- Email: [email protected]
Prognostic and clinical significance of claudin-1 in colorectal cancer: A systemic review and meta-analysis
International Journal of Surgery 39 (2017) 214e220
Contents lists available at ScienceDirect
International Journal of Surgery journal homepage: www.journal-surgery.net
Review
Prognostic and clinical significance of claudin-1 in colorectal cancer: A systemic review and meta-analysis Lu Jiang a, 1, Long Yang a, 1, He Huang a, Bai-ying Liu a, **, Guo Zu b, * a b
Dalian Medical University, China Department of Gastrointestinal Surgery, The Dalian Municipal Center Hospital Affiliated of Dalian Medical University, China
h i g h l i g h t s The studies of claudin-1 expression in colorectal cancer(CRC) have become more and more, whereas there still exist different arguments about the effects of claudin-1 in colorectal cancer. Low expression of claudin-1 is associated with TNM III-IV stage and poor prognosis of CRC patients. Low expression of claudin-1 is not associated with gender, depth of invasion, lymph node involvement, tumors’ differentiation of CRC patients. Claudin-1 is a candidate novel prognostic biomarker for CRC patients.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 10 November 2016 Received in revised form 5 February 2017 Accepted 6 February 2017 Available online 10 February 2017
Background: The current reports on the association of claudin-1 expression with colorectal cancer (CRC) result were controversial. Thus, we conducted a meta-analysis to assess the correlation between claudin1 expression and the clinical parameters and assess the prognostic value of claudin-1 in CRC. Methods: Systematic searches on PubMed, Embase, Elsevier, CNKI (China National Knowledge Infrastructure), Wanfang data and Cochrane Library prior to August 2016 were performed. The pooled odds ratio (OR) with its 95% confidence interval (95 %CI) was used to assess association between claudin-1 expression and clinical parameters of CRC patients, and to assess association between claudin-1 expression and the prognostic value of CRC patients. Results: Eight studies with a total of 1146 CRC patients were included. Overall, the pooled results showed that low expression of claudin-1 was associated with TNM III-IV stage of CRC patients (OR: 1.714, 95%CI: 1.215e2.418, P ¼ 0.002). Low expression of claudin-1 was also associated with a poor survival in CRC patients (one year survival rate: OR: 2.112, 95%CI: 1.028e4.339, P ¼ 0.042; three years survival rate: OR: 1.501, 95%CI: 1.030e2.186, P ¼ 0.035; five years survival rate: OR: 1.794, 95%CI: 1.139e2.439, P ¼ 0.000). Whereas, low expression of claudin-1 is not associated with gender (OR: 1.259, 95%CI: 0.957e1.657, P ¼ 0.100), tumors' differentiation (OR: 1.317, 95%CI: 0.916e1.892, P ¼ 0.137), depth of invasion (OR: 1.016, 95 %CI: 0.701e1.472, P ¼ 0.935) and lymph node metastasis group (OR: 1.286, 95% CI: 0.982e1.684, P ¼ 0.06) of CRC. Conclusions: Low expression of claudin-1 is associated with TNM III-IV stage and poor prognosis of CRC patients. Low expression of claudin-1 is not associated with gender, tumors' differentiation depth of invasion and lymph node involvement of CRC patients. © 2017 Published by Elsevier Ltd on behalf of IJS Publishing Group Ltd.
Keywords: Claudin-1 Colorectal cancer Prognosis Meta-analysis
1. Introduction
* Corresponding author. Gastrointestinal surgery, The Dalian Municipal Center Hospital Affiliated of Dalian Medical University, No. 826 Southwest Road, Shahekou District, Dalian, 116027, China. ** Corresponding author. E-mail address: [email protected] (G. Zu). 1 Co-first author. http://dx.doi.org/10.1016/j.ijsu.2017.02.005 1743-9191/© 2017 Published by Elsevier Ltd on behalf of IJS Publishing Group Ltd.
Colorectal cancer (CRC) is one of the most common malignant cancers of digestive system, it is estimated that an annual increase of one million new cases in the world [1,2]. Despite the advancement of medical treatment, approximately a half million of CRC patients die of CRC or primary metastatic disease [3]. It has been
L. Jiang et al. / International Journal of Surgery 39 (2017) 214e220
done in a try to understand the pathogenesis and assess the prognosis of CRC, So, there is an urgent need to find markers to estimate the prognosis of CRC that can stratify patients earlier and better by their risk of CRC recurrence and overall survival (OS) [4,5]. Claudins are major integral cell membrane tight junction proteins which play a vital role in the formation and function of tight junctions and claudin family consists of at least 24 members [6e8]. The main functions of claudin-1 are to regulate paracellular permeability and maintain the cell polarity [8]. These functions are believed to keep inter-cellular adhesion, which prevent the release of cancer cells from the primary cancer nests [9]. Simultaneously, it also decrease the access of tumor cells to nutrients and signaling peptides [10,11]. It is now clear that claudin-1 is frequently reside in various cancers. For example, some studies have demonstrated claudin-1 is frequently up-regulated in pancreatic and thyroid cancers [12e16]. However, some reports showed that claudin-1 also have abnormal expression in CRC [17]. But, the relationship between claudin-1 expression and clinical parameters, such as gender, tumors' differentiation, depth of invasion, lymph node metastasis, TNM stage and OS in CRC, are yet inconsistent. Masatsune et al. reported low claudin-1 expression is associated with lymph node metastasis, tumors' differentiation, extent of poorlydifferentiated component, reduced OS [6]. However, Youshida et al. indicated that the expression levels of claudin-1 is not significant differences with gender, depth of invasion, lymph node metastasis and TNM stage [11]. Considering the inconsistent results of current findings, we, therefore, performed a meta-analysis of all eligible studies available to explore the relationship between levelexpression of claudin-1 and the CRC. 2. Methods 2.1. Search strategy We searched PubMed, Embase, Elsevier, CNKI(China National Knowledge Infrastructure), Wanfang data and Cochrane Library for all articles and conference abstracts to identify relevant articles published prior to August 2016. No language restriction was applied. The following key words were used: “colorectal cancer (carcinoma)” or “rectal cancer (carcinoma)” or “colonic cancer (carcinoma)”, “claudin-1” or survival or prognosis. Reference lists of the identified articles were examined and the literature retrieval was performed in duplication by two independent reviewers. 2.2. Inclusion and exclusion criteria Criteria for eligibility of a study in this meta-analysis were: (1) the patients were diagnosed clearly with CRC and investigated for claudin-1 expression status. (2) study is focused on the relationship between claudin-1 expression and clinicopathologic parameters or/and overall survival rates in CRC; (3) immunohistochemistry (IHC) was used as the main method to determine claudin-1 expression in CRC specimens. (4) Hazard ratios (HR) for overall survival rates according to claudin-1 expression were reported or could be calculated from the published data. Studies were excluded if (1) studies without original data; animal or laboratory studies; (2) the samples came from lymph nodes or the peritoneal cavity; (3) repeated studies were based on the same database or patients; (4) none of these patients received adjuvant chemoradiation. 2.3. Quality assessment The quality of each study was assessed using the Newcastle Ottawa Quality Assessment Scale (NOQAS) by two independent
