CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

443 Phenome-Wide Association Study (PheWAS) Using Data From ACTG Clinical Trial A5202 AnuragVerma 1 ;Yuki Bradford 1 ; ShefaliVerma 1 ; Sarah Pendergrass 2 ; Eric Daar 3 ; CharlesVenuto 4 ; Gene D. Morse 5 ; Paul McLaren 6 ; Marylyn D. Ritchie 1 ; DavidW. Haas 7 1 The Cntr for Systems Genomics, Pennsylvania State Univ, University Park, PA, USA; 2 Geisinger Hlth System, Danville, PA, USA; 3 Harbor Univ of California Los Angeles Med Cntr, Torrance, CA, USA; 4 Univ of Rochester Med Cntr, Rochester, NY, USA; 5 State Univ of New York at Buffalo, Buffalo, NY, USA; 6 PH Agency of Canada, Winnipeg, MB, Canada; 7 Vanderbilt Univ, Nashville, TN, USA Background: Clinical trials datasets likely contain as yet undiscovered genetic associations, some of which are context dependent. Phenome-wide association studies (PheWAS) explore whether human genetic polymorphisms (SNPs) are associated with any trait (phenotype) across the “phenome”. This pilot PheWAS used data from antiretroviral therapy (ART)-naïve subjects who were randomized to initiate ART regimens in AIDS Clinical Trials Group (ACTG) protocol A5202. Methods: In A5202, 1858 HIV-infected subjects were randomized to tenofovir DF/emtricitabine or abacavir/lamivudine, with either atazanavir/ritonavir (ATVr) or efavirenz (EFV). We analyzed phenotypes across 4 clinical domains: immunology derived from CD4 counts; virology derived from HIV-1 RNA data (<200 copies/mL); metabolism derived from fasting LDL cholesterol and triglycerides (TG); and pharmacology derived from EFV and ATV pharmacokinetics (PK). From these, we derived 774 phenotypes based on context at baseline (sex, race/ethnicity, age category, randomized ART regimen, individual ART drug, HIV RNA, CD4 count, and body mass index strata). We considered absolute values and change from baseline. Using Illumina Human1M-Duo data from 1181 subjects we imputed ~5.8 million SNPs, fromwhich we assessed for association with 2374 SNPs in 761 genes. These SNPs were selected from PharmGKB based on published association with any drug. Statistical models with ≥100 subjects excluded SNPs with minor allele frequency (MAF) <0.05, and models with <100 subjects excluded SNPs with MAF <0.10. Analyses controlled for age, sex, and 3 principal components. By permutation testing, the majority of SNP- phenotype associations with p<1.5x10 -4 were not by chance alone. Results: Within each domain, we observed SNP-phenotype pairs with p<1.5x10 -4 . Some associations were not unexpected, including CYP2B6 rs3745274 with EFV PK phenotypes (p=1.1x10 -28 ), APOE rs7412 with LDL phenotypes (p=2.9x10 -10 ), and rs17482753 (near LPL ) with TG phenotypes (p=2.3x10 -6 ). We also observed potentially novel associations, such as rs1799964 (near TNF ) with CD4 counts (p=2.0x10 -6 ), ABCC4 rs57270423 with ATV PK (p=2.9x10 -5 ), and CDKN2B-AS1 rs7865618 with HIV RNA control (p=6.2x10 -7 ). Conclusions: PheWAS efficiently identified both expected and potentially novel SNP-phenotype associations across multiple clinical domains in a large, randomized clinical trial. These results encourage continued application of PheWAS to HIV clinical trials datasets. 