Cancer Letters 353 (2014) 87–94
Contents lists available at ScienceDirect
Cancer Letters journal homepage: www.elsevier.com/locate/canlet
FOXC2 promotes colorectal cancer proliferation through inhibition of FOXO3a and activation of MAPK and AKT signaling pathways Yan-Mei Cui a,b,1, Dan Jiang d,1, Shi-Hong Zhang c,1, Ping Wu a,b, Ya-Ping Ye a,b, Cui-Min Chen a,b, Na Tang a,b, Li Liang a,b, Ting-Ting Li a,b, Lu Qi a,b, Shu-Yang Wang a,b, Hong-Li Jiao a,b, Jie Lin a,b, Yan-Qing Ding a,b, Wen-Ting Liao a,b,⇑ a
Department of Pathology, Nanfang Hospital, Southern Medical University, Guangzhou 510515, People’s Republic of China Department of Pathology, School of Basic Medical Sciences, Southern Medical University, Guangzhou 510515, People’s Republic of China Department of Laboratory Medicine, The First Affiliated Hospital of Sun Yat-sen University, Guangzhou, Guangdong, China d Department of Pathology, West China Hospital, Sichuan University, Chengdu 610041, People’s Republic of China b c
a r t i c l e
i n f o
Article history: Received 27 February 2014 Received in revised form 22 June 2014 Accepted 7 July 2014
Keywords: FOXC2 Colorectal cancer Prognosis Nuclear localization Proliferation
a b s t r a c t Abnormal expression of FOXC2 has been found in several human cancers. However, the role of FOXC2 in the progression of colorectal cancer (CRC) has not been well characterized. In analysis of 206 CRC specimens, we revealed that both high expression and nuclear localization of FOXC2 were correlated to aggressive characteristics and poor survival of patients with CRC. FOXC2 promoted cell proliferation through activation of MAPK and AKT pathways, subsequently down-regulating p27, up-regulating cyclin D1 and p-FOXO3a. Furthermore, FOXC2 nuclear localization was required for its promotion of cell proliferation. These findings suggest that FOXC2 plays an essential role in CRC progression and may serve as a valuable clinical prognostic marker of this disease. Ó 2014 Elsevier Ireland Ltd. All rights reserved.
Introduction Colorectal cancer (CRC) is one of the most common malignancies in the world and is a multi-step process involving progressive disruption of epithelial-cell proliferation, apoptosis, differentiation, and survival mechanisms [1,2]. Despite of the advancements in diagnostic and therapeutic strategies, the clinical outcome and prognosis of CRC patients remain pessimistic [3]. Thus, understanding the mechanisms involved in the initiation and progression of CRC is a pivotal step for generation of more effective therapies. Although numerous genetic alterations have been well identified to be involved in the development of CRC, the other genetic and epigenetic changes responsible for this disease remain largely unknown. Mesenchyme forkhead 1 (also known as FOX protein C2, FOXC2) belongs to the FOX-containing transcription factor family, which is composed of at least 43 members. The FOX gene family possess a winged-helix DNA-binding domain known as the ⇑ Corresponding author at: Department of Pathology, Nanfang Hospital, Southern Medical University, Guangzhou 510515, People’s Republic of China. Tel.: +86 (20) 6164 8224; fax: +86 (20) 6164 2148. E-mail address: [email protected] (W.-T. Liao). 1 Equal contributors. http://dx.doi.org/10.1016/j.canlet.2014.07.008 0304-3835/Ó 2014 Elsevier Ireland Ltd. All rights reserved.
