You are here

Article Text

Article menu
Original article
ZEB1 expression is associated with prognosis of intrahepatic cholangiocarcinoma
  1. Katsumi Terashita1,
  2. Makoto Chuma1,2,
  3. Yutaka Hatanaka3,
  4. Kanako Hatanaka3,
  5. Tomoko Mitsuhashi3,
  6. Hideki Yokoo4,
  7. Takumi Ohmura5,
  8. Hiroyuki Ishizu6,
  9. Shunji Muraoka7,
  10. Atsushi Nagasaka8,
  11. Takahiro Tsuji9,
  12. Yoshiya Yamamoto10,
  13. Nobuaki Kurauchi11,
  14. Norihiko Shimoyama12,
  15. Hidenori Toyoda13,
  16. Takashi Kumada13,
  17. Yuji Kaneoka14,
  18. Atsuyuki Maeda14,
  19. Koji Ogawa1,
  20. Mitsuteru Natsuizaka1,
  21. Hirofumi Kamachi4,
  22. Tatsuhiko Kakisaka4,
  23. Toshiya Kamiyama4,
  24. Akinobu Taketomi4,
  25. Yoshihiro Matsuno3,
  26. Naoya Sakamoto1
  1. 1Departments of Gastroenterology and Hepatology, Hokkaido University Hospital, Sapporo, Japan
  2. 2Gastroenterological Center, Yokohama City University Medical Center, Yokohama, Kanagawa, Japan
  3. 3Department of Surgical Pathology, Hokkaido University Hospital, Sapporo, Japan
  4. 4Department of Gastroenterological Surgery I, Hokkaido University Hospital, Sapporo, Japan
  5. 5Department of Gastroenterology, Sapporo-Kosei General Hospital, Sapporo, Japan
  6. 6Department of Surgery, Sapporo-Kosei General Hospital, Sapporo, Japan
  7. 7Department of Pathology, Sapporo-Kosei General Hospital, Sapporo, Japan
  8. 8Department of Gastroenterology, Sapporo Municipal Hospital, Sapporo, Japan
  9. 9Department of Pathology, Sapporo Municipal Hospital, Sapporo, Japan
  10. 10Department of Gastroenterology, Hakodate Municipal Hospital, Hakodate, Japan
  11. 11Department of Surgery, Hakodate Municipal Hospital, Hakodate, Japan
  12. 12Department of Pathology, Hakodate Municipal Hospital, Hakodate, Japan
  13. 13Departments of Gastroenterology and Hepatology, Ogaki Municipal Hospital, Ogaki, Japan
  14. 14Department of Surgery, Ogaki Municipal Hospital, Ogaki, Japan
  1. Correspondence to Dr Makoto Chuma, Department of Gastroenterology and Hepatology, Hokkaido University, Kita 15, Nishi 7, Kita-ku, Sapporo 060-8638, Japan; chuma{at}yokohama-cu.ac.jp

Abstract

Background/Aim Intrahepatic cholangiocarcinoma (ICC) is one of the most aggressive malignant tumours, so the identification of molecular targets for ICC is an important issue. Zinc finger E-box binding homeobox 1 (ZEB1) is a key inducer of epithelial–mesenchymal transition (EMT). The aim of the present study was to clarify the clinical significance of ZEB1 in ICC and the associations between ZEB1 expression and EMT-related proteins.

Methods We immunohistochemically examined the expression of EMT-related proteins, namely ZEB1, vimentin and E-cadherin, in ICC specimens from 102 patients. The clinicopathological and prognostic values of these markers were evaluated.

Results ZEB1 and vimentin were expressed in 46.1% and 43.1% of tumours, respectively, and E-cadherin expression was lost in 44.1% of tumours. ZEB1 expression showed a significant inverse correlation with E-cadherin expression (p=0.004) and a positive correlation with vimentin expression (p=0.022). Altered expression of ZEB1 was associated with aggressive tumour characteristics, including advanced tumour stage (p=0.037), undifferentiated-type histology (p=0.017), lymph node metastasis (p=0.024) and portal vein invasion (p=0.037). Moreover, overall survival rates were significantly lower for patients with high ZEB1 expression than for patients with low ZEB1 expression (p=0.027). Kaplan–Meier analysis also identified E-cadherin expression (p=0.041) and vimentin expression (p=0.049) as prognostic indicators for overall survival.

