Early low-titer neutralizing antibodies impede HIV-1 replication and select for virus escape

Abstract

Single genome sequencing of early HIV-1 genomes provides a sensitive, dynamic assessment of virus evolution and insight into the earliest anti-viral immune responses in vivo. By using this approach, together with deep sequencing, site-directed mutagenesis, antibody adsorptions and virus-entry assays, we found evidence in three subjects of neutralizing antibody (Nab) responses as early as 2 weeks post-seroconversion, with Nab titers as low as 1∶20 to 1∶50 (IC(50)) selecting for virus escape. In each of the subjects, Nabs targeted different regions of the HIV-1 envelope (Env) in a strain-specific, conformationally sensitive manner. In subject CH40, virus escape was first mediated by mutations in the V1 region of the Env, followed by V3. HIV-1 specific monoclonal antibodies from this subject mapped to an immunodominant region at the base of V3 and exhibited neutralizing patterns indistinguishable from polyclonal antibody responses, indicating V1-V3 interactions within the Env trimer. In subject CH77, escape mutations mapped to the V2 region of Env, several of which selected for alterations of glycosylation. And in subject CH58, escape mutations mapped to the Env outer domain. In all three subjects, initial Nab recognition was followed by sequential rounds of virus escape and Nab elicitation, with Nab escape variants exhibiting variable costs to replication fitness. Although delayed in comparison with autologous CD8 T-cell responses, our findings show that Nabs appear earlier in HIV-1 infection than previously recognized, target diverse sites on HIV-1 Env, and impede virus replication at surprisingly low titers. The unexpected in vivo sensitivity of early transmitted/founder virus to Nabs raises the possibility that similarly low concentrations of vaccine-induced Nabs could impair virus acquisition in natural HIV-1 transmission, where the risk of infection is low and the number of viruses responsible for transmission and productive clinical infection is typically one.

Figure 1. Sequences and autologous neutralization sensitivities of consensus infectious molecular clones.

A. 6 mo and 6 mo-Nab IMC sequences are aligned to the T/F sequence with red and blue tics indicating non-synonymous changes implicated in CTL and Nab escape, respectively. Green tics denote synonymous changes and aqua tics changes in non-coding regions. B. Neutralization of IMCs by autologous 6 month plasma is reported as mean (+/− SD) reciprocal plasma dilutions (IC₅₀). Experiments were conducted in triplicate and repeated three times.

Figure 2. Highlighter analysis and env sequence alignments of putative Nab epitopes in subject CH40.

A. Highlighter plot traces acquired mutations in gp160 env against the T/F sequence at top. Nucleotide differences from the T/F sequence are indicated (red: non-synonymous, green: synonymous) according to days post-seroconversion. CTL epitopes previously confirmed in T cell assays, are indicated by blue triangles. Mutations responsible for Nab escape are highlighted in yellow. B. Amino acid alignments of the V1 and V3 regions (HXB2 numbering). The two amino acid span interrogated by PASS is underlined. SGA sequences were from 6 independent experiments.

Figure 3. Highlighter analysis and env sequence alignments of putative Nab epitopes in subject CH77.

A. Highlighter plots as described in Figure 2. B. Amino acid alignments of segments of V1, V2 and C2 regions as described in Figure 2. SGA sequences were obtained from 7 independent experiments.

Figure 4. Highlighter analysis and env sequence alignments of putative Nab epitopes in subject CH58.

A. Highlighter plots as described in Figure 2. B. Amino acid alignments of segments of the C2, C3, and V4 regions as described in Figure 2. SGA sequences were obtained from 12 independent experiments.

Figure 5. Adsorption of plasma Nabs by autologous T/F Env monomers and trimers.

Plasma from CH40 (A), CH77 (B), and CH58 (C) was incubated with magnetic-bead bound gp120 or tethered gp140 protein corresponding to the T/F sequence from each subject. Beads were removed and neutralization assessed by TZM-bl assay (BSA, bovine serum albumen; b12 broadly neutralizing mAb positive control). Results are the mean +/− SD of three independently performed experiments each performed in duplicate.

Figure 6. Env models.

Models of gp120 molecules from (A) CH40, (B) CH77, and (C) CH58 are shown as white surface projections, with the V3 and V4 loop regions colored in blue and cyan. Models of the V1/V2 regions of CH40 and CH77 are shown as yellow ribbons with parts of the variable V1 and V2 loops shown as dotted lines. Potential N-linked glycans are modeled as grey spheres. Nab escape mutations are shown in red (HXB2 numbering), with mutations removing or adding a potential N-linked glycosylation site marked with an asterisk. (D) Schematic of putative V1/V2 (yellow dot), V3 (blue dot), and V4 (cyan dot) locations on the Env trimer.

Figure 7. Sequence entropy and viral dynamics at the CH40 V1 Nab epitope.

(A) 454 sequence entropies inside (grey) and outside (white) of a 12-amino acid Nab epitope region in V1. Bar length indicates Shannon entropy computed bidirectionally (a and b) (one-sided Wilcoxon analysis). (B and C) Plasma viral RNA (black asterisks) consist of epitope variants in various abundances as determined by SGA or 454 sequencing. Error bars depict 95% confidence intervals from the binomial distribution.