Abstract
Jiansong Ji1,*, Min Xu1,*, Zhongwei Zhao1,*, Jianfei Tu1,*, Jun Gao2,*, Chenying Lu1, Jingjing Song1, Weiqian Chen1, Minjiang Chen1, Xiaoxi Fan1, Xingyao Cheng1, Xilin Lan1, Jie Li3
1Department of Radiology, The Fifth Affiliated Hospital of Wenzhou Medical University/Affiliated Lishui Hospital of Zhejiang University/The Central Hospital of Zhejiang Lishui, Lishui, Zhejiang, P.R. China
2Department of Hepatobiliary Surgery, Beijing Chao-yang Hospital Affiliated with Capital Medical University, Beijing, P.R. China
3The First Affiliated Hospital, Chongqing Medical University, Chongqing, P.R. China
*These authors contributed equally to this work
Correspondence to:
Jiansong Ji, e-mail: [email protected]
Keywords: HCC, mothers against decapentaplegic homolog 7 (SMAD7), transforming growth factor beta (TGF-β), risk, SNP
Received: November 05, 2015 Accepted: February 05, 2016 Published: March 14, 2016
ABSTRACT
Genome-wide association studies (GWAS) have identified three loci at 18q21 (rs4939827, rs7240004, and rs7229639), which maps to SMAD7 loci, were associated with risk of diseases of the digestive system. However, their associations with hepatocellular carcinoma (HCC) risk remain unknown. A case-control study was conducted to assess genetic associations with HCC risk and clinicopathologic development among Chinese Han population. Three SNPs were genotyped among 1,000 HCC cases and 1,000 controls using Sequenom Mass-ARRAY technology. We observed statistically significant associations for the three SMAD7 loci and HCC risk. Each copy of minor allele was associated with a 1.24–1.36 fold increased risk of HCC. We also found that significant differences were observed between rs4939827 and clinical TNM stage and vascular invasion, as well as rs7240004 and vascular invasion. We also established a genetic risk score (GRS) by summing the risk alleles. The GRS was significantly associated with increased risk of HCC and vascular invasion. Our data revealed the SMAD7 loci is associated with HCC susceptibility and its clinicopathologic development.
INTRODUCTION
Recent progress through the application of genome-wide association studies (GWAS) have identified a number of common variants involved in the etiology of hepatocellular carcinoma (HCC) [1, 2]. While various genome-wide significant findings have been reported previously, it remains likely that a substantial number of additional SNPs that did not satisfy the highly stringent (Bonferroni) statistical threshold may nonetheless be important factors in modifying disease risk, if for example, their main effects were operative only in certain sub-groups of the overall population. Understanding the effects of these variants in different populations is extremely important in terms of inferring the causality and mechanisms of HCC tumorigenesis, as well as for the translation of these results into risk prediction in different populations.
HCC is a disease with very different incidence rates between populations [3–5]. The risk variants may confer different magnitudes of increased risk in different populations for a variety of reasons, including differences in allele frequency and linkage disequilibrium (LD) structure, and differences in genetic and environmental backgrounds that interact with the variants [6, 7]. Recent GWASs have identified three loci at 18q21 (rs4939827, rs7240004, and rs7229639), which maps to SMAD7, were associated with risk of colorectal cancer [8–10]. SMAD7 is involved in inflammation-related pathways and has been shown to modulate transforming growth factor-β (TGF-β) and Wnt signaling, which are central to the development of carcinogenesis [11–15]. The Smad7 gene encodes an intracellular protein, which interacts with the transforming growth factor (TGF)-β type I receptor, targeting it for degradation in the proteasome, then inhibiting TGF-β1-induced phosphorylation of Smad2/Smad3 [16]. In vivo and in vitro studies also support the important role of SMAD7 in tumor progression of HCC [17, 18]. Given the role of SMAD7 in the TGF-β signaling pathway and carcinogenesis of HCC, we performed a case-control study to comprehensively examine 3 loci (rs4939827, rs7240004, and rs7229639), which located at SMAD7 loci, for their associations with HCC risk and clinicopathologic development in a Han Chinese population, which accounts for 92% of the Chinese population [19, 20].