215
reviewers [8]. These scales were used to allocate a maximum of nine points for quality of selection, comparability, exposure, and outcome of study participants. The studies considered to be of high methodological quality (score above 6) were included in this meta-analysis. 2.4. Data extraction All data were extracted by two independent reviewers. Discrepancies were resolved by discussions and referring to the contents of the articles. We extracted the name of first author, year of publication, region or country where the study was conducted, name of journal, IHC methodology, size of study population and quality score of each study. In studies that reported OR in both univariate and multivariate models, we extracted the latter because these results were more convincing, as there had been adjustment for potential confounders. 2.5. Statistical analysis The odds ratios (OR) was used to quantitatively determine the association between claudin-1 expression and clinicopathologic parameters of CRC, including age, gender, tumors' differentiation, depth of invasion, lymph node metastasis, TNM-stage, while the OR was used to quantitatively assesses the association between claudin-1 expression and survival rate of CRC. If the OR and 95% confidence intervals (CI) was not provided directly in the study, we can use method reported by Tierney [18]. To calculate these value available data in study. Heterogeneity assumption was evaluated using the chi-squared based Q-test and I2 test. The mild heterogeneity is I2 values < 25%, I2 values between 25 and 50% is a moderate heterogeneity. I2 values > 50% and P-values0.1 were considered to have a large heterogeneity [19]. For a large heterogeneity, we can use a fixed-effects model or random-effects model calculate the pooled estimates. A diagnosis of publication bias was provided by Egger test and inverted funnel plots. All statistical tests in this meta-analysis were performed using Stata10.0 software (Stata Corporation, College Station, TX, USA) with two-tail P values. P value < 0.05 was considered statistically significant. 3. Results 3.1. Search results and study characteristics The initial search yielded 35 studies. After screening of the titles, abstracts, 18 records were excluded. Among the other 17 articles, 9 articles were excluded, including cell assay only (N ¼ 3), lack of clinical data and statistical analysis (N ¼ 4), abstract only (N ¼ 1), review only (N ¼ 1). Thus a total of eight studies were eventually included in this study based on the predefined criteria [2,6,9,11,17,27e29]. Fig. 1 details the selection process. Among these eight studies, four were from Japan, two were from China, one was from USA, and one was from Egypt. Altogether, these eight studies recruited a total of 1164 CRC, with sample sizes ranged from 50 to 344. Eight articles were published between 2005 and 2013. All the studies used IHC methods for claudin-1 staining. Their basic characteristics and quality are summarized in Table 1. 3.2. Claudin-1 expression and gender The pooled results of five studies with 976 patients, including 570 males and 406 females, failed to show the significant association between expression of claudin-1 and gender of CRC (OR: 1.259, 95%CI: 0.957e1.657, P ¼ 0.100). See Fig. 2. The results of this metaanalysis showed low expression of claudin-1 was not associated with gender.
216
L. Jiang et al. / International Journal of Surgery 39 (2017) 214e220
Fig. 1. Flow chart for selection of studies for inclusion in this meta-analysis.
Table 1 Main characteristics and result of the eligible studies. Author
Year/country
Journal
Methods
Patient number
Gender (M/F)
Age
NOS score
Shibutani Yoshiday Resnick Matauoka Huang Wang Abdelzaher Nakagawa
2013/Japan 2011/Japan 2005/USA 2011/Japan 2012/China 2012/China 2011/Egypt 2011/Japan
Anticancer Research Anticancer Research Modern Pathology Journal of Surgical Oncology Chines Journal of Histochemistry and Cytochem National Medical Frontier of China Journal of the Egyptian National cancer International Journal of Oncology
IHC IHC IHC IHC IHC IHC IHC IHC
344 306 129 156 51 60 50 119
197/147 199/107 60/69 99/57 30/21 28/28 18/32 70/45
66.8 63.6 72.5 65.5 61.9 65 52.6 68.5
8 8 7 8 6 6 7 7
3.3. Claudin-1 expression and tumors' differentiation We found that expression of claudin-1 was not associated with tumor differentiation, after we pooled the data of five studies with
1164 patients of tumor differentiation (OR: 1.317, 95%CI: 0.916e1.892, P ¼ 0.137). See Fig. 5. The results of this meta-analysis showed low expression of claudin-1 was not associated with low differentiation of tumor. See Fig. 3.
Fig. 2. Pooled analysis for the association between claudin-1 expression and gender (a: forest plots; b: funnel plots).
L. Jiang et al. / International Journal of Surgery 39 (2017) 214e220
Regarding to large heterogeneity, we adopt random-effects model. (OR: 1.139, 95%CI: 0.511e2.535, P ¼ 0.750, I2 ¼ 68.7%). The results of this meta-analysis showed it is no significant correlation between low expression of claudin-1 and low differentiation of tumor. 3.4. Claudin-1 expression and depth of invasion The ORs for depth of invasion were in five studies, including 665 cases with T3-4 categories and 154 cases with T1-2 categories. Our results show that claudin-1 expression was not associated with depth of invasion (OR: 1.016, 95 %CI: 0.701e1.472, P ¼ 0.935). See Fig. 4. The results of this meta-analysis showed low expression of claudin-1 was not associated with positive depth of invasion. 3.5. Claudin-1 expression and lymph node involvement Also, in this study, claudin-1 expression was not associated with lymph node involvement after we pooled the data of five studies with 930 patients of lymph node metastasis group (OR: 1.286, 95% CI: 0.982e1.684, P ¼ 0.06). See Fig. 5. The results of this metaanalysis showed low expression of claudin-1 was not associated with positive lymph node involvement. Regarding to large heterogeneity, we adopt random-effects model. (OR: 1.278, 95%CI: 0.948e1.722, P ¼ 0.108, I2 ¼ 10.2%). However, no significant correlation were detected between the low expression of claudin-1 and positive lymph node involvement. 3.6. Claudin-1 expression and TNM stage Five studies with 621 patients investigated the association between claudin-1 expression and the tumor stage of CRC. The metaanalysis showed that expression of claudin-1 was associated with TNM stage (OR: 1.714, 95%CI: 1.215e2.418, P ¼ 0.002). See Fig. 6. The results of this meta-analysis showed low expression of claudin1 was significantly associated with TNM III-IV stage. Claudin-1 loss based on TNM results in each study are summarized in Table 2. 3.7. Claudin-1 expression and OS In a pooled analysis including five studies with 1054 patients on prognostic effect for claudin-1 expression in CRC, we found that low-expression of claudin-1 was significantly corrected with a poor survival in CRC patients (one year survival rate: OR: 2.112, 95%CI: 1.028e4.339, P ¼ 0.042; three years survival rate: OR: 1.501, 95%CI: 1.030e2.186, P ¼ 0.035; five years survival rate: OR: 1.794, 95%CI: 1.139e2.439, P ¼ 0.000). See Fig. 7. The results of this meta-analysis
217
showed low expression of claudin-1 tend to have a poor OS.
