利用全基因組關聯分析及全外顯子組定序於台灣B型肝炎病毒相關之家族型肝癌病人進行易感基因座研究

Abstract

根據癌症統計，肝癌全球發生率排名第五位，死亡率卻高居第二名，其中原發性的肝細胞癌佔絕大部分。在臺灣， B型肝炎病毒的慢性感染為最主要誘發肝細胞癌的危險因子，然而致癌的分子機轉仍不清楚。之前已有研究顯示若HBV帶原者有兩位以上的肝細胞癌家屬，其轉變為肝癌的危險率是其他HBV帶原者5倍之多，我們認為這些家族性肝癌患者中存在著可能會遺傳的致癌基因。因此我們想找出存在於患者的染色體當中，有哪些基因變異可能誘發肝癌的發生。在這篇研究之前，我們針對205位B型肝炎相關之家族性肝癌病例和對照組完成了兩階段的全基因體關聯性分析，已經找到8個表現有顯著差異的單核苷酸多型性(SNPs)坐落於染色體1q42.1的egln1 (phd2)基因附近。 在這份研究中，我們對phd2基因進行單倍體分析並研究其在肝癌發展中所扮演的角色。首先根據連鎖不平衡定律，我們利用全基因體關聯性分析所發現的顯著SNP建立phd2單倍體。比較B型肝炎相關之家族性肝癌病例和HBsAg陰性對照組的單倍體頻率後發現其中最主要的兩種單倍體在兩者之間有顯著差異(Fisher p-value < 0.005)，另外比較家族性肝癌患者和HapMap北京漢人族群的phd2單倍體也發現在最主要的兩種單倍體在兩者之間有顯著差異(Fisher p-value < 0.001)。 我們也嘗試在細胞株中研究PHD2如何影響B型肝炎相關之肝癌的發生，已知B型肝炎病毒所產生的X蛋白會穩定並提升HIF-1α的表現，而PHD2會對HIF-1α進行羥基化並促進其降解，因此我們想研究PHD2對HIF-1α進行羥基化的酵素活性是否受HBx所影響。我們分別在Huh7、HepG2、293T三種細胞株進行轉染，發現HBx的表現會使HIF-1α在細胞中的蛋白量增加，而hydroxylated HIF-1α則明顯減少。這種情況在293T細胞中最為明顯，在Huh7及HepG2細胞中則沒有什麼表現量上的差異。我們更進一步發現三種細胞株帶有不同的phd2單倍體，這或許就是PHD2對HIF-1α進行羥基化的酵素活性在這三種細胞間有所差異的原因。 另外為了找到B型肝炎相關之肝癌的罕見基因變異，我們挑了10個B型肝炎相關之家族性肝癌病例，分別對腫瘤及正常肝細胞組織進行全外顯子定序，經過分析後發現有128個體細胞變異同時存在於2個病例之中，並有4個同時存在於3個病例之中。另外在和1000 genome project之286位漢人比對過後，我們發現了161個有顯著差異的germline SNPs (p-value < 0.0001)，這其中包含了5個SNP群集分別位於CDK11A, MUC4, KCNJ12, OR7A10, SIRPA基因內。再和52位臺灣人比對過後，發現MUC4 和OR7A10群集中帶有顯著SNP (p-value ≤ 0.01)值得之後進一步研究。 總結來說，我們接續著全基因體關聯性分析的發現，透過單倍體分析及功能性研究對於PHD2在B型肝炎相關之肝癌中有了更多的了解。未來還需要實驗來探討phd2是否扮演腫瘤抑制基因的角色，全外顯子定序所發現的體細胞和生殖細胞變異也需要再進行驗證。基因變異的研究有助於肝癌的診斷及標靶治療，我們期望能在臨床上有所貢獻。

Liver cancer is the fifth most frequently diagnosed cancer and the second leading cause of cancer death worldwide. Hepatocellular carcinoma (HCC) is a primary malignancy of the liver and accounts for most cancer cases. Chronic hepatitis B virus infection has been recognized as the main risk factors for HCC in Taiwan, yet the molecular mechanism driving hepatocarcinogenesis remains largely unknown. Previous studies have found that HBV carriers with two or more affected HCC relatives carrying a cancer risk about 5 folds compared with HBV carriers without family history. It suggested a more possible hereditary transmission of susceptibility genes in these families. Therefore, we searched for genetic variations in the host genome that may affect the risk of HCC development. We already completed a two-stage genome-wide association study (GWAS) between 205 HBV-related familial HCCs and controls. 8 significant single nucleotide polymorphisms (SNPs) were discovered on chromosome 1q42.1 in egln1 (phd2) gene. In this study, we analyzed the SNPs from GWAS and investigated the role of PHD2 for HCC development. Based on the knowledge of linkage disequilibrium, we use the significant SNPs in different stage to construct the phd2 haplotype. Then we compared the haplotype frequency between HCCs and controls. There were significant differences in the top two major haplotypes (Fisher p-value < 0.005) between HBV-related familial HCCs and HBsAg negative controls. There were also significant differences in the top two major phd2 haplotypes (Fisher p-value < 0.001) between HBV-related familial HCCs and CHB population from HapMap project. We further explored the function of PHD2 during HBV-related hepatocarcinogenesis in cell culture. The HBV X protein has been shown to stabilize and to enhance the HIF-1α, which is hydroxylated by PHD2 under normoxic conditions then targeted for degradation. We investigated whether the enzymatic activity of PHD2 was affected by HBx in three cell line, Huh7, HepG2, and 293T. In 293T cell line, the protein level of HIF-1α was increased by HBx transfection and the level of hydroxylated HIF-1α was decreased. However, there was no significant difference after HBx expression in Huh7 and HepG2. We further found that the phd2 haplotype was different in these three cell lines. This may be the reason why hydroxylase activity of PHD2 was affected in 293T but not in other two cell lines. In order to identify the rare genetic variants associated with HBV-related HCC, we conducted whole exome sequencing in 10 familial HCCs matched tumor and normal pairs. There were 128 somatic variants detected in two samples simultaneously and 4 detected in three samples simultaneously. 161 germline variants were found with significant difference (p-value < 0.0001) between10 familial HCC and 286 ASIAN from 1000 genome project. We further discovered 5 germline SNPs clusters in CDK11A, MUC4, KCNJ12, OR7A10, SIRPA gene locus respectively. These SNPs were compared with 52 Taiwanese controls, several significant SNPs (p-value ≤ 0.01) were found in the MUC4 and OR7A10 clusters. These SNPs are going to be validated in the future. Based on the finding in GWAS, we investigate the role of PHD2 in HBV-related HCC development through haplotype analysis and functional study. Further studies are required to explore whether phd2 is the novel tumor suppressor gene. The somatic and germline variants found in exome sequencing also need to replicate. The validated familial HCC related genes may be applied in HCC risk prediction or as gene therapy target. We hope this study can provide some important clinical values.