Vaccination with peptide mimetics of the gp41 prehairpin fusion intermediate yields neutralizing antisera against HIV-1 isolates

Abstract

Eliciting a broadly neutralizing polyclonal antibody response against HIV-1 remains a major challenge. One approach to vaccine development is prevention of HIV-1 entry into cells by blocking the fusion of viral and cell membranes. More specifically, our goal is to elicit neutralizing antibodies that target a transient viral entry intermediate (the prehairpin intermediate) formed by the HIV-1 gp41 protein. Because this intermediate is transient, a stable mimetic is required to elicit an immune response. Previously, a series of engineered peptides was used to select a mAb (denoted D5) that binds to the surface of the gp41 prehairpin intermediate, as demonstrated by x-ray crystallographic studies. D5 inhibits the replication of HIV-1 clinical isolates, providing proof-of-principle for this vaccine approach. Here, we describe a series of peptide mimetics of the gp41 prehairpin intermediate designed to permit a systematic analysis of the immune response generated in animals. To improve the chances of detecting weak neutralizing polyclonal responses, two strategies were employed in the initial screening: use of a neutralization-hypersensitive virus and concentration of the IgG fraction from immunized animal sera. This allowed incremental improvements through iterative cycles of design, which led to vaccine candidates capable of generating a polyclonal antibody response, detectable in unfractionated sera, that neutralize tier 1 HIV-1 and simian HIV primary isolates in vitro. Our findings serve as a starting point for the design of more potent immunogens to elicit a broadly neutralizing response against the gp41 prehairpin intermediate.

Fig. 1.

Schematic model of engineered peptide immunogens. IZN helices forming a homotrimeric coiled-coil are represented by cylinders. Each chimeric N-peptide consist of an N-terminal solubilizing and trimerizing coiled-coil domain (IZ, blue) fused to a sequence from gp41 NHR (gray), designated N17 (residues 565–581 of HXB2) or N36 (residues 546–581 of HXB2). (CCIZN)₃ helices are covalently stabilized at the N termini of each peptide chain by three disulfide bonds between cysteine pairs located outside the coiled-coil domain. Only one of the possible combinations of the three disulfides is drawn. (Left Inset) “IP” scaffolds were shielded from immune recognition by LMWPEG chains, attached by amide linkage to the ε-amino group of lysine residues and represented by light-blue boxes in the main diagram. (Right Inset) Alignment of the sequences of each peptide or fragment to the NHR sequence (residues 540–588) and corresponding positions ”a” and “d” of the heptad repeats of the coiled-coil. HIV-1 residues in the hydrophobic pocket are shown in red, non-HIV-1 and scaffold residues are shown in blue, and lysine residues derivatized with LMWPEG are shown as bold.

Fig. 2.

Neutralization of HIV-1 infection with purified IgG. The IgG fraction from sera of individual guinea pigs (A and B) or rabbits (C) vaccinated with NHR immunogens was purified and tested in the p4-2/R5 single-cycle infectivity assay for its ability to neutralize either D5 hypersensitive HXB2-V570A virus (A and C) or parental HXB2 virus (B). The IC₅₀ is plotted on the y axis for each sample and grouped by NHR immunogen IZN17 (red ○), (CCIZN17)₃ (orange ▵), (CCIPN17)₃ (pink ▿), IZN36 (blue ○), (CCIZN36)₃ (green ▵), and (CCIPN36)₃ (turquoise ▿). All preimmune IC₅₀s were >3,000 μg/mL, with the exception of a single sample [1,800 μg/mL vs. HXB2-V570A for one guinea pig in the (CCIPN36)₃ group]. All rabbit sera had IC₅₀s >3,000 μg/mL vs. HXB2 virus. For guinea pigs, comparing the group of all N17 peptides vs. the group of all N36 peptides, N17 and N36 were highly statistically different (P < 0.001, two-tailed t test on ranked data) for HXB2-V570A but were not statistically significant (P = 0.08) for HXB2. For HXB2-V570A in guinea pigs, (CCIZN36)₃ and (CCIPN36)₃ were superior to all N17 immunogens (maximum P < 0.017, Dunnett's test with control on ranked neutralization values).

Fig. 3.

Neutralization potency correlates with D5-like antibody response. Neutralization IC₅₀ titers for purified IgG determined in the p4-2/R5 assay against HXB2-V570A from vaccinated guinea pig (A) or rabbit (B) antisera were plotted as a function of the antibody concentration that competes for the D5 epitope as determined by DCBA (scheme is shown in Fig. S2). Symbols for individual immunogens correspond to those in Fig. 2. Each data set is rank-correlated (P < 0.0001) with Spearman's ρ = 0.76 (A) and ρ = 0.85 (B).

Fig. 4.

Neutralization breadth and correlations to HXB2-V570A in unfractionated sera. Antisera from guinea pigs immunized with either (CCIZN36)₃ (identification numbers 1–8), or (CCIPN36)₃ (identification numbers 9–16) were tested in the p4-2/R5 neutralization assay against a small panel of HIV-1/SHIV viral isolates. (A) Reciprocal of IC₅₀ is plotted for individual animals. Note that sera with high neutralization of HXB2-V570A tend to higher neutralization of other viral isolates, an indication of broad cross-reactivity. (B) These data are presented as individual rank-correlation plots for each viral isolate relative to HXB2-V570A, in descending order of correlation. Rank correlations (Spearman's ρ) title each plot; all correlations were highly significant (P < 10⁻⁴ for all except P = 0.005 for 89.6 vs. HXB2-V570A). (CCIZN36)₃ (▴) and (CCIPN36)₃ (▾).