Immune correlates of cell-associated HIV RNA levels in infected individuals undergoing long-term antiretroviral therapy


BACKGROUND: Chronic low-level immune activation in people living with HIV on suppressive antiretroviral therapy (ART) is associated with elevated morbidity and mortality. Sources or immune activation are uncertain, but may be related to ongoing HIV RNA production from persistently infected cells. To investigate the role of HIV RNA levels in immune activation during ART, we used multi-modal statistical analyses to identify cellular immune subsets associated with HIV RNA levels.
METHODS: Clinical information and peripheral blood mononuclear cells (PBMC) were collected from individuals on suppressive ART '¥3 years. Cell-associated HIV DNA and RNA levels were measured by multiplexed droplet digital PCR assays that simultaneously quantified HIV LTR and gag regions. Levels of unspliced HIV RNA levels were quantified per total number of proviruses (normalizing gag RNA to levels of LTR DNA/2) and per total number of gag-containing proviruses (normalizing to HIV gag DNA. PBMC were analyzed in extended flow cytometry panels quantifying >20 lymphocyte and activation markers. To identify immune parameters most frequently associated with HIV gag production, we used 24 different statistical methods, including parametric and non-parametric correlation approaches, regression, and classification methods using varying assumptions about dataset structure.
RESULTS: PBMC samples (N=70; 90% male; median CD4=672 cells/µl) were analyzed (N=70 for DNA, N=60 for RNA). Participants had been infected for a median '¥19.75 years (range 3.9-34.2 years). All had detectable DNA (LTR 20-17327 copies/106 PBMC, gag 7-4779 copies/106 PBMC). HIV RNA levels were lower (LTR 5-2219 copies/106 PBMC, gag 0-883 copies/106 PBMC); HIV RNA and DNA were quantifiable in 81% of participants. For gag RNA:gag DNA levels, proportion of natural killer (NK) cells (CD16+CD56+) positively correlated in the most models (11/24 models), followed by nadir CD4 (9/24 models) and minimum duration of infection (7/24 models). When gag was normalized to LTR DNA, NK and minimum duration of infection were strongly correlated in 14/24 and 5/24 models, respectively.
CONCLUSIONS: Levels of cell associated HIV RNA were quantifiable in the majority of individuals undergoing ART. Detailed immunophenotyping revealed correlations between cell-associated HIV gag RNA levels and proportion of NK cells, suggesting a key role of innate immunity in proviral RNA expression.

Download the e-Poster (PDF)