E-mail Abstract Author Session Search Abstracts Program


Session 70 Poster Presentations
Resistance Testing: Methodology and Clinical Applications
Session Day and Time: Tuesday 1:30 - 3:30 pm
Room: Hall A


582
Performance of Celera Diagnostics ViroSeq HIV-1 Genotyping System on Genetically Diverse HIV-1 Strains
N. Marlowe*1, P. Swanson2, B. Drews1, J. Hackett2
1Celera Diagnostics, Alameda, CA and 2Abbott Labs, Abbott Park, IL

Background: Genotyping to assess drug resistance is critical to optimal management of HIV-infected patients (pts) undergoing antiretroviral therapy. Given the increasing prevalence of non-subtype B strains in the AIDS pandemic, it is important to evaluate performance of genotyping assays on genetically divergent HIV-1 strains. In this study, the performance of ViroSeq HIV-1 Genotyping System (pending 510(k) review) was evaluated using a well-characterized panel of genetically diverse HIV-1 strains.

Methods: A panel of 126 HIV-1 group M specimens was obtained from blood donors in Brazil, Cameroon, Uganda, Thailand, Argentina, and South Africa. The gag p24, pol integrase, and env gp41 genes (gag/integrase/env) were amplified and sequenced for each specimen to determine the HIV-1 subtype. Viral loads were determined using the LCx HIV-1 RNA Quantitative assay. RT-PCR reactions were performed using ViroSeq to amplify a 1.8 kb pol gene product that spans the entire protease and 335 codons of reverse transcriptase. Both strands were sequenced, and data were collected on an ABI Prism 3100 Genetic Analyzer.

Results: Based on results of the phylogenetic analysis, the panel consisted of: 18 subtype A, 11 B, 14 C, 9 D, 9 F, 7 G, 9 CRF01_AE, 34 CRF02_AG, and 15 mosaics (1 AD, 1 A/G/F2, 3 AF, 7 BF, 2 GA, 1 HA). Viral loads for the 126 RNA samples ranged from 2.92– 6 log10 copies/ml, and 122/126 (96.8%) specimens were successfully amplified and sequenced using ViroSeq. Four (4) specimens, for whom a genotype was not determined, included 3 Cameroonian subtype D samples (3.38, 4.73, 5.4 log10 copies/ml) and 1 CRF02 AG (3.49 log10 copies/ml). Notably, 2 specimens had viral loads close to the ViroSeq limit of detection (3.3 log10 copies/ml). Subtypes based on the pol sequences were concordant with gag/integrase/env-based genotyping for all but one sample. The K/G mosaic for pol, was subtype A in gag/integrase/env. Drug resistance analysis detected a high frequency of L10I, K20R, M36I, L63P in subtypes A, CRF01_AE, CRF02_ A/G and mosaics. 90% of samples presented M36I in protease gene. In contrast, no resistance mutations were detected in reverse transcriptase.

Conclusions: This study demonstrates that ViroSeq has the ability to accurately genotype geographically and genetically diverse HIV-1 strains. High frequency of natural polymorphisms in non-clade B HIV-1 could potentially affect a response to protease inhibitors.