ANTICANCER RESEARCH 33: 3983-3988 (2013)

TXNDC5 Polymorphism Contributes to Increased Risk of Hepatocellular Carcinoma in the Korean Male Population

MIN-SU PARK1*, SU KANG KIM2*, HYUN PHIL SHIN3, SANG-MOK LEE1 and JOO-HO CHUNG2

Departments of 1Surgery, 3Internal Medicine Internal Medicine, and 2Kohwang Medical Research Institute, School of Medicine, Kyung Hee University, Seoul, Republic of Korea

Abstract. Thioredoxin domain-containing -5 Hepatocellular carcinoma (HCC) is the most common (TXNDC5) has been found to be associated with cancer primary cancer of the liver and the third leading cause of development and growth. We investigated whether TXNDC5 cancer-related mortality worldwide (1). HCC is attributable gene polymorphisms are associated with hepatocellular to the complex interaction between lifestyle and carcinoma (HCC) in a Korean male population. Seven SNPs environmental carcinogenic factors and genetic factors. Viral were selected based on minor allelic frequency (≥5%). The hepatitis, alcoholic cirrhosis, hemochromatosis and non- SNPs consisted of three exonic SNPs (rs8643, rs7764128 and alcoholic steatohepatitis are major carcinogenic factors of rs1043784) and four intronic SNPs (rs1225944, rs1225943, HCC (2). However, HCC can also develop without these rs1225945 and rs1225958). We selected and assessed these factors; this suggests that gene susceptibility may play a key SNPs in 160 patients with HCC and 178 controls. Genetic role in HCC development. Recently, many studies have data were analyzed using SNPAnalyzer Pro, SNPStats, and revealed that single-nucleotide polymorphisms (SNPs) may Haploview programs. Two SNPs of the TXNDC5 gene were be associated with HCC development independently or in found to be associated with the risk of HCC development. combination with some carcinogenic factors (3-5). The genotypic frequency of rs1225944 was associated with Thioredoxin domain-containing protein-5 (TXNDC5) is a HCC in the recessive model [CC/CT vs. TT, p=0.43, Fisher’s protein that in humans is encoded by the TXNDC5 gene. The exact test p=0.032; odds ratio (OR)=0.54, 95% confidence TXNDC5 gene is located on 6p24 and encodes interval (CI)=0.11-2.71]. The genotypic frequency of a protein disulphide isomerase (PDI). As a novel PDI-like rs1225943 was associated with HCC in the co-dominant 2 protein, TXNDC5 is highly expressed in endothelial cells (AA vs. CC, p=0.07; Fisher’s exact test p=0.001, OR=0.23, and plays a critical role in regulating redox homeostasis in 95% CI=0.05-1.10), recessive (AA/AC vs. CC, p=0.044, the cell. Its expression is induced by hypoxia and its role is Fisher’s exact test p=0.001, OR=0.25, 95% CI=0.05-1.17), to protect hypoxic cells from apoptosis through its protein and log-additive models (p=0.08, Fisher’s exact test folding and chaperone activity (6). p=0.002, OR=0.68, 95% CI=0.44-1.05). The haplotype CA TXNDC5 has been implicated in cancer progression. The and TC of rs1229544 and rs1225943, demonstrated a up-regulation of TXNDC5 was demonstrated in cancer of the significant association with HCC. Our results suggest that cervix, uterus, stomach, liver and lung (6-9). This suggests TXNDC5 polymorphisms could be related to the development that TXNDC5 could be associated with tumor progression. of HCC in the Korean male population. However, the relationship between TXNDC5 polymorphisms and HCC has not been extensively studied. In this study, we analyzed TXNDC5 polymorphisms to explore their genetic relationship with HCC development in a Korean male population. *These Authors contributed equally to this work. Materials and Methods Correspondence to: Prof. Joo-Ho Chung, Kohwang Medical Research Institute, School of Medicine, Kyung Hee University, Study participants. A total of 160 male patients with HCC and 178 Seoul, Korea. E-mail: [email protected] healthy male individuals were enrolled in this study. Patients with HCC were recruited at the Kyung Hee University Medical Center, Key Words: Hepatocellular carcinoma, single nucleotide Seoul and Keimyung University Dongsan Medical Center, Daegu, polymorphism, thioredoxin domain containing 5, Korean, male Republic of Korea. Patients were divided into subgroups based on population. clinical features, such as tumor size, serum alpha-feto protein(AFP)

