Cthrc1 Overexpression Is an Independent Prognostic Marker In

Human Pathology (2014) 45, 1031–1038

www.elsevier.com/locate/humpath

Original contribution Cthrc1 overexpression is an independent prognostic marker in gastric cancer☆,☆☆ Lina Gu MD a,1, Lei Liu MD b,⁎,1, Lili Zhong MD c, Yuxian Bai MD b, Hong Sui MD b, Xiaoli Wei MD b, Wenjie Zhang MD b, Peng Huang MD b, Dandan Gao MD b, Ying Kong MD b, Ge Lou MD a,⁎,1 aDepartment of Gynecology, The Affiliated Tumor Hospital, Harbin Medical University, Haping Road 150 of Nangang District, Harbin, Heilongjiang Province, 150081, China bDepartment of Internal Medicine, The Affiliated Tumor Hospital, Harbin Medical University, Haping Road 150 of Nangang District, Harbin, Heilongjiang Province, 150081, China cDepartment of Pathology, The Affiliated First Hospital, Heilongjiang University of Chinese Medicine, Heping Road 24 of Xiangfang District, Harbin, Heilongjiang Province, 150081, China

Received 1 November 2013; revised 21 December 2013; accepted 27 December 2013

Keywords: Summary Collagen triple helix repeat containing 1 (CTHRC1) was identified as a novel gene expressed Collagen triple helix in the adventitia and neointima on arterial injury and was found to be overexpressed in several repeat containing 1; malignant tumors, such as breast cancer and malignant melanoma. However, the expression of Cthrc1 Immunohistochemistry; and its role in gastric cancer progression remain unknown. We investigated the expression of the Cthrc1 Gastric carcinoma; protein by immunohistochemistry in 30 normal tissues from the control subjects and 166 gastric Survival; carcinomas and analyzed its correlation with various clinicopathological features, including patient Prognosis outcome. Cthrc1 immunoreactivity was overexpressed in gastric carcinoma cases compared with normal tissues (P b .001). High Cthrc1 expression was found in 108 (65.06%) of these 166 carcinomas and was positively correlated with the American Joint Committee on Cancer stage classification, depth of gastric wall invasion, lymph node metastasis, lymphovascular space involvement, and recurrence but not with age, tumor site, and carcinoembryonic antigen level. Patients with high Cthrc1 expression had significantly poorer overall survival and disease-free survival compared with patients with low expression of Cthrc1 (P = .001 and P = .002, respectively). Multivariate analysis showed that high Cthrc1 expression was an independent prognostic factor for both overall survival and disease-free survival of patients with gastric carcinoma (both P = .005). These results showed that high Cthrc1 expression was associated with progression and prognosis of gastric carcinoma. © 2014 Elsevier Inc. All rights reserved.

☆ Competing interests: None of the authors has any financial arrangement with any company whose product or competing product plays a role in the manuscript. ☆☆ Funding/Support: The work was sponsored by the Medical Science and Technology Program of Health Department of Heilongjiang Province, China (No. 2011-136); Science and Technology Program of Heilongjiang Province, China (No. GC09C405-1); and Research Fund of the Affiliated Tumor Hospital, Harbin Medical University (No. JJ2009-03). ⁎ Corresponding authors. L. Liu is to be contacted at Department of Internal Medicine, The Affiliated Tumor Hospital, Harbin Medical University, Harbin 150081, China. G. Lou: Department of Gynecology, The Affiliated Tumor Hospital, Harbin Medical, University, Haping Road 150 of Nangang District, Harbin, Heilongjiang Province, 150081, China. E-mail addresses: [email protected] (L. Liu), [email protected] (G. Lou). 1 These authors contributed equally to this work.