forkhead box, and play multiple important roles in development, metabolism, immunology and longevity [4]. It has been reported that FOXC2 directly induces the transcription of chemokine (C– X–C motif) receptor 4 (CXCR4) and integrin-b3 by activating their promoters and is functionally associated with angiogenesis and lymphangiogenesis [5]. FOXC2 can also regulate osteoblast differentiation through induction of wingless-related MMTV integration site 5A (WNT5a) [6]. These findings suggest that FOXC2 could enable cancer cells to acquire invasive and metastatic properties. Abnormal expression of FOXC2 has been found in several types of human cancers, including breast cancer [7,8], esophageal cancer [9], gastric cancer [10], and non-small-cell lung cancer [11]. However, the clinicopathologic significance of FOXC2 protein and the role of FOXC2 in the progression of CRC have not been well characterized. Here, we sought to investigate the expression patterns and the potential role of FOXC2 in the development and progression of CRC. We reported here that both expression levels and nuclear localization of FOXC2 are correlated to aggressive characteristics (advanced clinical stage, T and N classifications, distant metastasis-positive tumors, and high proliferation index) and with poor survival of patients. In addition, FOXC2 might promote proliferation, tumor growth and invasion of CRC partly through regulating MAPK and AKT signaling pathways.
88
Y.-M. Cui et al. / Cancer Letters 353 (2014) 87–94
Materials and methods Patients and specimens A total of 206 archived, formalin-fixed paraffin-embedded human colorectal carcinoma samples were obtained from the Department of Pathology, NanFang hospital, Southern Medical University, China. All of the cases were clinically and histologically diagnosed between 2002 and 2005. The stage of disease was determined according to the tumor size, lymph node and metastasis (pTNM) classification system [12]. The patients included 114 males and 92 females, ranging in age from 18 to 73 years-old (mean, 56 years). The median follow-up time for overall survival was 64.2 months for patients still alive at the time of analysis (ranged, 3–83 months). A total of 74/206 (35.9%) patients died during follow-up. The 10 freshly collected colorectal cancer tissues and paired normal mucosal tissue specimens taken from sites distant to the cancerous lesion were obtained from patients with CRC undergone surgical resection, frozen and stored in liquid nitrogen till further use. Immunohistochemistry Immunohistochemistry (IHC) staining and scoring were done as previously described [13]. For details, please see Supplementary Methods.
Results FOXC2 is up-regulated in CRC Western blotting analysis and real-time PCR revealed that all nine CRC cell lines tested, including SW480, HT29, HCT116, Km12, SW620, Caco-2, HCT15, Ls174t, and DLD1 cells, exhibited varying levels of FOXC2 expression. Interestingly, FOXC2 expression was relatively lower in less aggressive cell lines HT29 and LS174t than that in aggressive cell lines HCT116 and SW620 (Fig. 1A) [16,17]. Western blot showed that the expression of FOXC2 protein was significantly up-regulated in all ten CRC tissues (T) compared with their adjacent normal intestine epithelial tissues (N) (Fig. 1B, left). Real-time RT–PCR revealed that the tumor/normal (T/N) ratio of FOXC2 mRNA level was >5-fold in all samples with the highest ratio up to 25-fold (Fig. 1B, right). In addition, immunohistochemical staining in these ten paired tissues further confirmed that FOXC2 was up-regulated in CRC tissues (Supplementary Fig. S1).
Cell culture and plasmids The human CRC cell lines (SW480, Caco-2, HCT116, KM12, SW620, HT29, HCT15, Ls174t, and DLD1) were originally purchased from the American Type Culture Collection (Manassas, VA, USA). SW620 and Caco-2 cells were cultured in DMEM medium (Invitrogen, Carlsbad, CA, USA) supplemented with 10% FBS (PAA Laboratories, Pasching, Austria) and 1% penicillin/streptomycin (Invitrogen). The FOXC2 constructs were generated by cloning PCR-amplified full-length or deletion mutant (del NLS) human FOXC2 cDNA into pBabe. To knockdown FOXC2, the human shRNA sequence (CCACACGTTTGCAACCCAA), as described by Mani [8], was cloned into pSuper-retro-neo (Oligo-Engine, Seattle, WA) to generate pSuperretro-FOXC2-shRNA. Real-time RT–PCR and western blotting analyses Real-time RT–PCR was carried out using the ABI PRISM 7500 Sequence Detection System (Applied Biosystems). See Supplementary Methods for further details. Western blotting was performed as previously described [14], using anti-FOXC2 (Bethyl