 directly transcriptionally activates STIM1 expression, particularly under cisplatin stress and in resistant cells

Given the consistent positive correlation between FOXM1 and STIM1 observed across our bioinformatic analyses and cell models, and considering the role of FOXM1 as a transcription factor in cancer progression, we hypothesized that FOXM1 directly regulates STIM1 transcription. To test this, STIM1 mRNA levels were first examined in FOXM1-knockdown cells (shF#1). Knockdown of FOXM1 led to a marked reduction in STIM1 mRNA (Figures 4A and 4C), consistent with the corresponding decrease in protein levels (Figures 3E and 3G), indicating that FOXM1 regulates STIM1 at the transcriptional level.

Figure 4.

FOXM1 directly transcriptionally regulates STIM1 expression in ovarian cancer cells

(A and C) qRT-PCR analysis of FOXM1 and STIM1 mRNA levels in parental A2780, ACR, and ACR cells with FOXM1 knockdown (shF) (A), and in parental IGROV1, ICR, and ICR shF cells (C). (B and D) Protein expression levels of FOXM1, STIM1, and ORAI1 in parental, cisplatin-treated (8 μM for 48 h in A2780; 2 μM for 48 h in IGROV1), and CR cells. (E) Immunofluorescence staining of FOXM1 in control (CTL) and cisplatin-treated ovarian cancer cells under different FBS conditions. Representative confocal images of A2780 cells treated with 8 μM cisplatin for 48 h (top) and IGROV1 cells treated with 2 μM cisplatin for 48 h (bottom) are shown. Quantification of the nuclear-to-cytoplasmic ratio of FOXM1 is presented in the images at the right. (F) Predicted FOXM1 binding motif (JASPAR database) and schematic diagram of the primer locations designed for ChIP-qPCR analysis on the STIM1 promoter region. (G) ChIP assay showing FOXM1 binding to the STIM1 promoter in parental (A), cisplatin-treated (Cis), and CR A2780 cells. IgG served as the negative control. Data are presented as mean ± SE from three independent experiments. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

We next investigated whether this regulatory effect is influenced by cisplatin treatment. Immunoblotting analyses revealed that both FOXM1 and STIM1 were upregulated in parental cells following cisplatin exposure, as well as in CR cells (Figures 4B and 4D), suggesting that cisplatin enhances FOXM1-mediated regulation of STIM1. To corroborate this, we assessed FOXM1 subcellular localization under both normal (10% FBS) and serum-free conditions to control for potential growth factor effects. Immunofluorescence demonstrated increased nuclear FOXM1 in parental cells cultured in 10% FBS compared with serum-free conditions. Importantly, cisplatin treatment further augmented nuclear FOXM1 signals under both culture conditions, confirming that cisplatin can induce FOXM1 nuclear translocation independently of serum factors (Figure 4E). These observations support the potential direct transcriptional activity of FOXM1 on STIM1, which is further enhanced by cisplatin.

To examine whether FOXM1 can bind directly to the STIM1 promoter, we analyzed the STIM1 promoter region using the Eukaryotic Promoter Database (EPD)²⁶ and screened for potential FOXM1 binding sites using JASPAR.²⁷ A high-affinity site (relative binding score = 0.874; Matrix ID: UN0802.1) was identified within the STIM1 promoter (Figure 4F). Based on this, qPCR primers were designed for chromatin immunoprecipitation (ChIP). ChIP-qPCR results demonstrated significantly increased FOXM1 enrichment at the STIM1 promoter in cisplatin-treated parental cells and CR cells compared with untreated parental cells (Figure 4G).

Together, these findings provide direct mechanistic evidence that FOXM1 transcriptionally regulates STIM1 expression. In addition, this regulatory axis is significantly enhanced under cisplatin stress and in CR cells, contributing to the elevated FOXM1 and STIM1 expression observed in chemoresistant ovarian cancer.

STIM1 and STIM1-mediated SOCE contribute to cisplatin resistance downstream of FOXM1

Although FOXM1 has been implicated in chemoresistance, the functional contribution of STIM1 and SOCE in this process remains largely unexplored. Having established the regulatory link between FOXM1 and STIM1, we next asked whether STIM1 and SOCE act as key mediators of FOXM1-driven chemoresistance.

CR cells were pretreated with sub-lethal doses of two well-characterized SOCE inhibitors SKF96365 and YM58483 prior to cisplatin exposure. Inhibition of SOCE significantly restored cisplatin sensitivity in CR cells (Figure 5A). Similar treatment in parental cells produced only modest effects on cisplatin sensitivity (Figure S6A), suggesting that SOCE activity is particularly important for the survival of resistant cells. To specifically address the role of STIM1, two independent STIM1-specific shRNAs (shS#6 and shS#7) were used to knockdown STIM1 expression (Figures S5A, S5B, and S5D). STIM1 depletion not only attenuated SOCE but also reduced cisplatin resistance in CR cells, while having limited impact in parental cells (Figures 5B, S5C, S5E, and S6B). These consistent results indicate that STIM1-mediated Ca²⁺ entry contributes to chemoresistance.

Figure 5.

