CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

microbiota abundance were examined for each visit separately (n=74) using zero-inflated negative binomial regression (ZINB). Further analysis of selected variables of interest was performed using propensity scores (PS) to better account for multiple confounders. Results: Using permutational multivariate analysis of variance (PERMANOVA) with Bray-Curtis distances we found the most significant drivers of microbiome variation were the individual (P=0.001), methamphetamine use (P=0.04), and oral semen exposure (P=0.003). Microbiome composition varied with sexual behavior; oral sex and receptive anal intercourse (RAI) displayed overlapping associations, likely attributable to multiple sex acts. Drug use similarly influenced microbiome composition, with drugs often used concurrently clustering together (prescription drugs, ecstasy, party drugs). Marijuana and methamphetamine displayed unique associations, most notable for increased Prevotella with methamphetamine. Further analysis using PS to better account for multiple confounders revealed clearer associations. Methamphetamine use was associated with increased Prevotella , Campylobacter , and Peptostreptococcus while marijuana use decreased Paraprevotella and Gemella . Recent RAI was associated with increased Gemella , Mycoplasma and Veillonella , with decreases in Parabacteroides , Blautia , and Ruminococcus . Conclusion: Drug use and sexual behavior influence intestinal dysbiosis during chronic HIV-1 infection among MSM, with specific inflammatory changes associated with methamphetamine use. This study highlights the importance of these factors in HIV-associated dysbiosis. 266 LIVER BACTERIAL DYSBIOSIS PERSISTS DURING ANTIRETROVIRAL THERAPY IN SIV+ MACAQUES Bridget Fisher 1 , Cole Fisher 1 , MatthewWood 1 , Melanie Gasper 2 , Elise Smith 3 , Richard Green 3 , Connor Driscoll 3 , Michael Gale 3 , Guido Silvestri 4 , Ann Chahroudi 5 , Nichole Klatt 6 , Donald Sodora 1 1 Center for Infectious Disease Research, Seattle, WA, USA, 2 Seattle Children’s Research Institute, Seattle, WA, USA, 3 University of Washington, Seattle, WA, USA, 4 Yerkes National Primate Research Center, Atlanta, GA, USA, 5 Emory University, Atlanta, GA, USA, 6 Washington National Primate Research Center, Seattle, WA, USA Background: Even during combination antiretroviral therapy (cART), liver disease is a significant contributor to morbidity and mortality in HIV+ individuals. Here, we present evidence in a SIV-macaque model that treatment with cART does not reduce bacteria levels or microbial dysbiosis in the liver, which may contribute to hepatic dysfunction and impact disease outcome. Methods: Liver tissue was acquired at necropsy from infant and adult rhesus macaques that were uninfected (n=7/4, infant/adult), SIV+ (n=9/6), or SIV+/ cART-treated (n=4/6). LPS-binding protein (LBP) levels were quantified in plasma (ELISA) and 16s bacterial DNA was quantified by in the liver (qPCR). The liver microbiome composition was evaluated by 16s microbial sequencing with operational taxonomic units determined by QIIME analysis. Dysbiotic liver bacteria were tested for the ability to induce liver-specific hepatocyte inflammation (gene expression analysis) and systemic immune activation using whole blood assays (flow cytometry). Results: An increase in bacterial translocation was identified via elevated plasma LBP levels in both SIV+ and SIV+cART macaques. Assessment of liver bacterial 16s DNA levels determined elevated bacterial load was present in SIV+cART macaques (compared to uninfected) (p=0.0006). Liver microbiome assessment of SIV+macaques revealed an abundance of bacteria generally associated with HIV dysbiosis in the gut (e.g. Gammaproteobacteria) and an enrichment of Mycobacteria, which persisted during cART. Using multi-gene sequencing, the liver-associated Mycobacteria was identified as M. smegmatis, an opportunistic pathogen. In vitro experiments demonstrated that M. smegmatis stimulates monocytes to produce high levels of TNFa, even more potently than other species of Mycobacteria (BCG, M. tuberculosis). Importantly, M. smegmatis also induced inflammatory responses in hepatocytes through an upregulation of inflammation-associated genes (e.g. CRP, IL-6, IFNb, CCL2), Conclusion: Our findings indicate that cART does not sufficiently reduce bacterial translocation as demonstrated by elevated levels of LBP in plasma and bacterial DNA in the liver. In addition, an abundance of Mycobacteria was observed in the liver that was associated with inflammatory responses in both monocytes and hepatocytes. These findings provide mechanistic insights regarding the factors that influence the high prevalence of liver disease in HIV+cART-treated individuals.

Poster Abstracts

267 PREGNANCY ASSOCIATED VAGINAL PROTEOME ALTERATIONS LINKED TO HIV ACQUISITION RISK Christina Farr Zuend 1 , Nicole Tobin 2 , Kenzie Birse 1 , Laura Noël-Romas 1 , Lani Kotyrba 1 , Trisha Vera 1 , Sarah Mutch 1 , Max Abou 3 , Stuart McCorrister 3 , Garrett Westmacott 3 , Grace M. Aldrovandi 2 , Adam Burgener 1 1 University of Manitoba, Winnipeg, MB, Canada, 2 University of California Los Angeles, Los Angeles, CA, USA, 3 Public Health Agency of Canada, Winnipeg, MB, Canada Background: Pregnant women are at increased risk of HIV acquisition, due to both behavioral and biological factors, but the mechanisms are not well understood. Immunological, structural and bacterial changes during pregnancy may all lead to increased HIV susceptibility. Here we assessed host immune pathways and bacterial differences in the vaginal mucosa of pregnant and non-pregnant women using a metaproteomics approach to identify potential mechanisms of HIV susceptibility. Methods: Cervicovaginal lavage (CVL) samples collected from 23 pregnant and 25 non-pregnant women were analyzed by mass spectrometry. Microbial abundance and taxa identifications were determined using non-homologous bacterial proteins. Differential protein expression was determined by t test and bacterial taxa by Mann-Whitney test. Functional information was assigned using the KEGG ontology and IPA knowledge databases. Gene set enrichment analysis (GSEA) was utilized to compare pregnancy-associated signatures with HIV acquisition risk using a proteomic dataset generated from 701 women, 63 of which were pre-seroconversion CVL samples. Results: 550 human proteins and 376 bacterial proteins from 9 genera were identified. Two major bacterial groups were identified, Lactobacillus dominant or non-Lactobacillus dominant. All pregnant women were Lactobacillus dominant (100%), compared to only 79% of non-pregnant women (p=1.66E-2). Pregnancy also associated with changes to the functional microbiome, including increases to carbohydrate metabolism (p=3.66E-2) and metabolism of cofactors and vitamins (p=3.95E-2). Host proteome analysis indicated 56 human proteins (10%) were differentially abundant (p<0.05) between pregnant and non- pregnant women, including alterations to complement (p=3.63E-3), leukocyte extravasation signaling (p=1.45E-2), blood vessel formation (3.36E-3) and tissue permeability (p=1.27E-4). GSEA analysis indicated that pregnant women with ectopy had the strongest overlap (p=0.003, p<0.001) to host proteome signatures predicting increased risk for HIV acquisition. Conclusion: Pregnant women, particularly those with cervical ectopy, have alterations to mucosal proteome pathways related to immune system, blood vessel formation and mucosal barrier function. The overlap of certain pregnancy-associated pathways with increased HIV acquisition risk in an independent cohort provides a plausible role of these functions in HIV susceptibility during pregnancy.

93

CROI 2018