1. Introduction were determined during and after surgery. None of our study patients had received preoperative chemotherapy and/ Gastric cancer (GC) is among the most common malig- or radiotherapy. Clinicopathological parameters of GCs were nancies worldwide and represents the third leading cause of determined according to the AJCC classification. The cancer-related death [1]. Although an improvement in treat- follow-up duration for disease-free survival (DFS) was ment effect has been achieved through curative resection defined as the period between the operation date and the day ’ and adjuvant therapies for some highly selected patients with of relapse, according to the patient s chart. The institutional GC [2], long-term survival remains dismal [3]. In recent ethics committee approval for the project was achieved years, there have been major advances in the treatment of before the study was started. patients with GC. Moreover, many studies focused on genes and their products involved in the carcinogenesis of GC, 2.2. Immunohistochemical staining such as p53 [4], KRAS [5], survivin [6,7], ERCC1 [8], and c-erbB2 [9,10]. Besides, some of these markers were shown Cthrc1 expression was detected using immunohistochem- to be of prognostic significance [6,11]. Obviously, these istry (IHC) for paraffin-embedded cancerous and noncancer- findings provided new clues for prognosis prediction of GC. ous specimens. Four-micrometer-thick sections were stained Collagen triple helix repeat containing 1 (CTHRC1) was by IHC using an anti-Cthrc1 antibody (rabbit polyclonal to originally identified as an overexpressed gene in injured rat Cthrc1, 1:100 dilution; Abcam, Cambridge MA). IHC arteries, and it encodes a secreted glycoprotein with 12 staining was performed according to the avidin-biotin immu- repeats of the Gly-X-Y motif, which is characteristic of noperoxidase technique. After collection of paraffin-embed- collagens [12]. Cthrc1 increases cellular motility to repair the ded tissue blocks of gastric specimens, all tissue blocks were injury by limiting the deposition of collagen matrix and cut in a microtome to the desired thickness (approximately promoting cell migration [12]. Aberrant expression of Cthrc1 4 μm is ideal for IHC) and affixed onto the slide. Tissue is also observed in some metastatic melanoma cells [13]. sections were dewaxed in xylene and rehydrated through Experiments in cancer research have shown that Cthrc1 graded alcohol concentrations using standard procedures. expression is highly up-regulated in solid human tumors and After washing in phosphate-buffered saline (3 times for seems to be associated with cancer tissue invasion and 5 minutes), endogenous peroxidase was blocked by incu- metastasis [13,14]. Recently, some studies that focused on bation in 0.3% hydrogen peroxide for 10 minutes at room digestive tract tumors found that Cthrc1 is associated with temperature. Then antigen retrieval with heating was per- Barrett esophagus and esophageal adenocarcinoma [15]. The formed for Cthrc1 (0.01 mol/L pH 6.0 citrate buffer, clinicopathological significance and biological role of microwave oven, 95°C, 30 minutes). The primary antibody Cthrc1 in gastric carcinoma, however, remain unclear. was used overnight at 4°C at a dilution of 1:100. The sections The aim of the present study was to investigate the expres- were incubated with biotin-labeled secondary antibody. sion of Cthrc1 in GC and to analyze its relationship to various After washing, sections were incubated with horseradish clinicopathological features, including patient outcome. peroxidase-conjugated streptavidin. 3,3′-Diaminobenzidine tetrahydrochloride was added to visualize the reaction, followed by counterstaining with Mayer hematoxylin. 2. Materials and methods 2.3. Evaluation procedures 2.1. Patients and tissue samples The expression levels of Cthrc1 were classified semi- Tissue samples were obtained from 166 patients with GC quantitatively based on the total combined scores of the who underwent surgical resection in the Affiliated Tumor percentage of positively staining tumor cells together with Hospital of Harbin Medical University from 2002 to 2005. the staining intensity. The immunostaining was evaluated Among them, 116 were males and 50 were females, with the according to the following standards: staining intensity was median age of 59 (range, 33-81) years. Resected specimens also divided into 4 subgroups (no staining = 0, weak = 1, were examined pathologically using the criteria of the moderate = 2, strong = 3), and the percentage of staining American Joint Committee on Cancer (AJCC) pathologic was designated 0 (b5%), 1 (5%-50%), or 2 (N50%). The TNM (pTNM) classification system [16]. Clinicopatholog- final score of Cthrc1 expression was defined as low ical parameters included patient age and sex, tumor location expression if the sum of the positive score and the and size, Lauren classification, wall invasion, tumor dif- corresponding intensity score lay in the range 0 to 2 and ferentiation, lymph node metastasis, vascular invasion, lym- high expression if the sum ranged from 3 to 4. The scoring phatic invasion, and perineural invasion. All cases in this procedure was carried out twice by 2 independent study received radical total or subtotal gastrectomy with D1 pathologists who were experienced in evaluating IHC and or D2 lymph node dissection. Histologic type, differentiation had no knowledge of the clinicopathological information grade, lymph node invasion, and TNM staging [17] of GC and corresponding hematoxylin-eosin slides. Cthrc1 and gastric cancer 1033