STIM1 and STIM1-mediated SOCE contribute to cisplatin resistance in ovarian cancer cells

(A) Cell viability following treatment with the SOCE inhibitors SKF96365 (SKF, 2 μM) or YM58483 (YM, 10 μM), alone or in combination with cisplatin (8 μM for ACR cells; 2 μM for ICR cells) for 48 h, was determined by cell counting. (B) Cell viability following 48 h cisplatin treatment (8 μM for ACR cells; 2 μM for ICR cells) was determined by cell counting in ACR and ICR cells following stable STIM1 knockdown using two independent shRNAs (shS#6 and shS#7). shC served as the non-targeting control. (C) Protein expression levels of STIM1 and FOXM1 in STIM1-overexpressing A2780 cells were analyzed by immunoblotting. Cells were transfected with EGFP-STIM1 or empty vector control (CTL). (D) Representative single-cell cytosolic Ca²⁺ traces showing SOCE activity in empty vector control (CTL)-transfected and STIM1-overexpressing (EGFP-STIM1) A2780 cells. Quantification of SOCE-mediated Ca²⁺ influx was calculated as the area under the curve and is presented in the images at the right. (E) Cisplatin sensitivity of empty vector control (CTL)-transfected and STIM1-overexpressing (EGFP-STIM1) A2780 cells was evaluated using the CCK-8 cell viability assay following 48 h cisplatin treatment. IC50 values were subsequently determined. (F) Protein expression levels of STIM1, FOXM1, and ORAI1 in parental A2780, ACR, ACR with FOXM1-knockdown (shF), and ACRshF with EGFP-STIM1 overexpression (ST1) cells were analyzed by immunoblotting. shC served as the non-targeting control. (G) Cell viability following 48 h cisplatin treatment was determined by cell counting in ACR and ICR cells following FOXM1 knockdown alone or FOXM1 knockdown combined with EGFP-STIM1 overexpression. Data are presented as mean ± SE from three independent experiments. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Conversely, ectopic overexpression of STIM1 in parental cells increased STIM1 protein levels, enhanced SOCE activity, and conferred greater tolerance to cisplatin, as evidenced by higher IC₅₀ values (Figures 5C–5E). Importantly, re-expression of STIM1 in FOXM1-knockdown CR cells restored STIM1 levels and rescued cisplatin sensitivity (Figures 5F and 5G), confirming that STIM1 functions as a critical downstream effector through which FOXM1 promotes chemoresistance.

Collectively, these loss-of-function, gain-of-function, and rescue experiments demonstrate that STIM1 and STIM1-mediated SOCE are essential for cell survival under cisplatin stress. Our findings provide functional evidence that the FOXM1-STIM1-SOCE axis is a novel and important signaling pathway contributing to chemoresistance in ovarian cancer.

Transcriptomic profiling reveals FOXM1-STIM1 axis-driven IFN signaling in resistant cells

To explore downstream pathways through which FOXM1 and STIM1 promote cisplatin resistance, transcriptomic profiling was performed to identify genes dependent on these factors. RNA-seq was conducted in parental IGROV1 (I), CR (ICR), ICR with FOXM1 knockdown (ICRshF, shF#1), and ICR with STIM1 knockdown (ICRshS, shS#6) cells. Principal component analysis (PCA) demonstrated clear separation of resistant versus parental and knockdown samples, confirming the distinct transcriptomic profiles associated with FOXM1/STIM1-mediated chemoresistance (Figure S7A). To further support the role of FOXM1 in regulating Ca²⁺ signaling, Gene Set Enrichment Analysis (GSEA) of the Ca²⁺ ion binding gene set (GO:0005509) was performed using RNA-seq data from ICRshF versus ICR cells. The results showed a significant downregulation of Ca²⁺-binding genes upon FOXM1 knockdown (Figure S7B), providing additional evidence that FOXM1 regulates Ca²⁺ signaling upstream.

To determine gene expression changes associated with resistance and reversed upon FOXM1 or STIM1 depletion, Venn diagram analysis identified 2,578 shared differentially expressed genes (DEGs) among the three comparisons (ICR vs. I, ICRshF vs. ICR, and ICRshS vs. ICR; p.adj <0.05). After applying a |log₂FC| ≥ 0.585 threshold and directional rescue filtering, a core set of 567 “axis-dependent activated genes” was identified, which were upregulated in resistant cells and downregulated upon FOXM1 or STIM1 knockdown (ICR↑ → KD↓) (Figure 6A). Pathway enrichment analysis of these axis-dependent activated genes using Metascape²⁸ with the Reactome_pathway_2024 database²⁹ revealed significant enrichment of IFN-related signaling pathways, including IFN-γ signaling and IFN-α/β signaling, which were among the top five enriched pathways (Figure 6B). IFN signaling has been reported to modulate DNA damage response and promote chemoresistance.^30,31 STAT1 and its transcriptional partner IRF9 form the ISGF3 complex, which drives expression of IFN-stimulated genes (ISGs) such as ISG15, IFIT3, and IFI6, influencing DNA repair and cell survival.^32,33 These observations suggested that activation of IFN-associated transcriptional programs may contribute to the FOXM1-STIM1-driven resistant phenotype.

Figure 6.

