CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

Methods: Specimen CG-0018a-01 was collected in DRC in 2001 as part of an HIV prevention of mother to child transmission (PMTCT) study. Previous subgenomic HIV-1 sequences branched closely with the proposed subtype L references, but small sample volume and low viral load limited efforts to expand genome coverage. In the present study, the complete genome was assembled through application of metagenomic Next-Generation-Sequencing (mNGS) and target enrichment (HIV-xGen) methods. Neighbor-joining phylogenetic and recombinant analyses were completed to classify the genome using Phylip v3.5 and Simplot v3.5.1. Results: The combined mNGS and HIV-xGen approach yielded 4,363,031 of 11,046,542 total reads (39.5%) that mapped to the final 9681 bp complete genome at an average coverage depth of 47,783x. The CG-0018a-01 genome branched with the putative subtype L references with a bootstrap value of 100 in a phylogenetic tree. Notably, the CG-0018a-01 branch was basal to the junction of 83CD003 and 90CD121E12, which suggests CG-0018a-01 may be more closely related to an ancestral strain. Recombinant analysis did not identify any breakpoints and indicated the putative subtype L references had the highest percent identity to CG-0018a-01 across the genome except in the well-conserved pol region. Subgenomic phylogenetic analysis of the pol region confirmed that CG-0018a-01 branched with L references with bootstrap support of 97. Conclusion: The subtype L classification has now been established by the non- recombinant HIV-1 genome of CG-0018a-01 as the third isolate in this divergent Group M branch. The identification of CG-0018a-01 decades after the first two subtype L strains were collected suggests rare transmission of subtype L may be ongoing in DRC. Although it was collected most recently, CG-0018a-01 appears to be more closely related to the ancestral subtype L strain than the other two isolates and will be important for determining the origins of subtype L. 936 WHOLE-GENOME SEQUENCING SHOWS INCREASING HIV-1 SUBTYPE COMPLEXITY AMONG MSM IN THE UK Jean L.Mbisa 1 , Juan Ledesma 1 , David F. Bibby 1 , Carmen M. Fernandez Manso 1 , Peter Kirwan 1 , Alison E. Brown 1 , Hodan Mohamed 1 , Yuen T. Chan 1 , Gary Murphy 1 , David Dunn 2 , Caroline Sabin 2 , Valerie Delpech 1 , Kate El Bouzidi 2 , Anna Maria Geretti 3 , for the UK HIV Drug Resistance Database 1 Public Health England, London, UK, 2 University College London, London, UK, 3 University of Liverpool, Liverpool, UK Background: HIV recombination can occur following co-infection with two or more different strains. HIV whole genome sequencing (WGS) provides a better understanding of the recombination process and characterization of circulating strains. This helps to better define virus evolution and transmission dynamics. Methods: HIV-1 WGS was undertaken on 382 samples frommen-who-have- sex-with-men (MSM) collected between 2000-2006 (n=201) and 2015-2016 (n=181). The former consisted of chronic (n=157) and recent (n=44) infections whereas the latter of recent infections only. Recency of infection was defined by avidity assay. More than 110,000 partial pol gene sequences from routine HIV-1 genotyping in the UK were obtained from the UK HIV Drug Resistance Database (UKHDRD). Subtyping was performed using REGA HIV subtyping tool and Cluster Picker was used for transmission cluster analysis (1.5% genetic distance and 90% bootstrap support). Linked clinical and demographic data were extracted from the HIV and AIDS Reporting System at PHE. Results: Partial pol gene sequence data shows a gradual increase in diagnosed infections involving complex recombinants among MSM in the UK from 0.8% (n=630) in 2000 to 9.3% (n=2655) in 2014 (p>0.001). Among recently infected MSM the proportion of complex recombinant infections was 11.0% (55/501) in 2014. WGS data shows even higher proportion of recent infections involving complex recombinants in 2015-2016 at 18.1% (33/181) compared to 2.3% (1/44) in 2000-2006. Furthermore, 32.4% (11/34) of WG sequences classified as complex recombinants were similarly classified using partial pol gene only. The most common subtypes involved in recombination were A and B (n=17 each; 50.0%). Most men infected with complex recombinants were born in the UK (63.6%; n=21) and probably acquired HIV in the UK (84.8%; n=28). Using WGS data only, 18.2% (n=6) of the complex recombinants formed 2 transmission clusters, containing 2 and 4 sequences. When analyses included partial pol sequences from the UKHDRD, 27.3% (n=9) of the complex recombinants were in 5 transmission clusters, each containing 2-9 sequences. Partial pol sequences were classified as pure subtypes or CRFs (B or CRF02_AG) in 3 clusters and complex recombinants in 2. Conclusion: WGS shows that routine HIV-1 genotyping significantly underestimates the prevalence and complexity of circulating recombinant