0250-7005/2013 $2.00+.40 3983 ANTICANCER RESEARCH 33: 3983-3988 (2013)

Table I. Demographic and clinical characteristics of the study Haploview version 4.2 (Broad Institute, Cambridge, MA, USA). population. SNPStats, SNPAnalyzer (ISTECH Inc., Goyang, Republic of Korea), and SPSS 18.0 statistical software (SPSS Inc., Chicago, HCC Controls IL, USA) were used to analyze the genetic data. Multiple logistic regression models were conducted to obtain the odds ratio (OR), Number 160 178 95% confidence interval (CI), and p-value, adjusting for age as covariate [co-dominant 1 (major allele homozygotes vs. Age (mean±SD), years 57.8±10.0 47.1±10.0 heterozygotes), co-dominant 2 (major allele homozygotes vs. minor allele homozygotes), dominant (major allele homozygotes Etiology of HCC vs. heterozygotes + minor allele homozygotes), recessive (major Alcohol 21 allele homozygotes+ heterozygotes vs. minor allele HBV 115 homozygotes), and log-additive (major allele homozygotes vs. HCV 13 HBV+HCV 1 heterozygotes vs. minor allele homozygotes)]. Statistical Tumor size significance was set at p<0.05. <5 cm 83 ≥5 cm 70 Results AFP ≤200 ng/ml 106 Association between TXNDC5 polymorphisms and HCC (Table >200 ng/ml 48 III). The genotypic distributions of five SNPs (rs1225944, Modified UICC stage rs1225943, rs8643, rs7764128 and rs1043784) were in HWE I/II 51 III/IV 103 (p>0.05). Two SNPs (rs1225945 and rs1225958) deviated Radiological morphology significantly (p<0.05) from HWE among the controls and were Nodular 101 therefore excluded from further analyses. Non-nodular 54 The T/T genotypic frequency of rs1225944 in the control Portal vein thrombosis group was higher than that of HCC (8.0% vs.1.8%). The Present 60 Absence 94 difference showed significant association with HCC in the recessive model (p=0.43, Fisher’s exact test p=0.032, OR- HCC: Hepatocellular carcinoma, SD: standard deviation, HBV: hepatitis 0.54, 95% CI=0.11-2.71). B, HCV: hepatitis C, AFP: alpha-fetoprotein, UICC: Union for The C/C genotypic frequency of rs1225943 in the control International Cancer Control. group was higher than that of HCC (9.7% vs. 1.3%). The difference showed significant association with HCC in the co-dominant 2 (p=0.07, Fisher’s exact test p=0.001, level, modified Union for International Cancer Control (UICC) OR=0.23, 95% CI=0.05-1.10), recessive (p=0.044, Fisher’s stage, radiological morphology, and portal vein thrombosis (Table exact test p=0.001, OR=0.25, 95% CI=0.05-1.17), and log- I). Patients with other types of cancers and severe diseases were additive models (p=0.08, Fisher’s exact test p=0.002, OR- excluded. Controls were recruited from among participants in a 0.68, 95% CI=0.44-1.05). general health check-up program and they had no clinical evidence In this study, SNPs of rs8643, rs7764128 and rs1043784 of HCC or other severe diseases. Informed consent was obtained from each participant. The study was approved by the Institutional of TXNDC5 were not significantly differently distributed Review Board of Kyung Hee University Medical Center (IRB between the HCC group and the control group. number: 040915). Association between TXNDC5 haplotypes and HCC. Two SNP selection and genotyping. The SNPs of the TXNDC5 gene were LD blocks were identified by the Gabriel method using retrieved from the NIH SNP database (www.ncbi.nlm.nih.gov/ SNP, Haploview 4.2 (Figure 1). Block 1 consists of rs1229544 dbSNP Build 135). Seven SNPs were selected based on minor allelic and rs1225943, and block 2 consists of rs8643, rs7764128, frequency (≥5%). The SNPs consisted of three exonic (rs8643, rs7764128 and rs1043784) and four intronic (rs1225944, rs1225943, and rs1043784. In haplotype analysis, haplotypes CA and rs1225945 and rs1225958) SNPs. Genomic DNA from blood was TC were significantly associated with HCC (p=0.021) extracted using Roche DNA Extraction kit (Roche, Indianapolis, IN, (Table IV). USA). DNAs were amplified using specific primers for each SNP (Table II). Polymerase Chain Reaction (PCR) products were Association between TXNDC5 polymorphisms and clinical sequenced using ABI PRISM 3730XL analyzer (PE Applied characteristics of HCC. Next, the relationship between Biosystems, Foster City, CA, USA) and sequence data were analyzed TXNDC5 SNPs and the clinical characteristics of HCC was by SeqManII software (DNASTAR Inc., Madison, WI, USA). investigated. There were no genotypic or phenotypic Statistical analysis. Hardy-Weinberg equilibrium (HWE) was frequencies for polymorphisms based on the tumor size, AFP evaluated using SNPStats (http://bioinfo.iconcologia.net/index). level, modified UICC stage, radiological morphology, and A linkage disequilibrium (LD) block was estimated using portal vein thrombosis. (data not shown).

