The Role of TWIST in Ovarian Epithelial Cancers

Abstract

Background: Epithelial-mesenchymal transition (EMT) is associated with tumor hypoxia. EMT is regulated, in part, by the action of TWIST, which inhibits of E-cadherin expression and may interfere with the p53 tumor-suppressor pathway. Methods: We examined the expression of TWIST, E-cadherin, hypoxia-inducible factor 1α (HIF1α), and p53 by immunohistochemistry in 123 cases of ovarian epithelial cancers (OEC) to evaluate the role of TWIST in OEC. We assessed the association between protein expression and clinicopathologic parameters. Results: The expression of TWIST, E-cadherin, HIF1α, and p53 proteins was found in 28.5%, 51.2%, 35.0%, and 29.3% of cases, respectively. TWIST expression was associated with higher histologic grade and unfavorable survival. TWIST expression was correlated with HIF1α expression and reduced E-cadherin expression. The altered HIF1α/TWIST/E-cadherin pathway was associated with lower overall survival (OS), while the co-expression of TWIST and p53 was correlated with lower progression-free survival. In the multivariate analyses, TWIST expression was an independent prognostic factor for OS. Conclusions: Our data imply that TWIST expression could be a useful predictor of unfavorable prognosis for OEC. TWIST may affect the p53 tumor-suppressor pathway. Moreover, hypoxia-mediated EMT, which involves the HIF1α/TWIST/E-cadherin pathway may play an important role in the progression of OEC.

Headnote

Background: Epithelial-mesenchymal transition (EMT) is associated with tumor hypoxia. EMT is regulated, in part, by the action of TWIST, which inhibits of E-cadherin expression and may interfere with the p53 tumor-suppressor pathway. Methods: We examined the expression of TWIST, E-cadherin, hypoxia-inducible factor 1α (HIF1α), and p53 by immunohistochemistry in 123 cases of ovarian epithelial cancers (OEC) to evaluate the role of TWIST in OEC. We assessed the association between protein expression and clinicopathologic parameters. Results: The expression of TWIST, E-cadherin, HIF1α, and p53 proteins was found in 28.5%, 51.2%, 35.0%, and 29.3% of cases, respectively. TWIST expression was associated with higher histologic grade and unfavorable survival. TWIST expression was correlated with HIF1α expression and reduced E-cadherin expression. The altered HIF1α/TWIST/E-cadherin pathway was associated with lower overall survival (OS), while the co-expression of TWIST and p53 was correlated with lower progression-free survival. In the multivariate analyses, TWIST expression was an independent prognostic factor for OS. Conclusions: Our data imply that TWIST expression could be a useful predictor of unfavorable prognosis for OEC. TWIST may affect the p53 tumor-suppressor pathway. Moreover, hypoxia-mediated EMT, which involves the HIF1α/TWIST/E-cadherin pathway may play an important role in the progression of OEC.

Key Words: Ovarian epithelial cancer; TWIST transcription factor; Cadherins; HIF1α; Tumor sup- pressor protein p53

Ovarian epithelial cancer (OEC) is the most common form of malignancy and the leading cause of cancer death involving the female genital tract as well as the fifth most common cause of cancer-related death among women in the United States.1 Al- though there have been major advances in cancer treatment, the survival rate of OEC has not improved significantly in the last few decades. OEC frequently remains clinically silent, and the majority of patients with OEC present for diagnosis at an ad- vanced stage.2 To improve the assessment of a patient's risk of disease progression, several parameters, including patient age, disease stage, residual tumor mass, tumor type, and histologic grade, have been shown to have a prognostic value in OEC.3

