Absence of HIV-1 evolution in the gut-associated lymphoid tissue from patients on combination antiviral therapy initiated during primary infection

Abstract

Mucosal mononuclear (MMC) CCR5+CD4+ T cells of the gastrointestinal (GI) tract are selectively infected and depleted during acute HIV-1 infection. Despite early initiation of combination antiretroviral therapy (cART), gut-associated lymphoid tissue (GALT) CD4+ T cell depletion and activation persist in the majority of HIV-1 positive individuals studied. This may result from ongoing HIV-1 replication and T-cell activation despite effective cART. We hypothesized that ongoing viral replication in the GI tract during cART would result in measurable viral evolution, with divergent populations emerging over time. Subjects treated during early HIV-1 infection underwent phlebotomy and flexible sigmoidoscopy with biopsies prior to and 15-24 months post initiation of cART. At the 2(nd) biopsy, three GALT phenotypes were noted, characterized by high, intermediate and low levels of immune activation. A representative case from each phenotype was analyzed. Each subject had plasma HIV-1 RNA levels <50 copies/ml at 2(nd) GI biopsy and CD4+ T cell reconstitution in the peripheral blood. Single genome amplification of full-length HIV-1 envelope was performed for each subject pre- and post-initiation of cART in GALT and PBMC. A total of 280 confirmed single genome sequences (SGS) were analyzed for experimental cases. For each subject, maximum likelihood phylogenetic trees derived from molecular sequence data showed no evidence of evolved forms in the GALT over the study period. During treatment, HIV-1 envelope diversity in GALT-derived SGS did not increase and post-treatment GALT-derived SGS showed no substantial genetic divergence from pre-treatment sequences within transmitted groups. Similar results were obtained from PBMC-derived SGS. Our results reveal that initiation of cART during acute/early HIV-1 infection can result in the interruption of measurable viral evolution in the GALT, suggesting the absence of de-novo rounds of HIV-1 replication in this compartment during suppressive cART.

Figure 1. Immunological profiles of experimental patients before and after initiation of cART.

(A) Representative flow plots from low immune activator patient (LIA) are depicted. CD8⁺ T cells are shown on the x-axis and CD4⁺ T cells on the y-axis. (B, C) Bar graphs depicting CD4∶CD8 ratio (B) and levels of immune activation (C) of PBMC (red bars) and MMC (blue bars) before (PBMC1 and MMC1) and after (PBMC2 and MMC2) initiation of antiretroviral therapy. High immune activator (HIA), intermediate immune activator (IIA) and low immune activator (LIA) are compared. (D, E) Immunohistochemical characterization of immune effector sites in rectosigmoid biopsies. Using a PC-based image-analysis system (KS 4000, Kontron) a standard area was set by the image analyzer. A total of between 10 and 15 consecutive non-overlapping fields were analyzed. In figure (D), a biopsy section (viewed at 40× magnification) compares CD4+ T cells (stained red) of subject IIA before (left panel) and after (right panel) initiation of antiretroviral therapy. Figure (E) depicts mean CD4+ T before (LP1) and after (LP2) initiation of antiretroviral therapy. HIA, IIA and LIA are compared.

Figure 2. Intra-Patient clustering of HIV-1 env quasi-species.

Maximum Likelihood (ML) tree depicting full-length HIV-1 env sequences from three experimental subjects (HIA, IIA, LIA) and 2 positive control subjects (POS1 and POS2). For each subject, all sequences from both time-points (baseline and on treatment) and from all compartments described are shown. Bootstrap values (>95%) are shown for inter-subject clusters. 1000 bootstrap replicates were run and Bootstrap values over 85% are shown. The horizontal scale bar represents 2.0% genetic distance. Each subject forms a tight cluster and is distinct from other experimental subjects or controls. HXB2 was used as an outgroup.

Figure 3. Evolution of HIV-1 env in the absence of cART – positive control patients.

ML tree of SGA sequences from (A) positive control subject 1 (POS1) and (B) positive control subject 2 (POS2). For both panels, PBMC time point #1 (open red circles) and PBMC time point #2 (closed red circles) are shown. 1000 bootstrap replicates were run and Bootstrap values over 85% are shown. The scale bar represents 0.005 nucleotide substitutions per site. HXB2 was used as an outgroup. Starred sequences represent those determined to be hypermutated. On the right is a Highlighter plot corresponding to the ML tree (outgroup sequence not shown). Individual nucleotide changes from a master sequence (TP1 sequence at the top of the ML tree) are shown. A-green; T-red; G-orange; C-light blue. For participant POS2, phylogenetic analysis reveals acquisition of more than one variant as shown.

Figure 4. HIV-1 env phylogenies do not suggest measurable evolution on suppressive cART.

ML tree of SGA sequences from participants (A) HIA (B) IIA and (C) LIA. For all panels, PBMC time point #1 (open red circles), PBMC time point #2 (closed red circles), GALT time point #1 (open blue squares) and GALT time point #2 (closed blue squares) are shown. 1000 bootstrap replicates were run and Bootstrap values over 85% are shown. The scale bar represents 0.005 nucleotide substitutions per site. HXB2 was used as an outgroup. Starred sequences represent those determined to be hypermutated. To the right of each ML tree is the corresponding Highlighter plot (outgroup sequence not shown). Individual nucleotide changes from a master sequence (TP1 sequence at the top of the ML tree) are shown. A-green; T-red; G-orange; C-light blue. For participant IIA, phylogenetic analysis reveals acquisition of more than one variant with in-vivo inter-lineage recombinants. Sequences surrounded by boxes represent HIV-1 env recombinants (red boxes denote a sequence p value of p<0.05 in Recco; black boxes denote a sequence p value of p<0.25 in Recco).