Vaccination induces HIV broadly neutralizing antibody precursors in humans

Abstract

Broadly neutralizing antibodies (bnAbs) can protect against HIV infection but have not been induced by human vaccination. A key barrier to bnAb induction is vaccine priming of rare bnAb-precursor B cells. In a randomized, double-blind, placebo-controlled phase 1 clinical trial, the HIV vaccine-priming candidate eOD-GT8 60mer adjuvanted with AS01_B had a favorable safety profile and induced VRC01-class bnAb precursors in 97% of vaccine recipients with median frequencies reaching 0.1% among immunoglobulin G B cells in blood. bnAb precursors shared properties with bnAbs and gained somatic hypermutation and affinity with the boost. The results establish clinical proof of concept for germline-targeting vaccine priming, support development of boosting regimens to induce bnAbs, and encourage application of the germline-targeting strategy to other targets in HIV and other pathogens.

Fig. 1.. Frequencies of antigen-specific and epitope-specific B cells.

(A) Schedule for immunization and B cell sampling. (B) Frequency of eOD-GT8-specific (GT8⁺⁺) IgG memory B cells (left), IgG GC B cells (middle), and IgD⁻ plasmablasts (right) shown over time for each participant, grouped by vaccine treatment. GT8⁺⁺ indicates binding to two different eOD-GT8 fluorescent probes. (C) Frequency of CD4bs-specific (KO⁻GT8⁺⁺) IgG B cells, displayed as in (B). KO⁻ indicates lack of binding to the eOD-GT8-KO11 probe. (D) Percent of eOD-GT8-specific IgG B cells that are CD4bs-specific (KO⁻), displayed as in (B). Each symbol represents frequency for one participant. Thick lines are median values; boxes indicate 25% and 75% quantiles.

Fig. 2.. Detection of VRC01-class IgG B cells in blood and lymph nodes.

(A) Number of VRC01-class IgG B cells detected over time in each participant. (B) Frequency of VRC01-class IgG B cells as a percent of IgG B cells in each participant. Median post-vaccination frequencies are stated as 1:number of IgG B cells. In (A) and (B), symbols represent participants, and the two placebo participants with pre-existing (week -4) VRC01-class B cells are indicated as a square and a triangle. Thick lines indicate medians, and boxes indicate 25% and 75% quantiles. Medians and quantiles were computed over non-zero values only, because non-responders are accounted for in (C). (C) Positivity of VRC01-class IgG B cell detection, the percent of participants in each group with at least one VRC01-class B cell detected, for each timepoint and sample type. Circles indicate median values, and lines indicate 95% confidence intervals computed using the Wilson score method. (D) Positivity of VRC01-class responses over all timepoints or only after the first or second vaccination.

Fig. 3.. VRC01-class BCR hierarchical clustering and genetic diversity.

(A) Number of clusters versus number of BCR sequences, for all post-vaccination VRC01-class BCR sequences in both vaccine groups. Number of sequences per cluster indicated at bottom. (B) Distributions of distances between BCR sequences in (A), including all versus all, intra-cluster (all versus all within cluster), and inter-centroid (between cluster centroids). (C) Number of clusters involving single donor and timepoint, single donor and multiple timepoints, and multiple donors, separately for low and high dose groups. (D) Histogram of cluster size for low and high dose groups. (E) Number of clusters versus number of VRC01-class BCR sequences, for each participant in the low and high dose groups. (F) VRC01-class BCR polyclonality over time. Each symbol reports the fraction of BCR sequences that cluster as unique clones within a single donor at a single timepoint. Thick lines represent median values; boxes indicate 25% and 75% quantiles; whiskers approximate 10% and 90% quantiles.

Fig. 4.. Frequency of VRC01-class B cells among CD4bs- or eOD-GT8-specific IgG B cells and plasmablasts.

(A) VRC01-class frequencies among CD4bs-specific B cells. (B) VRC01-class frequencies among eOD-GT8-specific B cells. Each symbol represents frequency for one participant. The two placebo participants with pre-existing (week -4) VRC01-class B cells are indicated as a square and a triangle. Thick lines are median values; boxes indicate 25% and 75% quantiles.

Fig. 5.. Amino acid mutation levels in heavy and light chain V genes over time, for VRC01-class and non-VRC01-class BCRs from vaccine recipients.

