• Open Access

Forkhead box protein C2 contributes to invasion and metastasis of extrahepatic cholangiocarcinoma, resulting in a poor prognosis



Extrahepatic cholangiocarcinoma (EHCC) is a cancer with a poor prognosis, and the postoperative survival of patients depends on the existence of invasion and metastasis. The epithelial-to-mesenchymal transition (EMT) is an important step in EHCC invasion and metastasis. Forkhead box protein C2 (FOXC2) is a transcription factor that has been reported to induce the EMT. Therefore we examined the correlation between FOXC2 expression and clinical pathological factors, and analysed the function of FOXC2. The expression of FOXC2 in 77 EHCC cases was investigated by immunohistochemical staining, and the relationship between FOXC2 expression and clinicopathological factor was assessed. Knockdown by small interfering RNA (siRNA) was performed to determine the roles of FOXC2 in EHCC cell line. FOXC2 expression correlated with lymph node metastasis (= 0.0205). Patients in the high FOXC2 expression group had a poorer prognosis than the patients in the low FOXC2 expression group. Moreover, FOXC2 knockdown inhibited cell motility and invasion, and decreased the expression of EMT markers (N-cadherin, and matrix metalloproteinase (MMP) -2) and Angiopietin-2 (Ang-2). The EMT inducer FOXC2 contributes to a poor prognosis and cancer progression. FOXC2 may be a promising molecular target for regulating EHCC metastasis.

The incidence and mortality of cholangiocarcinoma, a cancer with a poor prognosis, are rising worldwide.[1] The 5-year survival rates for cholangiocarcinoma is 10–40% overall.[2] Cholangiocarinoma is categorized into intrahepatic or extrahepatic cholangicarinoma, extrahepatic cholangiocarcinoma (EHCC), the latter of which consists of either a hilar tumor or a bile duct tumor. Surgical therapy is the only effective curative treatment of EHCC, and the postoperative survival has been shown to be dependent on the existence of invasion and lymph node metastasis.[3, 4] Therefore, to improve patient prognoses, it is essential to elucidate the mechanisms of invasion and metastasis in EHCC.

The epithelial to mesenchymal transition (EMT), whereby epithelial cells alter their morphology to resemble mesenchymal cells, is required for the invasion and metastasis of cancer cells. Epithelial to mesenchymal transition was first proposed as a central differentiation process in embryogenic morphogenesis.[5] Embryonic cells lose E-cadherin expression and acquire cellular rearrangements for their conversion into the motile fibroblastic cells that are integral to embryonic development, and these cells can migrate from their original position to establish new colonies. A number of reports have supported EMT as an essential mechanism that prompts the detachment of cancer cells from a primary site and permits their migration, invasion, and metastasis.[6, 7] However, several reports have demonstrated that N-cadherin expression is more important for cancer metastasis than the expression of E-cadherin and or EMT inducers.[8, 9]

Previously, we described EMT in a cholangiocarcinoma cell line using recombinant transforming growth factor β1 (TGF-β1) treatment. Transforming growth factor β1 induces a decrease in the expression of E-cadherin and an increase in N-cadherin expression, promoting cell motility. Therefore, we hypothesized that the switch from E-cadherin to N-cadherin was important for EMT in EHCC.[10]

Recently, several reports have suggested that mesoblastic developmental control genes, which contribute to the development of blood vessels and bone tissue, promote EMT and enhance the ability of cancer cells to invade and metastasis. Forkhead box protein C2 (FOXC2), also known as mesenchyme forkhead 1 is a gene encoding a transcription factor that controls the generation of mesodermal tissue such as vascular tissue and lymphatic tissue.[11, 12] FOXC2 has been reported to be involved in the EMT,[13-15] in tumor angiogenesis,[16, 17] and in various cancers. For example FOXC2 expression induces EMT in breast cancer,[13] and high expression of FOXC2 is related to a poor prognosis and cancer progression in esophageal cancer.[18] However, the relationship between FOXC2 expression and clinicopathological factors in EHCC has not yet been investigated.