4. Discussion CRC has become increasingly common, especially in big cities. Worldwide, more than one million individual will develop CRC every year, and recurrence, metastasis and cancer-related disease are chief cause of early death in patients of CRC [20,21]. Early diagnosis and treatment of CRC is becoming more and more significant. To our present knowledge, the results of this meta-analysis documented that the claudin-1 was associated with TNM stage and prognosis. Claudin-1 was not significant differences with gender, tumors' differentiation, depth of invasion and lymph node involvement [11]. Claudin-1 is an important marker for assessing prognosis of CRC, it can enable us to have an early intervence to prevention, diagnosis and treatment of CRC after undergoing surgery [4,21]. Recent results showed that expression levels of claudin-1 in high TNM stage (III-IV) of CRC patients was more frequently downregulated as compared to the low TNM stage (I-II) of CRC patients [22e24]. Eman Abdelzaher showed a statistically significant strong inverse correlation between claudin-1 expression and tumors' TNM stage [2]. Several studies reported that low expression of claudin-1 was significantly associated with lower differentiation of the tumor which is until now considered the most significant prognosis factor in CRC [25,26]. The results of these studies suggest that low expression of claudin-1 in CRC patience of lower differentiation was due to its down-regulation might contribute to the more aggressive behavior of the tumors. However, Yoshida has reported that the expression levels of claudin-1 was not significantly associated with TNM stage of CRC patients [2]. Our meta-analysis showed that low expression of claudin-1 was associated with TNM III-IV stage (OR: 1.714). Claudin-1 expression was demonstrated to associate with tumors' differentiation, depth of invasion and lymph node involvement [2,22,26]. The loss of the tight junction and structure may affect tumor penetration which is the first step in the progression of tumor invasion and metastasis [11,27]. Some studies indicate that claudin-1 expression is no significant differences with depth of invasion, lymph node involvement [11]. For example, Yoshida et al. reported that expression levels of claudin-1 was not associated with age, gender, depth of invasion, lymph node metastasis except for tumors' differentiation and perineural invasion. Simultaneously, our meta-analysis show claudin-1 is not associated with gender, tumors' differentiation, depth of invasion, lymph node involvement of CRC patients. Many previous studies have reported that the relationship
Fig. 3. Pooled analysis for the association between claudin-1 expression and tumors' differentiation (a: forest plots; b: funnel plots).
218
L. Jiang et al. / International Journal of Surgery 39 (2017) 214e220
Fig. 4. Pooled analysis for the association between claudin-1 expression and depth of invasion (a: forest plots; b: funnel plots).
Fig. 5. Pooled analysis for the association between claudin-1 expression lymph node involvement (a: forest plots; b: funnel plots).
Fig. 6. Pooled analysis for the association between claudin-1 expression and TNM stage (a: forest plots; b: funnel plots).
Table 2 Expression of claudin-1 based on TNM results in the patients of CRC. Author
Patients number
TNM stage I ~ II (number)
III ~ IV (number)
Shibutani Yoshiday Resnick Matauoka Huang Wang Abdelzaher Nakagawa
344 306 129 156 51 60 50 119
no 173 no 72 27 25 34 no
no 133 no 84 24 35 16 no
High expression of claudin-1 (number)
Low expression of claudin-1 (number)
234 143 96 32 35 46 21 59
110 163 31 124 16 14 29 60
L. Jiang et al. / International Journal of Surgery 39 (2017) 214e220
219
Fig. 7. Pooled analysis for the association between claudin-1 expression and one year survival rate (a: forest plots, b: funnel plots), three year survival rate (c: forest plots, d: funnel plots), five year survival rate (e: forest plots, f:funnel plots).
between low expression of claudin-1 and markers of prognosis of CRC. In the present study, the clinicopathologic analysis showed that high expression of claudin-1 have a better prognosis for OS than the low expression group [17]. The main reason for poor survival of CRC is that loss of claudin-1 expression contribute to tumor invasion and metastasis, and loss of claudin-1 expression was a strong predictor of disease recurrence and poor survival. Meanwhile, our data of meta-analysis show that low expression of claudin-1 is associated with one, three and five years' OS. Although our meta-analysis was robust in identifying a correlation between low expression of claudin-1 and poor clinical results in CRC. However, some limitations should be included in our metaanalysis. First, most of the populations of the studies is from East Asia, so this may not generalize global populations. Second, some individual data could not be obtained from articles, we gained these data by assessing survival curves in the articles, heterogeneity was not neglected completely. Moreover, this meta-analysis of sample
size of cases is limited, which need a large sample sizes to confirm these conclusions further. In conclusion, this meta-analysis shows that low expression of claudin-1 is associated with high TNM stage (III-IV) and poor prognosis of CRC patients. Simultaneously, low expression of claudin-1 was not associated with gender, tumors' differentiation, depth of invasion, lymph node metastasis. Claudin-1 is likely to become a novel marker for prognosis of CRC. Ethical approval No Ethical Approval was given. Sources of funding None.
