Hepatitis, Alcoholic study with AMPLICON data on dbGaP (1,213 participants)

phs003659 — Exome Sequencing of Alcohol-Associated Hepatitis

Alcoholic hepatitis (AH) is a clinically severe, acute disease that only afflicts a fraction of patients with alcohol use disorder (AUD). Although recent genomic studies of alcoholic cirrhosis (AC) and steatosis have revealed that several genes of large effect are shared between these disorders, the genetic and environmental sources of variation that lead to AH, AC or alcohol fatty liver disease (AFLD) in some AUD patients remain largely unknown. In order to investigate the genetic determinants of AH, we sequenced a 58 Mb exome target to high genotype accuracy (0.996) and completeness (0.941) in 1,735 patients with AH (N=784) and heavy drinking controls without alcohol-associated liver disease (ALD) (N=951), detecting 1,085,982 unique single nucleotide variants (SNVs). Analysis of common SNPs that comprised 17.5% of those SNVs, we observed exome-wide significant association of <a href="https://www.ncbi.nlm.nih.gov/gene/80339">PNPLA3</a><a href="https://www.ncbi.nlm.nih.gov/gene/80339"> </a>(rs738409-G allele: odds ratio [OR]=1.73, 95% confidence interval [CI]: 1.44-2.08, p =3.75e-9), as had previously been seen for AH in candidate gene contexts, and as previously observed for <a href="https://www.ncbi.nlm.nih.gov/gene/80339">PNPLA3</a><a href="https://www.ncbi.nlm.nih.gov/gene/80339"> </a>to AC and nonalcohol steatohepatitis in GWAS. Other genes previously implicated in AC did not strongly contribute to AH, and the only prominently implicated (but not exome wide significant) gene overlapping with AUD was <a href="https://www.ncbi.nlm.nih.gov/gene/125">ADH1B</a>, also a known liver disease gene. However, SNPs at <a href="https://www.ncbi.nlm.nih.gov/gene/23308">ICOSLG</a><a href="https://www.ncbi.nlm.nih.gov/gene/23308"> </a>(Chr 21) and <a href="https://www.ncbi.nlm.nih.gov/gene/9878">TOX4</a>/<a href="https://www.ncbi.nlm.nih.gov/gene/84932">RAB2B</a><a href="https://www.ncbi.nlm.nih.gov/gene/84932"> </a>(Chr 14), were also exome-wide significant and of large effect-size (<a href="http://ncbi.nlm.nih.gov/gene/23308">ICOSLG</a><a href="http://ncbi.nlm.nih.gov/gene/23308"> </a>rs13052840-C allele: OR=1.67, 95% CI: 1.39-2.02, p = 4.59e-8; <a href="https://www.ncbi.nlm.nih.gov/gene/9878">TOX4</a>/<a href="https://www.ncbi.nlm.nih.gov/gene/84932">RAB2B</a><a href="https://www.ncbi.nlm.nih.gov/gene/84932"> </a>rs17516362-C allele: OR = 0.61, 95% CI: 0.50-0.74, p=1.25e-7). <a href="http://ncbi.nlm.nih.gov/gene/23308">ICOSLG</a><a href="http://ncbi.nlm.nih.gov/gene/23308"> </a>encodes a co-stimulatory signal for T-cell proliferation and cytokine secretion and induces B-cell proliferation and differentiation, and thus could be tied to the inflammatory component of AH. <a href="https://www.ncbi.nlm.nih.gov/gene/9878">TOX4</a><a href="https://www.ncbi.nlm.nih.gov/gene/9878"> </a>was previously implicated in diabetes and immune system and its linked SNP is a cis eQTL for <a href="https://www.ncbi.nlm.nih.gov/gene/9878">TOX4</a><a href="https://www.ncbi.nlm.nih.gov/gene/9878"> </a>and <a href="https://www.ncbi.nlm.nih.gov/gene/84932">RAB2B</a>. A polygenic signal was visible on the exome-wide quantile-quantile graph, where there was modest p value inflation (lambda = 1.040). Ranking p values of all 18,705 genes evaluated versus a list of 227 genes objectively precompiled for relationship to liver diseases (via GWAS), we observed a strong (p&#x3C;0.0001) signal of polygenic contribution to AH, that signal being balanced between the reference (usually more abundant) and nonreference alleles at these 227 genes. There was no such polygenic overlap detected to heavy drinking or AUD. Via rare variant analysis, no gene or individual SNV was exome-wide significant, however the top genes implicated by rare SNVs included several candidates of interest due to their roles in inflammation or other processes, and several of the top SNVs were predicted to be pathogenic. The data generated here also represents a resource that can be used cumulatively or meta-analytically with future datasets.