Transcriptomic profiling reveals that the FOXM1/STIM1 axis drives IFN signaling in CR ovarian cancer cells

(A) Schematic workflow illustrating the identification of FOXM1/STIM1 axis-dependent activated genes from RNA-seq analyses. RNA-seq was performed in parental IGROV1 (I), ICR, ICR with FOXM1 knockdown (shF), and ICR with STIM1 knockdown (shS) cells. Differentially expressed genes (DEGs) satisfying |log₂FC| ≥ 0.585 in all comparisons were further filtered based on directionally reversed expression patterns. KD, knockdown. (B) Pathway enrichment analysis of FOXM1/STIM1 axis-dependent activated genes was performed using Metascape based on the Reactome Pathway 2024 database. (C) Heatmap showing the expression pattern of genes in the IF Signaling pathway (Reactome: R-HSA-913531) across the indicated groups, generated using ClustVis. (D) qRT-PCR validation of selected IFN-stimulated genes (IFIT3, IFI6, ISG15, and IRF9) in parental A2780, ACR, ACRshF, and ACRshS. Data are presented as mean ± SE from three independent experiments. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 versus respective control groups.

Therefore, to visualize IFN-signaling gene expression across conditions, we generated a heatmap of Reactome IFN signaling genes (R-HSA-913531) using Clustvis.³⁴ Parental and CR cells displayed distinct expression patterns, with many IFN-related genes elevated in CR cells and reduced following FOXM1 or STIM1 knockdown (Figure 6C). This coordinated expression pattern further supported the involvement of an IFN-associated transcriptional program downstream of the FOXM1-STIM1 axis. Among these genes, IFIT3, IFI6, ISG15, and IRF9 were selected for validation because they are components of the IFN-related DNA damage resistance signature (IRDS) and have been associated with cellular adaptation to genotoxic stress (Figure S7C). RT-qPCR in A2780, ACR, ACR-shFOXM1, and ACR-shSTIM1 cells confirmed that these genes are upregulated in chemoresistant cells and downregulated following FOXM1 or STIM1 knockdown (Figure 6D), consistent with the RNA-seq findings in IGROV1 cells. The coordinated upregulation of these ISGs suggests the establishment of an IFN-associated transcriptional state in resistant cells. Previous studies have linked IFN/STAT1-dependent gene programs to cellular adaptation to genotoxic stress and resistance to DNA-damaging therapies, with IRDS genes associated with enhanced DNA damage tolerance and prolonged cell survival following treatment.^30,31 Within this program, IFIT3 has been implicated in cancer cell survival and chemoresistance, while ISG15 has been linked to the regulation of DNA replication and repair responses under genotoxic stress.^31,32,33 Although the functions of individual ISGs are context-dependent, their coordinated induction in our resistant cells may contribute to a broader DNA damage resistance program that enables cells to tolerate persistent cisplatin-induced stress.

Collectively, these transcriptomic analyses identify IFN/STAT1 signaling as a downstream pathway regulated by the FOXM1-STIM1 axis, providing mechanistic insight into how this pathway contributes to cisplatin resistance in ovarian cancer.

The FOXM1-STIM1-SOCE axis activates STAT1 signaling to promote cisplatin resistance and represents a therapeutic vulnerability

Building on transcriptomic findings, we next sought to functionally validate this pathway in mediating cisplatin resistance. We first examined total STAT1 and phosphorylated STAT1 (p-STAT1 at Tyr701) levels in our cell models. Immunoblotting showed increased STAT1 activation in CR cells, which was markedly reduced upon FOXM1 or STIM1 knockdown, consistent with RNA-seq findings (Figure 7A). Furthermore, chelation of intracellular Ca²⁺ with BAPTA-AM significantly decreased p-STAT1 levels in CR cells (Figure 7B), indicating that STAT1 activation depends on Ca²⁺ signaling.

Figure 7.

FOXM1/STIM1-mediated Ca²⁺ signaling promotes STAT1 activation in CR ovarian cancer cells

(A) Protein expression levels of phosphorylated STAT1 (p-STAT1) and total STAT1 in parental, CR, CR with FOXM1-knockdown (CRshF), and CR with STIM1-knockdown (CRshS) cells were analyzed by immunoblotting. (B) Protein expression levels of p-STAT1 and total STAT1 in parental, CR, and CR cells treated with 20 μM BAPTA-AM (BAP) for 24 h were analyzed by immunoblotting. (C) Immunofluorescence staining of STAT1 in IGROV1, ICR, ICRshS, ICR cells treated with 20 μM BAPTA-AM for 24 h (BAP), and ICR cells treated with 2 μM SKF96365 for 24 h (SKF). Quantification of the nuclear-to-cytoplasmic ratio of STAT1 is presented in the images at the right. (D) Protein expression levels of FOXM1, STIM1, p-STAT1, and total STAT1 in A2780, ACR, and ACR cells treated with 10 μM Fludarabine for 48 h (FD), and ICR cells treated with 5 μM Ruxolitinib for 48 h (RU) were analyzed by immunoblotting. (E) Cell viability following treatment with Fludarabine (FD, 10 μM) or Ruxolitinib (RU, 5 μM), alone or in combination with cisplatin (2 μM), for 48 h in ICR cells was determined by cell counting. Data are presented as mean ± SE from three independent experiments. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 versus respective control groups.

We next assessed STAT1 nuclear translocation as a measure of its functional activation. Immunofluorescence analysis revealed elevated nuclear STAT1 in CR cells, whereas knockdown of STIM1, chelation of Ca²⁺ via BAPTA-AM, or treatment with the SOCE inhibitor SKF96365 markedly impaired STAT1 nuclear accumulation (Figure 7C). These results confirm that STAT1 activation is downstream of the FOXM1-STIM1-SOCE/Ca²⁺ signaling cascade in CR ovarian cancer cells.

