Tailored Immunogens Direct Affinity Maturation toward HIV Neutralizing Antibodies

Abstract

Induction of broadly neutralizing antibodies (bnAbs) is a primary goal of HIV vaccine development. VRC01-class bnAbs are important vaccine leads because their precursor B cells targeted by an engineered priming immunogen are relatively common among humans. This priming immunogen has demonstrated the ability to initiate a bnAb response in animal models, but recall and maturation toward bnAb development has not been shown. Here, we report the development of boosting immunogens designed to guide the genetic and functional maturation of previously primed VRC01-class precursors. Boosting a transgenic mouse model expressing germline VRC01 heavy chains produced broad neutralization of near-native isolates (N276A) and weak neutralization of fully native HIV. Functional and genetic characteristics indicate that the boosted mAbs are consistent with partially mature VRC01-class antibodies and place them on a maturation trajectory that leads toward mature VRC01-class bnAbs. The results show how reductionist sequential immunization can guide maturation of HIV bnAb responses.

Graphical abstract

Figure 1

Design of a Sequential Immunization Strategy Employing BG505 GT3 (A) Design of boosting immunogens (BG505 GT3 and SOSIP N276D) presenting a CD4bs epitope that is increasingly more native-like than the priming immunogen eOD-GT8. (B) BG505 core-GT3 and SOSIP-GT3 were also designed to minimize off-target responses. (C) Conservation of T-help between sequential immunogens. When using BG505 core-GT3 NP, the nanoparticle base is shared between the prime and first boost, while BG505 core gp120 is shared between the first and second boost. When using BG505 SOSIP-GT3, a PADRE peptide is conserved between the prime and first boost. (D) Affinity of germline-reverted (GLrev) and mature VRC01-class antibodies for BG505 core-GT3. See also Figures S1, S2, and S3 and Tables S1 and S2.

Figure 2

Immunization of VRC01 gH Mice (A) Schedule for priming and boosting VRC01 gH mice. Each immunization group consisted of five animals, two of which were sacrificed following the first boost (BG505 GT3 SOSIP or BG505 GT3 NP), with the remaining three animals receiving the entire immunization schedule before being sacrificed. (B) ELISA binding of VRC01 gH mouse serum following boost 3. The immunogen used for boost 1 (B1) and the adjuvant used for all three boosts is shown for each plot. Binding to r1-core-N276D (filled circles) and r1-core-KO (filled squares) is represented as area under the curve, and the difference ± SEM between binding to r1-core-N276D and r1-core-KO is shown beneath each plot. Each point represents a single animal that received the entire immunization schedule, of which there are three per immunization group. The value of each point is the mean of three technical replicates. The error bars represent mean ± SD. (C) Frequency of epitope-specific IgG memory B cells in VRC01-gH mice 14 days following boost 1, as measured by binding to GT3 and lack of binding to GT3-KO. Each point represents a single animal that was sacrificed following the first boost, of which there are two per group. The mean for each group is indicated by a single horizontal bar. (D) Frequency of epitope-specific IgG memory B cells after the full immunization regimen, measured by binding to r1-core-N276D and lack of binding to r1-core-KO (for animals boosted twice with GT3 and SOSIP N276D) or binding to GT8 and lack of binding to GT8-KO (for animals that were boosted twice with GT8). Each point represents a single animal that received the entire immunization schedule, of which there were three per group. The error bars represent mean ± SD. See also Figures S4 and S5.

