B型肝炎病毒全基因序列變異與肝細胞癌危險性分析:重疊病例對照研究

Abstract

背景：B型肝炎病毒全基因體的變異或許直接與肝細胞癌的發生相關，目前對於HBV全基因變異和肝細胞癌的相關性研究尚相當稀少。本研究利用重疊病例對照研究探討HBV全基因變異與肝細胞癌的關係。 材料與方法：研究個案來自一1988-1992年間收案之男性HBV帶原者世代，共計120名病例及128名對照個案加入研究。HBV全基因序列以直接定序分析獲得。利用HBV全基因體進行親緣演化樹分析以分辨基因型、次基因型。結合已知的肝炎指標，以非條件式邏輯斯迴歸，發展以單點核苷酸多型性(single-nucleotide polymorphisms)的組合來預測肝細胞癌的發生。 結果：在161個感染B基因型的個案中，次基因型Ba佔87.0%，發現少數個案屬Ba/Bj的重組型態(佔5.6%)，並發現4.3%屬於過去未分類的新次基因型；在76個感染C基因型的個案中，次基因型Ce佔82.9%，Cs僅佔5.3%，其餘是Cs/Ce的重組型態(佔11.8%)。B基因型群的次基因型和HBeAg、anti-HBe、T1858、A1896、病毒量有關，而C基因型群只和T1858、A1896有關。在全基因體序列中，總共有1815個位置具有核苷酸多型性，其中有291個位置的單核甘變異和肝細胞癌具顯著相關，這當中的166個位置與病毒量也有相關，82個位置是位在Polymerase基因區中未與其他基因重疊的部份，主要聚集在polymerase 區三個enzymatic domain。在利用邏輯斯回歸計算預測肝細胞癌的危險分數時，使用五個SNP signature和已知HBV相關的危險因子，在training set中建構預測肝細胞癌的模式，之後驗證於validation set中亦有良好的預測能力。 結論：自然狀況下和HBV病毒量有關的變異仍集中在polymerase區域。由全基因變異掃瞄，在控制了其他已知的病毒因子下，發現五個 SNP signature和肝細胞癌顯著相關。

Background & Aims：Development of hepatocellular carcinoma (HCC) may be associated with complex variants in full-length genome of HBV. The aim of this study was to identify the genetic variants that might affect the progression from healthy HBV carrier state to HCC. Materials and Methods：Baseline plasma samples were collected from 120 cases and 128 controls nested within a cohort of male HBV carriers. Phylogenetic analysis was performed with complete genome of HBV to determine the genotypes and subtypes. We used logistic regression and risk-score analysis to identify single-nucleotide polymorphisms (SNPs) signature to predict HCC. Results：Of the 161 subjects infected with genotype B HBV, 90.1% were infected with subtype Ba, 5.6% with Ba/Bj, and 4.3% with a new subtype. Of the 76 subjects infected with genotype C HBV, 82.9% were infected with subtype Ce, 11.8% with Ce/Cs, and 5.3% with Cs. For genotype B HBV, subtype was associated with HBeAg/anti-HBe status, the presence of T1858 or A1896 variants, and viral load. For genotype C, subtype was associated with the presence of T1858 or A1896 variants but not other viral factors. Among the 1815 nucleotide positions with SNPs in the complete genome of HBV, we identified 291 associated with HCC. One hundred sixty-six of the above SNPs were also correlated with viral load, 82 of which reside in the P gene. These P gene SNPs majorly cluster in the three enzymatic domains. Using logistic regression and risk score analysis, we identified a 5-SNP signature combined with known HBV-related factors for the prediction of HCC in the training set. This 5-SNP signature was validated by the testing set. Conclusions：In natural situation, SNPs that might affect viral load cluster in the P gene region. We identified a 5-SNP signature associated with HCC after taking account of other known viral factors.