WNT2 Promotes Cervical Carcinoma Metastasis and Induction of Epithelial-Mesenchymal Transition

Background Previously, we found an 11-gene signature could predict pelvic lymph node metastasis (PLNM), and WNT2 is one of the key genes in the signature. This study explored the expression and underlying mechanism of WNT2 in PLNM of cervical cancer. Methods WNT2 expression level in cervical cancer was detected using western blotting, quantitative PCR, and immunohistochemistry. Two WNT2-specific small interfering RNAs (siRNAs) were used to explore the effects of WNT2 on invasive and metastatic ability of cancer cells, and to reveal the possible mechanism of WNT2 affecting epithelial—mesenchymal transition (EMT). The correlation between WNT2 expression and PLNM was further investigated in clinical cervical specimens. Results Both WNT2 mRNA and protein expression was upregulated in cervical cancer. High WNT2 expression was significantly associated with tumor size, lymphovascular space involvement, positive parametrium, and most importantly, PLNM. PLNM and WNT2 expression were independent prognostic factors for overall survival and disease-free survival. WNT2 knockdown inhibited SiHa cell motility and invasion and reversed EMT by inhibiting the WNT2/β-catenin pathway. WNT2 overexpression in cervical cancer was associated with β-catenin activation and induction of EMT, which further contributed to metastasis in cervical cancer. Conclusion WNT2 might be a novel predictor of PLNM and a promising prognostic indicator in cervical cancer.


Conclusion
WNT2 might be a novel predictor of PLNM and a promising prognostic indicator in cervical cancer.

Background
Cervical cancer is the third most commonly diagnosed cancer in women worldwide [1]. Despite the availability of early screening programs, the morbidity of cervical cancer remains high. The morbidity of cervical cancer varies considerably between countries, where more than 77,100 deaths have been recorded in the developing countries [1,2]. Among the clinical prognostic factors for cervical cancer, peripheral lymph node metastasis (PLNM) is one important prognostic factor [3,4]. The rate of PLNM in stage IB-IIA cervical cancer (stage IB, T1BN0M0; stage IIA T2bN0M0; stage IB or IIA is a combination of tumor-nodes-metastasis [TNM] classification that meets the criteria for stage IB or IIA disease) is 22%-27% [4]. The number of PLNM sites is also the most important prognostic factor in cervical cancer. The 5-year overall survival (OS) rate is 93.3%-13.8% in patients with cervical cancer with 1 to 4 positive nodes. The OS rate of patients with cervical cancer with 1 or 2 PLNM sites is significantly better than that for patients with >2 PLNM sites [5]. Multivariate analysis has revealed that >2 PLNM sites is an independent prognostic factor for survival [6]. Although the presence of PLNM does not affect staging according to the International Federation of Obstetrics and Gynecology (FIGO) system, it modulates postoperative therapy and affects prognosis. However, computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET) preoperative evaluation of nodal status are insufficiently sensitive for detecting small PLNMs (< 1.0 cm in greatest dimension) [7]. Early-stage patients with undetected PLNM, still undergo lymph node dissection, which might lead to infection, increased incidence of complications, and immune system damage [6]. Therefore, searching for promising markers to accurately detect PLNM in patients with early-stage cervical carcinoma is of great clinical importance, as it would aid gynecologic oncologists in directing the choice of treatment, avoiding unnecessary surgical intervention.
In a previous study, we discovered an 11-gene signature that could predict PLNM. In this preliminary study, evaluated expression of WNT2 was frequently observed in samples with PLNM compared to that without PLNM, and thus stood out as a potential candidate gene in regulating PLNM [8]. The human WNT gene family consists of 19 members, which encode evolutionarily conserved glycoproteins with 22 or 24 Cys residues [9]. WNT signaling are involved in regulating cell proliferation, differentiation, apoptosis, and migration [10][11][12]. Recent reports have uncovered the role of WNT2 in the development and progression of various cancers. For example, WNT2 is up-regulated in gastric cancers, and could impact tumor formation, invasion and dissemination [13,14]. Tumor fibroblast cells also secrete WNT2, which acts as a growth and invasion-promoting factor by activating the canonical WNT/β-catenin signaling pathway in esophageal cancer cells [15]. However, the relationship between WNT2 expression and lymph node metastasis in cervical cancer, and the involved mechanism remains largely unknown.
In this study, we investigated the expression and biological significance of WNT2 in clinical samples from cervical cancer patients and a cervical cancer cell line. We found that expression of WNT2 was evaluated in cervical tumor samples, especially in samples with PLNM. The overexpression of WNT2 was an independent prognostic factor for both OS and DFS. Moreover, downregulation of WNT2 could inhibit cervical tumor cell motility and invasion, and reversed EMT by inhibiting the WNT2/β-catenin pathway.