The purpose of this study was to clarify the function of FOXC2 in cholangiocarcinoma cell lines in vitro and to determine the clinical significance of FOXC2 in primary EHCC. To this end, we performed an immunohistochemical analysis to evaluate the relationships between FOXC2 expression and clinicopathological factors in clinical EHCC samples. We also examined the in vitro effects of siRNA-mediated FOXC2 suppression on the proliferation, migration, and invasion of human EHCC cell lines.

Materials and Methods

Patients and samples

An immunohistochemical analysis was performed using samples from 77 EHCC patients who had undergone potentially curative surgery at our departments between 1995 and 2011. The ages of the patients ranged from 43 to 84 years. The tumor stage was classified according to the seventh tumor-node-metastasis (TMN) classification of the Union for International Cancer Control (UICC). All patients signed written informed consent forms as required by our institutional guidelines.

Immunohistochemical staining

Immunohistochemical staining of tissue sections for FOXC2 expression was performed using a standard streptavidin-biotin-peroxidase-complex method. Each 4-μm section was deparaffinized, rehydrated, and incubated with fresh 0.3% hydrogen peroxide in methanol for 30 min at room temperature to block endogenous peroxidase activity. The sections were then heated in a microwave at 700 W for 7.5 min in 10 mM citrate buffer (pH 6.5) and then cooled to 30°C. Nonspecific binding sites were blocked by incubation with 10% rabbit pre-immune serum for 30 min. The sections were then incubated with anti-FOXC2 primary monoclonal antibodies (Abnova, Taipei, Taiwan) at a dilution of 1:100 at 4°C overnight. The sections were washed in PBS and, incubated with biotinylated anti-mouse IgG for 60 min at room temperature (Nichirei, Tokyo, Japan). The chromogen 3,3′-diaminobenzidine tetrahydrochloride was applied as a 0.02% solution containing 0.005% H2O2 in 50 mM ammonium acetate-citrate acid buffer (pH 6.0). The sections were lightly counterstained with Mayer's haematoxylin and mounted. Negative controls were established by replacing the primary antibody with rabbit pre-immune serum; no detectable staining was evident in the negative controls.

The immunohistochemical evaluation of FOXC2 expression was confirmed independently by two observers. As FOXC2 was strongly expressed in advanced esophageal carcinomas[18]; therefore, these samples were used as a positive control (Fig. S1). Because FOXC2 was thought to play a role in invasion and metastasis, its expression was evaluated at the invasion front of EHCC tissues. In positive cases, staining was primarily cytoplasmic. The intensity of FOXC2 staining was scored as 0, 1+, or 2+. Additionally, tumor cells within the tissue sample were counted, and the percentages of positively stained cells were determined. All samples were then categorized as grade 0, 1, 2, or 3 according to the criteria presented in Table S1. Grade 0 staining was considered to be negative for FOXC2 expression, while grades 1, 2, and 3 were considered to be positive.

Cell culture

The human EHCC cell lines HuCCT-1 and TFK-1 and the human intrahepatic cholangiocarcinoma cell line HuH-28 were used in this study. All cells were obtained from RIKEN BRC through the National Bio-Resource Project of MEXT, Tokyo, Japan.

Protein extraction and western blot analysis

All cells were harvested at 80% confluence, and the total protein was extracted using the PRO-PREP Protein Extraction Solution Kit (iNtRON Biotechnology, Kyungki-Do, Korea). The total proteins were separated on 4–12% Bis-Tris Mini Gels (Life Technologies, Carlsbad, CA, USA) and transferred to membranes using an iBlot Dry Blotting System (Life Technologies). The membranes were incubated overnight at 4°C with mouse monoclonal antibodies against FOXC2 (1:400; Abnova), TGF-β1 (1:1000; R&D System, Minneapolis, MN, USA), Snail (1:500; WuXi Apptec, St Paul, MN, USA), E-cadherin (1:1000; Takara, Shiga, Japan), N-cadherin (1:500; Invitrogen, Carlsbad, CA, USA), matrix metalloproteinase-2 (MMP-2) (1:1000 Cell Signaling, Danvers, MA, USA), angiopoietin-2 (Ang-2) (1:500; Abcam, Cambridge, UK) and β-actin (1:1000; Sigma, St. Louis, MO, USA). The membranes were next treated with horseradish peroxidase-conjugated anti-mouse secondary antibodies, and the proteins were detected using the ECL Prime Western Blotting Detection System (GE Healthcare, Tokyo, Japan).