Laboratories, Inc, MT), anti-cyclinD1 (BD Biosciences, San Diego, CA, USA), anti-Bmi-1 (Cell Signaling Technology), anti-p27, anti-p-ERK, anti-ERK, pAKT, anti-AKT, p-FOXO3a and FOXO3a antibodies (Bioworld Technology, Louis Park, MN, USA). The membranes were stripped and re-probed with a mouse anti-a-Tubulin monoclonal antibody (Sigma, Saint Louis, MO, USA) as a loading control. MTT assay, colony formation assay, soft agar assay MTT assay, colony formation assay and soft agar assay were performed as previously described [15]. For details, please see Supplementary methods. Tumorigenesis in nude mice Xenograft tumors were generated by subcutaneous injection of cells (2 106 for HT29/Vector and HT29/FOXC2 (n = 6), 5 105 for SW620/Scramble and SW620/shFOXC2, n = 5) on the hindlimbs of each 4–6 week old Balb/C athymic nude mouse (nu/nu) obtained from the Animal Center of Southern Medical University, Guangzhou, China. All mice were housed and maintained under specific pathogen-free conditions and used in accordance with institutional guidelines and approved by the Use Committee for Animal Care. Tumor size was measured by a slide caliper and tumor volume was determined by the formula 0.44 A B2 (A indicates tumor base diameter one direction and B the corresponding perpendicular value). The tumors were fixed and 4 lm sections were cut and stained with haematoxylin and eosin according to standard protocols. Sections were further under IHC staining using antibody against Ki-67. Statistical analysis All statistical analyses were carried out using SPSS version 13.0. Mann–Whitney U tests were used to analyze the relationship between FOXC2 expression and the clinicopathologic features of CRC. Survival curves were plotted by the Kaplan–Meier method and compared using the log-rank test. The Cox proportional hazard model was used to calculate relative risk ratios. Univariate and multivariate survival distributions were compared using the log-rank test. P < 0.05 was considered significant.
Increased FOXC2 expression was associated with progression and poor prognosis in CRC To investigate the role of FOXC2 in the clinical progression of CRC, we examined expression and localization of FOXC2 protein by immunohistochemistry (IHC) in 206 paraffin-embedded, archival CRC tissues. As shown in Table 1 and Fig. 1C, FOXC2 protein was detected in 195 of 206 (94.6%) cases of tissue samples, 99 cases displayed low FOXC2 expression and 107 cases displayed high FOXC2 expression (Table 1). In addition, weak to moderate staining of FOXC2 was observed in the areas of adenoma. Interestingly, FOXC2 protein displayed two main expression patterns in adenocarcinomas: cytoplasmic localization (76/195, 39%) and nuclear localization (119/195, 61%). Mann–Whitney U tests revealed that the FOXC2 expression level strongly correlated with the clinical stage (P = 0.017), T classification (P = 0.015), lymph node metastasis (P = 0.004), and distant metastasis in human CRC (P = 0.021) in this cohort of 206 cases of CRC (Table 1). Kaplan–Meier survival analysis indicated that patients with high FOXC2 level had significantly poorer overall survival (P = 0.002) and disease-free survival (P = 0.006, Fig. 1D, left panels) than patients with low FOXC2 level. We also analyzed the clinicopathologic significance of the localization of FOXC2 in CRC. Table 1 showed that high expression level of FOXC2 was significantly associated with the nuclear localization of FOXC2 (P = 0.002). Moreover, nuclear localization of FOXC2 correlated significantly with poor tumor differentiation (P = 0.01), advanced clinical stage (P = 0.001), T classification (P = 0.001), N classification (P = 0.013) and distant metastasis (P = 0.013). Kaplan–Meier survival analysis confirmed that patients with FOXC2 nuclear localization had poorer overall survival (P = 0.011) and disease-free survival (P = 0.002) than patients with FOXC2 cytoplasmic localization (Fig. 1D, right panels). Cox regression analyses revealed that both FOXC2 expression levels and subcellular localization were significant prognostic factors in CRC (Supplementary Tables S1–S3). Additionally, there was a significant correlation between the Ki67 labeling index and either FOXC2 expression levels (P = 0.014, Table 1) or subcellular localization (P = 0.006) in CRC. Samples that had lower level of FOXC2 expression also had a lower Ki-67 labeling index, whereas samples that had higher level of FOXC2 expression had a higher Ki-67 labeling index (Fig. 2). These observations suggest that increased expression of FOXC2 or nuclear localization of FOXC2 was closely associated with aggressive phenotypes of CRC.