Samples and Patients
This study involved 314 patients diagnosed with early-stage cervical cancer (stage Ib-IIa) who underwent radical hysterectomy and lymphadenectomy at the Department of Gynecologic Oncology (Sun Yat-sen University Cancer Center, Guangzhou, China) from January 2002 to December 2006. Thirty fresh cervical cancer tissues, their paired adjacent noncancerous cervical tissues, and normal cervical tissues were collected for qPCR and western blotting. We obtained prior consent from the patients for the research use of the paraffinembedded tissues and fresh tissues. Informed consent was obtained written informed consent from the particients. The Sun Yat-Sen University Cancer Center Institutional Review Board approved this study. The clinicopathological features of the cervical cancer cohort are outlined in Table 1. All patients were followed up regularly, and the last follow-up was carried out in May 2012; the mean observation period was 73 months (1-124 months) and there were 33 cancer-related deaths. All 51 patients in whom lymph node metastasis was detected received adjuvant chemoradiotherapy postoperatively.

Cell Line and Antibodies
The SiHa cervical cancer cell line was obtained from the Shanghai Institutes for Biological Sciences Cell Bank (Shanghai, China), and was tested and authenticated by short tandem repeat genotyping. The cells were cultured in Roswell Park Memorial Institute 1640 medium (GIBCO BRL, Rockville, MD, USA) supplemented with 10% fetal bovine serum (FBS; HyClone Laboratories, Logan, UT, USA). Three human short hairpin RNA (shRNA) sequences for repressing WNT2 expression were cloned; S1 Table lists their sequences. Retroviral production and infection were performed as described previously [16]. Mouse anti-E-cadherin, anti-vimentin (BD Transduction Laboratories, Lexington, UK), and anti-β-tubulin (Abmart, Shanghai, China), and rabbit anti-WNT2 (Abcam, Cambridge, UK) and anti-β-catenin (Cell Signaling Technology, Danvers, MA, USA) antibodies were used in this study. Rhodamine-conjugated antibodies (Invitrogen, Carlsbad, CA, USA) and diaminophenylindole (Sigma-Aldrich, St Louis, MO, USA) were used in immunofluorescence analysis.

Tissue Microarray
Hematoxylin and eosin-stained sections from each paraffin-embedded, formalin-fixed block were used to define diagnostic areas; 2-5 (average, 4) random representative 0.6-mm cores were obtained from each case and inserted in a grid pattern into a recipient paraffin block using a tissue arrayer (Beecher Instruments, Silver Spring, MD, USA). Sections (5 μm) were then cut from each tissue microarray and stained with antibodies against WNT2, β-catenin, Ecadherin, and vimentin [18].

Wound Healing, Cell Invasion, and Migration Assays
The SiHa cells were plated to confluence in 6-well plates. Scratches were made in the monolayer with a pipette tip. The progression of cell migration was observed and photographed at 24h after wounding. The cell invasion and migration assays were performed using Transwell chambers (Corning Inc, Corning, NY, USA) with or without Matrigel coating (BD Biosciences, San Diego, CA, USA). The lower chambers were filled with 500 μl DMEM supplemented with 10% FBS. After 24-h incubation, cells invading to the bottom side of the inserts were fixed, stained, photographed, and quantified by counting in five random high-power fields [19].

Statistical Analysis
Statistical analyses were performed using SPSS v16.0 software (IBM, USA). The relationship between WNT2 expression and clinicopathological characteristics was assessed using Pearson's χ 2 test. Survival curves were plotted using the Kaplan-Meier method and compared using the log-rank test. Multivariate survival analysis was carried out for all parameters that were significant in univariate analysis using the Cox regression model. A two-sided probability value of <0.05 was considered statistically significant.

WNT2 Expression was Elevated in Cervical Cancer
We performed western blotting, qPCR, and immunohistochemistry assay on SiHa cervical cancer cells, fresh cervical tissues, and fresh cervical cancer tissues with paired adjacent noncancerous tissues to investigate the potential role of WNT2 in the tumorigenesis and lymphatic metastasis of cervical cancer. Both WNT2 protein expression and mRNA expression were upregulated in the SiHa cells and in cervical cancer specimens compared with that in the adjacent noncancerous tissues (Fig 1A-1C). The T/ANT ratios for WNT2 mRNA expression levels in the specimens without PLNM (0.00-17.69, Fig 1A) were lower than that in the specimens with PLNM (0.95-28.84, Fig 1A), and WNT2 mRNA expression was highly expressed in the SiHa cells as compared with the normal cervical tissues (Fig 1A). Western blotting confirmed the protein level of WNT2 was evaluated in tumors with PLNM ( Fig 2B). WNT2 was upregulated in 51.3% (161/314) of paraffin-embedded cervical cancer tissues. By contrast, WNT2 was not observable in the marginal areas surrounding the cancerous tissues. Moreover, WNT2 protein was highly expressed in 91% (20/22) of metastatic lymph node tissues following postoperative pathological confirmation (Fig 1C). Subsequently, we studied the association between WNT2 expression level and the patient clinicopathological characteristics. There was a significant correlation between WNT2 expression level and prognostic risk factors such as lymphovascular space involvement (LVSI, P = 0.001), positive parametrium (P = 0.031), and most importantly, lymph node metastasis (P < 0.0001) ( Table 2). Moreover, squamous cell carcinoma antigen, an tumor marker for predicting PLNM, was sinificantly correlated with WNT2 expression level (P = 0.029, Table 2). Univariate and multivariate analyses revealed that WNT2 expression (P < 0.001) and deep stromal invasion (P = 0.009) were independent factors for PLNM. Table 3 lists the results of Spearman correlation analysis between WNT2 expression and the clinicopathological factors.