RNA extraction and quantitative real-time RT-PCR

Total RNA was extracted from tissues and cells using the miRN-easy Kit (Qiagen, Hilden, Germany), and the quantity of isolated RNA was measured using an ND-10000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). Quantitative real-time RT-PCR (RT-qPCR) was performed using the GoTaq 1-Step RT-qPCR System (Promega, Madison, WI, USA) in a total volume of 20 μL. The program included four stages: reverse transcription at 37°C for 15 min; RT inactivation and hot-start activation at 95°C for 10 min; qPCR, of 40 cycles of 95°C for 10 s, 60°C for 30 s, and 72°C for 30 s; and dissociation at 60–95°C. The sequences of the primer pairs were as follows: FOXC2 forward, 5′-CCGAGAAGAAGATCACCTTGAA-3′; FOXC2 reverse, 5′-GACACGTCCTTCTTTTTGAAGC-3′; GAPDH forward, 5′-AAGGTGAAGGTCGGAGTCAAC-3′; and GAPDH reverse, 5′-CTTGATTTTGGAGGGATCTCG-3′.

Small-interfering RNA (siRNA) transfection

Forkhead box protein C2specific siRNA (Silencer Pre-designed siRNA) was purchased from Ambion (Austin, TX, USA). HuCCT-1 cells were seeded in 6-well, flat-bottomed microtitre plates at a density of 1 × 105 cells per well in a volume of 2 mL and incubated in a humidified atmosphere (37°C and 5% CO2). After incubation, 500 μL Opti-MEM I Reduced Serum Medium (Invitrogen), 5 μL Lipofectamine RNAi MAX (Invitrogen), and 3.75 μL FOXC2-specific siRNA (30 nM final concentration in each well) were mixed and incubated for 20 min to form chelate bonds. The siRNA reagents were then added to the cells. The experiments were performed after 24–72 h of incubation.

Wound healing assay

We examined migration using HuCCT1 cells that had been transfected with negative control or FOXC2 siRNA. Transfected HuCCT-1 cells were grown in 6-well plates until confluence, and a uniform straight wound was produced in the monolayer in each well using a pipette tip. The wells were washed with PBS to remove all the cell debris, and the cells were cultured in 5% CO2 at 37°C. Closure or filling in of the wound was evaluated at 24 h using bright-field microscopy (Nikon TMS; Nikon, Tokyo, Japan; 40× magnification).

Invasion assay

Cell invasion assays were performed using 24-well BD BioCoat Matrigel Invasion Chambers (Becton Dickinson, San Jose, CA, USA) to evaluate the cellular invasion ability. HuCCT-1 cells (5.0 × 104) were seeded in the upper chamber, and the lower chamber was filled with 750 μL of RPMI1960 supplemented with 10% FBS as a chemoattractant. After 48 h of incubation at 37°C, the cells were fixed with 70% ethanol and stained with Diff-Quik. The cells that invaded through the pores to the lower surface of the filter were counted using microscope. A total of 10 random fields were evaluated in triplicate assays.

Statistical analysis

The data for the continuous variables are expressed as the means ± SEM. The significance of the differences between values was determined using Student's t-tests. Statistical analysis of the immunohistochemical staining was performed using the Wilcoxon signed-rank test and χ2 test. Survival curves for the patients were calculated using the Kaplan–Meier method and analyzed using the log-rank test. Prognostic factors were examined by univariate and multivariate analyses using a Cox proportional hazards model. All differences were deemed significant at < 0.05, and all statistical analyses were performed with jmp software, version 5.01 (SAS Institute, Cary, NC, USA).