Epithelial-mesenchymal transition (EMT) is a vital process during embryonic development in which epithelial cells lose cell-cell adhesion and acquire mesenchymal properties and a migratory phenotype. It is essential for numerous developmen- tal processes, such as neural tube formation and wound healing in organ fibrosis. Nevertheless, an appreciation of the role EMT plays in tumor metastasis is relatively new.4 TWIST was first identified in Drosophila melanogaster and is a basic helix-loop-he- lix transcription factor that induces motility and metastatic po- tential in several tumors by promoting EMT.5 The role of TWIST proteins in cancer has been investigated over the course of the last decade. It has been found that TWIST is a central player in the EMT of many cancers,6 including breast,7 ovari- an,8,9 endometrial,10 bladder,11 lung,12 gastric,13 pancreatic,14 prostate,11 thyroid,15 and head and neck squamous cell carcino- mas (HNSCC).16 In all cancer types, the expression of TWIST is associated with poor prognosis, and high-grade, invasive, and metastatic lesions. TWIST protein promotes EMT by turning down the expression of epithelial-specific proteins, such as E- cadherin, a hallmark of EMT.17 Recently, a novel function of TWIST has been reported in the development of acquired che- moresistance in human cancer cells.18

Tumor hypoxia has been linked to an aggressive phenotype that correlates with cancer metastasis and chemoresistance in various solid malignancies.19 Hypoxia-inducible factor 1α (HIF1α) is one of the key factors regulating cellular responses to hypoxia. HIF1α overexpression has been reported in a variety of human cancers, and is correlated with aggressive tumor phenotypes, therapeutic resistance, and poor prognosis. HIF1α and HIF1α- regulated genes also promote EMT, a critical step in metastasis. Co-expression of HIF1α and TWIST has been shown, and could be used as a prognostic marker in cancer patients.12,20,21

TWIST may be involved in several pathways that contribute to carcinogenesis, including interference of the p53-associated pathway involved in apoptosis, as well as EMT-related path- ways.18 p53 is one of the most frequently mutated tumor sup- pressor genes and is more commonly altered in advanced cases of ovarian cancer.22 Mutation in p53 or p53 inactivation correlates with altered expression of EMT markers such as TWIST and E- cadherin.23 Recent evidence indicates that TWIST has been shown to reduce the activity of p53 through a complex network of post-translational modifications and molecular interactions influenced by several signaling pathways. These findings suggest the involvement of p53 in the regulation of the EMT pathway.24

Recent studies in OEC have supported the role of EMT in ovarian carcinogenesis.8,9 Overexpression of TWIST plays a role not only in the development of ovarian carcinoma, but also in patient survival. We examined the expression of TWIST by im- munohistochemistry in a panel of 123 OEC patients. We also evaluated the expression of E-cadherin, HIF1α, and p53. We aimed to investigate the association of these factors with clini- copathologic features and their impact on patient survival.

MATERIALS AND METHODS

Patients and tissue samples

The OEC tissue samples were obtained during surgical opera- tions performed at Pusan National University Hospital, Korea, from 1999 to 2008. A total of 123 patients (mean age, 49.5 years; range, 15 to 82 years) who underwent surgical resection for OEC were examined. Pathological data, such as pathologic stage, tumor grading, tumor size, mitosis, and histological tu- mor type, were obtained from the primary pathology reports, and patient chart reviews were performed retrospectively. His- tologic tumor type and grade were determined according to the World Health Organization criteria. The criteria of the Interna- tional Federation of Gynecology and Obstetrics (FIGO) were used to classify cases as stages I-IV. All patients, except those with grade 1, stage IA, were given adjuvant chemotherapy of platinum/taxol-containing drugs. The patients were followed up from the date of surgery until death or the last visit to the outpatient department. The follow-up period ranged from one to 175 months (median, 49 months). Overall survival (OS) was calculated from the date of surgery to the date of death or last followup visit. Progression-free survival (PFS) was calculated from the date of surgery to the date of tumor progression or re- lapse, including local recurrence and/or metastasis. The patients' demographic and tumor characteristics are presented in Table 1. This study was approved by the Ethics Committee of the Pusan National University Hospital, and informed consent was ob- tained from all the patients.