220
L. Jiang et al. / International Journal of Surgery 39 (2017) 214e220
Author contribution Lu Jiang, Long Yang and Guo Zu: study design. Lu Jiang, Long Yang, He Huang, Bai-ying Liu and Guo Zu: data collection, data analysis. Lu Jiang, Long Yang, Bai-ying Liu and Guo Zu: writing. Conflicts of interest None to declare. Research registration unique identifying number (UIN) Register a Systematic Review/Meta-Analysis and Review Registry UIN is reviewregistry151. References [1] K.H. Jun, J.H. Kim, J.H. Jung, H.J. Choi, H.M. Chin, Expression of claudin-7 and loss of claudin-18 correlate with poor prognosis in gastric cancer, Int. J. Surg. 12 (2) (2014) 156e162. [2] Eman Abdelzaher, Azza Mohamed Rizk, Samer Saad Bessa, Khalafalla Mustafa Omer, Predictive value of immunohistochemical expression of claudin-1 in colonic carcinoma, J. Egypt Natl. Cancer Inst. 23 (4) (2011 Dec) 123e131. [3] D.M. Parkin, F. Bray, J. Ferlay, P. Pisani, Global cancer statistics, CA Cancer J. Clin. 55 (2) (2002) (2005 Mar-Apr) 74e108. [4] M.S. Kahlenberg, J.M. Sullivan, D.D. Witmer, N.J. Petrelli, Molecular prognostics in colonic cancer, Surg. Oncol. 12 (3) (2003) 173e186. [5] A. Walther, E. Johnstone, C. Swanton, R. Midgley, I. Tomlinson, D. Kerr, Genetic prognostic and predictive markers in colonica cancer, Nat. Rev. Cancer 11 (4) (2011) 309. [6] Masatsune Shibutani, Eiji Noda, Kiyoshi Maeda, Hisashi Naganara, Hiroshi Ohtani, Kosei Hirakawa, Low expression of Claudin-1 and presence of poorly-differentiated tumor clusters correlate with poor prognosis in colorectal cancer, Anticancer Res. 33 (8) (2013 Aug) 3301e3306. [7] M. Furuse, K. Fujita, T. Hiiragi, K. Fujimoto, S. Tsukita, Claudin-1 and -2: novel integral membrane proteins localizing at tight junctions with no sequence similarity to occludin, J. Cell Biol. 141 (7) (1998 Jun) 1539e1550. [8] S. Tsukita, M. Furuse, M. Itoh, Multifunctional strands in tight junctions, Nat. Rev. Mol. Cell Biol. 2 (4) (2001 Apr) 285e293. [9] Takashi Matsuoka, Hiroyuki Mitomi, Naoshi Fukui, Hideki Kanazawa, Tsuyoshi, Saito, Takuo Hayashi, Takashi Yao, Cluster analysis of Claudin-1 and -4, e-cadherin, and b-catenin expression in colorectal cancers, J. Surg. Oncol. 103 (7) (2011 Jun 1) 674e686. [10] J.M. Mullin, Epithelial barriers, compartmentation, and cancer, Sci. STKE 2004 (216) (2004 Jan 13) 2. [11] Takefumi Yoshida, Testsushi Kinugasa, Yoshito Akagi, Akihiko Kawahara, Kansakar Romeo, Ichitarou Shiratsuchi, Yasuhiko Ryu, Yukito Gotanda, Kazuo Shirouzu, Decreased expression of Claudin-1 in rectal cancer: a factor for recurrence and poor prognosis, Anticancer Res. 31 (7) (2011 Jul)
2517e2525. [12] Safak Ersoz, Sevdegul Mungan, Umit Cobanoglu, Havvanur Turgutalp, Yavuz Ozoran, Prognostic importance of Claudin-4 and Claudin-4 expression in colon carcinomas, Pathology -Research and Practice 2011(207) 285-289, Pathol. Res. Pract. 207 (5) (2011 May 15) 285e289. [13] S.S. Oliveria, I.M. Oliveria, V. DeSouza, J.A. Morgado-Diaz, Claudins upregulation in human colorectal cancer, FEBS Lett. 579 (2005) 6179e6185. [14] N. Miwa, M. Furuse, S. Tsukita, N. Niikawa, Y. Nakamura, Y. Furukawa, Involvement of claudin-1 in the beta-catenin/Tcf signaling pathway and its frequent upregulation in human colorectal cancers, Oncol. Res. 12 (11e12) (2001) 469e476. [15] C.A. Iacobuzio-Donahue, A. Maitra, G.L. Shen-Ong, T. van Heek, R. Ashfaq, R. Meyer, et al., Discovery of novel tumor markers of pancreatic cancer using global gene expression technology, Am. J. Pathol. 160 (4) (2002 Apr) 1239e1249. [16] O. Fluge, O. Bruland, L.A. Akslen, J.R. Lillehaug, J.E. Varhaug, Gene expression in poorly differentiated papillary thyroid carcinomas, Thyroid 16 (2) (2006 Feb) 161e175. [17] Sumiko Nakagawa, Norikatsu Miyoshi, Hideshi Ishii, Koshi Mimori, Fumiaki Tanaka, Mitsugu Sekimoto, Yuichiro Doki, Masaki Mori, Expression of CLDN1 in coloractal cancer: a novel marker for prognosis, Int. J. Oncol. 39 (4) (2011 Oct) 791e796. [18] J.F. Tierney, L.A. Stewart, D. Ghersi, S. Burdett, M.R. Sydes, Practical methods for incorporating summary time-to-event data into meta-analysis, Trials 8 (2007 Jun 7) 16. [19] J.P. Higgins, S.G. Thompson, Quantifying heterogeneity in a meta-analysis, Stat. Med. 21 (11) (2002) 1539e1558. [20] D. Cunningham, Colonic cancer reply, Lancet 376 (9738) (2010) 331e332. [21] Guo Zu, Anlong Ji, Tingting Zhou, Ningwei Che, Clinicopathological significance of SIRT1 expression in coloretal cancer: a systematic review and meta analysis, Int. J. Surg. 26 (2016 Feb) 32e37. [22] T. Matsuoka, H. Mitomi, N. Fukui, H. Kanazawa, T. Saito, T. Hayashi, et al., Cluster analysis of claudin-1 and -4, Ecadherin,and beta-catenin expression in colonic cancers, J. Surg. Oncol. 103 (7) (2011 Jun 1) 674e686. [23] V.N. Tzelepi, A.C. Tsamandas, H.D. Vlotinou, C.E. Vagianos, C.D. Scopa, Tight junctions in thyroid carcinogenesis: diverse expression of claudin-1, claudin4, claudin-7 and occludin in thyroid neoplasms, Mod. Pathol. 21 (1) (2008 Jan) 22e30. [24] M. Resnick, T. Konkin, J. Routhier, E. Sabo, V. Pricolo, Claudin-1 is a strong prognostic indicator in stage II colonic cancer: a tissue microarray study, Mod. Pathol. 18 (4) (2005 Apr) 511e518. [25] J. Kehoe, V.P. Khatri, Staging and prognosis of colon cancer, Surg. Oncol. Clin. N. Am. 15 (1) (2006) 129. [26] S. Ersoz, S. Mungan, U. Cobanoglu, H. Turgutalp, Y. Ozoran, Prognostic importance of claudin-1 and claudin-4 expression in colon carcinomas, Pathol. Res. Pract. 207 (5) (2011 May 15) 285e289. [27] Murray B. Resnick, Tamako, Justin Routhier, Edmond Sabo, Victor E. Pricolo, Claudin-1 is a strong prognostic indicator in stage II colonic cancer: a tissue microarray study, Mod. Pathol. 18 (4) (2005 Apr) 511e518. [28] Wang Lin, Li shi-yong. Expression of Claudin-1, Claudin-4 and correlation study, in: Colorectal Cancer. Nation Medical Frontiers of China, 7(2), 2012, pp. 9e11. [29] Huang Chengcheng, Zhou Xuchan, Liu Wenxiu, Expression and significance of Claudin1, ZEB-1 and e-cadherin in colorectal carcinoma, Chin. J. Histochem. Cytochem. 21 (6) (2012) 578e583.