RESULTS
The demographic and clinical features of individuals in this population are listed in Table 1. There were no significant differences between cases and controls in terms of the distribution of age, sex, smoking and drinking status. All four SNPs conformed to Hardy–Weinberg proportions in the controls (p > 0.05).
Table 1: Characteristics of cases and controls in this study
Characteristic | Case (N = 1,000) | Control (N = 1,000) | p-value |
|---|---|---|---|
Sex, No.(%) |
|
|
|
Male | 620 (62.0%) | 612 (61.2%) | 0.713 |
Female | 380 (38.0%) | 388 (38.8%) |
|
Age, No |
|
|
|
< 50 | 523 (52.3%) | 521 (52.1%) | 0.929 |
≥ 50 | 477 (47.7%) | 479 (47.9%) |
|
family history of all cancers | 73 (7.3%) | 27 (2.7%) | P < 0.01 |
Ever smoker | 192 (19.2%) | 172 (17.2%) | 0.246 |
Ever drinker | 199 (19.9%) | 201 (20.1%) | 0.911 |
HBV infection | 221 (22.1%) | 84 (8.4%) | P < 0.01 |
*p < 0.05 indicates statistical significance.
The genotype distributions and the association between HCC and healthy controls with SMAD7 loci polymorphisms are shown in Table 2. Significant differences between the patients with HCC and the controls were detected for all three SNPs (Table 2). Compared with individuals with the major homogeneous genotype, the adjusted OR for developing HCC ranged from 1.41 (95% CI: 1.06–1.88) to 2.64 (95% CI: 1.30–4.39) among those with the minor homogeneous genotype or heterogeneous genotype. Each copy of minor allele was associated with a 1.24–1.36 fold increased risk of HCC. We also conducted sensitivity analyses to exclude the subject with family history of all cancers, as well as add smoking and drinking status to the adjustment variables, however, the results didn’t changed materially. Stratified analyses by HBV status, smoking and drinking status were presented in Table 3. All the significant trend kept during the subjects of HBV negative, non-smokers and non-drinkers. However, due to the insufficient statistical power, the trend didn’t keep in HBV positive subjects, smokers and drinkers.
Table 2: Association between SMAD7 SNPs and HCC risk
| Cases (N = 1,000) | Controls (N = 1,000) | adjusted OR* |
|---|---|---|---|
rs4939827 | |||
CC | 555 | 610 | 1.00 (reference) |
CT | 325 | 297 | 1.20 (0.99–1.46) |
TT | 120 | 93 | 1.42 (1.06–1.90) |
T vs C |
|
| 1.24 (1.07–1.42) |
P trend |
|
| 3.2 × 10−3 |
rs7240004 | |||
GG | 775 | 820 | 1.00 (reference) |
AG | 200 | 170 | 1.24 (0.99–1.56) |
AA | 25 | 10 | 2.64 (1.30–4.39) |
A vs G |
|
| 1.36 (1.12–1.66) |
P trend |
|
| 2.4 × 10−3 |
rs7229639 | |||
GG | 643 | 700 | 1.00 (reference) |
AG | 235 | 206 | 1.24 (1.00–1.54) |
AA | 122 | 94 | 1.41 (1.06–1.88) |
A vs G |
|
| 1.28 (1.10–1.49) |
P trend |
|
| 1.1 × 10−3 |
*Adjusted for age, gender, family history of cancer and HBV infection status.