Finally, to assess the functional importance of STAT1 signaling in chemoresistance, we treated resistant cells with two clinically relevant inhibitors, Fludarabine (STAT1 inhibitor) and Ruxolitinib (JAK inhibitor), at sublethal doses. Both inhibitors significantly sensitized CR cells to cisplatin (Figures 7D and 7E), indicating that STAT1 pathway activity is required for maintaining chemoresistance.

In conclusion, our findings demonstrate that the FOXM1-STIM1-SOCE axis promotes cisplatin resistance in ovarian cancer by activating IFN/STAT1 signaling. This previously unrecognized regulatory cascade not only provides mechanistic insight into chemoresistance but also highlights a potential therapeutic vulnerability.

Discussion

In this study, we demonstrate that FOXM1 promotes cisplatin resistance in ovarian cancer by directly transcriptionally activating STIM1, thereby enhancing SOCE and downstream IFN/STAT1 signaling. Through integrated bioinformatic analyses, CR cell models, functional perturbations, and transcriptomic profiling, we establish a novel FOXM1-STIM1/SOCE-STAT1 signaling axis that contributes to chemoresistance. This axis was consistently observed in both A2780-and IGROV1-derived resistant sublines and supported by clinical specimen data, providing a coherent mechanistic framework linking a master transcription factor to Ca²⁺-dependent survival signaling.

A key mechanistic insight from our work is the reciprocal regulation between FOXM1 and Ca²⁺ signaling. Previous studies primarily positioned Ca²⁺ as an upstream regulator of FOXM1, demonstrating that intracellular Ca²⁺ can alter FOXM1 post-translational modification, localization, and transcriptional activity.^21,35,36 In parallel, FOXM1 has been implicated in mitochondrial survival and anti-apoptotic responses associated with Ca²⁺ homeostasis, suggesting that FOXM1 and Ca²⁺ signaling may form a reciprocal regulatory axis in which Ca²⁺ modulates FOXM1 activity, while FOXM1 reinforces Ca²⁺-dependent survival programs.⁸ Consistent with this concept, bioinformatic analyses from ENCODE and Harmonizome datasets revealed that FOXM1 target gene networks overlap with genes involved in ion transport and signaling pathways, suggesting that FOXM1 may also function as an upstream regulator of Ca²⁺-related signaling.^37,38 Despite these observations, the mechanisms by which FOXM1 regulates Ca²⁺-dependent signaling remain largely unexplored. In our study, FOXM1 knockdown significantly downregulated the Ca²⁺ ion binding gene set (Figure S7B), and ChIP-qPCR confirmed direct FOXM1 binding to the STIM1 promoter, particularly under cisplatin-induced stress (Figure 4), leading to enhanced STIM1 expression and SOCE. Together with our Ca²⁺ imaging data showing enhanced SOCE in resistant cells, these findings support a model in which cisplatin-induced FOXM1 nuclear accumulation promotes STIM1 transcription, thereby enhancing SOCE and sustaining cytosolic Ca²⁺ signaling. This reciprocal relationship may establish a positive feedback loop that amplifies pro-survival Ca²⁺ signaling and contributes to persistent drug resistance. Based on previous evidence that intracellular Ca²⁺ can regulate FOXM1 localization and post-translational modification, enhanced SOCE may in turn promote FOXM1 activity or nuclear retention, thereby sustaining STIM1 transcription. Although this final Ca²⁺-to-FOXM1 step was not directly tested in the present study, the proposed feedback model provides a framework for future mechanistic investigation. In addition, although ChIP-qPCR demonstrated FOXM1 occupancy at the STIM1 promoter and FOXM1 perturbation altered STIM1 expression, the functional requirement of the predicted FOXM1-binding motif was not tested using a promoter-reporter assay or site-directed mutagenesis. Future studies comparing wild-type and binding-site-mutant STIM1 promoter constructs will be needed to define the motif-specific contribution of FOXM1 to STIM1 transcription.

Transcriptomic profiling further revealed that the FOXM1-STIM1 axis is associated with activation of IFN/STAT1 signaling in resistant cells. ISGs have been implicated in DNA damage tolerance, apoptosis resistance, and chemoresistance,^32,33 suggesting that activation of this pathway may provide cells with an adaptive survival advantage. Although SOCE has been widely linked to cancer progression, its downstream mechanisms in drug resistance remain incompletely understood. Previous studies have suggested that SOCE may support cancer cell survival through regulation of ER stress resistance,³⁹ apoptosis suppression,⁴⁰ and Ca²⁺ homeostasis.⁴¹ Our findings extend this concept by showing that STIM1-mediated SOCE also contributes to cisplatin resistance through activation of IFN/STAT1 signaling. In this context, SOCE may act not only as a Ca²⁺ entry mechanism but also as a signaling platform that connects FOXM1-driven transcriptional regulation to IFN-associated survival programs.

