Antibodies from Rabbits Immunized with HIV-1 Clade B SOSIP Trimers Can Neutralize Multiple Clade B Viruses by Destabilizing the Envelope Glycoprotein

Abstract

The high viral diversity of HIV-1 is a formidable hurdle for the development of an HIV-1 vaccine. Elicitation of broadly neutralizing antibodies (bNAbs) would offer a solution, but so far immunization strategies have failed to efficiently elicit bNAbs. To overcome these obstacles, it is important to understand the immune responses elicited by current HIV-1 envelope glycoprotein (Env) immunogens. To gain more insight, we characterized monoclonal antibodies (MAbs) isolated from rabbits immunized with Env SOSIP trimers based on the clade B isolate AMC008. Four rabbits that were immunized three times with AMC008 trimer developed robust autologous and sporadic low-titer heterologous neutralizing responses. Seventeen AMC008 trimer-reactive MAbs were isolated using antigen-specific single B-cell sorting. Four of these MAbs neutralized the autologous AMC008 virus and several other clade B viruses. When visualized by electron microscopy, the complex of the neutralizing MAbs with the AMC008 trimer showed binding to the gp41 subunit with unusual approach angles, and we observed that their neutralization ability depended on their capacity to induce Env trimer dissociation. Thus, AMC008 SOSIP trimer immunization induced clade B-neutralizing MAbs with unusual approach angles with neutralizing effects that involve trimer destabilization. Optimizing these responses might provide an avenue to the induction of trimer-dissociating bNAbs. IMPORTANCE Roughly 32 million people have died as a consequence of HIV-1 infection since the start of the epidemic, and 1.7 million people still get infected with HIV-1 annually. Therefore, a vaccine to prevent HIV-1 infection is urgently needed. Current HIV-1 immunogens are not able to elicit the broad immune responses needed to provide protection against the large variation of HIV-1 strains circulating globally. A better understanding of the humoral immune responses elicited by immunization with state-of-the-art HIV-1 immunogens should facilitate the design of improved HIV-1 vaccine candidates. We identified antibodies with the ability to neutralize multiple HIV-1 viruses by destabilization of the envelope glycoprotein. Their weak but consistent cross-neutralization ability indicates the potential of this epitope to elicit broad responses. The trimer-destabilizing effect of the neutralizing MAbs, combined with detailed characterization of the neutralization epitope, can be used to shape the next generation of HIV-1 immunogens to elicit improved humoral responses after vaccination.

Keywords: AMC008 SOSIP; HIV-1; approach angle; human immunodeficiency virus; monoclonal antibodies; trimer destabilization; vaccine.

FIG 1

Characteristics of monoclonal antibodies (MAbs) isolated from AMC008 SOSIP trimer-immunized rabbits. (A) Immunization and sample collection scheme (left) and autologous and heterologous serum neutralization titers at week 22 (right) for four AMC008 SOSIP trimer-immunized rabbits. Serum neutralization 50% infective dose (ID₅₀) is shown for each animal. (B) Absolute number of MAbs isolated per animal. (C) Individual CDR3 lengths and CDR amino acid junction sequence of heavy and light chains of all isolated MAbs. (D) Individual CDR3 amino acid lengths of the isolated MAbs. (E) Autologous neutralization ability of the isolated MAbs. Neutralization 50% inhibitory concentration (IC₅₀) values are shown for each of the four neutralizing antibodies (NAbs) in the accompanying table.

FIG 2

Competitive assays of NAbs with bNAbs and non-NAbs to specify an epitope. (A) Percent residual binding of NAbs to AMC008 SOSIP trimers in the presence of a competing NAb. The percent residual binding was calculated as follows: (shift in nm at 600 s of 2nd MAb binding × 100)/(shift in nm of 2nd MAb in the absence of 1st MAb binding). (B) Competition enzyme-limited immunosorbent assay (ELISA) and surface plasmon resonance (SPR) data of all NAbs with human broadly neutralizing antibodies (bNAbs). The percent residual binding was calculated as follows. For ELISA, (average optical density at 450 nm [OD₄₅₀] of a triplo in the presence of the 2nd MAb × 100)/(average OD₄₅₀ of a triplo in the absence of the 1st MAb binding); for SPR: [(response difference at 200 s for the second Ab)/(response difference at 200 s for the same, second Ab when injected as a single Ab in a separate cycle)] × 100 (%). Significant results are highlighted by asterisks (**, P < 0.005; ***, P < 0.0001). SPR was only performed for the 05A family of NAbs. (C) SPR binding curves of AMC008 SOSIP trimer binding competition between NAbs and the bNAb ACS2020 that shows the influence of assay directionality. Ab binding is recorded in real time; the x axis indicates the time in seconds. The y axis shows the response (response units [RU]), proportional to the mass bound. Dissociation starts at 0 s. The top three graphs show binding of ACS202 in the presence of competitors 05A1 to 05A3 (dark red, red, and purple lines), compared to ACS202 binding in the absence of these NAbs (blue lines). The lower three graphs show a reverse assay setup, showing binding ability of NAbs 05A1 to 05A3 in the presence of competitor ACS202 (blue lines) and in the absence of ACS202 (dark red, red, and purple lines). (D) Competition results of the four rabbit NAbs with the isolated non-NAbs, measured by BLI. Percent residual binding was calculated as stated for Fig. 2A.