2.4. Statistical analysis gastric carcinoma expression (P = .005). One patient was lost to follow-up, without recurrence having been diagnosed. The χ2 test was used to analyze the differences of cate- gorical variables. Overall survival (OS) or DFS was calculated according to the Kaplan-Meier method and evaluated by the 4. Discussion log-rank test. Cox regression (proportional hazard model) was performed for multivariate analysis of prognostic predictors. In this study, we analyzed Cthrc1 expression by IHC in P b .05 was considered significant. Statistical analysis was gastric carcinoma using 166 surgical specimens. We found carried out using SPSS 16.0 software (SPSS, Chicago, IL). that high Cthrc1 expression was associated with tumor differentiation, recurrence, metastasis, and unfavorable outcome, indicating that Cthrc1 may be an independent prognostic 3. Results factor. To the best of our knowledge, this is the first study to demonstrate in detail an association of clinicopathological 3.1. Cthrc1 expression parameters and prognostic significance of Cthrc1 expression in GC. Cthrc1 expression was overexpressed in gastric carcino- We analyzed the correlation between Cthrc1 expression and clinicopathological features in GC. High Cthrc1 ex- ma cases compared with normal tissues, and the difference pression was significantly correlated with pTNM stage, tumor was statistically significant (P b .001). As demonstrated in differentiation, depth of tumor invasion, lymph node Table 1, of the 166 specimens with gastric carcinoma metastasis, recurrence, vascular/lymphatic invasion, tumor examined, Cthrc1 expression was low in 58 (34.94%) of 166 size, and peritoneal seeding. Correlation analysis further dem- patients with gastric carcinoma and high in 108 (65.06%) of onstrated that high Cthrc1 expression was associated with 166 patients. As shown in Fig. 1A to D, the Cthrc1 pTNM stage, tumor differentiation, and Cthrc1 expression. expression was localized in the cytoplasm of tumor cells. Cthrc1 expression was significantly higher in cases assigned Moreover, our data also demonstrated that the patients with high Cthrc1 expression had significantly poorer OS and DFS a high AJCC pTNM classification (P b .001), recurrence by using the Kaplan-Meier method and log-rank test. (P = .04), and poor tumor differentiation (P = .0233); it was Multivariate analysis suggests that Cthrc1 expression was also significantly increased in patients with cancers showing an independent prognostic factor for both OS and DFS in lymph node metastasis (P b .001), peritoneal seeding (P b patients with gastric carcinoma. Taken together, our findings .001), larger tumor size (P = .006), a deep gastric wall in this study provided evidence that high Cthrc1 expression invasion (P b .001), and lymphovascular space involvement plays a critical role in the progression of GC and is sig- (P b .001). There was no significant association between Cthrc1 immunoreactivity and other clinicopathological nificantly associated with a poor prognosis independently of other factors. This raises the possibility that Cthrc1 may be factors, as shown in Table 2. a prognostic parameter for gastric carcinoma and suggests the possibility to use Cthrc1 in individualization of both patient 3.2. Survival analysis prognosis and therapy. Previous studies have found that many aberrantly expressed genes in gastric tumor can help classify In univariate Kaplan-Meier analysis, high expression of the risk of patient outcome [16,18–21]. However, whether Cthrc1 was related to a poor OS or DFS for patients with such markers are effective for clinical application gastric carcinoma (Fig. 2AandB;Table 3; P =.001andP = as replacements for or in addition to the prognostic param- .002, respectively). Multivariate analysis showed that pTNM eters currently in use is still unclear, and further investigation and tumor differentiation were the strongest independent is called for to identify whether combined detection of Cthrc1 prognostic factors for OS (P =.011andP = .003, respectively); together with some of these other molecules would be this was followed by high Cthrc1 expression (P =.005).As valuable in improving prognostic effectiveness. demonstrated in Table 4, tumor differentiation and pTNM Our results are consistent with the previous reports of the stage (P =.018andP = .003, respectively) were the strongest roles of Cthrc1 in tumor progression, including breast cancer independent prognostic factors for DFS, followed by high [22], melanoma [23], colorectal cancer [24], gastrointestinal tract, lung, thyroid, ovarian, cervix, liver [13], and the pan- creas [25] and support a relationship between high Table 1 Expression of Cthrc1 in gastric specimens expression of Cthrc1 protein and unfavorable biological Group n Cthrc1 expression P behavior in gastric carcinoma. All these findings suggest an High expression % important tumor biological role of Cthrc1 in carcinogenesis Normal 30 1 3.33 b.001 and tumor progression. Gastric carcinoma 166 108 65.06 It is possible that Cthrc1 produced by gastric cells promotes tumor invasion by regulating the surrounding micro- aGastric carcinoma versus normal, P b .001; χ2 test. environment. Currently, there have been some clues that are 1034 L. Gu et al.