Immunohistochemistry

Immunohistochemistry was performed on serial 4-µm-thick paraffin sections. The slides were deparaffinized in xylene and re- hydrated through four decreasing grades of ethanol (100%, 95%, 80%, and 70%) for 2 minutes each. Endogenous peroxi- dase activity was blocked by immersing the slides in 3% hydro- gen peroxide in methanol for 15 minutes at room temperature. Heat-induced antigen retrieval was performed for 5 minutes with 1× citrate buffer (pH 6.3) using a microwave and then cooled for 5 minutes. This process was performed three times. In order to reduce the nonspecific binding of antibodies to the tis- sues, the slides were preincubated with blocking serum in 1% bovine serum albumin for 30 minutes at room temperature. The slides were left to incubate with rabbit polyclonal anti-TWIST antibody (1:200, Santa Cruz Biotechnology, San Diego, CA, USA), mouse monoclonal anti-E-cadherin antibody (1:400, In- vitrogen, Carlsbad, CA, USA), mouse monoclonal anti-HIF1α antibody (1:1,000, Novus Biologicals, Littleton, CO, USA), and rabbit monoclonal anti-p53 antibody (1:200, Thermo Scientific, San Jose, CA, USA) overnight at 4°C. After incubation with the primary antibodies, the slides were rinsed with phosphate-buff- ered saline for 10 minutes. Thereafter, for the detection of the antibody reactions, the EnVision Detection System (Dako, Car- pinteria, CA, USA) was used. The color was developed using di- aminobenzidine. Finally, the sections were counterstained with Mayer's hematoxylin solution for 1 minute, and dehydrated with graded alcohols, dipped in two changes of xylene, and mounted.

Assessment of immunohistochemical staining

The immunohistochemical staining was assessed with a light microscope by two pathologists who had no prior knowledge of the clinical data. For TWIST and HIF1α, only nuclear staining was considered with a 10% cut-off value for positivity.7 For E- cadherin, membrane staining was graded from 0-3: 0, no stain- ing or weak intensity in 10% of the tumor cells; 1, low intensity in 10% of the tumor cells; 2, moderate intensity in 10% of the tumor cells; and 3, strong intensity in 10% of the tumor cells. The cases were divided into two groups: preserved E-cadherin expression (score 3) or reduced E-cadherin expression (scores 0, 1, and 2).21 p53 was recognized through nuclear staining of the positive cells. The results were divided into two groups: positive ( ≥ 30% positive tumor cells) and negative (<30% positive tu- mor cells).25

Statistical analysis

All statistical analyses were performed using SPSS ver. 19.0 (IBM Co., Armonk, NY, USA). Pearson's chi-squared test was used to study the associations between the clinicopathological factors and immunohistochemical markers. The linear correla- tions for protein expression were assessed using the Sperman's rank test. OS was calculated from the date of surgery to the date of death or last follow-up visit to the outpatient department. PFS was calculated from the date of surgery to the date of tumor relapse or progression. Data on patients who had survived until the end of the observation period were censored at their last fol- low-up visit. Death from a cause other than ovarian cancer and cases lost to follow-up were considered censoring events. The ef- fect of immunohistochemical markers on survival was assessed using Kaplan-Meier survival curves and a log-rank test. Multi- variate survival analyses were calculated using Cox's proportion- al hazards model. For all the tests, a p-value ≤ .05 was consid- ered to be significant.

RESULTS

Altered expression of HIF1α, E-cadherin, TWIST, and p53 in relation to clinicopathologic parameters

Expression of HIF1α, TWIST, and p53 proteins was detected in 35 (28.5%), 43 (35.0%), and 36 (29.3%) of the 123 OEC cases, respectively. Reduced expression of E-cadherin was ob- served in 60 cases (48.8%). Specific staining for TWIST was ob- served in the nucleus of tumor cells, but in some cases, both nu- clear and cytoplasmic staining was observed. There was little immunoreactivity for TWIST in the tumor stroma. E-Cadherin expression was membranous, and occasionally cytoplasmic. Rep- resentative images of immunostaining are shown in Fig. 1.