3984 Park et al: TXNDC5 Gene Polymorphism and HCC

Table II. Primer sequences for each single-nucleotide polymorphism (SNP).

SNP Forward (5’-3’) Reverse (5’-3’) rs1225958 TATAGGGGTCCATCAATAAAGCA TCTCTGCTCACTACCTGCTTCTT rs1225945 GTTGTGGAGATGAACCTTGAAAC AACTGTCATTTTGTCCAAAGCAT rs1225944 GTTGTGGAGATGAACCTTGAAAC AACTGTCATTTTGTCCAAAGCAT rs1225943 CCTCTGAAGCATACAGGTCTCTG AAGATCAACACCTTGAAAACCAA rs8643 CAGAGACCTTGACTCGTTACACC ATGGCATCAGCATTCTCTTACTC rs7764128 CACGAGGCTCATTCTCAGTATCT CTTGACAGTTGTGCTTCTCTAGG rs1043784 TCTAACTCATGCTGTCCTTGTGA CCTTGTAGAAAGTGAGGCTCAGA

Table III. Genotypic and allelic frequencies of TXNDC5 polymorphisms in patients with HCC and controls.

SNP Type Control HCC Model OR (95% CI) p-Value Fisher’s exact p n (%) n (%) rs1225944 C/C 99 (61.1) 71 (65.7) Co-dominant 1 1.03 (0.56-1.92) 0.92 Intron C/T 50 (30.9) 35 (32.4) Co-dominant 2 0.55 (0.11-2.78) 0.47 0.05 T/T 13 (8.0) 2 (1.8) Dominant 0.97 (0.53-1.76) 0.91 Recessive 0.54 (0.11-2.71) 0.43 0.032 Log-additive 0.91 (0.55-1.51) 0.72 0.09 C 248 (76.5) 177 (81.9) T 76 (23.5) 39 (18.1) Allele 1.17 (0.66-2.06) 0.59 rs1225943 A/A 101 (57.4) 104 (68.9) Co-dominant 1 0.82 (0.48-1.41) 0.48 Intron A/C 58 (33.0) 45 (29.8) Co-dominant 2 0.23 (0.05-1.10) 0.07 0.001 C/C 17 (9.7) 2 (1.3) Dominant 0.72 (0.43-1.21) 0.21 Recessive 0.25 (0.05-1.17) 0.044 0.001 Log-additive 0.68 (0.44-1.05) 0.08 0.002 A 260 (73.9) 253 (83.8) C 92 (26.1) 49 (16.2) Allele 1.17 (0.66-2.06) 0.59 rs8643 G/G 143 (81.2) 127 (81.4) Co-dominant 1 0.71 (0.36-1.39) 0.32 3’UTR G/A 31 (17.6) 27 (17.3) Co-dominant 2 1.12 (0.10-12.07) 0.92 1.00 A/A 2 (1.1) 2 (1.3) Dominant 0.73 (0.38-1.41) 0.34 Recessive 1.19 (0.11-12.65) 0.89 1.00 Log-additive 0.78 (0.43-1.42) 0.41 1.00 G 317 (90.1) 281 (90.1) A 35 (9.9) 31 (9.9) Allele 1.17 (0.66-2.06) 0.59 rs7764128 G/G 138 (81.2) 84 (80.8) Co-dominant 1 0.63 (0.28-1.37) 0.24 3’UTR G/A 30 (17.6) 19 (18.3) Co-dominant 2 0.71 (0.05-10.88) 0.81 1.00 A/A 2 (1.2) 1 (1.0) Dominant 0.63 (0.29-1.36) 0.23 Recessive 0.78 (0.05-11.78) 0.86 1.00 Log-additive 0.67 (0.33-1.35) 0.26 1.00 G 306 (90.0) 187 (89.9) A 34 (10.0) 21 (10.1) Allele 1.17 (0.66-2.06) 0.59 rs1043784 A/A 137 (81.1) 84 (80.0) Co-dominant 1 0.57 (0.25-1.27) 0.17 3’UTR A/G 30 (17.8) 16 (15.2) Co-dominant 2 3.03 (0.38-24.16) 0.29 0.11 G/G 2 (1.2) 5 (4.8) Dominant 0.71 (0.34-1.50) 0.36 Recessive 3.33 (0.43-26.05) 0.22 0.11 Log-additive 0.90 (0.48-1.66) 0.72 0.19 A 304 (89.9) 184 (87.6) G 34 (10.1) 26 (12.4) Allele 1.17 (0.66-2.06) 0.59