WNT2 Expression Predicted Poor Survival
Kaplan-Meier analysis was used to investigate whether WNT2 expression was of prognostic significance. In our study cohort, high WNT2 expression was significantly associated with both shorter OS (P < 0.001, log-rank test) and disease-free survival (DFS; P < 0.001, log-rank test ;  Fig 2A and 2B). The 10-year OS and DFS rates were 73% and 72%, respectively, in patients with high WNT2 expression as compared to that in patients with low WNT2 expression, in whom both the OS and DFS rate was 91%. Furthermore, when stratified based on PLNM, both the OS and DFS rates of patients with high WNT2 expression were significantly lower than that of patients with low or no WNT2 expression (both P < 0.001, log-rank test; Fig 2C-

WNT2 Knockdown Inhibited SiHa Cell Motility and Invasion
We used two WNT2-specific siRNAs to explore the functional significance and possible mechanism of WNT2 in the invasive capability of SiHa cells, knockdown of WNT2 inhibited the expression of WNT2 both in mRNA and protein expression levels ( Fig 3A). Boyden chamber invasion assay revealed that the motility and invasiveness of SiHa cells were dramatically hampered by the ablation of WNT2; the average number of Siha cell transfected with short hairpin RNA-1 and RNA-2 were respectively 17 and 26, which was more than Siha cell transfected without short hairpin RNA. Cell mobility was measured by calculating the rate of wound closure at 0, 8, and 16 h in ×200 magnification. After 16h, the distance between shRNA-transfected SiHa cells on both sides of the wound was shorter than that between shRNAuntransfected SiHa cells (Fig 3B and 3C). As expected, silencing WNT2 led to decreased cell motility and invasiveness.

WNT2 Downregulation Reversed EMT and Suppressed the WNT2/β-Catenin Pathway
Although, there were no EMT-associated morphological changes detected in WNT2 knockdown SiHa cells, the inhibition of endogenous WNT2 suppressed the expression of epithelial markers such as E-cadherin, but partially rescued the expression of mesenchymal markers such as vimentin and fibronectin. Both qPCR and immunofluorescence showed that E-cadherin expression was decreased in WNT2-suppressed SiHa cells; however, vimentin expression was increased (Fig 4A and 4B). SiRNA silencing of WNT2 expression inhibited WNT2/β-catenin pathway activity. The knockdown of endogenous WNT2 reduced EMT in the SiHa cells, which was observed as decreased E-cadherin expression and increased vimentin expression, and prevention of β-catenin/E-cadherin complex formation.

Clinical Association of WNT2 Expression with PLNM in Human Cervical Cancer
To further investigate the clinical relevance of the above findings in human cervical cancer, we examined WNT2 expression and EMT features in 314 human cervical cancer tissue specimens. Patients were first divided into those who developed PLNM and those who were PLNM-free within the follow-up period. The expression level of WNT2 was significantly elevated in the primary tumors of 60 patients with PLNM as compared with that in the primary tumors from 101 patients without PLNM (Fig 5A). Kaplan-Meier analysis showed that PLNM was significantly associated with both shorter OS (P < 0.0001, log-rank test) and DFS (P < 0.0001, logrank test; Fig 5B). Moreover, cervical cancer samples with PLNM had higher β-catenin activation (63/75; 84.2%) than PLNM-free samples (122/239; 51%; P < 0.05). Importantly, there was strong E-cadherin expression in 132 of 239 specimens without PLNM (55.2%); in contrast, high E-cadherin expression was observed in only 18 of 75 samples with PLNM (23.4%; Fig 5C). Concurrently, there was an inverse expression pattern of vimentin in the same patient cohort, suggesting PLNM is significantly correlated (P < 0.05) with WNT2 expression (Fig 5C). Taken together, WNT2 overexpression in cervical cancer is associated with β-catenin activation and induction of EMT, which further contributes to the metastasis of cervical cancer.

