| Publication Abstract Display | | Type: Published Abstract | | Title: Genetically predicted gene expression and HIV-associated neurocognitive disorder. | | Authors: Smieszek S, Bush W, Hulgan T, Samuels D, Franklin D, Heaton R, Letendre S, Ellis R, Kallianpur A | | Year: 2017 | | Publication: Annual Conference on Retroviruses and Opportunistic Infections | | Volume: Issue: Pages: | | Abstract:Background
HIV-Associated Neurocognitive Disorder (HAND) is a term established to capture a spectrum of neurocognitive (NC) deficits associated with HIV infection. The genetic underpinnings of HAND are poorly understood. CHARTER is a U.S.-based observational study of neuro-HIV outcomes in ambulatory, HIV+ adults who underwent standardized, comprehensive NC assessment (2003-7) and were assigned a Global Deficit Score (GDS) [normal (GDS<0.5) or impaired (GDS≥0.5)]. In this study, we investigated the impact of genetic variants known to alter the expression of genes in whole blood on GDS outcomes.
Methods
We used CHARTER genome-wide association study (GWAS) data (imputed to the 1000 Genomes reference) to predict gene expression using an approach called PrediXcan. It estimates the genetically regulated component of gene expression using reference panels from studies of expression quantitative trait loci (eQTL). In this study, Gene/Tissue Expression data and CHARTER genome-wide genotype data were used to model the expression profile of 11,000 genes. We then evaluated associations of these imputed gene expression traits with two CHARTER NC phenotypes (continuous GDS and GDS impairment). We performed regression analyses to identify predicted gene expression values that associate to continuous GDS and GDS impairment, adjusting for population effects and known clinical covariates (comorbidity status, current CART use, plasma viral load, nadir CD4+ T-cell count).
Results
Among the top genes were Ankyrin Repeat Domain 44 (ANKRD44), insulin receptor substrate 2 (IRS2), and Activating Transcription Factor 3 (ATF3). Using a set of 222 genes associated to CHARTER phenotypes at p < 0.01, we performed gene pathway enrichment analysis. Initial analysis suggests overrepresentation of iron ion binding, immune defense response, regulation of inflammatory process and mitochondrial membrane pathways. Additionally, among the most significant associations (p < 0.01), we found 17 genes with known HIV protein interactions.
Conclusions
Using PrediXcan, we have identified genes and pathways potentially influencing NC impairment in HIV-infected individuals. Based on these findings, we further hypothesize that individuals with altered regulation of HIV-interacting genes may be predisposed to HAND in the presence of HIV infection. This study provides further support for a role of iron transport in HAND pathogenesis and evaluates gene expression effects by modeling the mechanisms through which genetic variants influence NC impairment. |
return to publications listing
|