TXNDC5: Thioredoxin domain containing-5, SNP: single-nucleotide polymorphism, HCC: hepatocellular carcinoma, UTR: untranslated region, OR: odds ratio, CI: confidence interval. The p-values were calculated from logistic regression analysis adjusting for age.

3985 ANTICANCER RESEARCH 33: 3983-3988 (2013)

Table IV. Haplotype analysis of TXNDC5 polymorphisms in patients with HCC and controls.

Haplotype Frequency Control HCC Chi-square p-Value

+–+–

Block 1 CA 0.780 260.9 (74.1%) 91.1 (25.9%) 241.7 (82.8%) 50.3 (17.2%) 6.97 0.008 TC 0.207 84.6 (24.1%) 267.4 (75.9) 48.5 (16.6%) 243.5 (83.4%) 5.36 0.021 Block 2 GGA 0.893 318.9 (90.1%) 35.1 (9.9%) 263.3 (88.4%) 34.7 (11.6%) 0.50 0.48 AAG 0.100 35.0 (9.9%) 319.0 (90.1%) 30.0 (10.1%) 268.0 (89.9%) 0.006 0.94

HCC: Hepatocellular carcinoma. Haplotypes of LD block 1 are comprised of rs1229544, rs1225943, Haplotypes of LD block 2 are comprised of rs8643, rs7764128 and rs1043784. Bold number indicates significant association.

Discussion

TXNDC5 was first identified by analysis of the luminal environment of endoplasmic reticulum of hepatic tissue (10). Abnormal regulation of TXNDC5 was found in some malignancies, including HCC. TXNDC5 was significantly up-regulated in colorectal adenoma and cancer tissues as compared with that in normal mucosa (7). Zhang et al. showed that up-regulation of TXNDC5 can promote the growth and proliferation of gastric cancer cells. They suggested that TXNDC5 could be thought of as a tumor- enhancing gene in gastric cancer (11). Wei et al. showed that up-regulation of TXNDC5 is closely associated with androgen-independent prostate cancer (9). Nissome et al. demonstrated that TXNDC5 expression increased in poorly differentiated HCC but not in well-differentiated HCC, consistent with a causative role in the development of more aggressive tumors (8). Although accumulating evidence has demonstrated that TXNDC5 is important for cancer development and growth, there has been no report about TXNDC5 polymorphisms in cancer. One study was published on the genetic association between a TXNDC5 SNP and non-segmental vitiligo (NSV) (12). They suggested that TXNDC5 gene polymorphisms are Figure 1. Linkage disequilibrium block consists of rs1229544, associated with the development of NSV in the Korean rs1225943, rs8643, rs7764128 and rs1043784. population. In this study, we analyzed seven SNPs (rs1225958, rs1225945, rs1225944, rs1225943, rs8643, rs7764128 and rs1043784) of the TXNDC5 gene. The two examined SNPs (rs1225944 and rs1225943) were associated with the HCC. We found that TXNDC5 gene polymorphisms may development of HCC. We investigated LD between contribute to an increase susceptibility to HCC in the TXNDC5 SNPs and performed a haplotype analysis. Korean male population. Future studies are warranted to Haplotype analysis demonstrated that the haplotype CA and validate our results and define the functional effects of these TC were significantly associated with HCC. From our SNPs on HCC. results, we suggest that TXNDC5 polymorphisms could be concerned with the development of HCC. Conflicts of Interest In conclusion, this study investigated the relationship between TXNDC5 polymorphisms and the occurrence of The Authors indicate that no potential conflicts of interest exist.