Immunohistochemical staining of FOXC2 in EHCC tissues

The expression of FOXC2 was evaluated by immunohistochemistry in 77 EHCC samples. Overall, 59 EHCC samples (76.7%) were negative for FOXC2 expression (Fig. 1a), whereas 18 (23.3%) were positive for FOXC2 expression, with mainly cytoplasmic staining being mainly observed (Fig. 1b). In addition, FOXC2 expression tended to be enhanced at the invasive front (Fig. 1c,d).

Figure 1.

Immunohistochemical staining of Forkhead box protein C2 (FOXC2)in primary extrahepatic cholangiocarcinoma (EHCC) samples. (a) An example of low FOXC2 expression in a primary EHCC specimen (200×). (b) An example of high FOXC2 expression in a primary EHCC specimen (200×). (c) FOXC2 expression tended to increase near the invasive front (arrow), as visualised in a low-power field (50×) (d) and a high-power field (200×, arrow).

Expression of FOXC2 in EHCC tissues, and its correlation with clinicopathological findings

The correlations between FOXC2 expression in the EHCC samples and the patients’ ages, genders, tumor stages, lymph node metastasis, lymphatic invasion, venous invasion, nerve invasion, infiltrating types, and TMN stages are shown in Table 1. FOXC2-positive expression was correlated with the progression of lymph node metastasis (= 0.0205), lymphatic invasion (= 0.0478), and TMN stage (= 0.0009).

Table 1. Clinicopathological characteristics of extrahepatic cholangiocarcinoma (EHCC) patients according to Forkhead box protein C2 (FOXC2) expression
FactorFOXC2 low expressionFOXC2 high expressionP-value
= 59= 18
  1. a

    < 0.05.

Tumor stage
Lymph node metastasis
Lymphatic invasion
Venous invasion
Nerve invasion (pn)
TMN stage (UICC)
I II2920.0009a
III IV3019

Prognostic significance of FOXC2 expression in EHCC

The postoperative recurrence-free and cancer-specific survival rates of the EHCC patients are shown in Figure 2. The FOXC2-positive group had significantly poorer prognoses than the FOXC2-negative group, considering both recurrence-free survival (= 0.0034) and cancer specific survival (= 0.0002). For cancer-specific survival, FOXC2 expression was a prognostic factor for poor survival in the univariate analysis (Table 2; = 0.0009), tumor stage, lymph node metastasis and lymphatic invasion were also prognostic factors in the univariate analysis. In the multivariate analysis, FOXC2 expression was also a prognostic factor for poor survival (Table 2; = 0.0289).

Table 2. Univariate and multivariate analysis of prognostic factors using Cox proportional hazards model
FactorUnivariate analysisMultivariate analysis
RR95% CIP-valueRR95% CIP-value
  1. a

    < 0.05. 95% CI, 95% confidence interval.

Age (≤65/>66)1.080.76–1530.652
Sex (M/F)0.6880.44–1.010.0559
Tumor stage (T1–2/3–4)3.291.61–7.120.001a2.541.19–5.720.0148a
Lymph node metastasis (−/+)2.061.43–3.060.0001a1.651.11–2.510.0117a
Lymphatic invasion (−/+)1.991.10–4.980.00204a1.290.66–3.310.4880
Venous invesion (−/+)1.590.99–2.930.0527
Nerve invasion (−/+)1.600.71–6.810.297
FOXC2 expression (−/+)3.681.75–7.530.0009a2.381.09–5.040.0289a
Figure 2.

Relationships between postoperative survival and Forkhead box protein C2 (FOXC2) expression. Kaplan–Meier curves of the low FOXC2 expression and high FOXC2 expression groups are shown. (a) High FOXC2 expression indicated a poor prognosis for recurrence free survival (= 0.0034). (b) High FOXC2 expression also indicated a poor prognosis for cancer specific survival (P = 0.0002).