Fig. 1 Immunohistochemical staining of Cthrc1 in gastric specimens. A and B, High expression of Cthrc1 in gastric carcinoma. C, High expression of Cthrc1 and lymphatic invasion in gastric carcinoma. D and E, Low expression of Cthrc1 in gastric carcinoma. F, Low expression of Cthrc1 and lymphatic invasion in gastric carcinoma. G, Low expression of Cthrc1 in a normal gastric specimen. Cthrc1 and gastric cancer 1035

Table 2 Association analyses between the expression levels of Cthrc1 and the clinicopathological characteristics of gastric carcinoma Variables Patients Cthrc1 P a expression n Low High Sex Male 116 37 79 .814 Female 50 21 29 Age (y) b60 76 25 51 .992 ≥60 90 33 57 Tumor site Upper third 43 17 26 .401 Middle third 42 13 29 Lower third 76 27 49 Whole 5 1 4 Lauren classification Intestinal 57 31 26 .006 Diffuse/mixed 109 27 82 Peritoneal seeding No 109 56 53 b.001 Yes 57 2 55 Size b5cm 98 50 48 b.001 ≥5cm 68 8 60 Depth of invasion T1, 2 70 41 29 b.001 T3, 4 96 17 79 Lymph node metastasis No 50 34 16 b.001 Yes 116 24 92 Recurrence No 136 55 81 .040 Yes 30 3 27 Serum CEA level b N5 ng/mL 94 31 63 .985 ≤5 ng/mL 72 27 45 Pathologic stage Stages 1, 2 46 32 14 b.001 Stages 3, 4 120 26 94 Tumor differentiation Well differentiated 52 30 22 .0233 Poorly differentiated 114 28 86 Vascular/lymphatic invasion Fig. 2 Kaplan-Meier analysis of OS and DFS relates to the No 44 23 21 b.001 expression of Cthrc1. Patients with high expression of Cthrc1 had a Yes 122 35 87 poorer prognosis than that of patients with low expression of Cthrc1. A, Overall survival curves of the 165 gastric carcinoma Abbreviation: CEA, Carcinoembryonic antigen. a χ2 test. patients according to their Cthrc1 expression status, P = .001. B, b P b 0.01. DFS curves of the 165 gastric carcinoma patients according to their Cthrc1 expression status, P = .002. able to help explain its mechanisms. A previous report with increased Cthrc1 expression. These data suggested that showed that Cthrc1 stromal expression was enhanced in Cthrc1 and periostin contribute to the tumor microenviron- triple-negative cases and also in patients with bone metas- ment, with implications for lymph node and bone metastasis tasis. Periostin expression was high in primary tumors, in [26]. Wang et al [27] found that the up-regulated expression particular triple-negative ones and also in their lymph node of Cthrc1 in gastric carcinogenesis contributes to tumor cell metastases. Cox regression analysis showed that, in patients invasion and metastasis, and promoter demethylation and with high periostin, the risk of bone metastases increased transforming growth factor β1 may coregulate the expression 1036 L. Gu et al.