Conclusions ZEB1 expression is associated with tumour progression and poor prognosis in patients with ICC through positive correlations with vimentin and negative correlations with E-cadherin. ZEB1 expression is associated with a poor prognosis and might be an attractive target for the treatment of ICC.

  • LIVER CANCER
  • IMMUNOHISTOCHEMISTRY
  • CHOLANGIOCARCINOMA

https://doi.org/10.1136/jclinpath-2015-203115

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Intrahepatic cholangiocarcinoma (ICC) is the second most common subtype of primary hepatic cancer, following hepatocellular carcinoma (HCC).1 Worldwide data over long periods have shown a drastic increase in the incidence of ICC.1 ,2 The long-term survival of patients with ICC remains unsatisfactory because of high recurrence rates and early metastasis.3 Surgical resection at the early stages of ICC is the only approach with curative intent.4 Patients with unresectable tumours have a dismal median survival time, and systemic chemotherapies have limited success in ICC.5 Hence, a better understanding of the molecular mechanisms associated with progression of ICC would be beneficial for the development of effective therapies, and the identification of useful biomarkers for ICC is an important clinical issue.

    Recent evidence suggests that epithelial–mesenchymal transition (EMT) is a key mechanism that is often activated during cancer invasion and metastasis.6 ,7 EMT is a biological process through which an epithelial cell undergoes multiple biochemical changes that enable it to assume a mesenchymal cell phenotype. During this process, mesenchymal markers such as vimentin are induced, and epithelial markers such as E-cadherin, which is essential for the structural integrity of the epithelium disappear.6–8 The loss of E-cadherin expression causes the disassembly of intercellular adhesion complexes, thus loosening contacts between neighbouring epithelial cells and disrupting the overall tissue architecture.6–8 Vimentin expression is a late event in which the loss of epithelial features directly precedes and leads to upregulation of mesenchymal genes.9 Several transcription factors have been described as key inducers of EMT, including members of the snail superfamily,10 the basic helix-loop-helix family and the two zinc finger E-box binding homeobox (ZEB) factors (ZEB1 and ZEB2).10–13 ZEB1 has emerged as a key player in cancer progression, and its signal pathway is one of the most important pathways involved in the EMT process.10–13 ZEB1 represses the expression of epithelial genes and certain microRNAs (miRs), including the miR-200 family, which function as strong inducers of epithelial differentiation and also as inhibitors of stem cell properties.13–15 Aberrant expression of ZEB1 protein in HCC,16 gastric carcinoma,17 colon cancer,18 pancreatic cancer,19 breast carcinoma20 and cervical cancer21 has been associated with aggressive disease, poor differentiation, development of metastases and poor clinical prognosis; however, ZEB1 expression in ICC has not been reported.

    To determine the importance of the EMT in the pathogenesis of ICC and its clinical relevance, we examined the expression of ZEB1 and EMT markers in 102 cases of ICC by immunohistochemistry and analysed the correlations among these markers and clinical outcome.

    Methods

    Patient samples

    For immunohistochemical analysis, 102 adult patients with ICC who had surgically resectable tumours and fulfilled the inclusion criteria were included. Patients were treated between 1996 and 2012 at five institutions in Japan (Hokkaido University Hospital, Hakodate Municipal Hospital, Sapporo Kosei Hospital, Sapporo Municipal Hospital and Ogaki Municipal Hospital) (see online supplementary figure S1). Surgical resectability was considered to be the complete resection of the tumour mass in the hepatic lesion with tumour-free margins and the resection of the regional lymph nodes, including the hilar, hepatoduodenal ligament lymph nodes and caval lymph nodes. Briefly, inclusion criteria were as follows: (a) distinctive pathological diagnosis, (b) no preoperative anticancer treatment and (c) complete clinicopathological and follow-up data. This study was approved by the ethics committees of Hokkaido University Hospital and all other participating hospitals. The study protocol was approved by the institutional review board and performed in compliance with the Declaration of Helsinki. Written informed consent was obtained from as many patients as possible who were still alive (deceased cases were approved for use without written informed consent). Histological diagnosis was made according to WHO criteria.22 The same staging system were applied to all patients.23 The main clinicopathological features are presented in table 1. During follow-up, clinical evaluations and biochemical tests were performed every 1–3 months. The median follow-up was 35 months (range, 3–170 months). Patients underwent triphasic CT of the liver every 2–3 months.