Our data further establish STIM1-mediated SOCE as an upstream regulator of STAT1 activation in chemoresistant ovarian cancer cells. Previous studies have linked STIM1 to IFN-associated signaling, mainly through indirect mechanisms. In immune cells, STIM1 has been shown to restrain spontaneous STING activation by retaining STING at the ER, thereby modulating type-I IFN responses.⁴² In addition, activation of the cGAS-STING pathway can induce type-I IFN production, leading to IFN-α/β receptor (IFNAR)-dependent STAT1/STAT2 phosphorylation and ISG transcription.⁴³ These findings suggest a potential connection between STIM1-dependent signaling and IFN/STAT-associated transcriptional programs, but they do not directly establish Ca²⁺ influx or SOCE as an upstream driver of STAT1 activation. The relationship between STIM1 and IFN signaling reported here differs mechanistically from the previously described STIM1-STING interaction in immune cells. In immune cells, ER-localized STIM1 can retain STING and restrain spontaneous type I IFN activation. By contrast, our findings indicate that, in cisplatin-resistant ovarian cancer cells, increased STIM1 expression enhances SOCE and that the resulting Ca²⁺ influx is required for STAT1 phosphorylation and nuclear accumulation. Thus, rather than functioning primarily as an ER-retention factor for STING, STIM1 acts in our model as a component of a Ca²⁺ entry pathway that supports STAT1-dependent transcription. To our knowledge, this study provides the first functional evidence linking STIM1-mediated SOCE to STAT1 activation and platinum resistance in ovarian cancer cells.

Consistent with this concept, pharmacological chelation of intracellular Ca²⁺ by BAPTA-AM or inhibition of SOCE by SKF96365 significantly reduced STAT1 phosphorylation and nuclear translocation in our resistant cells (Figures 7B and 7C). These results provide functional evidence that SOCE-dependent Ca²⁺ signaling is required for STAT1 activation downstream of STIM1. Therefore, STIM1-mediated SOCE may serve as a central signaling node that activates transcriptional programs, including ISGs, to support cell survival and platinum resistance. The molecular intermediate connecting SOCE-mediated Ca²⁺ influx to STAT1 phosphorylation remains to be defined. STAT1 Tyr701 is commonly phosphorylated by receptor-associated JAK kinases; therefore, Ca²⁺-dependent modulation of JAK activity or upstream cytokine-receptor signaling represents one possible mechanism. Alternatively, Ca²⁺-responsive kinases and signaling proteins, including CaMK family members, PKC, Src-family kinases, or Ca²⁺-dependent phosphatases, could indirectly regulate STAT1 activation, nuclear transport, or the production of autocrine IFN signals. The current data establish a requirement for SOCE and intracellular Ca²⁺ in maintaining STAT1 activation but do not distinguish among these possible intermediates. Future phosphoproteomic analyses, kinase-inhibitor studies, and time-resolved measurements of JAK and STAT1 phosphorylation will be required to resolve this mechanism.

Several limitations should be considered. First, the functional experiments were performed using established ovarian cancer cell lines and their cisplatin-resistant derivatives. Although two independent cellular models were examined, validation in patient-derived organoids, xenografts, or genetically engineered tumor models will be required to determine whether the FOXM1-STIM1/SOCE-STAT1 axis operates in vivo. Second, the present study tested fixed-dose combinations of cisplatin with SOCE or JAK/STAT pathway inhibitors but did not systematically optimize drug sequence, dose ratio, therapeutic window, or combination synergy. Pharmacokinetic, toxicity, and schedule-dependent studies will therefore be necessary before clinical translation. Third, although public ovarian cancer cohorts supported the association of STIM1 with platinum resistance and poor progression-free survival, we did not directly evaluate FOXM1 and STIM1 protein expression in an independent clinical tissue cohort. Future studies should determine whether combined FOXM1/STIM1 expression can serve as a predictive biomarker for platinum response or patient outcome.

In conclusion, our study identifies the FOXM1-STIM1/SOCE-STAT1 axis as a previously unrecognized mechanism driving cisplatin resistance in ovarian cancer. This novel axis provides a molecular explanation for the co-upregulation of FOXM1 and Ca²⁺ signaling components in chemoresistant tumors and highlights several potential therapeutic intervention points. Inhibition of SOCE or JAK/STAT signaling sensitized resistant cells to cisplatin in our models, supporting the potential value of combining platinum-based chemotherapy with inhibitors targeting this pathway. In particular, patients with high FOXM1 and STIM1 expression may benefit from therapeutic strategies designed to disrupt Ca²⁺-dependent STAT1 activation. Future studies targeting this signaling axis may provide new opportunities to overcome platinum resistance and improve outcomes for patients with advanced ovarian cancer.

Materials and methods

Public database and bioinformatics analysis

Transcriptomic data of ovarian cancer were obtained from The Cancer Genome Atlas (TCGA) Ovarian Cancer (OV) dataset. The correlation between FOXM1 expression and SOCE signature (STIM1, STIM2, ORAI1, ORAI2, and ORAI3) was analyzed using the GEPIA3 web server.²⁴ The SOCE signature was defined as the average log₂(TPM+1) expression of five key SOCE-related genes. Gene expression datasets were downloaded from the NCBI Gene Expression Omnibus (GEO) database. The following datasets were analyzed: GSE142567 was used to compare FOXM1-depleted ovarian cancer cells with the corresponding control samples. GSE33482 was used to compare cisplatin-sensitive A2780 cells with A2780cis-resistant ovarian cancer cells. GSE103115 was used to compare the cisplatin-sensitive and resistant triple-negative breast cancer cell models annotated in the original dataset. For GSE142567, GSE33482, and GSE103115, differential expression analysis was performed using the GEO2R online tool with default limma settings and quantile normalization. For GSE51373, the raw expression matrix was downloaded from NCBI GEO, only tumors from patients with stage III or IV ovarian cancer and annotated platinum-response status were included. This analysis included 12 platinum-sensitive and 10 platinum-resistant tumors. For the GSE51373 dataset, the raw expression matrix was downloaded from NCBI GEO. Only advanced-stage (stage III and IV) samples were included (n = 12 sensitive and n = 10 resistant). Expression values were log₂-transformed prior to statistical analysis. Platinum-sensitive and platinum-resistant groups were defined according to the original study. GSE26712 and GSE9891 were used for survival analyses based on available STIM1 expression and progression-free survival information.