TWIST expression was associated with higher histologic tu- mor grades (p= .004). Clear cell types exhibited TWIST expres- sion more frequently than other histologic types, but this was not statistically significant (p= .240). Reduced E-cadherin ex- pression was more commonly found in the advanced stages of OEC (38 of 60 cases, 63.3%) than the early stages (22 of 63 case, 34.9%) (p=.002). HIF1α showed different expression ac- cording to histologic tumor type but was not associated with any clinicopathological parameters. There was a statistically significant increase in the expression of p53 in serous-type OEC compared with other types (p=.000). There were also positive correlations between p53 and the histologic tumor grade, FIGO stage, and mitotic rate (Table 2).

TWIST expression was significantly correlated with reduced E-cadherin expression (p=.000, correlation coefficient [c.c.] = -0.309) and increased HIF1α expression (p= .000, c.c.= 0.532). TWIST expression was not correlated with the expression of p53 (p= .142, c.c.=0.168). There was a negative correlation between p53 expression and the reduced expression of E-cad- herin (p= .041, c.c.=-0.219) (Table 3).

Correlation between TWIST, E-cadherin, HIF1α, and p53 expressions and survival

Follow-up data were available for 123 patients (median, 49 months; range, 1 to 175 months). Fifty patients (40.7% of 123 patients) died from the disease during the follow-up period. Sixty patients (48.8%) developed either recurrence or metastasis (progression group), whereas 63 patients (51.2%) remained free of progression (progression-free group). The univariate analysis showed that tumor histologic grade, tumor stage, mitotic rate, positive expression of TWIST and p53, and reduced expression of E-cadherin were associated with poor OS and PFS. Serous type and tumor size were also correlated with poor PFS (Table 4, Fig. 2).

We tried to explore whether the combined expression of these protein markers had any prognostic value. Eighteen of 27 pa- tients (66.7%) who had a pattern of positive TWIST expression and reduced expression of E-cadherin [TWIST(+)/E-cadherin(-) group] died of disease, compared to 32 of 96 patients (33.3%) without a pattern of TWIST(+)/E-cadherin(-) (p= .003). Of 19 patients showing an altered HIF1α/TWIST/E-cadherin pathway with positive expression of HIF1α and TWIST, and reduced E- cadherin expression, 13 (68.4%) died of disease, whereas 37 of 104 patients (35.6%) without an altered HIF1α/TWIST/E-cad- herin pathway died of disease (p=.011). Thirteen of 14 patients (92.9%) who displayed expression of both TWIST and p53 showed disease progression compared to 47 of 109 patients (43.1%) who had no expression of TWIST or p53. This differ- ence was statistically significant (p=.000). In a multivariate analysis with the variables defined in Table 5, positive expression of TWIST was associated with lower OS after adjustment for histologic grade, tumor stage, mitotic rate, p53 expression, and altered expression of the HIF1α/TWIST/E-cadherin pathway, which were significant parameters in the univariate analysis.

DISCUSSION

The EMT, which is considered to be one of the major mecha- nisms underlying tumor invasion and metastasis, has been studied to understand how epithelial cancer cells can acquire the ability to migrate and metastasize. EMT is triggered by a set of stimuli, which include hypoxia and growth factors, and tumor stromal cell interactions with transcription factors such as TWIST, SNAI1, SLUG, and ZWB1.4 TWIST overexpres- sion has also been identified in various kinds of epithelial tu- mors and plays a distinct role in tumor progression via the loss of E-cadherin expression.6-16

