In vivo emergence of HIV-1 highly sensitive to neutralizing antibodies

Abstract

Background: The rapid and continual viral escape from neutralizing antibodies is well documented in HIV-1 infection. Here we report in vivo emergence of viruses with heightened sensitivity to neutralizing antibodies, sometimes paralleling the development of neutralization escape.

Methodology/principal findings: Sequential viral envs were amplified from seven HIV-1 infected men monitored from seroconversion up to 5 years after infection. Env-recombinant infectious molecular clones were generated and tested for coreceptor use, macrophage tropism and neutralization sensitivity to homologous and heterologous serum, soluble CD4 and monoclonal antibodies IgG1b12, 2G12 and 17b. We found that HIV-1 evolves sensitivity to contemporaneous neutralizing antibodies during infection. Neutralization sensitive viruses grow out even when potent autologous neutralizing antibodies are present in patient serum. Increased sensitivity to neutralization was associated with susceptibility of the CD4 binding site or epitopes induced after CD4 binding, and mediated by complex envelope determinants including V3 and V4 residues. The development of neutralization sensitive viruses occurred without clinical progression, coreceptor switch or change in tropism for primary macrophages.

Conclusions: We propose that an interplay of selective forces for greater virus replication efficiency without the need to resist neutralizing antibodies in a compartment protected from immune surveillance may explain the temporal course described here for the in vivo emergence of HIV-1 isolates with high sensitivity to neutralizing antibodies.

Figure 1. Susceptibility of sequential Envs to neutralization by heterologous sera.

The neutralization phenotype of sequential Envs derived from seven HIV-1 infected patients (MM1, 2, 4, 8, 23, 27 and 28), at the indicated days following onset of PHI symptoms, was assessed against three heterologous sera (QC1, 2 and 6) with IC₉₀ titres (reciprocal serum end-point dilutions) being reported. Five reference strains were also assayed for comparative purposes. Where less than 90% reduction of infection was observed at the highest serum concentration tested (1:10 dilution) the data is plotted as 1. All patients were infected with neutralization resistant viruses, but highly neutralization sensitive variants (Tier 1 like) emerged during chronic infection in at least three individuals (MM4, 8 and 23). The autologous Nab response to the underlined clones is shown in Fig. 2.

Figure 2. Susceptibility of patient Envs to neutralization by sequential autologous sera.

All Envs cloned from patients MM4, MM8 and MM23 (i.e. the patients in which we identified neutralization sensitive virus) were assayed against sequential autologous sera. The IC₉₀ titres (reciprocal serum end-point dilutions) are reported for one representative clone from each time-point (underlined in Fig. 1) and viruses that are highly susceptible to neutralization (by heterologous sera) are labelled accordingly (i.e. 493-S, 608-S, 316-S and 1065-S). Due to lack of serum some clones could not be tested against all serum samples. These are labelled accordingly (nt, not tested) and include: MM4 day 493 and 833 viruses which were not tested against day 17 and 108 sera, and MM23 day 1534 viruses which only were assayed against sera from day 316 and 1534. Where less than 90% reduction of infection was observed at the highest serum concentration tested (1∶10 dilution) the data is plotted as 1.

Figure 3. Phylogenetic relationship between cloned envs.

Maximum-likelihood tree representing the phylogenetic relationship between 56 HIV-1 partial (1937 bp) env gene sequences from seven HIV-infected patients from the UK and 200 HIV-1 subtype B env gene sequences extracted from the LANL HIV database. The tree was reconstructed according to the GTR+I+G model of nucleotide substitution. Branch lengths are expressed as the number of nucleotide substitution per sites, with branches leading to the clones generated herein indicated in bold. Percent bootstrap support values above 90%, or of 100%, are indicated by one or two asterisks, respectively, on the corresponding branches. The envs are labelled with patient ID followed by isolation day (e.g. MM1-28, env from MM1 day 28). Clones from patient MM23 separates into two distinct clusters (marked with stars), indicating that the patient was infected with two different clade B strains.

Figure 4. MM23 strain specific nested PCR.

Strain specific primers for nested PCR were designed to amplify ∼500 bp segments of the env gene of MM23′s viruses. The PCR conditions were optimised on plasmids encoding previously cloned envs and then applied to PBMC proviral DNA (Fig. A), or plasma viral RNA (Fig. B), from indicated days after onset of PHI.symptoms. (A) Strain A is detectable in PBMC from all time-points assayed (expected size of PCR product 577 bp), whereas strain B is only detectable from day 204 onwards (expected size of PCR product 524 bp). (B) Strain A is detectable in plasma samples from all time-points assayed, and likewise is strain B. The marker (M) is the BenchTop 100 bp DNA ladder (Promega, UK), which contains a 500 bp band of triple intensity. Positive controls (+) were plasmid pHXB2env23.2.E (for strain A) and pHXB2env23.8.12 (for strain B). As a negative control (–) proviral DNA or viral RNA was replaced with dH₂O. Footnotes: ^a Plasma viral load (VL) determined using Chrion 3.0 (Emeryville, Cal., USA); ^b nd, not determined.

Figure 5. Neutralization phenotype of SGA derived gp160 Envs.

The neutralization phenotype of SGA derived full-length Envs was assessed against heterologous sera QC1 and 2 alongside neutralization sensitive HXB2-gp120 chimeras. Top panels display data for Envs from MM4 day 493; SGA clone 4.10_SGA9 was almost as neutralization sensitive as the original HXB2-gp120 chimera 4.10.3 whereas clone 4.10_SGA29 displayed a neutralization resistant phenotype (IC₉₀s<20). Lower panels display data for Envs from MM8 day 608; one out of three SGA clones tested (8.8_SGA11) was hyper sensitive to neutralization. The graphs display data from one representative titration, with error bars representing the standard deviations between replicates.

Figure 6. Molecular determinants of heightened neutralization sensitivity.

To map molecular determinants of the neutralization sensitive phenotype sections of the env gene was swapped between neutralization sensitive (black boxes) and neutralization resistant (white boxes) gp120 clones from MM4 day 493 (A) and MM4 day 608 (B). This was done by exploiting conserved Bgl II or PpuM I restriction enzyme sites (indicated in the figure) in combination with restriction sites incorporated in the primers used for gp120 cloning (see Figure S1 for details). Further mapping was done by SDM, altering amino acid residues indicated with asterisks. The phenotype of the chimeras and the mutants was assessed against heterologous sera (QC1, 2 and 6) with IC₉₀ titres (reciprocal serum end-point dilutions) being reported. (A) The neutralization sensitive phenotype of MM4 clone 4.10.1 was dependent on an unusual arginine (R) residue in the stem of V3 (4.10.1-R328Q). However, introduction of this residue was not sufficient to infer sensitivity on the resistant MM4 clone 4.10.7 (4.10.7-Q328R). (B) In MM8, neutralization sensitivity determinants were present on both sides of the PpuM I site (chimera 1 and 2) and included a threonine to isoleucine change at residue 399 in V4 (compare 8.8.4 and 8.8.4-T399I).