3986 Park et al: TXNDC5 Gene Polymorphism and HCC

References 7 Wang Y, Ma Y, Lu B, Xu E, Huang Q and Lai M: Differential expression of mimecan and thioredoxin domain-containing 1 Ferlay J, Shin HR, Bray F, Forman D, Mathers C and Parkin protein 5 in colorectal adenoma and cancer: a proteomic study. DM: Estimates of worldwide burden of cancer in 2008: Exp Biol Med 232: 1152-1159, 2007. GLOBOCAN 2008. Int J Cancer 127: 2893-2917, 2010. 8 Nissom PM, Lo SL, Lo JC, Ong PF, Lim JW, Ou K, Liang RC, 2 El-Serag HB, Davila JA, Petersen NJ and McGlynn KA: The Seow TK and Chung MC: Hcc-2, a novel mammalian ER continuing increase in the incidence of hepatocellular carcinoma thioredoxin that is differentially expressed in hepatocellular in the United States: an update. Ann Intern Med 139: 817-823, carcinoma. FEBS Lett 580: 2216-2226, 2006. 2003. 9 Wei Q, Li M, Fu X, Tang R, Na Y, Jiang M and Li Y: Global 3 Abu Dayyeh BK, Yang M, Fuchs BC, Karl DL, Yamada S, analysis of differentially expressed in androgen- Sninsky JJ, O’Brien TR, Dienstag JL, Tanabe KK and Chung independent prostate cancer. Prostate Cancer Prostatic Dis 10: RT: A functional polymorphism in the epidermal growth factor 167-174, 2007. gene is associated with risk for hepatocellular carcinoma. 10 Clissold PM and Bicknell R: The thioredoxin-like fold: hidden Gastroenterology 141: 141-149, 2011. domains in protein disulfide isomerases and other chaperone 4 Nahon P, Sutton A, Rufat P, Ziol M, Akouche H, Laguillier C, . Bioessays 25: 603-611, 2003. Charnaux N, Ganne-Carrie N, Grando-Lemaire V, N’Kontchou 11 Zhang L, Hou Y, Li N, Wu K and Zhai J: The influence of G, Trinchet JC, Gattegno L, Pessayre D and Beaugrand M: TXNDC5 gene on gastric cancer cell. J Cancer Res Clin Oncol Myeloperoxidase and superoxide dismutase 2 polymorphisms 136: 1497-1505, 2010. comodulate the risk of hepatocellular carcinoma and death in 12 Jeong KH, Shin MK, Uhm YK, Kim HJ, Chung JH and Lee alcoholic cirrhosis. Hepatology 50: 1484-1493, 2009. MH: Association of TXNDC5 gene polymorphisms and 5 Kim SK, Jeon JW, Park JJ, Cha JM, Joo KR, Lee JI, Chung JH, susceptibility to nonsegmental vitiligo in the Korean population. Shin DH and Shin HP: Associations of EPHB1 polymorphisms Br J Dermatol 162: 759-764, 2010. with hepatocellular carcinoma in the Korean population. Hum Immunol 72: 916-920, 2011. 6 Sullivan DC, Huminiecki L, Moore JW, Boyle JJ, Poulsom R, Creamer D, Barker J and Bicknell R: EndoPDI, a novel protein- disulfide isomerase-like protein that is preferentially expressed Received June 23, 2013 in endothelial cells acts as a stress survival factor. The J Biol Revised July 19, 2013 Chem 278: 47079-47088, 2003. Accepted July 23, 2013

3987