Y.-M. Cui et al. / Cancer Letters 353 (2014) 87–94
89
Fig. 1. FOXC2 was up-regulated and associated with progression and poor prognosis in CRC. (A and B) Expression of FOXC2 protein and mRNA in CRC cell lines (A) and ten primary CRC (T) and adjacent noncancerous tissues (N) paired from the same patient (B) by western blotting and real-time PCR. Expression levels were normalized with aTubulin or GAPDH. Error bars represent mean ± SD calculated from 3 parallel experiments. (C) Representative images of FOXC2 expression in normal intestinal epithelium, adenoma, and CRC specimens examined by IHC. (D) Influence of FOXC2 expression patterns on overall survival and disease-free survival by Kaplan–Meier analyses.
Overexpression of FOXC2 promotes the proliferation and tumorigenesis of human CRC cells To investigate whether FOXC2 plays a role in the proliferation of CRC cells, we established stable FOXC2-expressing CRC cells (HT29/FOXC2 and Ls174t/FOXC2) (Fig. 3A). Overexpression of FOXC2 accelerated the cell growth (Fig. 3B and C), and strongly enhanced the migratory ability of CRC cells (Supplementary Fig. S2), as compared with control cells. We next investigated the role of FOXC2 in tumorigenesis of CRC cells. Overexpression of FOXC2 in HT29 and Ls174t promoted cells to grow in soft agar and form more colonies in comparison with control cells (Fig. 3D). To confirm this effect in vivo, we performed tumorigenesis assays in nude mice by subcutaneous injection of control cells (HT29/pBabe) and HT29/FOXC2 cells. In contrast to control cells, the HT29/FOXC2 cells exhibited more
rapid tumor growth (Fig. 3E, upper) and remarkably larger tumor volumes (Fig. 3E, lower). Of note, IHC staining of the tumor histological sections revealed that tumors developed by HT29–FOXC2 cells presented a distinctly high expression of Ki-67, while tumors from control group just had a few scattered Ki-67-positive cells (Fig. 3F). To explore the association of FOXC2 subcellular localization with its role in tumor proliferation, we deleted the nuclear localization sequence (NLS) of FOXC2 to see if this deletion mutant could still promote cell proliferation. Immunofluorescence analysis showed that FOXC2 protein could not locate in the nuclear after deletion of the NLS. Transfection of the deletion mutant failed to accelerate cell proliferation of HT29 cells determined by MTT and soft agar assays (Supplementary Fig. S3), suggesting that FOXC2 nuclear localization is required for the promotion of tumor proliferation.