There have been a few other reports showing a correlation between TWIST expression and clinical significance in patients with OEC.8,9 These previous studies have examined the expres- sion of TWIST in OEC cases, reporting immunohistochemical detection of TWIST expression in about 30-40% of cases. TWIST expression was correlated with clinicopathological pa- rameters and found to be a predictor of poor survival. Hosono et al.8 also reported that positive TWIST expression was an inde- pendent prognostic factor for survival in their multivariate analyses. In this study of 123 OEC patients with follow-up, positive TWIST expression was detected in 35 of 123 OEC cases (28.5%). We found that TWIST expression was correlated with higher histologic grades and shorter OS and PFS in OEC patients, which is consistent with results from previous studies of other malignant tumors, including OEC. In the analysis of the histological type-stratified data, TWIST expression tended to be higher in the clear cell type of OEC (44.0%), although the difference was not significant (p= .240). Kajiyama et al.26 reported that TWIST expression was not only detected more frequently, but was also an independent prognostic factor in the clear cell type of OEC. They suggested that TWIST might play a role in the mechanisms for poor prognosis and chemoresis- tance in the clear cell type of OEC.

p53 is a tumor suppressor gene that is involved in the regula- tion of cell growth. Mutations in the p53 gene are among the most commonly detected genetic abnormalities in human neo- plasia. Mutations result in a conformational change in the pro- tein, which becomes stabilized, thus usually, but not always, al- lowing for immunohistochemical detection. Although not fully defined, recent evidence has indicated that TWIST may inter- fere with the p53 tumor suppressor pathway.23,24 TWIST has previously been shown to reduce the activity of the p53 pro- moter as well as mediate suppression on p53 target sequences. Wang et al.18 demonstrated that increased TWIST was respon- sible for the development of acquired resistance to chemothera- peutic drugs and microtubule-disrupting agents, including pa- clitaxel. TWIST may protect from apoptosis mediated by p53 in response to DNA damage induced by radiation and drugs. Chemoresistance, as well as the facilitation of metastasis by TWIST, may be the underlying mechanism linking TWIST expression with an unfavorable prognosis. In this study, we found that p53 expression showed no correlation with TWIST expression, but an inverse correlation with E-cadherin expres- sion was established. Interestingly, co-expression of TWIST and p53 (13/14 cases, 92.9%) showed unfavorable PFS compared with no co-expression (47/109 cases, 43.1%) (p= .000). These results suggest that TWIST and p53 could cooperate with each other to facilitate cancer progression, and p53 could have an important role in EMT.

E-Cadherin plays a critical role in regulating the balance of cell-cell adhesion and cell motility, placing it in a significant po- sition to regulate tumor cell proliferation and invasion. Some studies have identified that reduced E-cadherin expression is as- sociated with more aggressive epithelial tumors.9,12,17,21 TWIST binds the E-cadherin promoter and inhibits E-cadherin tran- scription levels, which modulate EMT in cancer progression. An inverse correlation between TWIST upregulation and E-cad- herin downregulation was also observed in several types of ma- lignancies.5,10 However, Yoshida et al.27 demonstrated that there was not an inverse correlation between the expression of E-cad- herin and TWIST in OEC. In this study, reduced E-cadherin expression was associated with advanced FIGO stage and poor OS and PFS. Our results are consistent with the observation that E-cadherin has been described in patients with advanced stage OEC and is associated with poor prognosis.9 Moreover, a signifi- cant inverse relationship was observed between these two mark- ers in this study, and we showed that TWIST(+)/E-cadherin(-) cases are also associated with poor prognosis.

