Background:  The pool of latently infected CD4⁺ T cells is established early in HIV infection and compromises the efficacy of ART. This latent pool contributes to ongoing viral replication on ART and pathogenesis of T-cell subsets. We therefore studied correlations between HIV-infected cellular reservoirs and T-cell subsets, and their association with viral suppression in the setting of a randomised controlled trial of ART in primary HIV infection (PULSE).

Methods:  Individuals with acute and early HIV infection (n = 68) were randomized to receive ART with (n = 35) or without hydroxyurea (n = 33). Following 6 to 12 months of continuous ART, all ART were ceased. If the HIV viral load reached >5000 copies/mL, ART was restarted. A good response (n = 18) was defined as maintenance of HIV RNA/mL >5000 copies for 24 weeks during any interruption. Cell associated HIV DNA and HIV unspliced RNA (usRNA) were quantified in peripheral blood mononuclear cells (PBMC) following 12 weeks of ART using real-time polymerase chain reaction (RT-PCR) in a representative sample (n = 19). Integrated HIV DNA was quantified prior to ARV using a nested Alu-LTR RT-PCR assay (n = 19). The lower limit of detection for each assay was 50 HIV RNA copies/mL, 200 copies usRNA/10⁶ PBMC, 20 copies total HIV DNA/10⁶ PBMC, and 16 copies integrated HIV DNA per input cellular DNA. Memory and activated CD4⁺ and CD8⁺ T-cell subsets were prospectively analyzed in fresh whole blood samples by flow cytometry from 25 individuals (n = 11 common).

Results:  A good response was more likely in individuals with plasma HIV RNA and cell-associated usRNA below the limit of detection at week 12 (p = 0.017 and 0.025, respectively, Fisher Exact). In a univariate analysis of all baseline T-cell subsets and plasma HIV RNA levels during the first year of therapy, week 12 HIV RNA was the best predictor of good response (p = 0.0007, Wilcoxon). Baseline integrated HIV DNA was highly correlated with plasma HIV RNA at week 12 (p = 0.002) but not with week 12 total HIV DNA. Integrated HIV DNA at baseline was inversely correlated with the number of CD4⁺CCR5⁺Ki67^– T cells at the end of 1 year of treatment (p = 0.03).

Conclusions:  The relationship between baseline integrated HIV DNA and plasma HIV RNA at week 12 on ART, and longer-term viral suppression suggest that the early establishment of the pool of latently infected cells greatly contributes to ongoing infection. Resting CCR5⁺CD4⁺ T cells may be vulnerable to depletion from this process even during therapy.