Neutralizing antibody VRC01 failed to select for HIV-1 mutations upon viral rebound

Abstract

Infusion of the broadly neutralizing antibody VRC01 has been evaluated in individuals chronically infected with HIV-1. Here, we studied how VRC01 infusions affected viral rebound after cessation of antiretroviral therapy (ART) in 18 acutely treated and durably suppressed individuals. Viral rebound occurred in all individuals, yet VRC01 infusions modestly delayed rebound and participants who showed a faster decay of VRC01 in serum rebounded more rapidly. Participants with strains most sensitive to VRC01 or with VRC01 epitope motifs similar to known VRC01-susceptible strains rebounded later. Upon rebound, HIV-1 sequences were indistinguishable from those sampled at diagnosis. Across the cohort, participant-derived Env showed different sensitivity to VRC01 neutralization (including 2 resistant viruses), yet neutralization sensitivity was similar at diagnosis and after rebound, indicating the lack of selection for VRC01 resistance during treatment interruption. Our results showed that viremia rebounded despite the absence of HIV-1 adaptation to VRC01 and an average VRC01 trough of 221 μg/mL. Although VRC01 levels were insufficient to prevent a resurgent infection, knowledge that they did not mediate Env mutations in acute-like viruses is relevant for antibody-based strategies in acute infection.

Keywords: AIDS/HIV; Adaptive immunity; Bioinformatics.

Figure 1. HIV-1 env sequences sampled at HIV-1 diagnosis and upon viral rebound were intermingled in phylogenetic trees.

Phylogenies were reconstructed for 14 participants based on env sequences obtained from plasma samples collected in acute HIV-1 infection (pre-ART, shown in black) and upon viral rebound (following simultaneous VRC01 infusion and treatment interruption, after an average of 3 years on ART, shown in gray [placebo] or colors [VRC01]). Asterisks denote sequences tested with in vitro neutralization assays. The horizontal bar represents the number of substitutions per site. Four participants had env sequences corresponding to subtype B (3183, 4011) or a CRF01_AE–containing recombinant (3499, 7242).

Figure 2. Sequence diversity, epitope distance, and VRC01 neutralization sensitivity values were similar in acute infection and upon viral rebound.

Nine participants who received VRC01 infusions showed no significant difference over time for the median pairwise diversity across env sequences (median of 10 sequences in acute infection and 15 sequences upon rebound) (A), VRC01-specific IC₅₀ values (B), and VRC01 epitope distances (C). Medians and standard deviations are represented and Wilcoxon’s signed-rank tests were performed.

Figure 3. Relationship between VRC01 decay rate, neutralization sensitivity, and time to rebound in participants who received VRC01 infusions.

Time to viral rebound was associated with the decay rate of VRC01 in serum after the first infusion (based on measurements on days 7, 14, and 21) (A). Time to viral rebound was also associated with the sensitivity to VRC01 using IC₅₀/IC₈₀ values corresponding to sequences sampled in acute infection (B and C) and to the VRC01 epitope distances (predicted from each participant’s sequences) (E and F). Predicted epitope distances were also correlated with the time to viral rebound (D). The time to rebound is the number of days between treatment cessation and an HIV-1 RNA test of 20 or more copies/mL. All 13 participants who received VRC01 infusions are represented; an asterisk before the participant ID indicates non-CRF01_AE infections.

Figure 4. Comparison of sequences from RV397 participants to sequences known to be most sensitive to VRC01 or CD4.

The top sequences correspond to the consensus residues found in the 5 sequences that were experimentally identified as most sensitive to VRC01 (VRC01-s) (9) and soluble CD4 (sCD4-s) (23). The importance of specific residues in the interaction with Env (measured by the number of VRC01 or sCD4 atoms that contact a residue) is indicated with darker colors for more influential residues. Core and rim epitope sites had a weight of 2 and 1, respectively. Sequences from the 13 participants who received VRC01 infusions are shown and labeled with a suffix corresponding to the time to rebound. Numbers in parentheses correspond to the number of sequences with the given epitope motif over the total sequences from a participant. Participants are ordered from longest to shortest time to rebound (from 296 to 9 days). Asterisks indicate participants with non-CRF01_AE infections.