90
Y.-M. Cui et al. / Cancer Letters 353 (2014) 87–94
Table 1 Correlation between clinicopathologic features and FOXC2 expression levels and localization. Characteristics
FOXC2 levels
P values
Low
High
Age 6mean(56) >mean (56)
50 49
43 64
Gender Male Female
56 43
Histology Columnar adenocarcinoma Mucinous adenocarcinoma Others
FOXC2 localization
P values
Cytoplasm
Nuclear
0.138
40 36
47 72
0.073
58 49
0.734
42 34
69 50
0.709
78 11 10
89 11 7
0.393
64 6 6
97 14 8
0.682
Differentiation Well and moderate Poor and undifferentiated
83 16
87 20
0.643
69 7
90 29
0.010
Dukes stage A B C D
9 45 21 24
6 39 29 33
0.017
8 38 13 17
5 38 29 47
0.001
T stage 1–2 3 4
22 57 20
16 53 38
0.015
20 42 14
14 62 43
0.001
N stage 0 1–3
68 31
52 55
0.004
52 24
60 59
0.013
Distant metastasis 0 (no) 1 (yes)
75 24
65 42
0.021
Ki-67 labeling index
Contents lists available at ScienceDirect
Cancer Letters journal homepage: www.elsevier.com/locate/canlet
FOXC2 promotes colorectal cancer proliferation through inhibition of FOXO3a and activation of MAPK and AKT signaling pathways Yan-Mei Cui a,b,1, Dan Jiang d,1, Shi-Hong Zhang c,1, Ping Wu a,b, Ya-Ping Ye a,b, Cui-Min Chen a,b, Na Tang a,b, Li Liang a,b, Ting-Ting Li a,b, Lu Qi a,b, Shu-Yang Wang a,b, Hong-Li Jiao a,b, Jie Lin a,b, Yan-Qing Ding a,b, Wen-Ting Liao a,b,⇑ a
Department of Pathology, Nanfang Hospital, Southern Medical University, Guangzhou 510515, People’s Republic of China Department of Pathology, School of Basic Medical Sciences, Southern Medical University, Guangzhou 510515, People’s Republic of China Department of Laboratory Medicine, The First Affiliated Hospital of Sun Yat-sen University, Guangzhou, Guangdong, China d Department of Pathology, West China Hospital, Sichuan University, Chengdu 610041, People’s Republic of China b c
a r t i c l e
i n f o
Article history: Received 27 February 2014 Received in revised form 22 June 2014 Accepted 7 July 2014
Keywords: FOXC2 Colorectal cancer Prognosis Nuclear localization Proliferation
a b s t r a c t Abnormal expression of FOXC2 has been found in several human cancers. However, the role of FOXC2 in the progression of colorectal cancer (CRC) has not been well characterized. In analysis of 206 CRC specimens, we revealed that both high expression and nuclear localization of FOXC2 were correlated to aggressive characteristics and poor survival of patients with CRC. FOXC2 promoted cell proliferation through activation of MAPK and AKT pathways, subsequently down-regulating p27, up-regulating cyclin D1 and p-FOXO3a. Furthermore, FOXC2 nuclear localization was required for its promotion of cell proliferation. These findings suggest that FOXC2 plays an essential role in CRC progression and may serve as a valuable clinical prognostic marker of this disease. Ó 2014 Elsevier Ireland Ltd. All rights reserved.
Introduction Colorectal cancer (CRC) is one of the most common malignancies in the world and is a multi-step process involving progressive disruption of epithelial-cell proliferation, apoptosis, differentiation, and survival mechanisms [1,2]. Despite of the advancements in diagnostic and therapeutic strategies, the clinical outcome and prognosis of CRC patients remain pessimistic [3]. Thus, understanding the mechanisms involved in the initiation and progression of CRC is a pivotal step for generation of more effective therapies. Although numerous genetic alterations have been well identified to be involved in the development of CRC, the other genetic and epigenetic changes responsible for this disease remain largely unknown. Mesenchyme forkhead 1 (also known as FOX protein C2, FOXC2) belongs to the FOX-containing transcription factor family, which is composed of at least 43 members. The FOX gene family possess a winged-helix DNA-binding domain known as the ⇑ Corresponding author at: Department of Pathology, Nanfang Hospital, Southern Medical University, Guangzhou 510515, People’s Republic of China. Tel.: +86 (20) 6164 8224; fax: +86 (20) 6164 2148. E-mail address: [email protected] (W.-T. Liao). 1 Equal contributors. http://dx.doi.org/10.1016/j.canlet.2014.07.008 0304-3835/Ó 2014 Elsevier Ireland Ltd. All rights reserved.