Table 3: Stratified analyses of association between SMAD7 SNPs and HCC risk
Variables | rs4939827 (T vs C) | rs7240004 (A vs G) | rs7229639 (A vs G) |
|---|---|---|---|
HBV infection | |||
Positive | 1.23 (0.82–1.86) | 1.36 (0.75–2.44) | 1.28 (0.82–1.99) |
P trend | 0.309 | 0.303 | 0.264 |
Negative | 1.24 (1.06–1.44) | 1.36 (1.09–1.69) | 1.28 (1.09–1.51) |
P trend | 6.7 × 10−3 | 5.2 × 10−3 | 2.8 × 10−3 |
Smoking status | |||
Smokers | 1.24 (0.89–1.72) | 1.36 (0.85–2.17) | 1.28 (0.90–1.83) |
P trend | 0.210 | 0.198 | 0.167 |
Non-smokers | 1.24 (1.06–1.46) | 1.36 (1.09–1.70) | 1.28 (1.09–1.52) |
P trend | 7.6 × 10−3 | 6.1 × 10−3 | 3.2 × 10−3 |
Drinking status | |||
Drinkers | 1.24 (0.90–1.70) | 1.36 (0.87–2.12) | 1.28 (0.88–1.90) |
P trend | 0.174 | 0.175 | 0.181 |
Non-drinkers | 1.24 (1.06–1.45) | 1.36 (1.09–1.70) | 1.28 (1.09–1.52) |
P trend | 7.7 × 10−3 | 6.7 × 10−3 | 3.3 × 10−3 |
As shown in Table 4, we also analyzed the role of SMAD7 loci polymorphisms in the clinical TNM stage, primary tumor size, lymph node involvement, distant metastasis, vascular invasion, and Child–Pugh grade. Significant differences were observed between rs4939827 and clinical TNM stage (OR = 1.44, 95% CI: 1.12–1.85, P = 4.19 × 10−3), vascular invasion (OR = 1.38, 95% CI: 1.03–1.85, P = 0.034), as well as rs7240004 and vascular invasion (OR = 1.88, 95% CI: 1.36–2.61, P = 1.55 × 10−4).
Table 4: Age and gender adjusted odds ratio and 95% confidence interval (CI) of HCC clinical status with SMAD7 SNPs
|
| rs4939827 | rs7240004 | rs7229639 | |||
|---|---|---|---|---|---|---|---|
CC/CT+TT | OR (95% CIs) | GG/AG+AA | OR (95% CIs) | GG/AG+AA | OR (95% CIs) | ||
Clinical stage | Stage I/II | 300/200 | 1.00 (reference) | 385/115 | 1.00 (reference) | 350/150 | 1.00 (reference) |
| Stage III/IV | 255/245 | 1.44 (1.12–1.85) | 390/110 | 0.94 (0.70–1.27) | 293/107 | 0.85 (0.64–1.14) |
Tumor size | ≤ T2 | 355/290 | 1.00 (reference) | 492/153 | 1.00 (reference) | 402/243 | 1.00 (reference) |
| > T2 | 200/155 | 0.95 (0.73–1.23) | 283/72 | 0.82 (0.60–1.12) | 241/114 | 0.77 (0.59–1.01) |
Lymph node metastasis | No | 530/424 | 1.00 (reference) | 736/218 | 1.00 (reference) | 614/340 | 1.00 (reference) |
| Yes | 25/21 | 1.05 (0.58–1.90) | 39/7 | 0.61 (0.27–1.36) | 29/17 | 1.06 (0.57–1.95) |
Distant metastasis | No | 527/423 | 1.00 (reference) | 740/210 | 1.00 (reference) | 613/337 | 1.00 (reference) |
| Yes | 28/22 | 0.97 (0.55–1.74) | 35/15 | 1.51 (0.81–2.81) | 30/20 | 1.21 (0.68–2.17) |
Vascular invasion | No | 443/330 | 1.00 (reference) | 620/153 | 1.00 (reference) | 506/267 | 1.00 (reference) |
| Yes | 112/115 | 1.38 (1.03–1.85) | 155/72 | 1.88 (1.36–2.61) | 137/90 | 1.24 (0.92–1.69) |
Child–Pugh grade | A | 435/333 | 1.00 (reference) | 600/168 | 1.00 (reference) | 495/283 | 1.00 (reference) |
| B or C | 120/112 | 1.22 (0.91–1.64) | 175/57 | 1.16 (0.82–1.64) | 148/82 | 0.97 (0.71–1.32) |
To explore the cumulative effect of the three susceptibility SNPs, we established a GRS by summing the risk alleles (Table 5). The GRS was significantly associated with increased risk of HCC and vascular invasion. Compared with subjects with GRs ≤ 3, those with GRS > 3 have an 1.56 fold increased risk of HCC (95% CI: 1.30–1.86, P = 7.84 × 10−7), as well as 1.63 fold increased risk of vascular invasion (95% CI: 1.21–2.20, P = 1.35 × 10−3).
