Background:  Previous reports showed the presence of increased levels of lipopolysaccharide (LPS) in HIV-1-infected progressors. We investigated the effect of short treatment interruptions on plasma LPS levels in chronically HIV-1-infected patients.

Methods:  A total of 76 subjects participated in this study (50 HIV⁺ and 26 healthy). Of 50 HIV⁺ subjects, 9 were viremic non-ART-treated patients, while 41 were chronically suppressed patients on ART (≥3 drugs, CD4 count >400, HIV-1 RNA <500 for >8 months, <50 at recruitment) followed for 40 weeks on continuous ART (20 of 41, non-TI group) or undergoing 2 treatment interruptions (21 of 41:  a 6-week and an open-ended treatment interruption [TI group]). T cell activation (CD8⁺/CD38⁺, CD8⁺/HLA-DR⁺, and CD3⁺/CD95⁺) by same day whole blood flow cytometry and plasma levels of LPS and immunoglobulin (Ig) M endotoxin-core antibodies (EndoCAb) by enzyme-linked immunosorbent assay (ELISA) were measured at:  a single visit (healthy and viremic non-ART-treated); sequential visits including follow-up during short-term 6-week and >20-week therapy interruptions (TI group); and beginning and on day 40 of follow-up on continuous ART (non-TI group). Variables levels between healthy and viremic were compared by t-test or the Wilcoxon/Kruskal-Wallis test (rank sums) depending on the data distribution. Correlations between variables were assessed using Spearman or pairwise correlation tests. All statistics were performed with JMP4.

Results:  LPS plasma levels were significantly lower in healthy patients than in steady-state viremic patients for >20 weeks off therapy (n = 18, p <0.0001). As a result of a treatment interruption-related acute viremia, we observed increased T cell activation and a negative association of plasma HIV RNA at week 6 with LPS levels at the same time-point, as well as with EndoCAb change between start and end of the 6-week treatment interruption (p = 0.0152, Spearman’s ρ = –0.6124 and p = 0.0204, correlation = –0.502, respectively). In addition, comparisons between the change in EndoCAb and LPS from time on therapy to a subsequent steady viremic time-point >12 weeks after treatment interruption showed a negative association (p = 0.0073, correlation = –0.8514). Finally, activation markers changes during acute viremia or CD4 count were not associated with LPS or EndoCAb.

Conclusions:  These data showing increased LPS levels in prolonged viremia compared to healthy subjects, and a negative association between EndoCab and LPS change during treatment interruption suggest an active association between HIV and microbial translocation in HIV-1-infected patients upon long-term viremia.