    View this table:
    Table 1

    Clinicopathological variables and the expression of EMT-related markers (n=102)

    Immunohistochemical staining

    Tissue sections were deparaffinised in xylene and rehydrated through a graded ethanol series. The deparaffinised sections were then heat treated with antigen retrieval solution (Target Retrieval Solution, pH 9.0; Dako, Glostrup, Denmark) at 95°C for 20 min by using the Dako PT Link system. After blocking endogenous peroxidase by using Dako Peroxidase-Blocking Reagent, the sections were incubated for 30 min at room temperature with specific antibodies to ZEB1 (GWB-B65B99 dilution 1:20, Genway Biotech), vimentin (CloneV9 1:1000; Dako) and E-cadherin (4A2C7 1:2, Invitrogen). Detection was performed by using a standard polymer method in accordance with the manufacturer's instructions (EnVision Flex system for mouse mAb, Dako). These immunohistochemical reactions were performed by using an automated immunostaining system (Autostainer Plus, Dako).

    Evaluation of immunostaining

    Pathological and immunohistochemical examinations were conducted by two experienced pathologists (KH and TM) who were unaware of the clinical courses of the patients. Staining for immunohistochemical markers for tumour cells was evaluated with regard to intensity and proportion. The intensity of immunohistochemical marker staining of tumour cells was evaluated by a sliding scale from 0 to 3+ (0=negative staining, 1+=weak, 2+=intermediate, 3+=strong). The percentage of each intensity score was determined (0%–100%). For each marker, p0, p1, p2 and p3 (p0+p1+p2+p3=100 and 0≤p0, p1, p2, p3≤100) were defined as the percentages of immunohistochemical intensity scores of 0, 1+, 2+ and 3+, respectively. A total score t=p0×0+p1×1+p2×2+p3×3 was then calculated. For example, the total score was 0 when p0=100%, while the total score was 300 when p3=100%; therefore, the range of total scores was 0–300. This method was adapted from a previous study.24 ,25 Stromal elements, especially endothelial cells, showed moderate to strong nuclear staining for ZEB1 in most cells and served as internal positive controls.19 Likewise, the internal control for E-cadherin staining was hepatocyte membranes, and the internal control for vimentin staining was hepatocyte cytoplasm.

    Statistical analysis

    The associations between ZEB1, E-cadherin, vimentin and clinicopathological features were assessed by using χ2 tests, and the correlations between the expressions of different proteins were examined with Spearman’s rank correlation tests. The overall survival curve was calculated by using the Kaplan–Meier method and analysed by using the log-rank test. Independent factors for survival were assessed with the Cox proportional hazard regression model. The level of statistical significance was set at p≦0.05. These analyses were performed with the SPSS V.20 statistical software package (SPSS Benelux, Gorinchem, The Netherlands).

    Results

    Expression of ZEB1, E-cadherin and vimentin in ICC by immunohistochemistry

    The frequencies of acquisition of markers ZEB1 and vimentin and the loss of marker E-cadherin in ICC were 46.1%, 43.1% and 44.1%, respectively. All non-cancerous liver cells were ZEB1 negative. Figure 1 shows expression of ZEB1, E-cadherin and vimentin at different levels in ICC, and figure 2 shows the immunohistochemistry scores for each marker. ZEB1 expression was localised in the nucleus, and the median score of ZEB1-positive cells was 40 (range, 0–170). E-cadherin expression was localised in the cytomembrane, and the median score of E-cadherin-positive cells was 140 (range, 0–300). Vimentin expression was localised in the cytoplasm, and the median score of vimentin-positive cells was 0 (range, 0–300). For each marker, we divided the cases into two groups (high expression and low expression) based on median expression scores.

