icon-folder.gif   Conference Reports for NATAP  
 
  XVII International HIV Drug Resistance Workshop
June 10-14, 2008
Sitges, Spain
Back grey_arrow_rt.gif
 
 
 
Identifying Low-Level CXCR4 Virus with Roche 454 Life Sciences Sensitive Assay
 
 
  Reported by Jules Levin
XVII HIV Drug Resistance Workshop
June 10-14, 2008
Sitges, Spain
 
In poster 105 T Le reported on using Roche 454 Life Sciences Ultra Deep sequencing to determine the prevalence and patterns of low abundance drug resistant variants in ARV-experienced subjects with viremia on ART. They concluded: "Low abundance drug resistant HIV variants in ARV-experienced individuals with viremia on ART are common. The overall burden of resistance in an ARV-experienced individual is often greater than that reflected in standard genotypes, with NRTI, NNRTI and PI low abundance resistance variants all being identified in subjects." Samples from 24 ARV-experienced subjects with detectable viral loads were evaluated. Low abundance drug resistant variants at levels ranging from 0.8% to <20% were identified in all 24 samples by Ultra Deep sequencing and were missed in 92% (22/24) samples by standard sequencing. The number of additional resistance mutations identified at a level <20% ranged from 1-10 per sample (mean of 4). The most common additional mutations identified were PI-79% (19/24 - the majority were "minor" PI mutations), followed by NRTI-75% (18/24) and NNRTI-50% (12/24 - 8 had either a K103N, Y181C, Y188C or 190A/E/S). Of these additional low abundance drug resistance variants, 38% imparted resistance to a single class of ARV, 25% to 2 classes, and 38% to 3 classes. The proportion of samples that harbored greater overall resistance by increased level of resistance to a drug(s) and by the number of additional resistant drugs were 85% (20/24) and 58% (14/24) respectively.
 
Dynamic HIV-1 Escape from Vicriviroc Therapy in vivo
 
Athe M.N. Tsibris1,2, Ramy Arnaout3,4, Bette Korber5,6, Carsten Russ7, Chien-Chi Lo5, Brian Gaschen5, Thomas Leitner5, Roger Paredes2,3,8, Zhaohui Su9, Michael D. Hughes9, Roy M. Gulick10, Wayne Greaves11, Eoin Coakley12, Charles Flexner13, Chad Nusbaum7, Bruce Birren7, Daniel R. Kuritzkes2,3 1Massachsetts General Hospital, Boston, MA, USA; 2Harvard Medical School, Boston, MA, USA; 3Brigham and Women's Hospital, Boston, MA, USA; 4Program for Evolutionary Dynamics, Harvard University, Cambridge, MA, USA; 5Los Alamos National Laboratories, Los Alamos, NM, USA; 6Santa Fe Institute, Santa Fe, NM, USA; 7Broad Institute of MIT and Harvard, Cambridge, MA, USA; 8Fundacions irsiCaixa i Lluita contra la SIDA, Hospital Universitari Germans Trias i Pujol, Universitat AutÚnoma de Barcelona, Badalona, Catalonia, Spain; 9Harvard School of Public Health, Boston, MA, USA; 10Weill Medical College, Cornell University, New York, NY, USA; 11Schering-Plough Research Institute, Kennilworth, NJ, USA; 12Monogram Biosciences, South San Francisco, CA, USA; 13Johns Hopkins University, Baltimore, MD.
 
METHODS
 
Subject Selection

We selected four subjects enrolled in ACTG 5211, a phase IIb clinical trial of VCV, who experienced protocol-defined virologic failure (VF) with a change in coreceptor usage as determined by a commercial phenotypic assay (5). Subjects with detectable X4 virus at baseline were excluded. All subjects were receiving VCV at the time of VF. Three time points were analyzed for each subject: study entry (wk 0), an intermediate time point on study drug, and VF.
 