Figure 3

Genetic Maturation of Immunogen-Induced Antibodies (A) Frequency of VRC01-class (VH1-2 heavy chain and a 5AA long LCDR3) antibodies among all paired heavy-light sequences recovered per mouse, for different mice in each immunization group. The numbers below each immunization group indicate the number of mice from which paired sequences were recovered. Each bubble represents antibody sequences from a single mouse and the area of the bubble is proportional to the total number of mAb sequences recovered. (B) Frequency of heavy-chain nucleotide mutations for all sequences from all animals in each immunization group. (C) Same as (B) but for amino acid mutations. (D) Two-dimensional histograms of the number of VRC01-class-like amino mutations (defined as those shared with VRC01, PGV04, PGV20, VRC-CH31, 3BNC60, and 12A12) (Jardine et al., 2015) versus the total number of amino acid mutations, for VH genes in all VRC01-class pairs recovered from each immunization group. The frequency of VRC01-class mutations expected by random SHM (black line) is shown on each plot, as well as the 95% confidence interval (gray shading). (E) Locations of VRC01-class mutations (blue) and other mutations (black) within heavy chains of all 130 VRC01-class pairs recovered from 15 animals receiving the complete immunization schedule (eOD-GT8 60-mer, BG505 GT3, and SOSIP N276D 2x). Positions at which the antibody sequence was identical to GLRev VRC01 are colored light gray. (F) Locations of VRC01-class mutations (blue) and other mutations (black) from 150 sequences randomly selected from a pool of 2,000 artificial antibody sequences that represent random SHM activity layered on the VRC01 gH sequence. Synthetic mutations were only generated in the variable gene region; hence, the lack of synthetic mutations in CDR3 and FR4 (see STAR Methods for a more detailed explanation). (G) Frequency of mouse light chain variable genes from paired VRC01-class sequences recovered from different immunization groups. Light-chain V-genes with germline-encoded short (≤6AA) LCDR1s are indicated with asterisks. See also Table S3. (H) Frequency of light chains encoding a short (≤6AA) LCDR1 for each mouse in different immunization groups. The numbers below each group indicate the number of mice from which paired sequences were recovered. Each bubble represents a single animal, and the area of the bubble is proportional to the total number of VRC01-like pairs. (I) LCDR3 sequence logos from each paired VRC01-class mAb in each immunization group compared to LCDR3 of VRC01. Increasing convergence on critical LCDR3 residues found in VRC01 is highlighted.

Figure 4

Serology of Immunized VRC01 gH Mice (A) Neutralization curves of total IgG purified from sera of immunized mice. Twelve mice were primed with eOD-GT8 60-mer and boosted with either core-GT3 NP or GT3 SOSIP (adjuvanted with either Ribi or PBS) followed by an additional two boosts with BG505 SOSIP N276D (with either Ribi or PBS). The leftmost column describes the immunization regimen of the three plots immediately to the right. Each neutralization plot shows neutralization activity against an 8-virus panel of near-native (N276A) virus isolates. Neutralization values with error bars are mean ± SD for two measurements. Total IgG concentrations are shown in μg/ml. Best fit curves were calculated with GraphPad Prism. (B) Sixteen additional VRC01 gH mice were primed with eOD-GT8 60-mer and boosted with BG505 core-GT3 NP, followed by two boosts with either BG505 SOSIP N276D in Ribi (top) or ABC SOSIP N276D cocktail in Ribi (bottom). Purified serum IgG was screened against a 7-virus panel of near-native (N276A) virus isolates, and ID₅₀ values are plotted for each mouse, as reciprocal serum titers.

Figure 5

Neutralization by mAbs from Immunized Mice (A) Neutralization breadth and potency of mAbs isolated from several mice receiving the entire immunization program and screened on a 25-virus cross-clade panel of near-native (N276A) isolates. (B) Neutralization activity of two mAbs isolated from mouse 286. Nem10 (circles) and Nem11 (squares) neutralized wild-type 191084 B7-19 virus grown in 293T cells (black) or 293S cells (blue), as well as N276A virus grown in 293T cells (red), with potency highest against N276A virus. In the bottom three panels, neutralization curves for mature VRC01, F105, and b12 are shown for comparison. Neutralization values with error bars are mean ± SD for two measurements. Best fit curves were calculated with GraphPad Prism. See also Figure S6 and Table S4.