    Figure 1
    Figure 1

    Representative microphotographs of zinc finger E-box binding homeobox 1 (ZEB1), E-cadherin and vimentin immunostaining by staining intensity in intrahepatic cholangiocarcinoma: strong expression of ZEB1, E-cadherin and vimentin (A, E and I); intermediate expression of ZEB1, E-cadherin and vimentin (B, F and J); weak expression of ZEB1, E-cadherin and vimentin (C, G and K) and negative expression of ZEB1, E-cadherin and vimentin (D, H and L). ZEB1, zinc finger E-box binding homeobox factor 1.

    Figure 2
    Figure 2

    Scores of immunohistochemical expression of zinc finger E-box binding homeobox 1 (ZEB1) (A), E-cadherin (B) and vimentin (C) in intrahepatic cholangiocarcinoma. The bar graph shows the number of cases with each score.

    Relationships between ZEB1, E-cadherin and vimentin expression and pathological characteristics of ICC

    We investigated the correlations among ZEB1, E-cadherin and vimentin expression. ZEB1 expression was inversely correlated with E-cadherin expression (p=0.004) and positively correlated with vimentin expression (p=0.022) (table 2). To further study the physiological or pathological relationship between expression of ZEB1, E-cadherin and vimentin in ICC, the clinicopathological parameters of the patients are summarised in table 1. ZEB1 overexpression was associated with aggressive tumour characteristics, including tumour node metastasis (TNM) tumour stage (p=0.037), undifferentiated-type histology (p=0.017), lymph node metastasis (p=0.024) and portal vein invasion (p=0.037). Likewise, overexpression of vimentin was associated with tumour progression, including tumour size (p=0.007) and portal vein invasion (p=0.027). However, no significant correlations were observed between E-cadherin expression and clinicopathological variables.

    View this table:
    Table 2

    Correlations among ZEB1, E-cadherin and vimentin expression in intrahepatic cholangiocarcinoma

    Association of ZEB1 expression with overall survival and case presentation

    We performed Kaplan–Meier analysis to determine whether ZEB1 expression was associated with overall survival of the patients with ICC (figure 3A). Postoperative 3-year survival rates of patients in the high ZEB1 expression group and low ZEB1 expression group were 33.5% and 64.9%, respectively. The overall survival rate was significantly lower in the patients in the high ZEB1 expression group than in the low ZEB1 expression group (p=0.027). Moreover, the overall survival rate was significantly lower in the patients in the low E-cadherin expression group than in the high E-cadherin expression group (figure 3B, 3-year 39.1% vs 58.8%, p=0.041) and in the high vimentin expression group than in the low vimentin expression group (figure 3C, 3-year 39.8% vs 57.3%, p=0.049).

    Figure 3
    Figure 3

    Kaplan–Meier curves illustrating zinc finger E-box binding homeobox 1 (ZEB1) expression (A), E-cadherin expression (B) and vimentin expression (C) for patients with intrahepatic cholangiocarcinoma. The 3-year survival rate was significantly lower in the ZEB1 high group than in the ZEB1 low group (33.5% vs 64.9%, p=0.027). (B) The 3-year survival rate was significantly lower in the E-cadherin low group than in the E-cadherin high group (39.1% vs 58.8%, p=0.041). (C) The 3-year survival rate was significantly lower in the ZEB1 high group than in the ZEB1 low group (39.8% vs 57.3%, p=0.049).

    Two cases of ICC with different expression levels of ZEB1, E-cadherin and vimentin are shown in figure 4. A 75-year-old man (case 1) presented with a massive, advanced ICC predominantly located in the right lobe of the liver, and a right hepatic lobectomy was performed. A histological examination revealed poorly differentiated adenocarcinoma (figure 4A) associated with portal invasion (vp1). ZEB1 and vimentin expression levels were markedly high (figure 4B, score 110; figure 4D, score 210), but E-cadherin expression was negative (figure 4C, score 0). Lung metastasis was diagnosed 8 months after hepatectomy, and the patient died 11 months after the operation. A 74-year-old woman (case 2) presented with massive, advanced ICC predominantly located in the right lobe of the liver and a right hepatic lobectomy was performed. A histological examination revealed moderately differentiated adenocarcinoma (figure 4E). ZEB1 and vimentin expression levels were low (figure 4F, score 20; figure 4H, score 0), but E-cadherin expression was high (figure 4G, score 230). There has been no evidence of recurrence in the 85 months since the operation.