Table 5: Association of genetic risk score with HCC risk and its clinical status
|
| GRS | |
|---|---|---|---|
≤ 3/> 3 | OR (95% CIs) | ||
HCC risk | Controls | 600/400 | 1.00 (reference) |
| Cases | 490/510 | 1.56 (1.30–1.86) |
Clinical stage | Stage I/II | 242/258 | 1.00 (reference) |
| Stage III/IV | 248/252 | 0.95 (0.74–1.22) |
Tumor size | ≤ T2 | 315/330 | 1.00 (reference) |
| > T2 | 175/180 | 0.98 (0.76–1.27) |
Lymph node metastasis | No | 369/385 | 1.00 (reference) |
| Yes | 21/25 | 1.14 (0.63–2.07) |
Distant metastasis | No | 365/385 | 1.00 (reference) |
| Yes | 25/25 | 0.94 (0.53–1.68) |
Vascular invasion | No | 400/373 | 1.00 (reference) |
| Yes | 90/137 | 1.63 (1.21–2.20) |
Child–Pugh grade | A | 377/391 | 1.00 (reference) |
| B or C | 113/119 | 1.02 (0.75–1.36) |
DISCUSSION
It is known that the contribution of risk alleles to HCC risk may vary between populations. This phenomenon may be due to differences in allelic frequencies or specific linkage disequilibrium (LD) structures, or because of additional genetic factors or environmental backgrounds may influence the effect of these genetic variants [25, 26]. In current study, we observed statistically significant associations for the three SMAD7 loci (rs4939827, rs7240004, and rs7229639) and HCC risk. We also found that significant differences were observed between rs4939827 and clinical TNM stage and vascular invasion, as well as rs7240004 and vascular invasion. To our knowledge, this should be the first to investigate the relationship between HCC risk and SMAD7 locipolymorphisms.
TGF-β pathway regulates growth inhibition and apoptosis and plays an important role in cancer initiation and progressions [27, 28]. This study highlights the potential importance of the TGF-β genetic polymorphisms was associated with HCC carcinogenesis. These data provide further evidence that common genetic variants in SMAD7 may confer susceptibility to HCC, particularly in the Chinese Han population. More research is warranted to confirm these findings and functionally characterize the SMAD7 variants. Among the three SMAD7 loci studied, both rs4939827 and rs7229639 were located in the intron region of the SMAD7 gene, while rs7240004 were located in the 3′ downstream of the SMAD7 gene. Using HaploReg V4.1 [29, 30], we found that about 16 Motifs changed for the three variants, especially for rs4939827, resulted in 11 altered motifs. While using RNA structure website (http://rna.urmc.rochester.edu/RNAstructureWeb/), we found the variation these loci with resulted in the change of secondary structures and influence on the stabilities of SMAD7 RNA, which will then influence the functions of SMAD7. Vascular invasion is the most important predictor of survival in HCC, thus, a link to vascular invasion means these 2 SNPs potentially could be the predictor of survival in HCC [31].
In vivo and in vitro studies also support the important role of SMAD7 in tumor progression of HCC. Feng et al. found YB-1/Smad7 could interfere with anti-proliferative/tumor-suppressive TGF-β actions in a subgroup of HCC cells, which may facilitate aspects of tumor progression [17]. High miR-520g expression promotes HCC cell mobility and EMT by targeting SMAD7, which is correlated with reduced survival in HCC patients [32]. Loss of Smad7 can enhance susceptibility to HCC, and SMAD7 suppresses HCC cell growth by inhibiting proliferation and G1-S phase transition and inducing apoptosis through attenuation of NFKB and TGFβ signaling [33].