Contents lists available at ScienceDirect
International Journal of Surgery journal homepage: www.journal-surgery.net
Review
Prognostic and clinical significance of claudin-1 in colorectal cancer: A systemic review and meta-analysis Lu Jiang a, 1, Long Yang a, 1, He Huang a, Bai-ying Liu a, **, Guo Zu b, * a b
Dalian Medical University, China Department of Gastrointestinal Surgery, The Dalian Municipal Center Hospital Affiliated of Dalian Medical University, China
h i g h l i g h t s The studies of claudin-1 expression in colorectal cancer(CRC) have become more and more, whereas there still exist different arguments about the effects of claudin-1 in colorectal cancer. Low expression of claudin-1 is associated with TNM III-IV stage and poor prognosis of CRC patients. Low expression of claudin-1 is not associated with gender, depth of invasion, lymph node involvement, tumors’ differentiation of CRC patients. Claudin-1 is a candidate novel prognostic biomarker for CRC patients.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 10 November 2016 Received in revised form 5 February 2017 Accepted 6 February 2017 Available online 10 February 2017
Background: The current reports on the association of claudin-1 expression with colorectal cancer (CRC) result were controversial. Thus, we conducted a meta-analysis to assess the correlation between claudin1 expression and the clinical parameters and assess the prognostic value of claudin-1 in CRC. Methods: Systematic searches on PubMed, Embase, Elsevier, CNKI (China National Knowledge Infrastructure), Wanfang data and Cochrane Library prior to August 2016 were performed. The pooled odds ratio (OR) with its 95% confidence interval (95 %CI) was used to assess association between claudin-1 expression and clinical parameters of CRC patients, and to assess association between claudin-1 expression and the prognostic value of CRC patients. Results: Eight studies with a total of 1146 CRC patients were included. Overall, the pooled results showed that low expression of claudin-1 was associated with TNM III-IV stage of CRC patients (OR: 1.714, 95%CI: 1.215e2.418, P ¼ 0.002). Low expression of claudin-1 was also associated with a poor survival in CRC patients (one year survival rate: OR: 2.112, 95%CI: 1.028e4.339, P ¼ 0.042; three years survival rate: OR: 1.501, 95%CI: 1.030e2.186, P ¼ 0.035; five years survival rate: OR: 1.794, 95%CI: 1.139e2.439, P ¼ 0.000). Whereas, low expression of claudin-1 is not associated with gender (OR: 1.259, 95%CI: 0.957e1.657, P ¼ 0.100), tumors' differentiation (OR: 1.317, 95%CI: 0.916e1.892, P ¼ 0.137), depth of invasion (OR: 1.016, 95 %CI: 0.701e1.472, P ¼ 0.935) and lymph node metastasis group (OR: 1.286, 95% CI: 0.982e1.684, P ¼ 0.06) of CRC. Conclusions: Low expression of claudin-1 is associated with TNM III-IV stage and poor prognosis of CRC patients. Low expression of claudin-1 is not associated with gender, tumors' differentiation depth of invasion and lymph node involvement of CRC patients. © 2017 Published by Elsevier Ltd on behalf of IJS Publishing Group Ltd.
Keywords: Claudin-1 Colorectal cancer Prognosis Meta-analysis
1. Introduction
* Corresponding author. Gastrointestinal surgery, The Dalian Municipal Center Hospital Affiliated of Dalian Medical University, No. 826 Southwest Road, Shahekou District, Dalian, 116027, China. ** Corresponding author. E-mail address: [email protected] (G. Zu). 1 Co-first author. http://dx.doi.org/10.1016/j.ijsu.2017.02.005 1743-9191/© 2017 Published by Elsevier Ltd on behalf of IJS Publishing Group Ltd.
Colorectal cancer (CRC) is one of the most common malignant cancers of digestive system, it is estimated that an annual increase of one million new cases in the world [1,2]. Despite the advancement of medical treatment, approximately a half million of CRC patients die of CRC or primary metastatic disease [3]. It has been
L. Jiang et al. / International Journal of Surgery 39 (2017) 214e220
done in a try to understand the pathogenesis and assess the prognosis of CRC, So, there is an urgent need to find markers to estimate the prognosis of CRC that can stratify patients earlier and better by their risk of CRC recurrence and overall survival (OS) [4,5]. Claudins are major integral cell membrane tight junction proteins which play a vital role in the formation and function of tight junctions and claudin family consists of at least 24 members [6e8]. The main functions of claudin-1 are to regulate paracellular permeability and maintain the cell polarity [8]. These functions are believed to keep inter-cellular adhesion, which prevent the release of cancer cells from the primary cancer nests [9]. Simultaneously, it also decrease the access of tumor cells to nutrients and signaling peptides [10,11]. It is now clear that claudin-1 is frequently reside in various cancers. For example, some studies have demonstrated claudin-1 is frequently up-regulated in pancreatic and thyroid cancers [12e16]. However, some reports showed that claudin-1 also have abnormal expression in CRC [17]. But, the relationship between claudin-1 expression and clinical parameters, such as gender, tumors' differentiation, depth of invasion, lymph node metastasis, TNM stage and OS in CRC, are yet inconsistent. Masatsune et al. reported low claudin-1 expression is associated with lymph node metastasis, tumors' differentiation, extent of poorlydifferentiated component, reduced OS [6]. However, Youshida et al. indicated that the expression levels of claudin-1 is not significant differences with gender, depth of invasion, lymph node metastasis and TNM stage [11]. Considering the inconsistent results of current findings, we, therefore, performed a meta-analysis of all eligible studies available to explore the relationship between levelexpression of claudin-1 and the CRC. 2. Methods 2.1. Search strategy We searched PubMed, Embase, Elsevier, CNKI(China National Knowledge Infrastructure), Wanfang data and Cochrane Library for all articles and conference abstracts to identify relevant articles published prior to August 2016. No language restriction was applied. The following key words were used: “colorectal cancer (carcinoma)” or “rectal cancer (carcinoma)” or “colonic cancer (carcinoma)”, “claudin-1” or survival or prognosis. Reference lists of the identified articles were examined and the literature retrieval was performed in duplication by two independent reviewers. 2.2. Inclusion and exclusion criteria Criteria for eligibility of a study in this meta-analysis were: (1) the patients were diagnosed clearly with CRC and investigated for claudin-1 expression status. (2) study is focused on the relationship between claudin-1 expression and clinicopathologic parameters or/and overall survival rates in CRC; (3) immunohistochemistry (IHC) was used as the main method to determine claudin-1 expression in CRC specimens. (4) Hazard ratios (HR) for overall survival rates according to claudin-1 expression were reported or could be calculated from the published data. Studies were excluded if (1) studies without original data; animal or laboratory studies; (2) the samples came from lymph nodes or the peritoneal cavity; (3) repeated studies were based on the same database or patients; (4) none of these patients received adjuvant chemoradiation. 2.3. Quality assessment The quality of each study was assessed using the Newcastle Ottawa Quality Assessment Scale (NOQAS) by two independent
