CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

228

Intrahost Viral Competition and Archiving Dynamics in HIV-1 Superinfection Antoine Chaillon; Davey M. Smith; Susan J. Little; Gemma Caballero; Douglas D. Richman; Gabriel Wagner Univ of California San Diego, San Diego, CA, USA Background: The frequent occurrence of HIV-1 superinfection has been increasingly described, but little is known about the competition dynamics of the initially infecting and superinfecting virus strains over time and which variants are archived. We hypothesized that characterizing HIV-1 RNA and DNA populations of superinfected individuals with next generation sequencing (NGS) would determine the relative efficiencies of the initial and superinfecting viruses to be archived into the latent reservoir. Methods: Blood plasma and peripheral blood mononuclear cell (PBMC) samples were longitudinally collected from antiretroviral (ARV) naive participants of the San Diego Primary Infection Resource Consortiumwith previously documented superinfection (n=10). RNA and DNA were extracted, and coding regions within HIV-1 env , gag , and pol were PCR-amplified. Amplified products were then deep sequenced (454 FLX, Roche). Phylogeographic techniques were utilized to infer intrahost viral population migration of sampled viral lineages in the sampled PBMC and plasma. Results: HIV-1 DNA NGS data derived from 19 PBMC samples of 10 individuals (2 median samples per subject; IQR: 2.0-2.8 samples) with HIV-1 intrasubtype B superinfection were analyzed in conjunction with viral sequences generated from HIV-1 RNA populations in the participants’ blood plasma (median 6 samples per subject; IQR: 4.3-7.8 samples). At the latest sampled timepoint (median 18.8 months post-estimated date of infection; IQR: 12.9- 28.1 months), only 4/10 subjects (40%) had archived the initial and superinfecting strains. Among the remaining 6 individuals: 2 (20%) only had evidence of the superinfecting strain in PBMC DNA, concurrent with replacement of initial virus by superinfecting virus in blood plasma, while the other 4 (40%) only archived the initial strain despite ongoing blood plasma circulation of superinfecting virus in 3 of the 4. Viral gene flow analyses between HIV-1 RNA and DNA populations showed bidirectional migration predominantly from plasma to PBMC ( p = 0.04). Conclusions: Overall, archiving of both initial and superinfecting strains was common (40%) in the cohort, although only a single variant was archived in the remaining individuals with the initial virus predominating. As might be expected with phylodynamic

analysis, viral populations flowed freely from plasma to PBMC in these ARV naive participants. It is not clear what mechanisms determine whether superinfecting virus ultimately becomes archived in PBMC.

229 Spatiotemporal Dynamics of Drug Resistance Evolution and Persistence of RT-SHIV Alison F. Feder 1 ; Christopher Kline 2 ; Shiu-Lok Hu 3 ; Brandon Keele 4 ; Dmitri A. Petrov 1 ; Pleuni S. Pennings 5 ; Zandrea Ambrose 2 1 Stanford Univ, Stanford, CA, USA; 2 Univ of Pittsburgh, Pittsburgh, PA, USA; 3 Univ of Washington, Seattle, WA, USA; 4 Frederick Natl Lab, Frederick, MD, USA; 5 San Fransisco State Univ, San Francisco, CA, USA Background: To better design therapeutic approaches that prevent the evolution of drug resistance in HIV-1, it is important to understand how drug resistance spreads and establishes within a patient. There is mounting evidence that intrahost viral evolution is a non-homogenous process within the body and therefore must be understood spatially and temporally. Methods: We examined >3300 single-genome sequences from four macaques infected with RT-SHIV, a SIV with an HIV reverse transcriptase (RT). Macaques were given monotherapies to induce the emergence of drug resistance within RT between weeks 12-20 post-infection. Both viral RNA and DNA (vRNA and vDNA) were sampled from four different compartments (lymph node, vagina, gut and PBMC) in addition to plasma vRNA between 13 and 30 weeks post-infection. Compartmentalization was assessed using the variance partitioning statistic Φ ST from population genetics. Results: Acquisition of drug resistance is a highly dynamic process with periods of stasis. Although compartmentalization is observed, both natural selection within and migration between compartments are important for establishing drug resistant lineages. We observe differential dynamics between compartments, with the gut being particularly notable. In all four macaques, gut vRNA has the greatest rate of turnover, both compared to gut vDNA and to vRNA in other tissues. In addition, drug resistance increases in frequency more quickly in gut vDNA compared to vDNA other compartments. We find that RT-SHIV does not acquire detectable drug resistance immediately after treatment initiation, but takes more than one week. Upon treatment removal, plasma vRNA does not change significantly over the following 6 weeks, suggesting that drug resistance can have limited fitness cost. Conclusions: Drug resistance evolution depends on the interactions between partially independent compartments, some of which have SHIV populations that showmore persistence than others or respond more quickly to pressures imposed by treatment. Characterizing the differences and relationships between the compartments could be important in understanding howmultidrug resistance emerges, particularly outside of the blood, and could inform the administration of treatment. That many of these findings are reliant on the sampling scheme of both vDNA and vRNA, across compartments and time demonstrates the importance of varied sampling to obtain a more complete picture of how intrahost evolution proceeds. 230 Limited Evidence for a Bias Toward Consensus Residues Upon HIV-1 Transmission Morgane Rolland 1 ; Julie Ake 2 ; Merlin L. Robb 2 ; Nelson L. Michael 3 1 US Military HIV Rsr Prog, Silver Spring, MD, USA; 2 US Military HIV Rsr Prog, Walter Reed Army Inst of Rsr, Silver Spring, MD, USA; 3 US Military HIV Rsr Prog, Bethesda, MD, USA Background: Understanding selective mechanisms governing HIV transmission can help design anti-HIV strategies. Carlson and colleagues recently observed that amino acids (AA) identical to the population consensus were preferentially transmitted among transmission pairs (TP), implying a selection bias for increased fitness at the HIV heterosexual transmission bottleneck (Science 2014). Methods: Phylogenetic trees were reconstructed using TP-derived sequences along with two consensus sequences (independent sequences in the Zambian cohort (Con.ZM); subtype C 2004 consensus (Con.C)) and the subtype B sequence HXB2, used as an outgroup to infer the most recent common ancestor (MRCA) of TP sequences.

Poster Abstracts

89

CROI 2016