Identification of susceptibility gene to severe influenza in humans : an integrated study of genetics, molecular biology and bioinformatics

Abstract

The 2009 influenza A(H1N1) virus (A(H1N1)pdm09) caused the first influenza pandemic in the 21st century. In 2013, a novel avian influenza virus A(H7N9) was identified to cause human infection with case-fatality rate of over 30%. The possible host genetic determinants predisposing to human A(H7N9) influenza and severe A(H1N1)pdm09 influenza have been speculated. We performed two genome-wide association studies (GWAS) to investigate the genetic susceptibility to A(H7N9) influenza and severe A(H1N1)pdm09 influenza. A bioinformatics pipeline, IndelLDplot, has been created to integrate genetic variants with expression quantitative trait loci (eQTLs) data and publicly available genome-wide functional annotation data. IndelLDplot can be utilized to identify potentially functional variants in high linkage disequilibrium (LD) with GWAS SNPs for further validation.

Firstly, we demonstrated that four JAK2 SNPs including rs1887429, rs17425819, rs7850484 and rs7034539 were associated with the susceptibility to severe A(H1N1)pdm09 influenza in 436 patients. Two potential functional variants rs59384377 and rs59679286, located in a putative enhancer can tag the two association SNPs rs17425819 and rs7850484. In reporter gene luciferase assay, these two variants could encode regulatory polymorphisms for the enhancer activity. Additionally, the involvement of JAK2 in the pathogenesis of A(H1N1)pdm09 influenza was evidenced with the observation that JAK2 inhibition can attenuate the exuberant immune activation triggered by A(H1N1)pdm09 virus. Thus, JAK2 is a novel host factor whose genetic variations may cause the diversified outcome in A(H1N1)pdm09 patients.

Secondly, we conducted a GWAS to investigate the genetic susceptibility to human A(H7N9) influenza in 102 patients and 106 healthy controls. We applied an artificial imputation program to improve the association signals. We identified three variants of LGALS1 represented by three anchor SNPs rs13057866, rs4820294 and rs62236673 coupled with their partner SNPs. A previous study indicated that LGALS1 can bind to various subtypes of influenza A viruses, inhibiting viral infectivity and production. Our in-depth analysis indicated that the protective variants against A(H7N9) infection correlated to the significantly higher levels of LGALS1 expression in lymphoblastoid cell lines, human primary monocytes and lung tissues. Therefore, we conclude that functional variants of LGALS1 causing the expression variations contribute to the differential susceptibility to human A(H7N9) influenza. We further integrated the results of A(H1N1)pdm09 GWAS with a lung eQTL dataset. We demonstrated that the G allele of rs2070788, a higher-expression variant of TMPRSS2, was a risk variant to severe A(H1N1)pdm09 influenza in 409 patients. A potentially functional SNP rs383510 in a putative regulatory region can tag rs2070788. Luciferase assay results showed that rs383510 may dictate the activity of the regulatory element. Due to the shared usage of TMPRSS2 for the activation of A(H1N1)pdm09 and A(H7N9) viruses, we validated the genetic predispositions of rs2070788 and rs383510 to A(H7N9) influenza. Therefore, we demonstrate that genetic variants with higher TMPRSS2 expression confer higher risk to severe A(H1N1)pdm09 influenza and human A(H7N9) influenza.

In summary, integrating the genetic association studies, molecular and cell biology studies and bioinformatics analysis, we identified and characterized a number of susceptibility genes for human A(H7N9) influenza and severe A(H1N1)pdm09 influenza.