CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

225 Probabilistic Estimation of HIV-1 Subtype B Transmission Rates FromViral Phylogenies Stephane Hue 1 ; Roberto MurcioVillanueva 2 ; AnnaTostevin 3 ;Valerie Delpech 4 ; Anton N. Pozniak 5 ; Lynne Ashton 6 ; Deenan Pillay 7 ; Richard Goldstein 2 ; for the UK HIV Drug Resistance Database 1 London Sch of Hygiene & Trop Med, London, UK; 2 Univ Coll London, London, UK; 3 Med Rsr Council Clinical Trials Unit at Univ Coll London, London, UK; 4 PH England, London, UK; 5 Chelsea and Westminster Hosp NHS Fndn Trust, London, UK; 6 Royal Liverpool Hosp, Liverpool, UK; 7 Africa Cntr for Hlth and Pop Studies, Mtubatuba, South Africa Background: Identifying the determinants of an epidemic is of critical importance for devising prevention strategies. In HIV epidemics, incomplete sampling, large numbers of undiagnosed transmitters and stigma-related misclassifications currently limit our understanding of the transmission dynamics of the virus. We developed a probabilistic model to estimate relative rates of HIV transmission from annotated viral phylogenies that can effectively account for missing or incorrect data, and investigated the rates of HIV-1 subtype B sexual transmission in the UK. Methods: 22,481 HIV-1 subtype B pol sequences from the UK HIV Drug Resistance Database linked to epidemiological data from national and cohort clinic data were analysed. Maximum Likelihood phylogenies were reconstructed and 2,860 monophyletic transmission clusters were identified. A maximum likelihood probabilistic model was used to estimate rates of HIV sexual transmission between and amongst men who have sex with men (MSM), heterosexual males (H M ) and heterosexual females (H F ) within the identified clusters. We also estimated the proportion and most likely risk category of unobserved/undiagnosed transmitters. Results: The highest rate of HIV infection was amongst MSM. Transmission rates from H M to H F was also high, 11 times the rate from H F to H M . High infection rates fromMSM to H M were observed which, when combined with the large number of infected MSM, indicates that a substantial proportion of H M in the clusters were infected through sex with men despite reporting sex with a woman as their route of infection. The infection rate between H M was higher than fromMSM to H M , identifying transmission amongst self-declared H M as a sub-epidemic distinct from that between MSM. Transmission from H M to H F was 58 times higher than that fromMSM to H F , highlighting the linking role of H M between MSM and women. Despite their importance in the infection dynamics, H M were the most likely group to be unobserved. Conclusions: Rates of HIV transmission estimated from self-reported acquisition routes are likely to underestimate homosexually acquired subtype B infections. Heterosexual men who have sex with men are more likely to do so with other self-identified heterosexual men rather than declared MSM. H M provide the dominant link between MSM and heterosexuals, while also being the group most likely to be missing from the cohort. This work highlights the power of likelihood based statistical models to analyse pathogen transmission patterns. 226 HIV-1 Subtypes, 5’LTR-Leader Sequence Variants, and Late HIV Seroconversion Background: A minority within the Pumwani sex worker cohort in Kenya remain persistently seronegative despite frequent exposure to HIV-1. A small proportion of this group seroconverted several years later. This study attempts to identify viral factors, specifically 5’ leader sequence variations that might contribute to the late seroconversion. The 5’ leader sequence contains sites essential for replication and genome packaging, viz, primer binding site (PBS), major splice donor (SD1) and major packaging signal (PS). Methods: Partial proviral 5’ long terminal repeat (LTR) region of 20 late seroconverters and 122 early seroconverters were amplified, cloned and sequenced. Sequencher, MEGA, SPSS and HelixTree SNP and Variation Suite version 6.4.3 were employed to classify HIV subtypes and specific sequence variants correlated with the late seroconversion. Results: Phylogenetic analysis of 3667 5’LTR-leader sequence clones showed that HIV-1 subtype A1 and subtype D of Ugandan origin