215
reviewers [8]. These scales were used to allocate a maximum of nine points for quality of selection, comparability, exposure, and outcome of study participants. The studies considered to be of high methodological quality (score above 6) were included in this meta-analysis. 2.4. Data extraction All data were extracted by two independent reviewers. Discrepancies were resolved by discussions and referring to the contents of the articles. We extracted the name of first author, year of publication, region or country where the study was conducted, name of journal, IHC methodology, size of study population and quality score of each study. In studies that reported OR in both univariate and multivariate models, we extracted the latter because these results were more convincing, as there had been adjustment for potential confounders. 2.5. Statistical analysis The odds ratios (OR) was used to quantitatively determine the association between claudin-1 expression and clinicopathologic parameters of CRC, including age, gender, tumors' differentiation, depth of invasion, lymph node metastasis, TNM-stage, while the OR was used to quantitatively assesses the association between claudin-1 expression and survival rate of CRC. If the OR and 95% confidence intervals (CI) was not provided directly in the study, we can use method reported by Tierney [18]. To calculate these value available data in study. Heterogeneity assumption was evaluated using the chi-squared based Q-test and I2 test. The mild heterogeneity is I2 values < 25%, I2 values between 25 and 50% is a moderate heterogeneity. I2 values > 50% and P-values0.1 were considered to have a large heterogeneity [19]. For a large heterogeneity, we can use a fixed-effects model or random-effects model calculate the pooled estimates. A diagnosis of publication bias was provided by Egger test and inverted funnel plots. All statistical tests in this meta-analysis were performed using Stata10.0 software (Stata Corporation, College Station, TX, USA) with two-tail P values. P value < 0.05 was considered statistically significant. 3. Results 3.1. Search results and study characteristics The initial search yielded 35 studies. After screening of the titles, abstracts, 18 records were excluded. Among the other 17 articles, 9 articles were excluded, including cell assay only (N ¼ 3), lack of clinical data and statistical analysis (N ¼ 4), abstract only (N ¼ 1), review only (N ¼ 1). Thus a total of eight studies were eventually included in this study based on the predefined criteria [2,6,9,11,17,27e29]. Fig. 1 details the selection process. Among these eight studies, four were from Japan, two were from China, one was from USA, and one was from Egypt. Altogether, these eight studies recruited a total of 1164 CRC, with sample sizes ranged from 50 to 344. Eight articles were published between 2005 and 2013. All the studies used IHC methods for claudin-1 staining. Their basic characteristics and quality are summarized in Table 1. 3.2. Claudin-1 expression and gender The pooled results of five studies with 976 patients, including 570 males and 406 females, failed to show the significant association between expression of claudin-1 and gender of CRC (OR: 1.259, 95%CI: 0.957e1.657, P ¼ 0.100). See Fig. 2. The results of this metaanalysis showed low expression of claudin-1 was not associated with gender.
216
L. Jiang et al. / International Journal of Surgery 39 (2017) 214e220
Fig. 1. Flow chart for selection of studies for inclusion in this meta-analysis.
Table 1 Main characteristics and result of the eligible studies. Author
Year/country
Journal
Methods
Patient number
Gender (M/F)
Age
NOS score
Shibutani Yoshiday Resnick Matauoka Huang Wang Abdelzaher Nakagawa
2013/Japan 2011/Japan 2005/USA 2011/Japan 2012/China 2012/China 2011/Egypt 2011/Japan
Anticancer Research Anticancer Research Modern Pathology Journal of Surgical Oncology Chines Journal of Histochemistry and Cytochem National Medical Frontier of China Journal of the Egyptian National cancer International Journal of Oncology
IHC IHC IHC IHC IHC IHC IHC IHC
344 306 129 156 51 60 50 119
197/147 199/107 60/69 99/57 30/21 28/28 18/32 70/45
66.8 63.6 72.5 65.5 61.9 65 52.6 68.5
8 8 7 8 6 6 7 7
3.3. Claudin-1 expression and tumors' differentiation We found that expression of claudin-1 was not associated with tumor differentiation, after we pooled the data of five studies with
1164 patients of tumor differentiation (OR: 1.317, 95%CI: 0.916e1.892, P ¼ 0.137). See Fig. 5. The results of this meta-analysis showed low expression of claudin-1 was not associated with low differentiation of tumor. See Fig. 3.
Fig. 2. Pooled analysis for the association between claudin-1 expression and gender (a: forest plots; b: funnel plots).
L. Jiang et al. / International Journal of Surgery 39 (2017) 214e220
Regarding to large heterogeneity, we adopt random-effects model. (OR: 1.139, 95%CI: 0.511e2.535, P ¼ 0.750, I2 ¼ 68.7%). The results of this meta-analysis showed it is no significant correlation between low expression of claudin-1 and low differentiation of tumor. 3.4. Claudin-1 expression and depth of invasion The ORs for depth of invasion were in five studies, including 665 cases with T3-4 categories and 154 cases with T1-2 categories. Our results show that claudin-1 expression was not associated with depth of invasion (OR: 1.016, 95 %CI: 0.701e1.472, P ¼ 0.935). See Fig. 4. The results of this meta-analysis showed low expression of claudin-1 was not associated with positive depth of invasion. 3.5. Claudin-1 expression and lymph node involvement Also, in this study, claudin-1 expression was not associated with lymph node involvement after we pooled the data of five studies with 930 patients of lymph node metastasis group (OR: 1.286, 95% CI: 0.982e1.684, P ¼ 0.06). See Fig. 5. The results of this metaanalysis showed low expression of claudin-1 was not associated with positive lymph node involvement. Regarding to large heterogeneity, we adopt random-effects model. (OR: 1.278, 95%CI: 0.948e1.722, P ¼ 0.108, I2 ¼ 10.2%). However, no significant correlation were detected between the low expression of claudin-1 and positive lymph node involvement. 3.6. Claudin-1 expression and TNM stage Five studies with 621 patients investigated the association between claudin-1 expression and the tumor stage of CRC. The metaanalysis showed that expression of claudin-1 was associated with TNM stage (OR: 1.714, 95%CI: 1.215e2.418, P ¼ 0.002). See Fig. 6. The results of this meta-analysis showed low expression of claudin1 was significantly associated with TNM III-IV stage. Claudin-1 loss based on TNM results in each study are summarized in Table 2. 3.7. Claudin-1 expression and OS In a pooled analysis including five studies with 1054 patients on prognostic effect for claudin-1 expression in CRC, we found that low-expression of claudin-1 was significantly corrected with a poor survival in CRC patients (one year survival rate: OR: 2.112, 95%CI: 1.028e4.339, P ¼ 0.042; three years survival rate: OR: 1.501, 95%CI: 1.030e2.186, P ¼ 0.035; five years survival rate: OR: 1.794, 95%CI: 1.139e2.439, P ¼ 0.000). See Fig. 7. The results of this meta-analysis
217
showed low expression of claudin-1 tend to have a poor OS.