    Figure 4
    Figure 4

    (A and E) H&E and (B–D and F–H) immunohistochemical staining of two intrahepatic cholangiocarcinoma specimens. Case 1 showed high levels of zinc finger E-box binding homeobox 1 (ZEB1) expression (B, score 110) and vimentin expression (D, score 210); however, E-cadherin expression was low (C, score 0). Case 2 showed low levels of ZEB1 expression (F, score 20) and vimentin expression (H, score 0), but E-cadherin expression was high (G, score 230).

    Analysis of prognostic factors in patients with ICC

    Table 3 shows the results of the univariate and multivariate analyses. Among the clinicopathological factors, male sex, tumour size, number of tumour nodules, extrahepatic metastasis, TNM stage, differentiation, lymph node metastasis, portal vein invasion, high ZEB1 expression, low E-cadherin expression and high vimentin expression, there is a detrimental association between these factors and overall survival that is statistically significant. In the multivariate analysis, the number of tumours (HR 2.266, 95% CI 1.011 to 5.082, p=0.047), differentiation (HR 2.155, 95% CI 1.175 to 3.952, p=0.013) and male sex (HR 2.405, 95% CI 1.234 to 4.687, p=0.01) were independent factors for poor prognosis among the significant prognostic factors.

    Discussion

    The EMT has been proposed as a key step during carcinoma progression and metastasis development.6 ,7 The hallmark of EMT is the loss of epithelial marker expression (eg, E-cadherin, catenin) and the induction of mesenchymal marker expression (eg, N-cadherin, vimentin, alpha-smooth muscle-actin).6–8 ZEB1 is a transcriptional repressor that functions as an inducer of EMT.10–13

    The expression of ZEB1 has a crucial impact on patient survival. ZEB1 expression in HCC,16 gastric carcinoma,17 colon cancer,18 pancreatic cancer,19 breast carcinoma20 and cervical cancer21 has been associated with aggressive disease, poor differentiation, development of metastases and poor clinical prognosis. In contrast, there have been no previous reports of the clinical implications of ZEB1 expression in ICC.

    In the present study, we performed immunohistochemical staining to detect the expression of ZEB1 in ICC specimens. Strong expression of ZEB1 was significantly associated with lymph node metastasis (p=0.024), portal vein invasion (p=0.037), undifferentiated-type histology (p=0.017) and advanced tumour stage (p=0.037). These results suggest that the overexpression of ZEB1 may play an important role in tumour progression and metastasis in ICC. In basic research studies of ZEB1 in ICC cells, Oishi showed that miR-200c was a common molecular note linking to EMT, and ZEB1 was significantly implicated in miR-200c by mRNA–miR pathway analysis in ICC samples.14 Furthermore, they showed that mesenchymal markers such as ZEB1 and vimentin were more abundantly expressed, whereas an epithelial marker, E-cadherin, was much less abundantly expressed in hepatic stem cell-like ICC cases as compared with mature hepatocyte-like ICC cases in mRNA levels. Another finding reported that ZEB1 correlates with the differentiation status of ICC cells, furthermore, Mizuguchi showed that small proline-rich protein 2a complexes with ZEB1 repress mirR-200c/141 transcription in ICC cells, thereby inducing EMT.15 These results were support of our finding of the clinical significance of ZEB1 in ICC, and ZEB1 is recognised as a key molecule of the EMT in ICC. Acquisition of an invasive phenotype through EMT, which enables cancer cells to break away from the primary tumour and invade surrounding tissues, may strongly promote the spread of cancer cells into the portal venous circulation and promote lymph node metastasis; thus, ZEB1 has been strongly implicated in the metastasis process.