This study had several limitations. First, selection bias, which is an intrinsic defect of case-control study, might have occurred when the sampling is not random within the subpopulations of cancer and cancer-free subjects; Second, in spite of the relatively large sample size, the power to elucidate gene–environment interactions was limited because of the small magnitudes of the overall associations. In conclusion, polymorphisms in the SMAD7 were loci highly associated with HCC risk among Chinese population. The combined effects of SMAD7 loci polymorphisms with environmental carcinogens significantly increase the risk of developing HCC, as well as clinicopathologic development.
MATERIALS AND METHODS
Study population
This study was conducted as a population-based case-control study among Chinese Han population. The case population was comprised of newly diagnosed HCC patients which were recruited from affiliated Lishui Hospital of Zhejiang University and the first affiliated hospital of Chongqing Medical University, while control subjects were randomly recruited from the health centers during the same period. The control population was matched with the case population based upon age and gender. All cases had histologically confirmed HCC. For each participant, a standard questionnaire was used to collect demographic information, including age, sex, HBV infection status, smoking status, alcohol use, and family history of all cancer. All subjects signed informed consent forms. Blood (5 ml) was collected from each subject according to the study protocol approved by the Clinical Research Ethics Committee.
Genotyping
Genomic DNA was extracted from the peripheral blood using the GoldMag Whole Blood Genomic DNA Extraction kit according to the manufacturer’s instructions. DNA concentrations were measured using a NanoDrop 2000 (Thermo Scientific, Waltham, MA, USA). A Sequenom Mass ARRAY mass spectrometry analyzer (Sequenom, San Diego, CA, USA) was used for genotyping, and data were managed using Sequenom Typer 4.0 Software (Sequenom, San Diego, CA, USA) [21, 22].
Statistical analyses
Differences in the distribution of selected demographic variables between HCC cases and cancer-free controls were evaluated using the Student’s t-test for continuous variables or Pearson’s χ2 test for categorical variables. The association for each of the SMAD7 loci genotypes and haplotypes was evaluated using unconditional logistic regression models. In controls, each SNP was tested to determine whether it fit with the Hardy–Weinberg equilibrium (HWE). Odds ratios (OR) and 95% confidence intervals (CI) were calculated using unconditional logistic regression analyses adjusted for age, gender, family history of all cancer and HBV infection status [23], and the most common control homozygote was used as reference. LD of the candidate SNPs was analyzed using Haploview v4.2 [24]. To measure the cumulative effect of multiple genetic risk variants, we calculated a genetic risk score (GRS) by summing the number of risk alleles at each locus (0, 1, or 2). All p-values reported in this study were two-tailed and p-values less than 0.05 were considered statistically significant.
ACKNOWLEDGMENTS AND FUNDING
The work was sponsored by Zheng Shu Medical Elite Scholarship Fund, the Major Social Development Program of Major Science and Technology Project of Zhejiang (Grant No. 2013C03010), and The National Natural Science Foundation of China (81573657).
CONFLICTS OF INTEREST
None.
REFERENCES
1. Jiang DK, Sun J, Cao G, Liu Y, Lin D, Gao YZ, Ren WH, Long XD, Zhang H, Ma XP, Wang Z, Jiang W, Chen TY, et al. Genetic variants in STAT4 and HLA-DQ genes confer risk of hepatitis B virus-related hepatocellular carcinoma. Nat Genet. 2013; 45:72–75.
2. Kumar V, Kato N, Urabe Y, Takahashi A, Muroyama R, Hosono N, Otsuka M, Tateishi R, Omata M, Nakagawa H, Koike K, Kamatani N, Kubo M, et al. Genome-wide association study identifies a susceptibility locus for HCV-induced hepatocellular carcinoma. Nat Genet. 2011; 43:455–458.
3. Di Bisceglie AM. Epidemiology and clinical presentation of hepatocellular carcinoma. Journal of vascular and interventional radiology. 2002; 13:S169–171.