Survival curves were initially explored on the KM-plotter platform and subsequently independently re-plotted in R (version 4.3.0) using the survival and survminer packages. Patients were stratified into high- and low-expression groups based on the median STIM1 expression value. Statistical significance was assessed by the log rank test. Statistical significance for all expression comparisons was determined by Welch’s t test. A p value < 0.05 was considered statistically significant. All analyses were conducted according to the original publications and platform guidelines.

Cell culture and establishment of CR cells

A2780 and IGROV1 cells and their derived CR sublines (ACR and ICR) were maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS; A5256701, Gibco) and 1% penicillin-streptomycin (P/S; PSL01, Caisson Laboratories) at 37°C in a humidified atmosphere containing 5% CO₂. To establish CR sublines, the half-maximal inhibitory concentration (IC₅₀) of cisplatin (P4394, Sigma-Aldrich) was first determined using a standard cell viability assay. Parental cells were initially exposed to a cisplatin concentration corresponding to approximately the IC₂₅ and maintained at this level for at least two weeks, during which surviving cells were monitored for normal proliferation. Once stable growth was achieved, the drug concentration was gradually increased in a stepwise manner. Each concentration was maintained until the cells exhibited consistent proliferation. This process was continued until the IC₅₀ of the resistant cells reached approximately 10-fold that of the parental cells. After establishment, the resistant sublines (CR cells) were pulsed with cisplatin once per week to maintain resistance.

Cell viability assay

Cell viability was assessed using the Cell Counting Kit-8 (CCK-8) assay (C0005, TargetMol) and direct cell number counting. For the CCK-8 assay, cells were seeded in 96-well plates and treated with the indicated chemicals according to the experimental design. After treatment, the culture medium was replaced with fresh medium containing CCK-8 reagent, and the plates were incubated at 37°C in a humidified atmosphere containing 5% CO₂ for 2 h. Absorbance was measured at 450 nm using a microplate reader (800 TS, Biotek). For cell number counting, cells were seeded in 6-well plates and treated with the indicated chemicals as described earlier. After treatment, cells were harvested by trypsinization, stained with trypan blue, and counted using a hemocytometer.

Drugs and inhibitors

Thiostrepton (B0090) was purchased from MedChemExpress. YM58483 (Y4895), cisplatin (P4394), and BAPTA-AM (A1076) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Fludarabine (sc-204755) and Ruxolitinib (sc-364729) were obtained from Santa Cruz Biotechnology (Dallas, TX, USA). SKF96365 (10009312) was purchased from Cayman Chemical (Ann Arbor, MI, USA).

Single-cell intracellular Ca²⁺ measurements

Cells grown on glass-bottom dishes were loaded with 2 μM Fura-2 AM (F1221, Thermo Fisher Scientific) for 30 min at 37 °C. Intracellular Ca²⁺ dynamics were recorded using a single-cell fluorimeter (TILL Photonics, Germany). To measure SOCE, cells were first perfused with Ca²⁺-free buffer (145 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 10 mM glucose, 5 mM HEPES; pH 7.4) for 30 s to record the basal Ca²⁺ signal. Subsequently, 2 μM thapsigargin (T7458, Sigma-Aldrich) was added in Ca²⁺-free buffer, and the signal was recorded for 10 min. The buffer was then replaced with 2 mM Ca²⁺ buffer (same ionic composition with the addition of 2 mM CaCl₂) to induce SOCE, and Ca²⁺ influx was recorded for an additional 5 min as the SOCE signal. Fura-2 was excited alternately at 340 and 380 nm, and emission fluorescence at 510 nm was collected. The 340/380 nm fluorescence ratio was calculated using TILLvisION 4.0 software (TILL Photonics, Germany) to represent relative cytosolic Ca²⁺ levels. The magnitude of SOCE was quantified by measuring the area under the curve of the Ca²⁺ trace. At least 50 cells were analyzed per group in each independent experiment.

Immunoblotting analysis

Cellular proteins were extracted on ice using cold radioimmunoprecipitation assay (RIPA) buffer (150 mM NaCl, 10 mM EDTA, 50 mM Tris-HCl pH 7.4, 1% NP-40, 0.004% sodium azide, 0.5% sodium deoxycholate, 0.1% SDS) supplemented with a protease inhibitor cocktail (5892791001, cOmplete ULTRA, Roche), 1 mM NaF, 1 mM PMSF, and 1 mM Na₃VO₄. Protein concentrations were determined using a DC protein assay kit (Bio-Rad). Equal amounts of total protein were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes. Membranes were incubated overnight at 4°C with the following primary antibodies: anti-FOXM1 (#ab245309, Abcam), anti-STIM1 (#610954, BD Biosciences), anti-ORAI1 (#4041, ProSci), anti-STAT1 (#9172, Cell Signaling Technology), anti-phospho-STAT1 (Tyr701) (#9167, Cell Signaling Technology), anti-β-actin (#A2066, Sigma-Aldrich), and anti-GAPDH (#GTX100118, GeneTex). After washing, membranes were incubated with appropriate HRP-conjugated secondary antibodies at room temperature for 1 h. Protein bands were visualized using an enhanced chemiluminescence (ECL) detection kit (PerkinElmer) and imaged with an Amersham Imager 600 system (GE Healthcare). β-Actin and GAPDH were used as loading controls.