4. Discussion CRC has become increasingly common, especially in big cities. Worldwide, more than one million individual will develop CRC every year, and recurrence, metastasis and cancer-related disease are chief cause of early death in patients of CRC [20,21]. Early diagnosis and treatment of CRC is becoming more and more significant. To our present knowledge, the results of this meta-analysis documented that the claudin-1 was associated with TNM stage and prognosis. Claudin-1 was not significant differences with gender, tumors' differentiation, depth of invasion and lymph node involvement [11]. Claudin-1 is an important marker for assessing prognosis of CRC, it can enable us to have an early intervence to prevention, diagnosis and treatment of CRC after undergoing surgery [4,21]. Recent results showed that expression levels of claudin-1 in high TNM stage (III-IV) of CRC patients was more frequently downregulated as compared to the low TNM stage (I-II) of CRC patients [22e24]. Eman Abdelzaher showed a statistically significant strong inverse correlation between claudin-1 expression and tumors' TNM stage [2]. Several studies reported that low expression of claudin-1 was significantly associated with lower differentiation of the tumor which is until now considered the most significant prognosis factor in CRC [25,26]. The results of these studies suggest that low expression of claudin-1 in CRC patience of lower differentiation was due to its down-regulation might contribute to the more aggressive behavior of the tumors. However, Yoshida has reported that the expression levels of claudin-1 was not significantly associated with TNM stage of CRC patients [2]. Our meta-analysis showed that low expression of claudin-1 was associated with TNM III-IV stage (OR: 1.714). Claudin-1 expression was demonstrated to associate with tumors' differentiation, depth of invasion and lymph node involvement [2,22,26]. The loss of the tight junction and structure may affect tumor penetration which is the first step in the progression of tumor invasion and metastasis [11,27]. Some studies indicate that claudin-1 expression is no significant differences with depth of invasion, lymph node involvement [11]. For example, Yoshida et al. reported that expression levels of claudin-1 was not associated with age, gender, depth of invasion, lymph node metastasis except for tumors' differentiation and perineural invasion. Simultaneously, our meta-analysis show claudin-1 is not associated with gender, tumors' differentiation, depth of invasion, lymph node involvement of CRC patients. Many previous studies have reported that the relationship
Fig. 3. Pooled analysis for the association between claudin-1 expression and tumors' differentiation (a: forest plots; b: funnel plots).
218
L. Jiang et al. / International Journal of Surgery 39 (2017) 214e220
Fig. 4. Pooled analysis for the association between claudin-1 expression and depth of invasion (a: forest plots; b: funnel plots).
Fig. 5. Pooled analysis for the association between claudin-1 expression lymph node involvement (a: forest plots; b: funnel plots).
Fig. 6. Pooled analysis for the association between claudin-1 expression and TNM stage (a: forest plots; b: funnel plots).
Table 2 Expression of claudin-1 based on TNM results in the patients of CRC. Author
Patients number
TNM stage I ~ II (number)
III ~ IV (number)
Shibutani Yoshiday Resnick Matauoka Huang Wang Abdelzaher Nakagawa
344 306 129 156 51 60 50 119
no 173 no 72 27 25 34 no
no 133 no 84 24 35 16 no
High expression of claudin-1 (number)
Low expression of claudin-1 (number)
234 143 96 32 35 46 21 59
110 163 31 124 16 14 29 60
L. Jiang et al. / International Journal of Surgery 39 (2017) 214e220
219
Fig. 7. Pooled analysis for the association between claudin-1 expression and one year survival rate (a: forest plots, b: funnel plots), three year survival rate (c: forest plots, d: funnel plots), five year survival rate (e: forest plots, f:funnel plots).
between low expression of claudin-1 and markers of prognosis of CRC. In the present study, the clinicopathologic analysis showed that high expression of claudin-1 have a better prognosis for OS than the low expression group [17]. The main reason for poor survival of CRC is that loss of claudin-1 expression contribute to tumor invasion and metastasis, and loss of claudin-1 expression was a strong predictor of disease recurrence and poor survival. Meanwhile, our data of meta-analysis show that low expression of claudin-1 is associated with one, three and five years' OS. Although our meta-analysis was robust in identifying a correlation between low expression of claudin-1 and poor clinical results in CRC. However, some limitations should be included in our metaanalysis. First, most of the populations of the studies is from East Asia, so this may not generalize global populations. Second, some individual data could not be obtained from articles, we gained these data by assessing survival curves in the articles, heterogeneity was not neglected completely. Moreover, this meta-analysis of sample
size of cases is limited, which need a large sample sizes to confirm these conclusions further. In conclusion, this meta-analysis shows that low expression of claudin-1 is associated with high TNM stage (III-IV) and poor prognosis of CRC patients. Simultaneously, low expression of claudin-1 was not associated with gender, tumors' differentiation, depth of invasion, lymph node metastasis. Claudin-1 is likely to become a novel marker for prognosis of CRC. Ethical approval No Ethical Approval was given. Sources of funding None.
