Analysis Of Multiple Gene Polymorphisms For Molecular Prediction Of HBV-Related Cirrhosis Risk

Background and objectives: HBV-related hepatic cirrhosis is usually irreversible; therefore the key is prevention in cirrhosis control. However, only minority of chronic HBV infectors finally develops cirrhosis, most of them will be asymptomatic HBsAg carriers for a long time. Identifying the infectors with high risk of cirrhosis development and then implementing individualized preventive measures is an economic and effective strategy for preventing cirrhosis. According to clinical risk factors, it is difficult to quantitatively predict the risk of cirrhosis development in chronic HBV infectors. Single nucleotide polymorphisms (SNPs) which account for over 90% genetic variations in human genome are the best genetic markers for explaining the hereditary susceptibility of diseases and suitable for quantitatively prediction. Different genetic background has been believed to be the main cause of cirrhosis development difference in chronic HBV infectors. Since HBV-related cirrhosis is a polygenic disease, a panel of SNPs should be screened out for cirrhosis risk prediction. The aims of the present study were to analyze the relationships between the HBV-related cirrhosis susceptibility and genetic polymorphisms involved in different stages of cirrhosis development and establish mathematic models for quantitatively molecular prediction of cirrhosis risk in chronic HBV infectors individually. So far, there is no report about systematic analysis of multiple genetic polymorphisms associated with HBV-related hepatic cirrhosis susceptibility and cirrhosis risk prediction in terms of them.Methods: A grouped case-control study was designed for the study, in which the case was HBV infectors with cirrhosis, and the control was asymptomatic HBsAg carriers. Seventeen polymorphism sites of 11 genes were selected for analysis: IL-1A-889C/T, IL-1B+3953C/T and -511C/T, IL-10-592A/C, IFN-γ+874T/A and +2109A/G, TNF-α-308G/A and -238G/A, CD14-159C/T, CTLA-4-318C/T, MEH Tyr113His and -613C/T, GSTM1null/normal, MMP-9-1562C/T, ER1+29T/C, AGT-20A/C and -6A/G. The genomic DNA was extracted from peripheral blood clot of the patients by phenol-chloroform method. PCR-RFLP or PCR-SSP methods were applied to genotype each polymorphism site. Genotype and allele frequencies of each site were calculated in two groups. Univariate and multivariate non-conditional Logistic regression analyses were conducted to observe the association and its intensity of each genotype or allele with the risk of HBV-related cirrhosis. By means of multivariate stepwise Logistic regression analysis, the independent related genotypes or alleles were screened out and mathematic models were created for the molecular prediction of cirrhosis risk in chronic HBV infectors, and their predictive performances were evaluated.Results: Three hundred and twenty three patients were recruited, including male 226, female 97, with the average age 51.0 years old. Of them, there were 168 patients with HBV-related hepatic cirrhosis and 155 asymptomatic HBsAg carriers. The differences of constituent ratios among age groups were not significant between case and control groups in total or different genders. The results about the genetic polymorphisms and cirrhosis were as follows:(1) In univariate analyses of total sample, IFN-γ+874T/A site was associated with HBV-related cirrhosis, in which TT were protective genotype (OR=0.55, P=0.039), TA were risk genotype (OR=1.93, P=0.025) and A was risk allele (OR=1.82, P=0.036). IL-10-592A/C site was weakly associated with cirrhosis (P= 0.057~0.077). The other polymorphism sites were not correlated with cirrhosis (P>0.1). Multivariate analysis indicated that ER1+29TC was a risk genotype (OR=2.00, P=0.024) and ER1+29T allele, AGT-20AA genotype and C allele were weakly correlated with cirrhosis (P=0.059~0.072).(2) In univariate analyses of male sample, IL-10-592 A/C site was associated with male HBV-related cirrhosis, in which AC was a risk genotype (OR=1.86, P=0.042), genotype AA and allele C were weakly correlated (P=0.076). Genotype TA of IFN-γ+874T/A site was also weakly correlated with male cirrhosis (P=0.074), other polymorphism sites were not correlated (P>0.1). Multivariate analyses screened out three risk genotypes or alleles of IL-10-592 site: genotype AC (OR=3.79, P=0.002) or allele A (OR=3.70, P=0.049), C allele (OR=3.70, P=0.049).(3) In univariate analysis of female sample, ER1+29T/C site was associated with female HBV-related cirrhosis, in which TC was a risk genotype (OR=3.28, P=0.006) , T was risk allele (OR=3.10, P=0.013) and CC was protective genotype (OR=0.32, P=0.013) ; IL-1B+3953C/T and MEH-613C/T sites were weakly correlated with female cirrhosis (P=0.076-0.089); Other polymorphism sites were not correlated (P>0.1). Multivariate analyses showed that the risk genotypes and alleles were ER1+29 TC genotype (OR=5.33, P=0.019) or T allele (OR=7.63, P=0.017), CTLA-4-318 CT genotype (OR=8.84, P=0.017) or T allele (OR=6.33, P=0.031), and IFN-γ+2109 AA genotype (OR=4.38, P=0.032); The protective genotypes and alleles were IL-1B3593 CT genotype (OR=0.076, P=0.036) or T allele (OR=0061, P=0.026) and IFN-γ+2109 G allele (OR=0.174, P=0.012).(4) Based on the three sets of genetic polymorphism data (gender-mixed patients, male patients and female patients), three prediction models were developed respectively. The prediction accuracies of the three models for their own training sample were 70.1%, 69.3% and 79.4%, respectively; the prediction accuracies for their own total sample were 67.1%, 67.5% and 77.9%, respectively. The areas under the ROC curves created according to the predictive probability values of the three models to corresponding total samples were 0.692, 0.670 and 0.813, respectively.Conclusions: According to the results above, we conclude:(1) In the gene polymorphism sites detected, only minority is correlated with cirrhosis susceptibility in chronic HBV infectors, most of them are weakly or not correlated.(2) Genetic polymorphisms related to HBV-related hepatic cirrhosis are mainly involved in immune-related genes and estrogen receptor-alpha gene, especially the cytokine genes.(3) There are differences between male and female patients in the gene polymorphisms related to HBV-related cirrhosis. In the male, mainly genetic polymorphisms involved are cytokine genes, while in the female, except cytokine genes, estrogen receptor-alpha gene and cellular-immune related genes were involved.(4) The association of female HBV-related cirrhosis with genetic variation is closer than that of male patients, more genetic polymorphism sites associated and more intensely correlated in female patients,indicating that environmental risk factor control might be more important for prevention of cirrhosis in male HBV infectors, while in female HBV infectors, more attention should be paid to genetic risk factors.(5) Based on the different genetic polymorphism data (total patients, male patients and female patients), corresponding models could be developed successfully for predicting HBV-related cirrhosis, respectively. The female-specific predictive model shows a good performance and better than male-specific and gender-mixed predictive models.