Immunofluorescence staining and confocal microscopy

Cells were grown on glass-bottom dishes, fixed with 4% paraformaldehyde (PFA) in PBS for 15 min at room temperature, permeabilized with 0.5% Triton X-100 in PBS for 15 min, and blocked with CAS-Blocking reagent (#8120, Thermo Fisher Scientific) for 1 h at room temperature. The cells were then incubated overnight at 4°C with primary antibodies against FOXM1 (#H00002305-M02, Abnova), STIM1 (#GTX113558, GeneTex), STAT1 (#9172, Cell Signaling Technology), and NFAT (#GTX127932, GeneTex). After washing three times with PBS, the cells were incubated with Alexa Fluor 488- and Alexa Fluor 594-conjugated secondary antibodies (Abcam) and DAPI (1 μg/mL, Invitrogen) for 1 h at room temperature. Images were acquired using an inverted fluorescence microscope or a laser scanning confocal microscope (FLUOVIEW FV3000, Olympus). Fluorescence intensity was quantified using FV10-ASW software (Olympus). At least 50 cells were analyzed per group in each independent experiment.

Plasmids and transfection

The plasmids pcDNA3.1-FOXM1, pLXSN-FOXM1,⁴⁴ and EGFP-STIM1⁴⁵ used in this study were constructed or obtained as previously described. For transient transfection, Opti-MEM Reduced Serum Medium (31985062, Gibco) was mixed separately with Lipofectamine 3000 (L3000015, Invitrogen) and with the indicated plasmids plus P3000 Reagent (L3000015, Invitrogen). The two mixtures were then combined and incubated at room temperature for 15 min to form transfection complexes. The complexes were subsequently added to the cells, which were incubated at 37°C in a humidified atmosphere containing 5% CO₂. Transfected cells were further incubated for 48 or 72 h before being subjected to downstream experiments.

RNA extraction, reverse transcription, and real-time qPCR

Total RNA was extracted from cells using the Total RNA Isolation Kit (NA017-0100, GeneDirex) according to the manufacturer’s instructions. RNA concentration and purity were assessed using a NanoDrop spectrophotometer. 1 μg of total RNA was reverse-transcribed into cDNA using the PrimeScript RT reagent Kit (RR037A, Takara) according to the manufacturer’s protocol. For quantitative real-time PCR, reactions were prepared in a total volume of 20 μL containing 7 μL nuclease-free water, 1 μL forward primer (5 μM), 1 μL reverse primer (5 μM), 1 μL cDNA template (20 ng), and 10 μL PowerUp SYBR Green Master Mix (A25918, Thermo Fisher Scientific). All reactions were performed in triplicate on an Applied Biosystems StepOnePlus Real-Time PCR System (Thermo Fisher Scientific). The thermal cycling conditions were as follows: initial denaturation at 95 °C for 2 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. Melting curve analysis was performed after amplification to confirm primer specificity. Primer sequences used in this study are listed in Table S1. Relative gene expression levels were calculated using the 2^−ΔΔCt method and normalized to GAPDH as the internal control.

ChIP assay

ChIP was performed using the SimpleChIP Enzymatic Chromatin IP Kit (Magnetic Beads) (9003S, Cell Signaling Technology) according to the manufacturer’s protocol with minor modifications. Briefly, cells (approximately 4 × 10⁶ cells per immunoprecipitation) were cross-linked with 1% formaldehyde for 10 min at room temperature, and the reaction was quenched with glycine. After washing with ice-cold PBS, nuclei were isolated and lysed. Chromatin was then sheared by sonication to obtain DNA fragments of 200–1,000 bp in length. The sheared chromatin was incubated overnight at 4°C with 2 μg of anti-FOXM1 antibody (GTX102170, GeneTex). Normal rabbit IgG was used as a negative control, and 10% of the digested chromatin was saved as input for each sample. ChIP-Grade Protein G Magnetic Beads were then added to the antibody-chromatin mixtures, and the immunocomplexes were incubated for 2 h at 4°C with rotation. After extensive washing, the cross-links were reversed, and the immunoprecipitated DNA as well as input DNA were purified using the spin columns provided in the kit.

The enrichment of the STIM1 promoter region in the immunoprecipitated DNA was analyzed by real-time qPCR using specific primers listed in Table S2. Results were calculated as percentage of input or fold enrichment relative to the IgG control.