Table 3 Univariate survival analysis of OS and DFS in 165 patients with gastric carcinoma Variables n OS, mo DFS, mo Mean ± SE 95% CI P a Mean ± SE 95%CI P a Sex Male 116 55 ± 2 51-58 .795 57 ± 2 53-61 .803 Female 49 53 ± 2 50-56 54 ± 2 50-56 Age (y) b60 76 57 ± 3 52-61 .663 59 ± 3 57-63 .691 ≥60 89 55 ± 1 52-58 57 ± 1 54-60 Peritoneal seeding No 108 59 ± 2 57-62 b.001 59 ± 2 57-62 b.001 Yes 57 28 ± 2 25-31 29 ± 2 26-33 Size b5 cm 98 60 ± 2 57-63 .006 60 ± 2 57-63 .007 ≥5 cm 67 51 ± 3 46-53 52 ± 3 48-55 Depth of invasion T1, 2 70 61 ± 3 57-66 .005 61 ± 3 57-66 .007 T3, 4 95 48 ± 2 45-52 51 ± 2 48-54 Lymph node metastasis No 49 64 ± 1 62-67 b.001 64 ± 1 62-67 b.001 Yes 116 42 ± 4 37-47 47 ± 4 42-52 Recurrence No 136 60 ± 2 55-64 b.001 60 ± 2 55-64 b.001 Yes 29 24 ± 7 15-32 31 ± 5 24-38 Tumor differentiation Well differentiated 52 59 ± 2 56-62 b.001 60 ± 2 57-62 b.001 Poorly differentiated 113 36 ± 3 33-40 37 ± 3 32-41 Vascular/lymphatic invasion Yes 121 57 ± 2 54-60 b.001 59 ± 2 55-62 b.001 No 44 25 ± 2 22-28 25 ± 2 22-28 Cthrc1 expression Low expression 58 40 ± 6 33-47 b.001 40 ± 6 33-47 b.001 High expression 107 59 ± 3 55-63 60 ± 3 56-64 Abbreviations: CI, confidence interval; DFS, disease-free survival; OS, overall survival. a Log-rank test. of Cthrc1. Their research showed that Cthrc1 could inferred that Cthrc1 can activate several key signaling mole- activate and enhance the promoter demethylation and cules, including Src, focal adhesion kinase, paxillin, mitogen- transforming growth factor β1 signaling pathways to activated protein kinase, extracellular signal–regulated kinase, improve a tumor cell invasion and metastasis, which and Rac1, which results in cancer progression and metastasis results in the tumorigenesis. by regulating migration and adhesion activities of cancer cells Moreover, recent data indicate an involvement of Cthrc1 in [25]. These data provide important evidence to explicate the signal transduction pathways, thus enhancing cell proliferation mechanism by which Cthrc1 expression contributes to carci- or survival. Kelley [28] showed that Cthrc1 may play an nogenesis and tumor progression, although the role of Cthrc1 important role in the Wnt-planar cell polarity pathway, which in signal transduction pathways in tumor cells is certainly regulates invasion and metastasis [29]. Another study has worth further research.

Table 4 Multivariate survival analysis of OS and DFS in 165 patients with gastric carcinoma Variables OS DFS Exp(B) 95% CI P a Exp(B) 95% CI P a Cthrc1 5.751 1.593-19.778 .005 5.534 1.512-17.662 .005 pTNM 2.926 1.548-4.973 .011 2.988 1.562-5.299 .018 Tumor differentiation 5.386 1.099-13.831 .003 5.659 1.652-13.751 .003 Abbreviations: Exp(B), CI, confidence interval; DFS, disease-free survival; OS, overall survival. a Cox regression test. Cthrc1 and gastric cancer 1037