Expression of FOXC2 in cholangiocarcinoma cell lines, and depletion of FOXC2 using siRNA in HuCCT-1 cells

Forkhead box protein C2 expression was observed in all cholangiocarcinoma cell lines (i.e., HuCCT-1, TFK-1, and HuH-28, Fig. 3a). We used siRNA to knockdown FOXC2 expression in the EHCC cell line HuCCT-1 to determine the contribution of FOXC2 to invasion, migration, and proliferation. The suppression of FOXC2 by siRNA1 and siRNA2 was demonstrated by both RT-qPCR and western blotting (Fig. 3b,c).

Figure 3.

Suppression of Forkhead box protein C2 (FOXC2)by siRNA. (a) FOXC2 expression in each EHCC cell lines. (b) FOXC2 mRNA expression in HuCCT-1 cells transfected with FOXC2 siRNA was measured by RT-qPCR and compared to the expression in parent and negative control cells. (c) The FOXC2 protein levels were measured by western blotting after transfection with FOXC2 siRNAs. The expression of other proteins in FOXC2 siRNA-transfected and untreated HuCCT-1 cells was assessed by western blot analysis. *P < 0.05.

Depletion of FOXC2 altered E-cadherin, N-cadherin, MMP-2 and Ang-2 expression in HuCCT-1 cells

Next, we examined the relevance of FOXC2 to the expression of EMT markers (i.e., E-cadherin, N-cadherin, and MMP-2) and lymphangiogenic factors (i.e., Ang-2, and VEGF-C). The depletion of FOXC2 significantly decreased N-cadherin, MMP-2, and Ang-2 expression in HuCCT-1 cells, whereas E-cadherin expression was slightly increased. In contrast, the expression levels of TGFβ1 and Snail were unaltered (Fig. 3c).

FOXC2 regulated migration and invasion in HuCCT-1 cells

Lastly, we assessed the role of FOXC2 in cell migration and invasion. As revealed in the wound-healing assay, FOXC2 knockdown suppressed cell migration in comparison to the migration of the parent and negative-control cells (< 0.05; Fig. 4). Similarly, FOXC2 knockdown significantly reduced cell invasiveness when compared with the invasiveness of the parent and negative-control cells (< 0.05; Fig. 5). The cellular proliferation ability was evaluated using the water-soluble tetrazolium (WST) assay, but the FOXC2 knockdown cells did not exhibit alter proliferation ability compared to the parent and negative control cells (data not shown).

Figure 4.

Wound healing assay. HuCCT-1 cells were transfected with Forkhead box protein C2 (FOXC2)siRNA, and wound healing assays were performed as described to the measure migration activity in comparison to untransfected cells. *P < 0.05.

Figure 5.

Invasion assay. HuCCT-1 cells were trans-fected with Forkhead box protein C2 (FOXC2)siRNA, and invasion assays were performed. The invasion of transfected cells was compared with that of untransfected cells. *P < 0.05.


In this study, we showed that the high expression of FOXC2 in primary EHCC samples was associated with cancer progression and a poor prognosis. In our in vitro FOXC2 suppression analysis, reduced invasive and migratory capacities were observed in FOXC2 siRNA-transfected cells compared with control cells. Moreover, FOXC2 suppression decreased N-cadherin, MMP-2, and Ang-2 expression and increased E-cadherin expression.

In the immunohistochemical analysis, a high level of FOXC2 expression also tended to be associated with local invasion, an effect that was not observed in the low-expression group. Indeed, nearly all the cancer cells at the invasive front of the stroma exhibited high FOXC2 expression. Mani et al.[13] reported that high FOXC2 expression in cancer cells activates other invasion factors, such as MMP-2. Moreover, Nishida et al.[18] reported that FOXC2 expression was positively correlated with MMP2 expression in clinical samples of esophageal cancer. These data proved that FOXC2 regulates the transcriptional activity of MMP-2. In the present study, FOXC2 suppression decreased the expression of MMP-2 in vitro. Indeed, FOXC2 may regulate MMP-2 to trigger invasion in EHCC.