forkhead box, and play multiple important roles in development, metabolism, immunology and longevity [4]. It has been reported that FOXC2 directly induces the transcription of chemokine (C– X–C motif) receptor 4 (CXCR4) and integrin-b3 by activating their promoters and is functionally associated with angiogenesis and lymphangiogenesis [5]. FOXC2 can also regulate osteoblast differentiation through induction of wingless-related MMTV integration site 5A (WNT5a) [6]. These findings suggest that FOXC2 could enable cancer cells to acquire invasive and metastatic properties. Abnormal expression of FOXC2 has been found in several types of human cancers, including breast cancer [7,8], esophageal cancer [9], gastric cancer [10], and non-small-cell lung cancer [11]. However, the clinicopathologic significance of FOXC2 protein and the role of FOXC2 in the progression of CRC have not been well characterized. Here, we sought to investigate the expression patterns and the potential role of FOXC2 in the development and progression of CRC. We reported here that both expression levels and nuclear localization of FOXC2 are correlated to aggressive characteristics (advanced clinical stage, T and N classifications, distant metastasis-positive tumors, and high proliferation index) and with poor survival of patients. In addition, FOXC2 might promote proliferation, tumor growth and invasion of CRC partly through regulating MAPK and AKT signaling pathways.
88
Y.-M. Cui et al. / Cancer Letters 353 (2014) 87–94
Materials and methods Patients and specimens A total of 206 archived, formalin-fixed paraffin-embedded human colorectal carcinoma samples were obtained from the Department of Pathology, NanFang hospital, Southern Medical University, China. All of the cases were clinically and histologically diagnosed between 2002 and 2005. The stage of disease was determined according to the tumor size, lymph node and metastasis (pTNM) classification system [12]. The patients included 114 males and 92 females, ranging in age from 18 to 73 years-old (mean, 56 years). The median follow-up time for overall survival was 64.2 months for patients still alive at the time of analysis (ranged, 3–83 months). A total of 74/206 (35.9%) patients died during follow-up. The 10 freshly collected colorectal cancer tissues and paired normal mucosal tissue specimens taken from sites distant to the cancerous lesion were obtained from patients with CRC undergone surgical resection, frozen and stored in liquid nitrogen till further use. Immunohistochemistry Immunohistochemistry (IHC) staining and scoring were done as previously described [13]. For details, please see Supplementary Methods.
Results FOXC2 is up-regulated in CRC Western blotting analysis and real-time PCR revealed that all nine CRC cell lines tested, including SW480, HT29, HCT116, Km12, SW620, Caco-2, HCT15, Ls174t, and DLD1 cells, exhibited varying levels of FOXC2 expression. Interestingly, FOXC2 expression was relatively lower in less aggressive cell lines HT29 and LS174t than that in aggressive cell lines HCT116 and SW620 (Fig. 1A) [16,17]. Western blot showed that the expression of FOXC2 protein was significantly up-regulated in all ten CRC tissues (T) compared with their adjacent normal intestine epithelial tissues (N) (Fig. 1B, left). Real-time RT–PCR revealed that the tumor/normal (T/N) ratio of FOXC2 mRNA level was >5-fold in all samples with the highest ratio up to 25-fold (Fig. 1B, right). In addition, immunohistochemical staining in these ten paired tissues further confirmed that FOXC2 was up-regulated in CRC tissues (Supplementary Fig. S1).
Cell culture and plasmids The human CRC cell lines (SW480, Caco-2, HCT116, KM12, SW620, HT29, HCT15, Ls174t, and DLD1) were originally purchased from the American Type Culture Collection (Manassas, VA, USA). SW620 and Caco-2 cells were cultured in DMEM medium (Invitrogen, Carlsbad, CA, USA) supplemented with 10% FBS (PAA Laboratories, Pasching, Austria) and 1% penicillin/streptomycin (Invitrogen). The FOXC2 constructs were generated by cloning PCR-amplified full-length or deletion mutant (del NLS) human FOXC2 cDNA into pBabe. To knockdown FOXC2, the human shRNA sequence (CCACACGTTTGCAACCCAA), as described by Mani [8], was cloned into pSuper-retro-neo (Oligo-Engine, Seattle, WA) to generate pSuperretro-FOXC2-shRNA. Real-time RT–PCR and western blotting analyses Real-time RT–PCR was carried out using the ABI PRISM 7500 Sequence Detection System (Applied Biosystems). See Supplementary Methods for further details. Western blotting was performed as previously described [14], using anti-FOXC2 (Bethyl