GC is one of the most common causes of cancer-related [9] Moelans CB, Milne AN, Morsink FH, Offerhaus GJ, van Diest PJ. death worldwide [30], especially in East Asian countries such Low frequency of HER2 amplification and overexpression in early onset gastric cancer. Cell Oncol (Dordr) 2011;34:89-95. as China and Japan [31]. Therefore, many articles focused on [10] Lee MS, Cha EY, Thuong PT, Kim JY, Ahn MS, Sul JY. Down- the identification of its significant prognostic factors, regulation of human epidermal growth factor receptor 2/neu oncogene including clinicopathological variables and specific markers by corosolic acid induces cell cycle arrest and apoptosis in NCI-N87 [32–34]. The current study provided a novel candidate, human gastric cancer cells. Biol Pharm Bull 2010;33:931-7. Cthrc1, as a molecular marker of both biological behaviors [11] Bellini MF, Cadamuro AC, Succi M, Proenca MA, Silva AE. Alterations of the TP53 gene in gastric and esophageal carcinogenesis. and survival outcome of GC. The finding that Cthrc1 J Biomed Biotechnol 2012;2012:891961. expression was closely correlated with late stage, unfavorable [12] Pyagay P, Heroult M, Wang Q, et al. Collagen triple helix repeat differentiation grade, and some invasion modes of breast containing 1, a novel secreted protein in injured and diseased arteries, cancer made its impact on postresectional survival easily inhibits collagen expression and promotes cell migration. Circ Res understood, similar with that in gastric carcinoma [35].Thus, 2005;96:261-8. [13] Tang L, Dai DL, Su M, Martinka M, Li G, Zhou Y. Aberrant the CTHRC1 gene may be focused the as a candidate of expression of collagen triple helix repeat containing 1 in human solid targets for inhibition against GC in the future because of its cancers. Clin Cancer Res 2006;12:3716-22. comprehensive roles in progression, invasion, and survival. [14] Ip W, Wellman-Labadie O, Tang L, et al. Collagen triple helix repeat In conclusion, we demonstrated that Cthrc1 is over- containing 1 promotes melanoma cell adhesion and survival. J Cutan expressed in a great proportion of patients with gastric car- Med Surg 2011;15:103-10. [15] Orloff M, Peterson C, He X, et al. Germline mutations in MSR1, cinoma and high Cthrc1 expression correlated with the ASCC1, and CTHRC1 in patients with Barrett esophagus and eso- disease progression and poor prognosis in gastric carcinoma. phageal adenocarcinoma. JAMA 2011;306:410-9. Furthermore, Cthrc1 might be an attractive target for GC [16] Zhao ZS, Wang YY, Chu YQ, Ye ZY, Tao HQ. SPARC is associated therapy. However, these findings remain to be confirmed by with gastric cancer progression and poor survival of patients. Clin a much larger study. Cancer Res 2010;16:260-8. [17] Hermanek P. The superiority of the new International Union Against Cancer and American Joint Committee on Cancer TNM staging of gastric carcinoma. Cancer 2000;88:1763-5. [18] de Maat MF, van de Velde CJ, Umetani N, et al. Epigenetic silencing Acknowledgments of cyclooxygenase-2 affects clinical outcome in gastric cancer. J Clin Oncol 2007;25:4887-94. The authors thank Falin Zhao for statistical analysis. [19] Hayashi M, Inokuchi M, Takagi Y, et al. High expression of HER3 is associated with a decreased survival in gastric cancer. Clin Cancer Res 2008;14:7843-9. [20] Chatterjee D, Sabo E, Tavares R, Resnick MB. Inverse association between Raf kinase inhibitory protein and signal transducers and acti- References vators of transcription 3 expression in gastric adenocarcinoma patients: implications for clinical outcome. Clin Cancer Res 2008;14:2994-3001. [1] Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. [21] De Vita F, Giuliani F, Silvestris N, Catalano G, Ciardiello F, Orditura Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. M. Human epidermal growth factor receptor 2 (HER2) in gastric Int J Cancer 2010;127:2893-917. cancer: a new therapeutic target. Cancer Treat Rev 2010;36(Suppl. 3): [2] Smalley SR, Benedetti JK, Haller DG, et al. Updated analysis of S11-5. SWOG-directed intergroup study 0116: a phase III trial of adjuvant [22] Turashvili G, Bouchal J, Baumforth K, et al. Novel markers for radiochemotherapy versus observation after curative gastric cancer differentiation of lobular and ductal invasive breast carcinomas by laser resection. J Clin Oncol 2012;30:2327-33. microdissection and microarray analysis. BMC Cancer 2007;7:55. [3] Jeung HC, Moon YW, Rha SY, et al. Phase III trial of adjuvant [23] Li J, Zhang Z, Li G. Patient outcome prediction using multiple 5-fluorouracil and adriamycin versus 5-fluorouracil, adriamycin, and biomarkers in human melanoma: a clinicopathological study of 118 polyadenylic-polyuridylic acid (poly A:U) for locally advanced gastric cases. Exp Ther Med 2011;2:131-5. cancer after curative surgery: final results of 15-year follow-up. Ann [24] Tan F, Liu F, Liu H, Hu Y, Liu D, Li G. CTHRC1 is associated with Oncol 2008;19:520-6. peritoneal carcinomatosis in colorectal cancer: a new predictor for [4] Cheng LL, Itahana Y, Lei ZD, et al. TP53 genomic status regulates prognosis. Med Oncol 2013;30:473. sensitivity of gastric cancer cells to the histone methylation inhibitor [25] Park EH, Kim S, Jo JY, et al. Collagen triple helix repeat containing-1 3-deazaneplanocin A (DZNep). Clin Cancer Res 2012;18:4201-12. promotes pancreatic cancer progression by regulating migration and [5] van Rees BP, Musler A, Caspers E, et al. K-ras mutations in gastric adhesion of tumor cells. Carcinogenesis 2013;34:694-702. stump carcinomas and in carcinomas from the non-operated stomach. [26] Kharaishvili G, Cizkova M, Bouchalova K, Mgebrishvili G, Kolar Z, Hepatogastroenterology 1999;46:2063-8. Bouchal J. Collagen triple helix repeat containing 1 protein, periostin [6] Lee GH, Joo YE, Koh YS, et al. Expression of survivin in gastric and versican in primary and metastatic breast cancer: an immunohis- cancer and its relationship with tumor angiogenesis. Eur J Gastro- tochemical study. J Clin Pathol 2011;64:977-82. enterol Hepatol 2006;18:957-63. [27] Wang P, Wang YC, Chen XY, et al. CTHRC1 is upregulated by [7] Tu SP, Jiang XH, Lin MC, et al. Suppression of survivin expression promoter demethylation and transforming growth factor-beta1 and inhibits in vivo tumorigenicity and angiogenesis in gastric cancer. may be associated with metastasis in human gastric cancer. Cancer Sci Cancer Res 2003;63:7724-32. 2012;103:1327-33. [8] Liu YP, Ling Y, Qi QF, et al. The effects of ERCC1 expression levels [28] Kelley MW. Leading Wnt down a PCP path: Cthrc1 acts as a on the chemosensitivity of gastric cancer cells to platinum agents and coreceptor in the Wnt-PCP pathway. Dev Cell 2008;15:7-8. survival in gastric cancer patients treated with oxaliplatin-based adju- [29] Golubkov VS, Chekanov AV, Cieplak P, et al. The Wnt/planar cell vant chemotherapy. Oncol Lett 2013;5:935-42. polarity protein-tyrosine kinase-7 (PTK7) is a highly efficient proteolytic 1038 L. Gu et al.