Tropism Assay
Subjects enrolled in A5211 had HIV coreceptor testing performed by Monogram Biosciences (5).
 
V3 amplicon preparation
HIV-1 RNA was extracted from plasma and subject-specific primer sets were used to reverse transcribe and amplify plasma V3 loop-coding regions of env.
 
Deep Sequencing
V3 amplicons were submitted in a blinded fashion to the Broad Institute for sequencing (454 Life Sciences, Roche) and custom analysis.
 
AUTHOR CONCLUSIONS
 
Deep sequencing provided multiple orders of magnitude greater coverage than conventional sequencing and revealed far greater sequence heterogeneity in the HIV-1 V3 loop than previously documented
 
Minor predicted CXCR4-using populations were identified in a largely CCR5-using swarm.
 
Minor variants present at <1% of the starting population emerge rapidly in response to VCV treatment and contribute to virologic failure. Deep sequencing provided a detailed view of both extreme shifts in population frequencies toward these forms and the rapid evolutionary impact of VCV selection.
 
Deep sequencing platforms can be used to comprehensively assess viral diversity and quantify minor sequence variants.
 
BACKGROUND
The advent of antiretrovirals targeting the gp120-CCR5 interaction, coupled with the observation in clinical trials that changes in coreceptor usage represent the major in vivo pathway of HIV escape from these drugs, have re-emphasized the need to improve our understanding of coreceptor usage at the sequence population level (1). Sequencing by synthesis technologies, originally developed for genomics applications, generate sequence data by repetitive sequencing, or oversampling, of a given DNA segment (2). This oversampling can be adapted to sequence one particular DNA region at great depth, rather than sequencing an entire genome (3,4). To quantify V3 loop diversity over time and under drug selection pressure, we used this approach to resequence V3 loop amplicons derived from plasma HIV-1 RNA of infected subjects who failed therapy VCV, an investigational CCR5 antagonist.
 
Abstract
Novel high-throughput sequencing platforms provide an alternative approach for detecting minor HIV-1 variants. We applied this technology to sequence the gp120 V3 loop-coding region of env in serial plasma samples from 4 chronically HIV-infected subjects who failed vicriviroc (VCV) therapy.
 
We selected subjects from ACTG 5211, a phase IIb clinical trial of VCV, who experienced virologic failure (VF) with a change in coreceptor usage as determined by a commercial phenotypic assay (Trofile, Monogram Biosciences).
 
Three subjects met criteria: subjects 18 and 19 had an early change (week 2 after starting VCV) in coreceptor usage whereas subject 07 developed high-level VCV resistance and late emergence of a minor X4 population. A fourth subject, subject 47, had VF with no detectable change in coreceptor usage or VCV susceptibility and served as a control. All subjects were receiving VCV at the time of VF.
 
Three time points were analyzed for each subject: study entry (wk 0), an intermediate time point on study drug, and VF. HIV-1 RNA was extracted from plasma and subject- specific primer sets were used to reverse transcribe and amplify plasma V3 loop-coding regions of env. V3 amplicons were then submitted in a blinded fashion for deep sequencing and custom analysis.
 
Between 25,000-140,000 single genome sequences were obtained per subject per time point. Profound baseline V3 loop sequence heterogeneity existed. Subjects 18 and 19 failed with a dominant predicted CXCR4-using sequence that was present at 0.8% of baseline sequences; VCV noncompliance in subject 19 led to a regression of CXCR4-using V3 forms. Subject 07 showed emergence of VCV resistance at week 19 (VF); the predominant week 19 sequences comprised 0.2-3.0% of baseline sequences. Subject 47 had a majority CCR5-using sequence that was replaced by another CCR5-using sequence, also present at baseline.
 
V3 loop forms associated with VCV escape, either through CXCR4 use or the emergence of high-level VCV resistance, exist prior to therapy. Minor variants present at <1% of the starting population emerge rapidly in response to VCV treatment and contribute to virologic failure.