Laboratories, Inc, MT), anti-cyclinD1 (BD Biosciences, San Diego, CA, USA), anti-Bmi-1 (Cell Signaling Technology), anti-p27, anti-p-ERK, anti-ERK, pAKT, anti-AKT, p-FOXO3a and FOXO3a antibodies (Bioworld Technology, Louis Park, MN, USA). The membranes were stripped and re-probed with a mouse anti-a-Tubulin monoclonal antibody (Sigma, Saint Louis, MO, USA) as a loading control. MTT assay, colony formation assay, soft agar assay MTT assay, colony formation assay and soft agar assay were performed as previously described [15]. For details, please see Supplementary methods. Tumorigenesis in nude mice Xenograft tumors were generated by subcutaneous injection of cells (2 106 for HT29/Vector and HT29/FOXC2 (n = 6), 5 105 for SW620/Scramble and SW620/shFOXC2, n = 5) on the hindlimbs of each 4–6 week old Balb/C athymic nude mouse (nu/nu) obtained from the Animal Center of Southern Medical University, Guangzhou, China. All mice were housed and maintained under specific pathogen-free conditions and used in accordance with institutional guidelines and approved by the Use Committee for Animal Care. Tumor size was measured by a slide caliper and tumor volume was determined by the formula 0.44 A B2 (A indicates tumor base diameter one direction and B the corresponding perpendicular value). The tumors were fixed and 4 lm sections were cut and stained with haematoxylin and eosin according to standard protocols. Sections were further under IHC staining using antibody against Ki-67. Statistical analysis All statistical analyses were carried out using SPSS version 13.0. Mann–Whitney U tests were used to analyze the relationship between FOXC2 expression and the clinicopathologic features of CRC. Survival curves were plotted by the Kaplan–Meier method and compared using the log-rank test. The Cox proportional hazard model was used to calculate relative risk ratios. Univariate and multivariate survival distributions were compared using the log-rank test. P < 0.05 was considered significant.
Increased FOXC2 expression was associated with progression and poor prognosis in CRC To investigate the role of FOXC2 in the clinical progression of CRC, we examined expression and localization of FOXC2 protein by immunohistochemistry (IHC) in 206 paraffin-embedded, archival CRC tissues. As shown in Table 1 and Fig. 1C, FOXC2 protein was detected in 195 of 206 (94.6%) cases of tissue samples, 99 cases displayed low FOXC2 expression and 107 cases displayed high FOXC2 expression (Table 1). In addition, weak to moderate staining of FOXC2 was observed in the areas of adenoma. Interestingly, FOXC2 protein displayed two main expression patterns in adenocarcinomas: cytoplasmic localization (76/195, 39%) and nuclear localization (119/195, 61%). Mann–Whitney U tests revealed that the FOXC2 expression level strongly correlated with the clinical stage (P = 0.017), T classification (P = 0.015), lymph node metastasis (P = 0.004), and distant metastasis in human CRC (P = 0.021) in this cohort of 206 cases of CRC (Table 1). Kaplan–Meier survival analysis indicated that patients with high FOXC2 level had significantly poorer overall survival (P = 0.002) and disease-free survival (P = 0.006, Fig. 1D, left panels) than patients with low FOXC2 level. We also analyzed the clinicopathologic significance of the localization of FOXC2 in CRC. Table 1 showed that high expression level of FOXC2 was significantly associated with the nuclear localization of FOXC2 (P = 0.002). Moreover, nuclear localization of FOXC2 correlated significantly with poor tumor differentiation (P = 0.01), advanced clinical stage (P = 0.001), T classification (P = 0.001), N classification (P = 0.013) and distant metastasis (P = 0.013). Kaplan–Meier survival analysis confirmed that patients with FOXC2 nuclear localization had poorer overall survival (P = 0.011) and disease-free survival (P = 0.002) than patients with FOXC2 cytoplasmic localization (Fig. 1D, right panels). Cox regression analyses revealed that both FOXC2 expression levels and subcellular localization were significant prognostic factors in CRC (Supplementary Tables S1–S3). Additionally, there was a significant correlation between the Ki67 labeling index and either FOXC2 expression levels (P = 0.014, Table 1) or subcellular localization (P = 0.006) in CRC. Samples that had lower level of FOXC2 expression also had a lower Ki-67 labeling index, whereas samples that had higher level of FOXC2 expression had a higher Ki-67 labeling index (Fig. 2). These observations suggest that increased expression of FOXC2 or nuclear localization of FOXC2 was closely associated with aggressive phenotypes of CRC.