For cancer progression, intratumoral hypoxia followed by the activation of HIF1α is one of the most important mechanisms promoting tumor aggressiveness, leading to metastasis and pa- tient mortality. Recent results have shown the upregulation of mRNA and protein levels of TWIST in different cell lines under hypoxia and constitutive expression of HIF1α.19,20 TWIST can be a direct target of HIF1α due to the functional presence of a hypoxia-related element located in the proximal promoter of TWIST.20 The survival analysis in the aforesaid report demon- strated the worst prognosis and shortest metastasis-free period in HNSCC patients with HIF1α, TWIST, and Snail co-expression, indicating that the co-expression of these markers in tumor tis- sues could be used as a valuable marker to predict metastasis and prognosis in HNSCC patients. In a study by Feng et al.,21 the overexpression of HIF1α and TWIST was frequently detected in endometrial endometrioid carcinoma (EEC) samples, and they were associated with aggressive tumor characteristics, such as higher histologic grades, deeper myometrial invasion, and lymph node metastasis. The interesting findings of the study were that TWIST-positive cases exhibited increased expression of HIF1α and reduced expression of E-cadherin in EEC. The researchers suggested that the HIF1α/TWIST/E-cadherin pathway may play a critical role in the progression of EEC. Our study showed a positive association between HIF1α and TWIST expression (?=0.407, p= .000) and a negative association between TWIST and E-cadherin expression (? =-0.358, p=.000). Moreover, al- tered expression of the HIF1α/TWIST/E-cadherin pathway was associated with poor OS in OEC in the univariate analysis.

In summary, we show that increased expression of TWIST, HIF1α, and p53 and reduced expression of E-cadherin are fre- quently observed in OEC cases. TWIST expression was associ- ated with poor OS in both the univariate and multivariate anal- yses. Therefore, TWIST may serve as a useful marker in pa- tients with OEC who are likely to have an unfavorable clinical outcome. Our results support the critical role of the HIF1α/ TWIST/E-cadherin pathway in cancer progression and the ac- quisition of the metastatic phenotype in OEC. Therefore, path- ways involved in hypoxia-mediated EMT, such as the HIF1α/ TWIST/E-cadherin pathway, may be possible candidates for molecular targeting in the future treatment of OEC. Further studies are needed to clarify the role and molecular mechanisms underlying of the HIF1α/TWIST/E-cadherin pathway in tu- mor progression.

Conflicts of Interest

No potential conflict of interest relevant to this article was reported.

Acknowledgments

This work was supported by Pusan National University Grant, 2011.

References

REFERENCES

1. Siegel R, Ward E, Brawley O, Jemal A. Cancer statistics, 2011: the impact of eliminating socioeconomic and racial disparities on pre- mature cancer deaths. CA Cancer J Clin 2011; 61: 212-36.

2. Naora H, Montell DJ. Ovarian cancer metastasis: integrating in- sights from disparate model organisms. Nat Rev Cancer 2005; 5: 355-66.

3. Friedlander ML. Prognostic factors in ovarian cancer. Semin Oncol 1998; 25: 305-14.

4. Thiery JP, Sleeman JP. Complex networks orchestrate epithelial- mesenchymal transitions. Nat Rev Mol Cell Biol 2006; 7: 131-42.

5. Yang J, Mani SA, Donaher JL, et al. Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis. Cell 2004; 117: 927-39.

6. Puisieux A, Valsesia-Wittmann S, Ansieau S. A twist for survival and cancer progression. Br J Cancer 2006; 94: 13-7.

7. Markiewicz A, Ahrends T, Welnicka-Jaskiewicz M, et al. Expression of epithelial to mesenchymal transition-related markers in lymph node metastases as a surrogate for primary tumor metastatic po- tential in breast cancer. J Transl Med 2012; 10: 226.

8. Hosono S, Kajiyama H, Terauchi M, et al. Expression of Twist in- creases the risk for recurrence and for poor survival in epithelial ovarian carcinoma patients. Br J Cancer 2007; 96: 314-20.

9. Terauchi M, Kajiyama H, Yamashita M, et al. Possible involvement of TWIST in enhanced peritoneal metastasis of epithelial ovarian carcinoma. Clin Exp Metastasis 2007; 24: 329-39.

10. Kyo S, Sakaguchi J, Ohno S, et al. High Twist expression is involved in infiltrative endometrial cancer and affects patient survival. Hum Pathol 2006; 37: 431-8.

11. Wallerand H, Robert G, Pasticier G, et al. The epithelial-mesenchy- mal transition-inducing factor TWIST is an attractive target in ad- vanced and/or metastatic bladder and prostate cancers. Urol Oncol 2010; 28: 473-9.