FIG 3

Epitope mapping of NAbs to the gp41 subunit of Env. (A) Negative-stain electron microscopy (NS-EM) 3D reconstructions of the NAbs Fab fragments in complex with the AMC008 SOSIP trimer. The AMC008 SOSIP trimer structure is modeled as the ribboned density. The NAbs are shown as white densities. Side and bottom views are depicted. (B) Alignment of viral sequences tested for neutralization and binding by 05A1 to 05A3 and 07A1. AMC008, AMC009, and SHIV162p3 could be neutralized by the NAbs, whereas the viruses below the line were not neutralized by the NAbs. HXB2 amino acid numbering is indicated on top, and lines indicate the NAb binding sites predicted by NS-EM imaging. Red boxes show which amino acid residues were mutated in the predicted binding sites to specify the epitope, namely, D620N, N624D, and E662A. (C) Model of the glycan shield present on the AMC008 SOSIP trimer with a strain-specific glycan hole indicated in cyan and yellow due to the absence of PNGS 230 and 234. Image was created with the glycan shield mapping tool on the Los Alamos database (57). (D) Neutralization ability of NAbs to multiple viral variants. Fold decrease in IC₅₀ values is plotted for each of the NAbs and the VRC01 control, each represented by different colors. The dotted lines indicate a 3-fold threshold compared to AMC008 wild-type neutralization IC₅₀ values. (E) Neutralization ability of the NAbs for various AMC008 mutants. IC₅₀ values are indicated in μg/ml. The bNAb VRC01 is taken along as a positive control.

FIG 4

Epitope mapping of the nonneutralizing antibodies. (A) Binding ability of all isolated MAbs to various AMC008 mutants and linear V3 peptide. Fold change in area under the curve is displayed relative to AMC008 binding. Binding to the V3 peptide is indicated as “yes” or “no.” (B) NAbs 05A1 to 05A3 (purple and pink) and 07D2 (gray) were complexed with the AMC008 SOSIP Env and analyzed by NS-EM with the AMC008 SOSIP trimer to show differences in their binding angle.

FIG 5

Analysis of neutralizing and nonneutralizing antibody characteristics that might influence neutralization. (A) Binding kinetics of NAbs and non-NAbs to the autologous AMC008 SOSIP and heterologous AMC016 SOSIP Env trimers. K_D, on-rate (K_a), and off-rate (K_d) constants are shown. NAbs were tested for binding to AMC008 and AMC016 SOSIP Env trimers. Non-NAbs were only tested for binding to the autologous AMC008 SOSIP trimer. Results were fitted to a 1:1 kinetics model for analysis. (B) Binding to full-length surface-expressed AMC008 SOSIP gp160 Env by MAbs. Maximum mean fluorescent intensity (MFI) is plotted for each MAb tested. (Left) Binding and expression controls (2G12 and PGT145, respectively). (Middle) Binding of the NAbs to full-length AMC008 gp160 Env trimers. (Right) Binding of a selection of non-NAbs to full-length AMC008 gp160 Env trimers.

FIG 6

NAbs destabilize the trimer as a mechanism of neutralization. (A) Negative-stain electron microscopy images of AMC008 SOSIP trimers incubated overnight with the MAb 05A3. The majority of the images displayed MAb binding with a stoichiometry of 2 MAbs per trimer (middle). MAbs binding the monomeric form of the trimer were also observed (right). (B) Preincubation neutralization experiment to determine the in vivo destabilization ability of the isolated Abs. A neutralization assay was performed in which virus and Ab mix were incubated for different time periods, up to 24 h. The fold change in IC₅₀ is shown for each of the Abs. The autologous AMC008 virus and heterologous SHIV162p3 viruses were tested. 3BC315 and PGT126 were tested as positive-control and negative-control Abs, respectively. Increased neutralization potency after the 24-h incubation step is seen for most of the NAbs and for 3BC315, but not for 06A1 and PGT126. NN, nonneutralizing.