Y.-M. Cui et al. / Cancer Letters 353 (2014) 87–94
89
Fig. 1. FOXC2 was up-regulated and associated with progression and poor prognosis in CRC. (A and B) Expression of FOXC2 protein and mRNA in CRC cell lines (A) and ten primary CRC (T) and adjacent noncancerous tissues (N) paired from the same patient (B) by western blotting and real-time PCR. Expression levels were normalized with aTubulin or GAPDH. Error bars represent mean ± SD calculated from 3 parallel experiments. (C) Representative images of FOXC2 expression in normal intestinal epithelium, adenoma, and CRC specimens examined by IHC. (D) Influence of FOXC2 expression patterns on overall survival and disease-free survival by Kaplan–Meier analyses.
Overexpression of FOXC2 promotes the proliferation and tumorigenesis of human CRC cells To investigate whether FOXC2 plays a role in the proliferation of CRC cells, we established stable FOXC2-expressing CRC cells (HT29/FOXC2 and Ls174t/FOXC2) (Fig. 3A). Overexpression of FOXC2 accelerated the cell growth (Fig. 3B and C), and strongly enhanced the migratory ability of CRC cells (Supplementary Fig. S2), as compared with control cells. We next investigated the role of FOXC2 in tumorigenesis of CRC cells. Overexpression of FOXC2 in HT29 and Ls174t promoted cells to grow in soft agar and form more colonies in comparison with control cells (Fig. 3D). To confirm this effect in vivo, we performed tumorigenesis assays in nude mice by subcutaneous injection of control cells (HT29/pBabe) and HT29/FOXC2 cells. In contrast to control cells, the HT29/FOXC2 cells exhibited more
rapid tumor growth (Fig. 3E, upper) and remarkably larger tumor volumes (Fig. 3E, lower). Of note, IHC staining of the tumor histological sections revealed that tumors developed by HT29–FOXC2 cells presented a distinctly high expression of Ki-67, while tumors from control group just had a few scattered Ki-67-positive cells (Fig. 3F). To explore the association of FOXC2 subcellular localization with its role in tumor proliferation, we deleted the nuclear localization sequence (NLS) of FOXC2 to see if this deletion mutant could still promote cell proliferation. Immunofluorescence analysis showed that FOXC2 protein could not locate in the nuclear after deletion of the NLS. Transfection of the deletion mutant failed to accelerate cell proliferation of HT29 cells determined by MTT and soft agar assays (Supplementary Fig. S3), suggesting that FOXC2 nuclear localization is required for the promotion of tumor proliferation.
90
Y.-M. Cui et al. / Cancer Letters 353 (2014) 87–94
Table 1 Correlation between clinicopathologic features and FOXC2 expression levels and localization. Characteristics
FOXC2 levels
P values
Low
High
Age 6mean(56) >mean (56)
50 49
43 64
Gender Male Female
56 43
Histology Columnar adenocarcinoma Mucinous adenocarcinoma Others
FOXC2 localization
P values
Cytoplasm
Nuclear
0.138
40 36
47 72
0.073
58 49
0.734
42 34
69 50
0.709
78 11 10
89 11 7
0.393
64 6 6
97 14 8
0.682
Differentiation Well and moderate Poor and undifferentiated
83 16
87 20
0.643
69 7
90 29
0.010
Dukes stage A B C D
9 45 21 24
6 39 29 33
0.017
8 38 13 17
5 38 29 47
0.001
T stage 1–2 3 4
22 57 20
16 53 38
0.015
20 42 14
14 62 43
0.001
N stage 0 1–3
68 31
52 55
0.004
52 24
60 59
0.013
Distant metastasis 0 (no) 1 (yes)
75 24
65 42
0.021
Ki-67 labeling index