Discussion
In this study, we demonstrate that both mRNA and protein level of WNT2 expression is upregulated in cervical cancer and in comparison with that in the paired adjacent noncancerous cervical tissues for the first time. There was a significant associstion between high WNT2 expression and tumor size, LVSI, positive parametrium. Most importantly, PLNM and WNT2 expression level were independent prognostic factors for poor OS and DFS rates in patients with early-stage cervical cancer. Moreover, knockdown WNT2 led to decreased cell motility and invasiveness reversed EMT by suppressing the WNT2/β-catenin pathway. In view of these findings, we believe that WNT2 is a potential novel predictor of PLNM and a promising therapeutic target in cervical cancer. So far, the standard treatment for early-stage cervical cancer is radical hysterectomy plus lymphadenectomy or chemoradiation, which have similar survival rates [20][21][22][23]. Chemoradiotherapy is required for patients with lymph node metastasis, which renders the initial surgical procedure unnecessary in retrospect [24]. However, as it is not possible to clinically detect pelvic and para-aortic lymph node metastasis accurately and efficiently, gynecologic   oncologists cannot direct the choice of treatment, and avoid unnecessary surgical intervention and reduce morbidity [24]. The use of MRI and CT in recent years for determining lymph node status may have been inaccurate; meta-analysis found that the pooled positive likelihood ratio for PET, MRI, and CT was 15.3, 6.4, and 4.3, respectively [25]. The sentinel concept is another promising technique, but a multi-center study reported its low sensitivity [26].
Our study demonstrated that high WNT2 expression levels were significantly correlated with poor prognostic risk factors and recurrence, especially PLNM. This is consistent with previous clinical observations where WNT2 expression was associated with increased risk of metastasis and worse prognosis [13,14].
The accuracy of predicting PLNM by observing WNT2 expression (high versus low) in our cohort was 36% (60/166). However, the recurrence rate was 10.8% in patients with high WNT2 expression but in whom lymph node metastasis was not observed, whereas recurrence was investigated in only 2.6% of patients (P < 0.01) in the low WNT2 expression group. As the reported pelvic lymph node micrometastasis rate is 15% in patients with early-stage cervical cancer [27,28], it is highly possibly that the actual accuracy of predicting PLNM could be 51% (36% + 15%) by detecting WNT2 expression. However, in the present study, only 75 of 314 (23.8%) patients were diagnosed with lymph node metastasis. To address this discrepancy, a further study extended to a larger cohort of patients with lymph node metastasis is currently undertaken, and with micrometastasis taken into account. In our cohort, it is more meaningful that the experiment found high WNT2 expression in 20 cases of lymph node metastasis, where the positive expression rate was 91% (20/22).
EMT is involved in a number of developmental events and contributes to tumor progression, invasion, metastasis, and dissemination [29]. The mechanism of cancer metastasis has been intensively studied recently and it may provide vital therapeutic targets for preventing metastasis. The major signaling pathways involved in EMT include the transforming growth factor-beta (TGF-β), WNT, Notch, and Hedgehog pathways [30]. EMT has been implicated in primary tumor metastasis in cervical cancer and the molecular mechanisms have been described [31]. Cervical cancer cells undergo EMT, and invasion and metastasis are promoted in response to stimulation from overexpressed EMT activators. The mechanism of EMT are divided in four major groups: (i) receptor-mediated signaling by TGF-β1, epidermal growth factor (EGF), and α5β1 integrin, and Jagged1-specific Notch signaling; (ii) an ion transport system containing KCl cotransporter-3; (iii) cytoskeletal modulators comprising gelsolin, and Rho C signaling; (iv) EMT-related transcription factors, including Six1 homeoprotein, Snail, and Twist [31].
Hence, we showed that suppressing WNT2 decreases cytoplasmic and nuclear accumulation of β-catenin in human cervical cells, downregulating WNT2 in cervical cancer cells removes the β-catenin/E-cadherin association, and that decreasing β-catenin leads to the disappearance of E-cadherin from the cell surface. We observed this based on the ability of WNT2 to specifically induce β-catenin cytoplasmic and nuclear localization by regulating β-catenin/E-cadherin interaction. Moreover, downregulating WNT2 remarkably inhibited cell motility and invasion and reversed EMT by inhibiting the WNT2/β-catenin pathway. Consistent with our findings, previously reported findings suggest that WNT2 expression in human colorectal carcinoma and esophageal cancer may also have a biological function [13][14][15][32][33].
Finally, we present a clinical correlation, where high WNT2 expression in human cervical cancer tissue specimens strongly correlates with cytoplasmic and nuclear accumulation of βcatenin and increased metastasis potential. WNT2 might be an identifying biomarker of PLNM for early cervical cancer, and a new potential therapeutic target.
Supporting Information S1 Table. The sequences of the shRNA used in the study.