(n=13), and 33 complex types). No reported circulating recombinant forms were identified. Almost all of the NFL genomes (91.8%) contained at least one breakpoint, either intra- or inter-subtype. The frequency of recombination breakpoints along the genome was similar in intra- and inter-subtype recombinants. K-means clustering of recombinant A1/D genomes revealed a particular genome region which was often inherited intact, extending from C2 in gp120 to TM in gp41. In addition, a generalized linear model showed significantly fewer breakpoints in the gag-pol and envelope C2-TM regions compared with accessory gene regions. There was little evidence of large-scale transmission of recombinants within this sample: almost all (153/164; 93%) of the A1/D recombinants are unique recombinant forms. Conclusion: Recombination in HIV genomes is pervasive within and between subtypes in the populations studied and exhibits clear biases in breakpoint location. Its distorting effect on genealogical inference should therefore be acknowledged and taken into account more widely. 934 JACKHAMMER RT-PCR RECOVERS DIVERSE ARCHIVAL VIRAL GENOMES FROM KINSHASA, 1983 Sophie Gryseels 1 , Thomas D.Watts 1 , Thomas Quinn 2 , Philippe Lemey 3 , Henry Francis 4 , Oliver Laeyendecker 5 , Michael Worobey 1 1 University of Arizona, Tucson, AZ, USA, 2 Johns Hopkins University, Baltimore, MD, USA, 3 Katholieke University Leuven, Leuven, Belgium, 4 FDA, Silver Spring, MD, USA, 5 Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA Background: Viral genome data are a key means for characterizing current epidemics as well as reconstructing past epidemiological and evolutionary histories. To reliably infer the past, genome data stored in archival samples can provide essential calibration points for dated phylodynamic analyses. In the case of epidemics with a long pre-discovery history, such as the HIV/AIDS pandemic, specimens from early phases in the epidemic are very scarce, and the remaining viral genetic material are by now often degraded, thus warranting very sensitive sample-to-sequence procedures. Methods: Here we expand our ‘jackhammer’ multiplex PCR approach to amplify and sequence HIV-1 RNA from 45 serum/plasma specimens from Democratic Republic of Congo (DRC) sampled in 1983 from the very first diagnosed AIDS patients of Africa. A sequential set of 63 primer pairs, compatible with most known subtypes, were designed that target 63 150-300 nucleotide overlapping regions across the coding HIV-1 genome. Primers with non-overlapping targets were combined into six pools, so that reverse transcription and a pre-amplification PCR could efficiently be performed in only six reactions per sample, before the final amplifications in 63 reactions. Results: On average 80 % of PCRs produced reliable (Sanger) sequences after this <2 day general procedure, resulting in an average of +- 7000 nt of HIV-1 sequence data per sample. Performing additional PCRs with shuffled primers from an augmented primer set resulted in complete coding genomes for all samples. Twenty of the sequenced genomes were designated as a ‘pure’ subtype (A1, D, C, F1), two genomes were of an unknown subtype, six were known circulating recombinant forms (01_AE, 02_AG, 13_CPX, 25_cpx), and the remaining seventeen were each unique recombinants. Conclusion: The recovered diversity spans essentially the entire global HIV-1 group M diversity, which (1) provides direct evidence that the breadth of HIV-1 group M diversity was already present when AIDS was first identified in Africa, and (2) indicates our method can efficiently recover virtually any (even degraded) HIV-1 group M genome. We analyze these genomes together with other time-stamped sequences from central Africa in a phylodynamic framework to refine the timings of the major early growth phases of the HIV-1 epidemic in this region. Julie Yamaguchi 1 , Ana Vallari 1 , Carole McArthur 2 , Larry Sthreshley 3 , Gavin Cloherty 1 , Michael Berg 1 , Mary A. Rodgers 1 1 Abbott Labs, Abbott Park, IL, USA, 2 University of Missouri–Kansas City, Kansas City, MO, USA, 3 Presbyterian Church (USA), Louisville, KY, USA Background: As part of the region where HIV-1 initially expanded early in the global pandemic, the Democratic Republic of Congo (DRC) is where the most diverse HIV isolates have been found, including all recognized Group M subtypes and many unclassifiable sequences. Two divergent non-recombinant sequences, 83CD003 and 90CD121E12, collected in 1983 and 1990 in DRC were previously proposed as a new subtype, L. However, HIV nomenclature standards require three epidemiologically distinct genomes for a new classification. 935 A THIRD COMPLETE GENOME ESTABLISHES HIV-1 SUBTYPE L

Poster Abstracts

CROI 2020 351