444 Effect of CYP2B6 Gene Variants on Levonorgestrel PKWhen CombinedWith EFV-Based ART Megan Neary 1 ; Mohammed Lamorde 2 ; Adeniyi Olagunju 1 ; Kristin M. Darin 3 ; Pauline Byakika-Kibwika 4 ; Concepta Merry 4 ; David J. Back 1 ; Marco Siccardi 1 ; Andrew Owen 1 ; Kimberly K. Scarsi 5 1 Univ of Liverpool, Liverpool, UK; 2 Infectious Diseases Inst, Makerere Univ, Kampala, Uganda; 3 Northwestern Univ, Chicago, IL, USA; 4 Makerere Univ Coll of Hlth Scis, Kampala, Uganda; 5 Univ of Nebraska, Omaha, NE, USA Background: The subdermal levonorgestrel (LNG) implant is a highly effective and desirable method for contraception. Our group described a significant drug interaction between LNG and efavirenz (EFV)-based antiretroviral therapy (ART), resulting in suboptimal LNG concentrations and a high rate of unintended pregnancy. Herein we describe the association between pharmacogenetic variants and LNG pharmacokinetic (PK) parameters in these study participants.< Methods: In this prospective PK evaluation of HIV-infected Ugandan women receiving EFV-based ART (n=20), a LNG implant was inserted at study entry and a whole blood sample was collected for pharmacogenetic analysis. At each study visit over 48 weeks, a plasma sample was collected to assess LNG and mid-dose EFV PK. The primary endpoint was LNG PK at week 24. Follow-up was interrupted for 9 subjects between weeks 36-44 after 3 pregnancies were identified. LNG and EFV concentrations were analysed by a validated LC-MS/MS and HPLC method, respectively. SNPs in CYP2B6 (rs3745274, rs28399499, rs4803419), CYP2A6 (rs28399433, *9B rs8192726), NR1I2 (rs2472677) and NR1I3 (rs2307424, rs3003596) were analysed using TaqMan assays. Associations between patient genotype and LNG PK were determined through univariate and backwards multivariate linear regression. Results: All women were Black African and received EFV for a median of 10 (range 5-66) months prior to entry. Allele frequencies for associated SNPs in CYP2B6 (rs3745274, 516G>T; rs4803419, C>T) are described in Table 1. CYP2B6 rs4803419 ( X 2 =20.62, P =<0.001) and NR1I3 rs2307424 ( X 2 =12.36, P =<0.001) were not in Hardy-Weinberg equilibrium, which compromises their interpretation. The presence of a T allele for CYP2B6 rs3745274 was associated with lower log 10 LNG C max ( P =0.02 , β=-0.20). The presence of a T allele for CYP2B6 rs4803419 was associated with lower log 10 LNG AUC 0-24wk ( P =0.006 , β=-0.18) and log 10 LNG AUC 0-last ( P =0.007 , β=-0.25). LNG and EFV PK results summarized by CYP2B6 variant are presented in Table 1. No other genetic associations were observed; however, the sample size was too low to robustly assess the CYP2B6 983T>C variant (rs28399499). Conclusions: These data demonstrate that pharmacogenetic variations in CYP2B6 influence LNG pharmacokinetics when combined with EFV-based ART. This supports the need for further investigation in a larger population to assess whether a pharmacogenetic approach could be used to identify patients on EFV who are at highest risk of suboptimal LNG concentrations.

Poster Abstracts

445

Transporter Genetics and TFV-DP/FTC-TP Cellular Pharmacology In Vivo Sharon M. Seifert 1 ; Xinhui Chen 2 ; CarolynW. Clayton 3 ;Taylor Alford 1 ; Amie L. Meditz 4 ; Jose R. Castillo-Mancilla 5 ; Lane R. Bushman 5 ; Christina Aquilante 1 ; Samantha MaWhinney 3 ; Peter L. Anderson 5 1 Univ of Colorado Skaggs Sch of Pharm and Pharmaceutical Scis, Aurora, CO, USA; 2 Univ of Colorado, Aurora, CO, USA; 3 Colorado Sch of PH, Aurora, CO, USA; 4 Beacon Cntr for Infectious Diseases, Boulder, CO, USA; 5 Univ of Colorado, Denver, CO, USA Background: The intracellular anabolites, tenofovir-diphosphate (TFV-DP) and emtricitabine-triphosphate (FTC-TP), exhibit high inter-individual variability in vivo, but the sources of this variability have not been fully elucidated. The purpose of this study was to assess the association between single nucleotide polymorphisms (SNPs) in candidate

170

CROI 2016

Made with FlippingBook - Online catalogs