The patients in the high FOXC2 expression group exhibited significant increases in lymph node metastasis and lymphatic invasion as compared to the patients in the low-expression group. FOXC2 may play a key role in lymphangiogenesis,[19] and Kume reported that FOXC2 functions as a transcription factor during lymphangiogenesis and induces such lymphangiogenic factors, as angiopoietin-2 (Ang-2).[11, 16] In addition, FOXC2 mutations have been identified in patients with lymphedema-distichiasis, and Kriederman et al.[12] reported that FOXC2-knockout mice exhibited lymphedema-distichiasis syndrome. In the present study, FOXC2 knockdown in EHCC cells decreased Ang-2 expression. In a study of lung cancer by Yu YH et al.[20] and in a study of esophageal cancer by Nishida et al.[18] high FOXC2 expression was correlated to lymphatic metastasis, as based on clinicopathological assessments. In our study, FOXC2 expression showed a significant difference in lymph node metastasis in the univariate and multivariate analyses (Table S2). FOXC2 is an independent determinant of lymph node metastasis. Although it is still unclear how FOXC2 expression specifically contributes to lymphangiogenesis, these data suggested that FOXC2 plays an important role in lymphatic invasion and lymph node metastasis via Ang-2 induction in EHCC.

The FOXC2 gene has also been reported to control the generation of mesodermal tissue, such as blood vessels and bone tissue, and FOXC2 has been reported to contribute to vascularization, lymphangiogenesis, and wound healing in normal tissue.[14, 19, 21-23] Therefore, in our IHC study, FOXC2 expression was also observed in intestinal cells, such as endothelial cells. Recently, studies have clarified that FOXC2 is involved in the invasion and metastasis of cancer cells.[24] FOXC2 is activated by EMT triggering factors, such as TGF-β,[25] and functions as a transcription factor that promotes the expression of various mesenchymal proteins.[13, 26, 27] In the present study, FOXC2 knockdown in EHCC cells caused a reduction in motility and invasion. However, there was no change in the proliferation rate following FOXC2 knockdown (data not shown), indicating that FOXC2 did not mediate proliferation, but instead induced motility and invasion. Moreover, our western blotting results demonstrated that the suppression of FOXC2 expression increased the expression of E-cadherin but decreased the expression of N-cadherin and, MMP-2. In contrast, FOXC2 knockdown had no effect on TGF-β1 and Snail expression. Mani et al.[13] reported that the administration of EMT inducing factors (i.e., TGF-β1 and, Snail) triggered an increase in FOXC2 expression, causing the cells to change their morphology into a spindle shape and enhancing invasive capacity. FOXC2 expression is positively correlated with the expression of N-cadherin, vimentin, and MMP-2, but inversely correlated with the expression of E-cadherin. These previous studies, combined with the results of our study, suggest that FOXC2 is located in downstream of EMT inducers (TGF-β1 and, Snail) and is involved in the induction of EMT by regulating N-cadherin and MMP-2 in EHCC.

In conclusion, the EMT inducer FOXC2 contributed to a shorter duration of recurrence-free survival and reduced cancer specific survival in patients with EHCC. The evaluation of FOXC2 expression in EHCC might be a useful predictor of recurrence and a poor prognosis. Moreover, FOXC2 expression was associated with lymphatic invasion and lymph node metastasis in clinical EHCC samples, and FOXC2 suppression regulated ability for migration and invasion in EHCC cell lines. Our results suggest that FOXC2 may be a promising molecular target for regulating EHCC metastasis.


We thank Ms Yukie Saito, Ms Tomoko Yano, Ms Tomoko Ubukata, Ms Midori Ohno, Ms Sayaka Kosaka, Ms Yuka Matsui, Ms Akaya Ishida and Ms Rieko Notegi for their excellent assistance.

Disclosure Statement

The authors have no conflict of interest.