4. Ricca Rosellini S, Miglio F. [Epidemiology of hepatocellular carcinoma]. Minerva Gastroenterol Dietol. 1991; 37:73–84.
5. Yeh FS, Mo CC, Yen RC. Risk factors for hepatocellular carcinoma in Guangxi, People’s Republic of China. Natl Canc Inst Monogr. 1985; 69:47–48.
6. LaFramboise T, Weir BA, Zhao X, Beroukhim R, Li C, Harrington D, Sellers WR, Meyerson M. Allele-specific amplification in cancer revealed by SNP array analysis. Plos Comput Biol. 2005; 1:e65.
7. Lerand SJ, Ireland M, Blum RW. Individual and environmental impacts on sexual health of Caribbean youth. The Scientific World J. 2006; 6:707–717.
8. Zhang B, Jia WH, Matsuda K, Kweon SS, Matsuo K, Xiang YB, Shin A, Jee SH, Kim DH, Cai Q, Long J, Shi J, Wen W, et al. Large-scale genetic study in East Asians identifies six new loci associated with colorectal cancer risk. Nat Genet. 2014; 46:533–542.
9. Peters U, Jiao S, Schumacher FR, Hutter CM, Aragaki AK, Baron JA, Berndt SI, Bezieau S, Brenner H, Butterbach K, Caan BJ, Campbell PT, Carlson CS, et al. Identification of Genetic Susceptibility Loci for Colorectal Tumors in a Genome-Wide Meta-analysis. Gastroenterology. 2013; 144:799–807 e724.
10. Broderick P, Carvajal-Carmona L, Pittman AM, Webb E, Howarth K, Rowan A, Lubbe S, Spain S, Sullivan K, Fielding S, Jaeger E, Vijayakrishnan J, Kemp Z, et al. A genome-wide association study shows that common alleles of SMAD7 influence colorectal cancer risk. Nat Genet. 2007; 39:1315–1317.
11. Ma L, Wang X, Jia T, Wei W, Chua MS, So S. Tankyrase inhibitors attenuate WNT/beta-catenin signaling and inhibit growth of hepatocellular carcinoma cells. Oncotarget. 2015; 6:25390–401. doi: 10.18632/oncotarget.4455.
12. Zhou DS, Wang HB, Zhou ZG, Zhang YJ, Zhong Q, Xu L, Huang YH, Yeung SC, Chen MS, Zeng MS. TACC3 promotes stemness and is a potential therapeutic target in hepatocellular carcinoma. Oncotarget. 2015; 6:24163–77. doi: 10.18632/oncotarget.4643.
13. Jiang HL, Sun HF, Gao SP, Li LD, Hu X, Wu J, Jin W. Loss of RAB1B promotes triple-negative breast cancer metastasis by activating TGF-beta/SMAD signaling. Oncotarget. 2015; 6:16352–16365. doi: 10.18632/oncotarget.3877.
14. Geng J, Fan J, Ouyang Q, Zhang X, Yu J, Xu Z, Li Q, Yao X, Liu X, Zheng J. Loss of PPM1A expression enhances invasion and the epithelial-to-mesenchymal transition in bladder cancer by activating the TGF-beta/Smad signaling pathway. Oncotarget. 2014; 5:5700–5711. doi: 10.18632/oncotarget.2144.
15. Ungefroren H, Sebens S, Giehl K, Helm O, Groth S, Fandrich F, Rocken C, Sipos B, Lehnert H, Gieseler F. Rac1b negatively regulates TGF-beta1-induced cell motility in pancreatic ductal epithelial cells by suppressing Smad signalling. Oncotarget. 2014; 5:277–290. doi: 10.18632/oncotarget.1696.
16. Stolfi C, De Simone V, Colantoni A, Franze E, Ribichini E, Fantini MC, Caruso R, Monteleone I, Sica GS, Sileri P, MacDonald TT, Pallone F, Monteleone G. A functional role for Smad7 in sustaining colon cancer cell growth and survival. Cell Death Dis. 2014; 5:e1073.