were over-represented in the viral population infecting the late seroconverters ( P =3.68x10 -93 ). Recursive partitioning analysis of sequence variants of PBS, SD1 and PS showed that specific sequence variants of PBS (Pc=2.12x10 - 55 ), SD1 (Pc=1.31x10 -137 ) and PS (1.90x10 -19 ) were identified only in the viral population of either early or late seroconverters. Furthermore, combinations of sequence variants of PBS [PBS-2 (P=0.001; Pc=0.04) and PBS-3 (P=6.24x10 -211 ; Pc=1.57x10 -173 )] variants with specific SD1 sequences were only present in late seroconverters. Molecular Phylogenetic analysis PBS sequence variants identified only in late (PBS-1) or early (PBS-4) seroconverters showed that PBS sequence variants identified only in late seroconverters co-clustered with PBS reference sequences utilizing tRNA Arg molecules. Whereas, the PBS sequence variants identified only in early seroconverters (PBS-4) co-clustered with PBS wild type references PBS- tRNA Lys3 and its variants. Conclusions: This study reports, for the first time, specific PBS, SD1 and PS sequence variants within 5’leader sequence are associated with clinical outcome upon HIV-1 exposure and the findings could aid designing effective anti-HIV strategies. 227 HIV-1 Stop Codon Usage and Its Clinical Impacts in Rural Ugandan Patients Guinevere Q. Lee 1 ;Viviane D. Lima 2 ;Wendy Zhang 2 ;Yap Boum 3 ; Simone H. Kigozi 3 ; A. Rain Mocello 4 ; PeterW. Hunt 4 ; Jeffrey N. Martin 4 ; David R. Bangsberg 5 ; P. Richard Harrigan 2 1 Massachusetts General Hosp, Boston, MA, USA; 2 BC Cntr for Excellence in HIV/AIDS, Vancouver, BC, Canada; 3 Mbarara Univ of Sci and Tech, Mbarara, Uganda; 4 Univ of California San Francisco, San Francisco, CA, USA; 5 Harvard Med Sch, Boston, MA, USA Background: Retroviruses have high mutation rates resulting in highly variable codon usage. Stop codon usage (TAA, TAG, TGA) is known to affect translation termination efficiency. In this study, we hypothesize that despite its high mutation rate, HIV-1 stop codon usage in clinical samples is conserved, and variability is associated with pre- and post- therapy clinical parameters. Methods: Pre-treatment baseline plasma samples were collected from initially treatment-naïve HIV-infected patients (n=454) in Mbarara, Uganda where subtype A1 and D co-circulate (44% vs 37% by pol ). Near-full-genome Sanger sequences were obtained by nested PCR and ABI 3730. Subtypes were inferred by Los Alamos RIP. Clinical correlates examined included: Pre-therapy viral load and CD4 count, post-therapy time to virologic suppression (<=400 copies/mL) and time to virologic rebound (first of two consecutive viral load >400 copies/mL post-suppression). Statistical significance was defined as p<0.003 after Bonferroni correction. As a comparator, stop codons were extracted from all available sequences in the Los Alamos HIV sequence database. Results: In the 454 Ugandan patients, HIV-1 stop codon usage was highly conserved and each gene had a specific preference: gag (90% used TAA), pol (85% TAG), vif ( 99% TAG), vpr (84% TAA) and vpu (81% TAG), env (94% TAA), nef (96% TGA). Stop codon usage was not associated with subtype except in gag : A1 (93% TAA, 202/217) and D (83% TAA, 111/134), p<0.001 Fisher’s exact. Stop codon usage was also not significantly associated with any pre-/post-therapy clinical correlates in all genes examined after Bonferroni correction (all p>=0.01). HIV sequences archived in the Los Alamos database that were predominantly subtype B (58%) and C (10%) showed similar pattern of usage conservation: gag (99% TAA), pol (82% TAG), vif (100% TAG), vpr (100% TAG), vpu (70% TAG), env (100% TAA), nef (98% TGA). Conclusions: HIV-1 stop codon usage was highly conserved, implying strong selection pressure(s) on stop codon usage and/or on codon positions in genes with overlapping reading frames. However, stop codon variability was not associated with the pre- and post-therapy clinical parameters examined in this study. Ma Luo 1 ; Raghavan Sampathkumar 2 ; Ian MacArthur 2 ; Joel Scott-Herridge 2 ; Paul Sandstrom 1 ; Joshua Kimani 3 ; Francis Plummer 1 1 Natl Microbiology Lab of Canada, Winnipeg, MB, Canada; 2 Univ of Manitoba, Winnipeg, MB, Canada; 3 Univ of Nairobi, Nairobi, Kenya

Poster Abstracts

88

CROI 2016