    In the present study, ZEB1 expression showed a significant inverse correlation with E-cadherin expression (p=0.004) and a positive correlation with vimentin expression (p=0.022). These results indicate that some tumour cells in ICC might undergo EMT, as defined by the loss of E-cadherin and the upregulation of the mesenchymal markers ZEB1 and vimentin. ZEB1 induces several EMT-related markers; ZEB1 not only represses E-cadherin, but it upregulates vimentin. Previous clinical reports have correlated ZEB1 expression, E-cadherin loss and vimentin expression with poor prognosis in several cancers.16 ,17 ,20 Moreover, the loss of E-cadherin and overexpression of vimentin were associated with poor prognosis in ICC.26 ,27 However, to our knowledge, this is the first report to implicate ZEB1 protein expression in tumour progression and poor prognosis in patients with ICC. Based on our results as well as the role of ZEB1 as an inducible transcriptional factor of EMT, ZEB1 could be a molecular target for therapy against ICC.

    In the present study, multivariate analysis identified lymph node metastases, low differentiation and male to be independent predictors of mortality (table 3). Lymph node metastases and low differentiation were reported to be prognostic factors of ICC in a large-scale study,28 and our results are consistent with these previous results. Although ZEB1 expression was associated with prognosis in univariate analysis, it was not an independent prognostic factor in multivariate analysis. This result is likely because the number of tumours, which was a significant factor for prognosis, is not correlated with ZEB1 expression. Furthermore, several clinicopathological factors are thought to be confounding variables of ZEB1 expression in the prognosis of ICC. In a subgroup analysis for prognostic factors in ICC, ZEB1 expression was also not an independent prognostic factor, because there were too few cases in these subgroups. Although it was not statically significant, patients with high ZEB1 expression had a worse prognosis than patients with low ZEB1 expression in the subgroup without lymph node metastasis (data not shown; p=0.051). While lymph node metastasis remains the most relevant clinical prognosticator,27 the determination of appropriate management for high-risk individuals with no detectable lymph node metastasis at the point of diagnosis is important. Therefore, ZEB1 expression could be important in identifying patients without detectable regional metastasis who should nevertheless be treated aggressively. Further prospective studies should be able to determine the clinical value of ZEB1 expression in identifying patients with poor survival for appropriate management.

    View this table:
    Table 3

    Univariate and multivariate analysis of predictive factors for overall survival

    The present study has a few limitations. First, only patients with a medical record sufficient for evaluation were enrolled. Additionally, postoperative adjuvant chemotherapy was received by only 19 patients but was not assigned with a comprehensive organised strategy; therefore, the influence of adjuvant chemotherapy29 ,30 on clinical outcome is not clear.

    In summary, we found that ZEB1 expression is closely related to the malignant progression of ICC, and we also confirmed that the EMT phenomenon occurred in ICC. ZEB1 expression is associated with a poor prognosis, and therefore, might be an attractive target for the treatment of ICC. Further studies are needed to clarify the functional role of ZEB1 in EMT in ICC.

    Take home messages

    • The frequencies of acquisition of ZEB1 in intrahepatic cholangiocarcinoma (ICC) was 46.1%.

    • ZEB1 expression is associated with tumour progression and poor prognosis in patients with ICC through positive correlations with vimentin and negative correlations with E-cadherin.

    • ZEB1 proved to be a significant prognostic factor and therefore might be an attractive target for the treatment of patients with ICC.