17. Feng T, Dzieran J, Gu X, Marhenke S, Vogel A, Machida K, Weiss TS, Ruemmele P, Kollmar O, Hoffmann P, Grasser F, Allgayer H, Fabian J, et al. Smad7 regulates compensatory hepatocyte proliferation in damaged mouse liver and positively relates to better clinical outcome in human hepatocellular carcinoma. Clin Sci (Lond). 2015; 128:761–774.
18. Xia H, Ooi LL, Hui KM. MicroRNA-216a/217-induced epithelial-mesenchymal transition targets PTEN and SMAD7 to promote drug resistance and recurrence of liver cancer. Hepatology. 2013; 58:629–641.
19. Wang Y, Yu Y, Ye R, Zhang D, Li Q, An D, Fang L, Lin Y, Hou Y, Xu A, Fu Y, Lu W, Chen X, et al. An epigenetic biomarker combination of PCDH17 and POU4F2 detects bladder cancer accurately by methylation analyses of urine sediment DNA in Han Chinese. Oncotarget. 2016; 7:2754–64. doi: 10.18632/oncotarget.6666.
20. Li N, Luo H, Liu X, Ma S, Lin H, Chen R, Hao F, Zhang D. Association study of polymorphisms in FOXO3, AKT1 and IGF-2R genes with human longevity in a Han Chinese population. Oncotarget. 2016; 7:23–32. doi: 10.18632/oncotarget.6627.
21. Gabriel S, Ziaugra L, Tabbaa D. SNP genotyping using the Sequenom MassARRAY iPLEX platform. Current protocols in human genetics / editorial board, Jonathan L Haines [et al]. 2009; Chapter 2:Unit 2 12.
22. Thomas RK, Baker AC, Debiasi RM, Winckler W, Laframboise T, Lin WM, Wang M, Feng W, Zander T, MacConaill L, Lee JC, Nicoletti R, Hatton C, et al. High-throughput oncogene mutation profiling in human cancer. Nat Genet. 2007; 39:347–351.
23. Bland JM, Altman DG. Statistics notes. The odds ratio. Bmj. 2000; 320:1468.
24. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics. 2005; 21:263–265.
25. Ioannidis JP. Non-replication and inconsistency in the genome-wide association setting. Hum Hered. 2007; 64:203–213.
26. Sawyer SL, Mukherjee N, Pakstis AJ, Feuk L, Kidd JR, Brookes AJ, Kidd KK. Linkage disequilibrium patterns vary substantially among populations. European journal of human genetics. 2005; 13:677–686.
27. Halder SK, Beauchamp RD, Datta PK. Smad7 induces tumorigenicity by blocking TGF-beta-induced growth inhibition and apoptosis. Exp Cell Res. 2005; 307:231–246.
28. Siegel PM, Massague J. Cytostatic and apoptotic actions of TGF-beta in homeostasis and cancer. Nat Rev Cancer. 2003; 3:807–821.
29. Ward LD, Kellis M. HaploReg v4: systematic mining of putative causal variants, cell types, regulators and target genes for human complex traits and disease. Nucleic Acids Res. 2016; 44:D877–881.
30. Ward LD, Kellis M. HaploReg: a resource for exploring chromatin states, conservation, and regulatory motif alterations within sets of genetically linked variants. Nucleic Acids Res. 2012; 40:D930–934.
31. Thuluvath PJ. Vascular invasion is the most important predictor of survival in HCC, but how do we find it? J Clin Gastroenterol. 2009; 43:101–102.
32. Kan H, Guo W, Huang Y, Liu D. MicroRNA-520g induces epithelial-mesenchymal transition and promotes metastasis of hepatocellular carcinoma by targeting SMAD7. FEBS Lett. 2015; 589:102–109.
33. Wang J, Zhao J, Chu ES, Mok MT, Go MY, Man K, Heuchel R, Lan HY, Chang Z, Sung JJ, Yu J. Inhibitory role of Smad7 in hepatocarcinogenesis in mice and in vitro. J Pathol. 2013; 230:441–452.