220
L. Jiang et al. / International Journal of Surgery 39 (2017) 214e220
Author contribution Lu Jiang, Long Yang and Guo Zu: study design. Lu Jiang, Long Yang, He Huang, Bai-ying Liu and Guo Zu: data collection, data analysis. Lu Jiang, Long Yang, Bai-ying Liu and Guo Zu: writing. Conflicts of interest None to declare. Research registration unique identifying number (UIN) Register a Systematic Review/Meta-Analysis and Review Registry UIN is reviewregistry151. References [1] K.H. Jun, J.H. Kim, J.H. Jung, H.J. Choi, H.M. Chin, Expression of claudin-7 and loss of claudin-18 correlate with poor prognosis in gastric cancer, Int. J. Surg. 12 (2) (2014) 156e162. [2] Eman Abdelzaher, Azza Mohamed Rizk, Samer Saad Bessa, Khalafalla Mustafa Omer, Predictive value of immunohistochemical expression of claudin-1 in colonic carcinoma, J. Egypt Natl. Cancer Inst. 23 (4) (2011 Dec) 123e131. [3] D.M. Parkin, F. Bray, J. Ferlay, P. Pisani, Global cancer statistics, CA Cancer J. Clin. 55 (2) (2002) (2005 Mar-Apr) 74e108. [4] M.S. Kahlenberg, J.M. Sullivan, D.D. Witmer, N.J. Petrelli, Molecular prognostics in colonic cancer, Surg. Oncol. 12 (3) (2003) 173e186. [5] A. Walther, E. Johnstone, C. Swanton, R. Midgley, I. Tomlinson, D. Kerr, Genetic prognostic and predictive markers in colonica cancer, Nat. Rev. Cancer 11 (4) (2011) 309. [6] Masatsune Shibutani, Eiji Noda, Kiyoshi Maeda, Hisashi Naganara, Hiroshi Ohtani, Kosei Hirakawa, Low expression of Claudin-1 and presence of poorly-differentiated tumor clusters correlate with poor prognosis in colorectal cancer, Anticancer Res. 33 (8) (2013 Aug) 3301e3306. [7] M. Furuse, K. Fujita, T. Hiiragi, K. Fujimoto, S. Tsukita, Claudin-1 and -2: novel integral membrane proteins localizing at tight junctions with no sequence similarity to occludin, J. Cell Biol. 141 (7) (1998 Jun) 1539e1550. [8] S. Tsukita, M. Furuse, M. Itoh, Multifunctional strands in tight junctions, Nat. Rev. Mol. Cell Biol. 2 (4) (2001 Apr) 285e293. [9] Takashi Matsuoka, Hiroyuki Mitomi, Naoshi Fukui, Hideki Kanazawa, Tsuyoshi, Saito, Takuo Hayashi, Takashi Yao, Cluster analysis of Claudin-1 and -4, e-cadherin, and b-catenin expression in colorectal cancers, J. Surg. Oncol. 103 (7) (2011 Jun 1) 674e686. [10] J.M. Mullin, Epithelial barriers, compartmentation, and cancer, Sci. STKE 2004 (216) (2004 Jan 13) 2. [11] Takefumi Yoshida, Testsushi Kinugasa, Yoshito Akagi, Akihiko Kawahara, Kansakar Romeo, Ichitarou Shiratsuchi, Yasuhiko Ryu, Yukito Gotanda, Kazuo Shirouzu, Decreased expression of Claudin-1 in rectal cancer: a factor for recurrence and poor prognosis, Anticancer Res. 31 (7) (2011 Jul)
2517e2525. [12] Safak Ersoz, Sevdegul Mungan, Umit Cobanoglu, Havvanur Turgutalp, Yavuz Ozoran, Prognostic importance of Claudin-4 and Claudin-4 expression in colon carcinomas, Pathology -Research and Practice 2011(207) 285-289, Pathol. Res. Pract. 207 (5) (2011 May 15) 285e289. [13] S.S. Oliveria, I.M. Oliveria, V. DeSouza, J.A. Morgado-Diaz, Claudins upregulation in human colorectal cancer, FEBS Lett. 579 (2005) 6179e6185. [14] N. Miwa, M. Furuse, S. Tsukita, N. Niikawa, Y. Nakamura, Y. Furukawa, Involvement of claudin-1 in the beta-catenin/Tcf signaling pathway and its frequent upregulation in human colorectal cancers, Oncol. Res. 12 (11e12) (2001) 469e476. [15] C.A. Iacobuzio-Donahue, A. Maitra, G.L. Shen-Ong, T. van Heek, R. Ashfaq, R. Meyer, et al., Discovery of novel tumor markers of pancreatic cancer using global gene expression technology, Am. J. Pathol. 160 (4) (2002 Apr) 1239e1249. [16] O. Fluge, O. Bruland, L.A. Akslen, J.R. Lillehaug, J.E. Varhaug, Gene expression in poorly differentiated papillary thyroid carcinomas, Thyroid 16 (2) (2006 Feb) 161e175. [17] Sumiko Nakagawa, Norikatsu Miyoshi, Hideshi Ishii, Koshi Mimori, Fumiaki Tanaka, Mitsugu Sekimoto, Yuichiro Doki, Masaki Mori, Expression of CLDN1 in coloractal cancer: a novel marker for prognosis, Int. J. Oncol. 39 (4) (2011 Oct) 791e796. [18] J.F. Tierney, L.A. Stewart, D. Ghersi, S. Burdett, M.R. Sydes, Practical methods for incorporating summary time-to-event data into meta-analysis, Trials 8 (2007 Jun 7) 16. [19] J.P. Higgins, S.G. Thompson, Quantifying heterogeneity in a meta-analysis, Stat. Med. 21 (11) (2002) 1539e1558. [20] D. Cunningham, Colonic cancer reply, Lancet 376 (9738) (2010) 331e332. [21] Guo Zu, Anlong Ji, Tingting Zhou, Ningwei Che, Clinicopathological significance of SIRT1 expression in coloretal cancer: a systematic review and meta analysis, Int. J. Surg. 26 (2016 Feb) 32e37. [22] T. Matsuoka, H. Mitomi, N. Fukui, H. Kanazawa, T. Saito, T. Hayashi, et al., Cluster analysis of claudin-1 and -4, Ecadherin,and beta-catenin expression in colonic cancers, J. Surg. Oncol. 103 (7) (2011 Jun 1) 674e686. [23] V.N. Tzelepi, A.C. Tsamandas, H.D. Vlotinou, C.E. Vagianos, C.D. Scopa, Tight junctions in thyroid carcinogenesis: diverse expression of claudin-1, claudin4, claudin-7 and occludin in thyroid neoplasms, Mod. Pathol. 21 (1) (2008 Jan) 22e30. [24] M. Resnick, T. Konkin, J. Routhier, E. Sabo, V. Pricolo, Claudin-1 is a strong prognostic indicator in stage II colonic cancer: a tissue microarray study, Mod. Pathol. 18 (4) (2005 Apr) 511e518. [25] J. Kehoe, V.P. Khatri, Staging and prognosis of colon cancer, Surg. Oncol. Clin. N. Am. 15 (1) (2006) 129. [26] S. Ersoz, S. Mungan, U. Cobanoglu, H. Turgutalp, Y. Ozoran, Prognostic importance of claudin-1 and claudin-4 expression in colon carcinomas, Pathol. Res. Pract. 207 (5) (2011 May 15) 285e289. [27] Murray B. Resnick, Tamako, Justin Routhier, Edmond Sabo, Victor E. Pricolo, Claudin-1 is a strong prognostic indicator in stage II colonic cancer: a tissue microarray study, Mod. Pathol. 18 (4) (2005 Apr) 511e518. [28] Wang Lin, Li shi-yong. Expression of Claudin-1, Claudin-4 and correlation study, in: Colorectal Cancer. Nation Medical Frontiers of China, 7(2), 2012, pp. 9e11. [29] Huang Chengcheng, Zhou Xuchan, Liu Wenxiu, Expression and significance of Claudin1, ZEB-1 and e-cadherin in colorectal carcinoma, Chin. J. Histochem. Cytochem. 21 (6) (2012) 578e583.