    References

    1. ↵
      1. Jemal A,
      2. Bray F,
      3. Center M, et al
      . Global cancer statistics. CA Cancer J Clin 2011;61:69–90. doi:10.3322/caac.20107
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Rizvi S,
      2. Gores GJ
      . Pathogenesis, diagnosis, and management of cholangiocarcinoma. Gastroenterology 2013;145:1215–29. doi:10.1053/j.gastro.2013.10.013
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Tyson GL,
      2. El-Serag HB
      . Risk factors for cholangiocarcinoma. Hepatology 2011;54:173–84. doi:10.1002/hep.24351
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Mavros MN,
      2. Economopoulos KP,
      3. Alexiou VG, et al
      . Treatment and prognosis for patients with intrahepatic cholangiocarcinoma: systematic review and meta-analysis. JAMA Surg 2014;149:565–74. doi:10.1001/jamasurg.2013.5137
      OpenUrlCrossRefPubMed
    5. ↵
      1. Maithel SK,
      2. Gamblin TC,
      3. Kamel I, et al
      . Multidisciplinary approaches to intrahepatic cholangiocarcinoma. Cancer 2013;119:3929–42. doi:10.1002/cncr.28312
      OpenUrlCrossRefPubMed
    6. ↵
      1. Thiery JP,
      2. Acloque H,
      3. Huang RY, et al
      . Epithelial-mesenchymal transitions in development and disease. Cell 2009;139:871–90. doi:10.1016/j.cell.2009.11.007
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Kalluri R,
      2. Weinberg RA
      . The basics of epithelial-mesenchymal transition. J Clin Invest 2009;119:1420–8. doi:10.1172/JCI39104
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Wendt MK,
      2. Allington TM,
      3. Schiemann WP
      . Mechanisms of the epithelial-mesenchymal transition by TGF-beta. Future Oncol 2009;5:1145–68. doi:10.2217/fon.09.90
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Iwatsuki M,
      2. Mimori K,
      3. Fukagawa T, et al
      . The clinical significance of vimentin-expressing gastric cancer cells in bone marrow. Ann Surg Oncol 2010;17:2526–33. doi:10.1245/s10434-010-1041-0
      OpenUrlCrossRefPubMed
    10. ↵
      1. Zheng H,
      2. Kang Y
      . Multilayer control of the EMT master regulators. Oncogene 2014;33:1755–63. doi:10.1038/onc.2013.128
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Schmalhofer O,
      2. Brabletz S,
      3. Brabletz T
      . E-cadherin, beta-catenin, and ZEB1 in malignant progression of cancer. Cancer Metastasis Rev 2009;28:151–66. doi:10.1007/s10555-008-9179-y
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Yang Y,
      2. Ahn YH,
      3. Chen Y, et al
      . ZEB1 sensitizes lung adenocarcinoma to metastasis suppression by PI3K antagonism. J Clin Invest 2014;124:2696–708. doi:10.1172/JCI72171
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Wellner U,
      2. Schubert J,
      3. Burk UC, et al
      . The EMT-activator ZEB1 promotes tumorigenicity by repressing stemness-inhibiting microRNAs. Nat Cell Biol 2009;11:1487–95. doi:10.1038/ncb1998
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Oishi N,
      2. Kumar MR,
      3. Roessler S, et al
      . Transcriptomic profiling reveals hepatic stem-like gene signatures and interplay of miR-200c and epithelial-mesenchymal transition in intrahepatic cholangiocarcinoma. Hepatology 2012;56:1792–803. doi:10.1002/hep.25890
      OpenUrlCrossRefPubMed
    15. ↵
      1. Mizuguchi Y,
      2. Isse K,
      3. Specht S, et al
      . Small proline rich protein 2a in benign and malignant liver disease. Hepatology 2014;59:1130–43. doi:10.1002/hep.26889
      OpenUrlCrossRefPubMed
    16. ↵
      1. Zhou YM,
      2. Cao L,
      3. Li B, et al
      . Clinicopathological significance of ZEB1 protein in patients with hepatocellular carcinoma. Ann Surg Oncol 2012;19:1700–6. doi:10.1245/s10434-011-1772-6
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Jia B,
      2. Liu H,
      3. Kong Q, et al
      . Overexpression of ZEB1 associated with metastasis and invasion in patients with gastric carcinoma. Mol Cell Biochem 2012;366:223–9. doi:10.1007/s11010-012-1299-6
      OpenUrlCrossRefPubMed
    18. ↵
      1. Singh AB,
      2. Sharma A,
      3. Smith JJ, et al
      . Claudin-1 up-regulates the repressor ZEB-1 to inhibit E-cadherin expression in colon cancer cells. Gastroenterology 2011;141:2140–53. doi:10.1053/j.gastro.2011.08.038
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Bronsert P,