target of membrane type-1 matrix metalloproteinase: implications in [33] Tani N, Ichikawa D, Ikoma D, et al. Circulating cell-free mRNA in cancer and embryogenesis. J Biol Chem 2010;285:35740-9. plasma as a tumor marker for patients with primary and recurrent [30] Dicken BJ, Bigam DL, Cass C, Mackey JR, Joy AA, Hamilton SM. gastric cancer. Anticancer Res 2007;27:1207-12. Gastric adenocarcinoma: review and considerations for future di- [34] Li N, Zhang J, Liang Y, et al. A controversial tumor marker: is SM22 a rections. Ann Surg 2005;241:27-39. proper biomarker for gastric cancer cells? J Proteome Res 2007;6: [31] Lin Y, Ueda J, Kikuchi S, et al. Comparative epidemiology of gastric 3304-12. cancer between Japan and China. World J Gastroenterol 2011;17:4421-8. [35] Kim JH, Baek TH, Yim HS, et al. Collagen triple helix repeat [32] Ichikawa D, Koike H, Ikoma H, et al. Detection of aberrant containing-1 (CTHRC1) expression in invasive ductal carcinoma of methylation as a tumor marker in serum of patients with gastric the breast: the impact on prognosis and correlation to clinicopathologic cancer. Anticancer Res 2004;24:2477-81. features. Pathol Oncol Res 2013;19:731-7.