Identification of novel susceptibility loci associated with hepatitis B surface antigen seroclearance in chronic hepatitis B

Abstract

Background/aims: The seroclearance of hepatitis B virus (HBV) surface antigen (HBsAg) is regarded as a functional cure of chronic hepatitis B (CHB) although it occurs rarely. Recently, several genome-wide association studies (GWASs) revealed various genetic alterations related to the clinical course of HBV infection. However, all of these studies focused on the progression of HBV infection to chronicity and had limited application because of the heterogeneity of HBV genotypes. In the present study, we aimed to determine susceptibility genetic markers for seroclearance of HBsAg in CHB patients with a homogenous viral genotype.

Methods: One hundred patients with CHB who had experienced HBsAg seroclearance before 60 years of age and another 100 with CHB showing high serum levels of HBsAg even after 60 years of age were enrolled. Extreme-phenotype GWAS was conducted using blood samples of participants.

Results: We identified three single nucleotide polymorphisms, rs7944135 (P = 4.17 × 10-6, odds ratio [OR] = 4.16, 95% confidence interval [CI] = 2.27-7.63) at 11q12.1, rs171941 (P = 3.52×10-6, OR = 3.69, 95% CI = 2.13-6.42) at 5q14.1, and rs6462008 (P = 3.40×10-6, OR = 0.34, 95% CI = 0.22-0.54) at 7p15.2 as novel susceptibility loci associated with HBsAg seroclearance in patients with CHB. The flanking genes at these loci including MPEG1, DTX4, MTX3, and HOXA13 were suggested to have functional significance. In addition, through functional analysis, CXCL13 was also presumed to be related.

Conclusions: To the best of our knowledge, this study is the first GWAS regarding the seroclearance of HBsAg in CHB patients. We identify new susceptibility loci for cure of CHB, providing new insights into its pathophysiology.

Fig 1. Chromosomal distribution of the single-nucleotide polymorphisms (SNPs) associated with the seroclearance of hepatitis B surface antigen (HBsAg) in patients with chronic hepatitis B (CHB).

The Manhattan plot shows the chromosomal distribution of the–log10 (P value) of SNPs. The threshold cutoff for the significant SNPs (P <10⁻⁵) is shown with horizontal line.

Fig 2. Three regions of interest (ROIs) for susceptibility loci of HBsAg seroclearance.

Three ROIs at 7p15.2, 5q14.1, and 11q12.1 are shown using LocusZoom (top). For the SNPs in the ROI, −log10 (P value) are plotted, and the significant SNPs are indicated. The section of the ROI is indicated with yellow bars. The recombination rates expressed in centimorgans (cM) per Mb (NCBI Build GRCh37) are also shown. The linkage disequilibrium values of the SNPs in the ROIs plotted using Haploview are shown (bottom). D′-based LD maps are shown for the two different data sets of our Korean cohort (KR) and the public data of the Beijing Han Chinese and Tokyo Japanese cohorts form 1000 Genomes Project data. The D′ scores for each paired SNP are indicated using different colors. As the LD increases, the color of the diamond becomes closer to red. The results show that both cohorts have significant LD among the SNPs within ROI. ROI1 and 2 was excluded for further analyses due to low LD (r² < 0.5).

Fig 3. Gene-set enrichment study of the significantly associated SNPs.

The functional enrichment among the flanking genes of identified susceptibility-associated SNPs (P < 5.0×10⁻⁴⁾ was estimated by gene ontology and pathway analyses implemented using DAVID and i-GSEA4GWAS softwares, respectively. The significance of the susceptibility-associated SNPs was estimated with a cutoff of P < 5.0×10⁻⁴. * P-values obtained from the pathway analysis are indicated.

Fig 4. Genotype-tissue expression analysis of rs7944135.

A. A cis-expression quantitative trait locus analysis and multi-tissue posterior probability for rs7944135 and DTX4 are shown using the GTEx portal database. The five tissues with the highest statistical significance in the cis-expression quantitative trait locus analysis (P-value < 10⁻⁵) are indicated. This means that the expression level of DTX4 significantly differs according to the allele of rs7944135 in those five tissues. B. The graphs for expression level of DTX4 according to the genotypes of rs7944135 in five tissues. With the alteration to A, minor allele, of rs7944135, that was associated with acquisition of HBsAg seroclearance, the expression of DTX4 is decreased.