Lentivirus-mediated FOXM1 and STIM1 knockdown

Knockdown of FOXM1 was conducted by lentivirus using a pLKO.1-puro vector carrying shRNA sequences (shF#1, clone ID: TRCN0000273982, 5′-GCCAATCGTTCTCTGACAGAA-3′; and shF#2, clone ID: TRCN0000273984, 5′-TTGCAGGGTGGTCCGTGTAAA-3′) that target consensus regions of human FOXM1 (Table S3). Knockdown of STIM1 was conducted by lentivirus using either pLKO.1-puro vector carrying shRNA sequences (shS#6, clone ID: TRCN0000146686, 5′-CCTGGATGATGTAGATCATAA-3′) or pLKO_005-puro vector carrying shRNA sequences (shS#7, clone ID: TRCN0000358716, 5′-CCTCTCTTGACTCGCCATAAT-3′) that target consensus regions of human STIM1. Another pLKO.1-puro vector carrying shRNA sequences targeting firefly luciferase (shC, clone ID: TRCN0000231719, 5′-GCGGTTGCCAAGAGGTTCCAT-3′) or pLKO_005-puro vector carrying shRNA sequences targeting firefly luciferase (shC, clone ID: TRCN0000072249, 5′-GCGGTTGCCAAGAGGTTCCAT-3′) were used as a control (National RNAi Core Facility, Academia Sinica, Taipei, Taiwan). A2780 and IGROV1 cells were infected individually with each lentivirus in the presence of 8 μg/mL Polybrene (H9268, Sigma-Aldrich) for 48 h. After infection, the medium was replaced with fresh growth medium containing puromycin (3 μg/mL; A11138-03, Invitrogen), and the cells were subjected to selection for an additional 48 h. Stable knockdown cells were maintained in puromycin-containing medium thereafter. The two stable lines thus obtained were then maintained in complete growth medium containing puromycin (1 μg/mL). The KD efficiency of FOXM1 and STIM1 was assessed by Western blotting and qPCR.

RNA sequencing library preparation and sequencing

Total RNA was extracted as described earlier. RNA integrity was assessed, and only samples with RNA Integrity Number (RIN) ≥ 8.0 were used for library preparation. Sequencing libraries were prepared from three independent biological replicates per group using the TruSeq Stranded mRNA Library Prep Kit (Illumina, San Diego, CA, USA) according to the manufacturer’s recommendations. Briefly, mRNA was purified from 1 μg of total RNA using oligo(dT)-coupled magnetic beads and fragmented under elevated temperature. First-strand cDNA was synthesized using reverse transcriptase and random primers. Second-strand cDNA synthesis was performed with dUTP incorporation for strand specificity. After end-repair, adenylation of 3′ ends, and adapter ligation, the products were enriched by PCR and purified using the AMPure XP system (Beckman Coulter, Brea, CA, USA). Library quality was assessed with the Qsep400 System (Bioptic Inc., Taiwan), and library concentration was quantified using a Qubit 2.0 Fluorometer (Thermo Fisher Scientific). The qualified libraries were sequenced on an Illumina NovaSeq platform with 150 bp paired-end reads by Genomics, BioSci & Tech Co. (New Taipei City, Taiwan).

RNA-seq data analysis

Raw sequencing reads were processed and aligned using standard bioinformatics pipelines. Differential gene expression analysis was performed using DESeq2. Genes with an adjusted p value (padj) < 0.05 were considered DEGs. To identify genes regulated by the FOXM1-STIM1 axis, three pairwise comparisons were performed: ICR versus I, ICR-shFOXM1 versus ICR, and ICR-shSTIM1 versus ICR. DEGs shared among all three comparisons were identified by Venn diagram analysis. Genes with an absolute log₂ fold change (|log₂FC|) ≥ 0.585 in all comparisons were further selected. Genes showing opposite directional changes after FOXM1 or STIM1 knockdown (KD) were considered rescue-associated genes and were classified into activated genes (ICR↑ → KD↓) or repressed genes (ICR↓ → KD↑). Pathway enrichment analysis was performed using Metascape. Reactome gene set enrichment (Reactome Pathway 2024) analysis was subsequently performed to identify biological pathways associated with FOXM1-STIM1 axis-dependent genes.

Statistical analysis

All data are presented as the mean ± standard error (SE) from at least three independent experiments. Statistical significance was evaluated using Student’s t test for comparisons between two groups or one-way analysis of variance followed by Dunnett’s post hoc test for comparisons involving multiple treatment groups, as appropriate. For RNA-seq analyses, genes with an adjusted p value of less than 0.05 were initially defined as differentially expressed. FOXM1-STIM1 axis-dependent genes were subsequently selected using an absolute log₂ fold-change threshold of at least 0.585 in each relevant comparison, together with directional-rescue filtering. All statistical analyses were performed using SPSS version 17.0 (IBM Corp., Armonk, NY, USA) and Origin software. A p value <0.05 was considered statistically significant.

Data and code availability

All data generated or analyzed during this study are included in this published article and its supplemental information files.

Acknowledgments

The authors gratefully acknowledge the technical services provided by the Bioimaging Core Facility of the National Core Facility for Biopharmaceuticals, National Science and Technology Council, Taiwan. This work was supported by the National Science and Technology Council, Taiwan (grant nos. NSTC 115-2740-B-006-002 and NSTC 114-2311-B-006-008). Ethics approval was not applicable to this study.

Author contributions

T.C.L. and W.-T.C. conceived and designed the experiments. T.-C.L., M.-R.H., and H.D.H.N. performed the experiments. T.-C.L. and W.-T.C. drafted the manuscript. S.V. and W.-T.C. provided technical and material support. W.-T.C. critically revised and finalized it. All authors read and approved the final version of the manuscript.

Declaration of interests

The authors declare no competing interests.