Figure S1

Development of Boosting Immunogen GT3.1, Related to Figure 1 Sequence alignment of BG505 SOSIP.664, BG505 GT3.1 SOSIP, BG505 core, BG505 GT3.1 core, GT6, and GT8. Engineered mutations in GT3.1 are highlighted in green.

Figure S2

Sequence and Characterization of BG505 Core-GT3.1 Nanoparticles, Related to Figure 1 (A) Sequences of the two genes co-transfected to produce BG505 core-GT3 nanoparticles (NPs). Co-transfection is 80% BG505 core GT3 60mer and 20% LumSyn_1hqk_naked. In the BG505 core-GT3 60-mer gene, BG505 core-GT3 (blue) is fused to the lumazine synthase gene (red) via a flexible linker (cyan) as shown. (B) SECMALS analysis of BG505 core-GT3 NP. (C) ELISA binding of mature VRC01 (black) or GLRev VRC01 (red) to BG505 core-GT3 NP.

Figure S3

Characterization and Structural Analysis of BG505 GT3.1 SOSIP and ABC SOSIP Trimers, Related to Figure 1 (A) Negative-stain 2D class average images of the GT3.1 SOSIP, CD4bs-B SOSIP and CD4bs-C SOSIP are shown. 37% of GT3.1 SOSIP trimers were classified as closed native-like trimers, 63% as open native-like trimers and 0% as non-native-like trimers. (B) DSC thermogram of GT3.1 SOSIP. (C) ELISA analysis of mature and GLRev VRC01-class Ab binding to BG505 SOSIP and BG505 GT3.1 SOSIP. (D) SPR sensograms of BG505 SOSIP and BG505 GT3.1 SOSIP as analytes and non-nAbs IgGs as ligands. Maximum analyte concentrations were 990 nM and 1.3 uM for BG505 SOSIP and BG505 GT3 SOSIP, respectively, and lower concentrations teseted were 4-fold dilutions.

Figure S4

Sequence and Antigenicity of the r1-Core Resurfaced Core gp120, Related to Figure 2 (A) Sequence alignment of r1-core with gp120core-e-2CC HxB2. The r1-core is derived from the core-e-2CC HxB2 protein that we previously described (Jardine et al., 2015) and is also known as gp120core-e-2CC_HxB2_r1. A total of 78 surface positions have been modified, out of 358 total residues. Mutations are classified in the third row of the alignment, with “”: denoting a conservative mutation between amino acids with similar physicochemical properties; ” ” denoting a non-conservative mutation; and “.” denoting a semi-conservative mutation. (B) Dissociation constants for core-e-2CC HxB2, r1-core, and RSC3 (Wu et al., 2010) with VRC01-class bnAbs. (C) ELISA binding of mature VRC01 to core-N276D (core-e-2CC HxB2 with the N276D mutation; inverted triangles), r1-core-N276D (circles), r1-core (squares) and r1-core-KO (triangles). Starting concentration of mature VRC01 was 2 μg/ml.

Figure S5

Serum Binding to r1-Core and r1-Core-KO, Related to Figure 2 Serum binding to r1-core (red, also referred to as HXcore) and r1-core-KO (blue, also referred to as HXcore KO). Representative binding curves from twelve mice are shown. Best fit curves for binding to r1-core and r1-core-KO were used to compute the AUC shown in Figure 2B.

Figure S6

Broad Neutralization of Near-Native Viruses by VRC01-gH mAbs, Related to Figure 5 Neutralization of a 25-virus panel consisting of near-native (N276A) viruses, with IC50 values shown in μg/ml. Antibodies are grouped by the mouse from which they were isolated. Blank entries indicate lack of neutralization (> 10 μg/ml), and antibody/virus combinations that were not tested are noted (ND). Neutralization by mature VRC01 is shown for comparison. Tier phenotypes for the parent viruses lacking the N276A mutation were taken from (Binley et al., 2004, Seaman et al., 2010, Sellhorn et al., 2012, Hraber et al., 2014).