Back to the future: covalent epitope-based HIV vaccine development

Abstract

Traditional HIV vaccine approaches have proved ineffective because the immunodominant viral epitopes are mutable and the conserved epitopes necessary for infection are not sufficiently immunogenic. The CD4 binding site expressed by the HIV envelope protein of glycoprotein 120 is essential for viral entry into host cells. In this article, we review the B-cell superantigenic character of the CD4 binding site as the cause of its poor immunogenicity. We summarize evidence supporting development of covalent immunization as the first vaccine strategy with the potential to induce an antibody response to a conserved HIV epitope that neutralizes genetically divergent HIV strains.

Figure 1. Exposure of the 421–433 epitope

Surface representation of glycoprotein 120 crystal structure showing the 421–433 epitope in green (PDB 2B4C; after stripping away bound soluble CD4 and Fab X5). Of the total epitope surface area, approximately 85% is fully accessible. The neighboring β15/β2 strands are shown in purple.

Figure 2. Neutralization of genetically diverse primary HIV strains and SHIVSF162P3 by IgA purified from three long-term survivors of HIV infection (19–21 years, LTS patients 2857, 2866 and 2886)

(A) Neutralization potency scatter plots (IC₅₀) of the 3 IgA preparations and the reference anti-CD4 binding site IgG b12. Each point represents an individual virus strain. (B) Phylogenic relationship of the strains neutralized by LTS IgA preparations and the autologous viruses from IgA donors (designated strains 2857. P1, 2866. P4 and 2886. P1). The bootstrap values of each node represent the percentage of 1,000 bootstrap replicates that support the branching order. Only bootstrap values at 80% or higher are shown. Horizontal branch lengths are drawn to scale. Bar indicates 0.02 nucleotide substitution per site. Percent infections caused by strains drawn from individual HIV subtypes is indicated on the right. Data for (A) from [48]. Data for (B) from [2]. IC₅₀: Half maximal inhibitory concentration; LTS: Long-term survivor; SHIV: Simian human immunodeficiency virus.

Figure 3. Covalent vaccine principle

(A) Traditional nonelectrophilic immunogens induce a transient Ab response to the 421–433 CD4 binding site epitope limited mostly to the IgM compartment because of deficient class-switching. (B) The electrophilic phosphonate of the prototype E-vaccine binds BCR nucleophiles (Nu) via the highly energetic covalent reaction, bypassing constraints on B-cell differentiation. This generates memory B cells and plasma cells producing neutralizing Abs. (C) An optional epitope 2 that generates a positive signal by binding the CDRs can be incorporated in the E-vaccine to counteract negative B-cell signaling due to 416–433 epitope binding at the FRs. (D) Electrophile-driven clonal selection of the B celIs results in adaptive strengthening of Ab nucleophilic reactivity, improving the innate catalytic activity of Abs. Specificity is derived from noncovalent epitope–paratope binding. Covalent immune complex 1 is a resonant stable complex prior to expulsion of C-terminal antigen fragment. Covalent immune complex 2 is an acyl-Ab complex. Ag′ and Ag″ are components of the epitope recognized by the Ab. Ag′ Lys-OH is the N-terminal antigen fragment and NH2-Ag″ is the C-terminal antigen fragment. Ab: Antibody; BCR: B-cell receptor; CDR: Complementarity determining region; FR: Framework region; gp120: Glycoprotein 120.

Figure 4. Structural aspects of CD4 binding site 421–433 epitope recognition by antibodies

(A) Split-site model explaining proteolytic Ab epitope specificity. Two different Ab subsites are responsible for the initial noncovalent antigen binding and the subsequent peptide bond hydrolysis process. In the initial immune complex (left), the antigen region not involved in noncovalent Ab binding enjoys conformational flexibility. Consequently, peptide bonds remote from the noncovalent binding site that are in register with the Ab nucleophilic subsite can be hydrolyzed (right). If the antigen contains an electrophilic phosphonate group, it can form a covalent bond with the nucleophile (not shown). The triangle represents a nucleophile, the circle represents a neighboring general base that activates the nucleophile. (B) Surface model of anti-E-glycoprotein 120 Fab YZ23 crystal solved at 2.5 Å resolution (PDB 3CLE). The V_L domain is shown in pink, the V_H domain in cyan. Complementarity-determining region (CDR)L1 and L3 are shown in yellow and orange, respectively, and CDRH1, H2 and H3 are shown in blue, light green and dark green, respectively. For clarity, the CDR cavity (CDR-cavity)and V_H framework region-dominated cavity (framework region-cavity) are fitted, respectively, with the gray-meshed object and white-meshed object. The inter-cavity centroid-to-centroid distance and cavity surface areas are indicated. For (A), see [54,57] for further details; figure taken from [57]. For (B) data are from [69]. Ab: Antibody.

Figure 5. Importance of conformational rigidity in CD4BS mimicry

(A) An immunogen expressing a flexible CD4BS can adopt a conformation complementary to the BCR combining site by the induced-fit mechanism, thereby losing its CD4BS-mimicking conformation and resulting in induction of non-neutralizing Abs. (B) A rigid CD4BS mimetic will induce the correct BCR conformation, which can be affinity-matured further by immunogen-driven selection, resulting in synthesis of neutralizing antibodies. Ab: Antibody; BCR: B-cell receptor; CD4BS: CD4 binding site.

Figure 6. Reliability of antibody neutralization in tissue culture

(A) Variability of scFV JL427 and IgG YZ23 neutralization potency (IC₅₀) in PBMC/native HIV assay (clinical strain subtype C 97ZA009) using the same virus preparation batch and same Ab preparation batches. PBMC host cells were either from the same batch of frozen cells pooled from eight donors or different batches of cells pooled from different sets of eight donors each. Each symbol shows the IC₅₀ extracted from the Ab dose–response curve of individual assays. The dashed line shows the arithmetic mean and the black line shows the geometric mean. (B) scFv JL427 dose–response using PBMCs isolated from four individual donors without pooling of the cells or the PBMC pool from the same four donors. HIV strain: subtype C 97ZA009. (C) IgG YZ23 dose–response using PBMCs isolated from four individual donors without pooling of the cells or the PBMC pool from the same four donors. HIV strain: subtype C 97ZA009. (D) Antibody-neutralizing potency observed in the PBMC/native HIV (clinical virus strains) and TZM-bl/pseudovirion assays. IC₅₀ values extracted from dose–response curves are shown for one subtype B strain and one subtype C strain each for Abs to the 421–433 epitope (scFv JL427, IgG YZ23 and LTS IgA 2857) and the reference IgG b12. The native clinical HIV strains and pseudovirus strains express gp120 with identical sequence. Ab: Antibody; IC₅₀: Half maximal inhibitory concentration; PBMC: Peripheral blood mononuclear cell; scFv: Single-chain variable fragment.