12. Lee JH, Park WY, Jeong SM, et al. Relationship between the endog- enous hypoxic markers hypoxia inducible factor-1alpha, carbonic anhydrase IX, and epithelial mesenchymal transition regulator TWIST expression in non-small cell lung cancer. Korean J Pathol 2010; 44: 469-76.

13. Gao XH, Yang XQ, Wang BC, Liu SP, Wang FB. Overexpression of twist and matrix metalloproteinase-9 with metastasis and progno- sis in gastric cancer. Asian Pac J Cancer Prev 2013; 14: 5055-60.

14. Arumugam T, Ramachandran V, Fournier KF, et al. Epithelial to mesenchymal transition contributes to drug resistance in pancreatic cancer. Cancer Res 2009; 69: 5820-8.

15. Buehler D, Hardin H, Shan W, et al. Expression of epithelial-mesen- chymal transition regulators SNAI2 and TWIST1 in thyroid carci- nomas. Mod Pathol 2013; 26: 54-61.

16. Jouppila-Mättö A, Närkiö-Mäkelä M, Soini Y, et al. Twist and snai1 expression in pharyngeal squamous cell carcinoma stroma is relat- ed to cancer progression. BMC Cancer 2011; 11: 350.

17. Zhang Z, Xie D, Li X, et al. Significance of TWIST expression and its association with E-cadherin in bladder cancer. Hum Pathol 2007; 38: 598-606.

18. Wang X, Ling MT, Guan XY, et al. Identification of a novel function of TWIST, a bHLH protein, in the development of acquired taxol resistance in human cancer cells. Oncogene 2004; 23: 474-82.

19. Ruan K, Song G, Ouyang G. Role of hypoxia in the hallmarks of human cancer. J Cell Biochem 2009; 107: 1053-62.

20. Yang MH, Wu MZ, Chiou SH, et al. Direct regulation of TWIST by HIF-1alpha promotes metastasis. Nat Cell Biol 2008; 10: 295-305.

21. Feng Z, Gan H, Cai Z, et al. Aberrant expression of hypoxia-induc- ible factor 1alpha, TWIST and E-cadherin is associated with aggres- sive tumor phenotypes in endometrioid endometrial carcinoma. Jpn J Clin Oncol 2013; 43: 396-403.

22. O'Neill CJ, Deavers MT, Malpica A, Foster H, McCluggage WG. An immunohistochemical comparison between low-grade and high-grade ovarian serous carcinomas: significantly higher expres- sion of p53, MIB1, BCL2, HER-2/neu, and C-KIT in high-grade neo- plasms. Am J Surg Pathol 2005; 29: 1034-41.

23. Stasinopoulos IA, Mironchik Y, Raman A, Wildes F, Winnard P Jr, Raman V. HOXA5-twist interaction alters p53 homeostasis in breast cancer cells. J Biol Chem 2005; 280: 2294-9.

24. Maestro R, Dei Tos AP, Hamamori Y, et al. Twist is a potential on- cogene that inhibits apoptosis. Genes Dev 1999; 13: 2207-17.

25. Havrilesky L, Darcy KM, Hamdan H, et al. Prognostic significance of p53 mutation and p53 overexpression in advanced epithelial ovarian cancer: a Gynecologic Oncology Group Study. J Clin Oncol 2003; 21: 3814-25.

26. Kajiyama H, Hosono S, Terauchi M, et al. Twist expression predicts poor clinical outcome of patients with clear cell carcinoma of the ovary. Oncology 2006; 71: 394-401.

27. Yoshida J, Horiuchi A, Kikuchi N, et al. Changes in the expression of E-cadherin repressors, Snail, Slug, SIP1, and Twist, in the devel- opment and progression of ovarian carcinoma: the important role of Snail in ovarian tumorigenesis and progression. Med Mol Mor- phol 2009; 42: 82-91.