      2. Kohler I,
      3. Timme S, et al
      . Prognostic significance of Zinc finger E box binding homeobox 1 (ZEB1) expression in cancer cells and cancer-associated fibroblasts in pancreatic head cancer. Surgery 2014;156:97–108. doi:10.1016/j.surg.2014.02.018
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Geradts J,
      2. de Herreros AG,
      3. Su Z, et al
      . Nuclear Snail1 and nuclear ZEB1 protein expression in invasive and intraductal human breast carcinomas. Hum Pathol 2011;42:1125–31. doi:10.1016/j.humpath.2010.11.004
      OpenUrlCrossRefPubMed
    21. ↵
      1. Chen Z,
      2. Li S,
      3. Huang K
      . The nuclear protein expression levels of SNAI1 and ZEB1 are involved in the progression and lymph node metastasis of cervical cancer via the epithelial-mesenchymal transition pathway. Hum Pathol 2013;44:2097–105. doi:10.1016/j.humpath.2013.04.001
      OpenUrlCrossRefPubMed
    22. ↵
      1. Nakanuma Y,
      2. Sripa B,
      3. Vatanasapt V, et al
      . Intrahepatic cholangiocarcinoma. In Hamilton SR, Aaltonen LA eds. World Health Organization classification of tumors. Pathology and genetics of tumors of the digestive system. Lyon: IARC Press, 2000:173–80.
    23. ↵
      The Japan Liver Cancer Study Group. Classification of primary liver cancer. 3rd edn. Tokyo: Kanehara, 2010.
    24. ↵
      1. Behrens C,
      2. Lin HY,
      3. Lee JJ, et al
      . Immunohistochemical expression of basic fibroblast growth factor and fibroblast growth factor receptors 1 and 2 in the pathogenesis of lung cancer. Clin Cancer Res 2008;14:6014–22. doi:10.1158/1078-0432.CCR-08-0167
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. Yonemori K,
      2. Tsuta K,
      3. Ando M, et al
      . Contrasting prognostic implications of platelet-derived growth factor receptor-beta and vascular endothelial growth factor receptor-2 in patients with angiosarcoma. Ann Surg Oncol 2011;18:2841–50. doi:10.1245/s10434-011-1640-4
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. Ryu HS,
      2. Chung JH,
      3. Lee K, et al
      . Overexpression of epithelial-mesenchymal transition-related markers according to cell dedifferentiation: clinical implications as an independent predictor of poor prognosis in cholangiocarcinoma. Hum Pathol 2012;43:2360–70. doi:10.1016/j.humpath.2012.07.004
      OpenUrlCrossRefPubMed
    27. ↵
      1. Yao X,
      2. Wang X,
      3. Wang Z, et al
      . Clinicopathological and prognostic significance of epithelial mesenchymal transition-related protein expression in intrahepatic cholangiocarcinoma. Onco Targets Ther 2012;5:255–61. doi:10.2147/OTT.S36213
      OpenUrlPubMed
    28. ↵
      1. Uchiyama K,
      2. Yamamoto M,
      3. Yamaue H, et al
      . Impact of nodal involvement on surgical outcomes of intrahepatic cholangiocarcinoma: a multicenter analysis by the Study Group for Hepatic Surgery of the Japanese Society of Hepato-Biliary-Pancreatic Surgery. J Hepatobiliary Pancreat Sci 2011;18:443–52. doi:10.1007/s00534-010-0349-2
      OpenUrlCrossRefPubMed
    29. ↵
      1. Murakami Y,
      2. Uemura K,
      3. Sudo T, et al
      . Prognostic factors after surgical resection for intrahepatic, hilar, and distal cholangiocarcinoma. Ann Surg Oncol 2011;18:651–8. doi:10.1245/s10434-010-1325-4
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      1. Hoehn RS,
      2. Wima K,
      3. Ertel AE, et al
      . Adjuvant chemotherapy and radiation therapy is associated with improved survival for patients with extrahepatic cholangiocarcinoma. Ann Surg Oncol 2015. Published Online First.

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Handling editor Cheok Soon Lee

    • Contributors KT: immunohistochemical examinations and writing. MC: direction and statistical analysis. YH: immunohistochemical examinations. KH and TM: evaluation of immunostaining. HY, HK and TKak: surgery and follow-up survey. TO, AN, YY, HT and TKu: medical examination and follow-up survey. HI, NK, YK, AM, TKam and AT: surgery and medical examination. SM, TT, NSh and YM: pathological diagnosis. KO: medical examination. MN: direction and medical examination. NSa: direction.

    • Competing interests None declared.

    • Patient consent Obtained.

    • Provenance and peer review Not commissioned; externally peer reviewed.