Overexpression of forkhead box protein M1 (FOXM1) in ovarian cancer correlates with poor patient survival and contributes to paclitaxel resistance

Abstract

Aim: Deregulation of FOXM1 has been documented in various cancers. The aim of this study was to evaluate the role of FOXM1 in ovarian cancer tumorigenesis and paclitaxel resistance.

Experimental design: Expression of FOXM1 was examined in 119 clinical samples by immunohistochemistry and correlated with clinicopathological parameters. Effects of FOXM1 knockdown on ovarian cancer cell migration, invasion and mitotic catastrophe were also studied. qPCR and ChIP-qPCR were used to establish KIF2C as a novel FOXM1 target gene implicated in chemoresistance.

Results: High nuclear FOXM1 expression in ovarian cancer patient samples was significantly associated with advanced stages (P = 0.035), shorter overall (P = 0.019) and disease-free (P = 0.014) survival. Multivariate analysis confirmed FOXM1 expression as an independent prognostic factor for ovarian cancer. FOXM1 knockdown significantly inhibited migration and invasion of ovarian cancer cells and enhanced paclitaxel-mediated cell death and mitotic catastrophe in a p53-independent manner. Bioinformatics analysis suggested a number of potential transcription targets of FOXM1. One of the potential targets, KIF2C, exhibited similar expression pattern to FOXM1 in chemosensitive and chemoresistant cells in response to paclitaxel treatment. FOXM1 could be detected at the promoter of KIF2C and FOXM1 silencing significantly down-regulated KIF2C.

Conclusion: Our findings suggest that FOXM1 is associated with poor patient outcome and contributes to paclitaxel resistance by blocking mitotic catastrophe. KIF2C is identified as a novel FOXM1 transcriptional target that may be implicated in the acquisition of chemoresistance. FOXM1 should be further investigated as a potential prognostic marker and therapeutic target for ovarian cancer.

Figure 1. Elevated nuclear FOXM1 expression was associated with advanced stages of ovarian cancer.

A, Representative images of immunoreactivity of nuclear FOXM1 in (I) benign cystadenoma, (II) borderline tumour, (III) stage I invasive cancer, (IV) stage II invasive cancer, (V) stage III invasive cancer and (VI) stage IV invasive cancer. Magnifications X400. Insets: Regions with higher magnifications of nuclear FOXM1 staining. B, Cumulative overall and disease-free survival plots using the Kaplan-Meier approach.

Figure 2. Silencing of FOXM1 reduced migration and invasion of SKOV-3.

A. Representative images showing cells migrated (gelatin-coated membrane) or invaded (matrigel-coated membrane) after 24 h. B. Representative Western blot analysis demonstrating the effectiveness of FOXM1 transient knockdown in SKOV-3. C. Graphic representation of migration (left panel) and invasion (right panel) results as fold change of migrated and invaded cells relative to the control, respectively, in five fields of triplicate wells from three independent experiments; * P<0.05, significant; Mann-Whitney U-test.

Figure 3. Paclitaxel treatment down-regulates FOXM1 expression in SKOV-3 but not in SKOV3-TR cells.

The paclitaxel sensitive SKOV-3 and resistant SKOV3-TR ovarian cancer cells were treated with 100 nM paclitaxel and harvested at times indicated for Western blot analysis. Paclitaxel treatment down-regulated FOXM1 expression at time points 48 h and 72 h in SKOV-3 but not in SKOV3-TR as shown by immunoblotting. There were also no marked changes in the cleaved Caspase-9 and Caspase-7 expression.

Figure 4. Transient FOXM1 knockdown significantly enhanced paclitaxel-induced mitotic catastrophe in SKOV-3 and OVCAR3 cells.

SKOV-3 and OVCAR3 cells transfected with either siRNA pools against FOXM1 or control siRNA pools were treated with paclitaxel (100 nM) and stained with DAPI. A. Representative staining results were shown showing that transient FOXM1 knockdown significantly enhanced paclitaxel-induced mitotic catastrophe (arrow) in SKOV-3 and OVCAR3 cells respectively (Upper panel). B. Graphs represent the results of three independent experiments, showing the percentage of cells undergoing mitotic catastrophe, * P<0.05, significant; Mann-Whitney U-test.

Figure 5. Flow cytometric analysis of SKOV-3 and SKOV-3-TR cells with and without FOXM1 depletion in the presence or absence of paclitacel treatment.

A. Flow cytometric analysis was performed following propidium iodide staining on SKOV-3 and SKOV-3-TR cells treated with paclitaxel (100 nM) or remained untreated after transfection with siRNA pools against FOXM1 or control siRNA pools. Representative data are shown indicating that FOXM1 silencing is capable of increasing the number of dead cells and cells blocked at G2/M cell cycle phase in SKOV-3-TR as compared to cells treated with control siRNA. B. Bar charts of different phases of cell cycle in SKOV-3 and SKOV3-TR treated with paclitaxel after transfection with control or siRNA against FOXM1. Results represent data from two independent experiments.

Figure 6. Identification of KIF2C as a novel FOXM1 transcriptional target that might be implicated in the acquisition of paclitaxel resistance.

A, Paclitaxel treatment (50 nM) down-regulated FOXM1 and KIF2C expressions at 48 h and 72 h in PEO1 but not in PEO1-TaxR. B, FOXM1 knockdown significantly reduced the transcript level of KIF2C in PEO1 and PEO1-TaxR. Data represent triplicates from three experiments. *P = 0.04, ***P = 0.0003. siControl: Non-specific control. siFOXM1: FOXM1 knockdown. C, ChIP-qPCR showed FOXM1 significantly pull down KIF2C promoter region in PEO1 and PEO1-TaxR as compared to the negative IgG control. Data represent triplicates from three experiments. *P = 0.04, ***P = 0.0001. D, Schematic diagram depicting locations of forkhead response element (FHRE) and the binding site of primers used in ChIP-qPCR upstream of the transcription start site of KIF2C.