HIV-Associated Gut Microbiome Differences and Immune Activation are Dependent on Host Context

Abstract

Even with antiretroviral therapy, individuals infected with Human Immunodeficiency Virus (HIV) experience greater mortality than uninfected individuals. This mortality is due in large part to inflammation-related non-communicable diseases, such as cardiovascular disease and kidney dysfunction, but the causes of the underlying inflammation are not fully understood. One proposed driver is the gut microbiome, which is essential to human health and involved in many metabolic and immune interactions. Recent studies have found HIV-associated disruptions in the gut microbial community, but few of these data are derived from subjects in sub-Saharan Africa where HIV burden is greatest. We hypothesized that HIV-associated gut microbiome differences vary by geography, influencing the relationship with elevated host inflammation. To test this, we profiled the gut microbiota and measured serum inflammatory markers in more than 700 subjects in rural Mbarara, Uganda; urban Gaborone, Botswana; and Boston, USA. Using computational and statistical methods, we found that within each geographic cohort, the microbiomes of HIV-infected and -uninfected individuals exhibited significant differences. In the Ugandan cohort, HIV-uninfected individuals had communities dominated by the genus Prevotella, while HIV-infected individuals had relatively lower abundance of Prevotella and a greater abundance of taxa including Bifidobacterium and Akkermansia. In contrast, gut communities from HIV-infected individuals in Boston showed increased Prevotella and decreased Akkermansia and other taxa relative to HIV-uninfected subjects. Compared to the other cohorts, individuals in the Botswanan cohort had both common and distinct HIV-associated differences. Within the Boston cohort, sexual practices were an interacting factor as men who have sex with men (MSM) exhibited greater HIV-associated gut microbiome differences. Inflammatory markers were elevated in HIV-infected individuals in all cohorts and this inflammation was associated with unique bacterial taxa in each cohort. We isolated bacteria within the species Prevotella copri from individuals across the cohorts and found geographically associated functional differences, demonstrating that host context can contribute to microbial differences both at high taxonomic levels and within sub-species diversity. In order to most effectively translate HIV-gut microbiome research into opportunities for therapeutic interventions, studies should include populations in multiple geographic locations, as we found HIV-associated changes to be highly context dependent.