(A) and (B) VRC01-class (A) and non-VRC01-class (B) BCR V_H percent amino acid (aa) mutation (top) and V_K/L %aa mutation (bottom) for low dose (left) and high dose (right), with symbols representing the median per participant per timepoint; (C) and (D) VRC01-class (C) and non-VRC01-class (D) BCR V_H %aa mutation (top) and V_K/L %aa mutation (bottom) for low dose (left) and high dose (right), with violin plots representing the distribution of all BCRs per group per timepoint. In (A) and (B), thick lines are medians, and box plots show 25% and 75% quantiles, for each dose group at each timepoint. Statistical significance is indicated for all comparisons with pairs of measurements from at least 8 participants (tables S46 and S48). Significance testing was done using Wilcoxon signed-rank test for paired data (two-sided, α = 0.05); ns, not significant; *, P<0.05; **, P<0.01; ***, P<0.001. In (C) and (D), solid lines are medians, and dashed lines show 25% and 75% quantiles, for each dose group at each timepoint.

Fig. 6.. Properties of post-vaccination BCRs shared with VRC01-class bnAbs.

(A) Percent of BCRs using VRC01-class bnAb V_K/L genes, for VRC01-class and VH1–2-using non-VRC01-class BCRs, and for control VH1–2 BCRs from HIV-unexposed individuals from DeKosky et al. (99); VRC01-class bnAb V_K/L are indicated in the color key. (B) Sequence logos for 5-amino acid LCDR3s from VRC01-class BCRs, for bnAbs (top row), low dose (second row) and high dose (third row) groups, and human naive precursors from prior studies (39, 43, 44) (bottom row), distinguishing kappa (left) and lambda (right) LCs. (C) Sequence logos for 5-amino acid LCDR3s from non-VRC01-class BCRs from the low dose (second row) and high dose (third row) groups, and control data human LCs from HIV-unexposed individuals from the Observed Antibody Space (OAS) (92, 93) (bottom row), with VRC01-class bnAbs (top row) shown for reference, distinguishing kappa (left) and lambda (right) LCs. (D) Percent of BCRs using Glu or Gln at LC position 96, for 5-amino acid LCDR3s from VRC01-class BCRs, non-VRC01-class BCRs, and OAS control data LCs (92, 93). (E) Percent of BCRs with LCDR3 matching a VRC01-class bnAb sequence, for 5-amino acid LCDR3s from VRC01-class BCRs, non-VRC01-class BCRs, and OAS control data LCs (92, 93), distinguishing IGK and IGL LCs. (F) Percent of BCRs with Trp_105–3, for VRC01-class and VH1–2-using non-VRC01-class BCRs, and for OAS (92, 93) control data VH1–2 HCs. (G) Number of key VRC01-class residues in VRC01-class HCs, for all timepoints in low dose (left) and high dose (middle) groups, with symbols indicating the 90% quantile per participant per timepoint, and for VRC01-class bnAbs (right, different y-axis scale), with symbols denoting bnAbs. 20 μg, low dose; 100 μg, high dose. Prime, data from weeks 3, 4, and 8; Boost, data from weeks 9, 10, 11, and 16. In (D)-(G), symbols represent individual participants [except the bnAb controls in (G)]; thick lines are medians; boxes show 25% and 75% quantiles; whiskers approximate 10% and 90% quantiles.

Fig. 7.. SPR analysis of BCR affinities for eOD-GT8.

Monovalent K_D values for antibody ligands binding to eOD-GT8 monomer analyte, for VRC01-class and non-VRC01-class BCRs from the low dose group post-vaccination (weeks 4–16) and pre-vaccination (week -4 and week 0 iGL), and for human naive BCRs isolated by prior human B cell sorting studies (21, 43) (week 0, far left). Thick lines are medians; boxes show 25% and 75% quantiles; whiskers approximate 10% and 90% quantiles For median K_D, “−” indicates median ≥ 100 μM. K_D values were generally representative of multiple measurements. For the 778 K_Ds in this figure, 163 (21%) were measured once, 170 (21.9%) were measured twice, and 445 (57.2%) were measured three or more times.

Fig. 8.. SPR analysis of VRC01-class BCR affinities for eOD-GT6 and variants.

(A) Monovalent K_D values for VRC01-class antibody ligands binding to monomeric eOD-GT6 and eOD-GT6 variant analytes (described in fig. S38), for memory BCRs from the low dose group post-vaccination (weeks 4–16) and pre-vaccination (week -4 and week 0 iGL), and for human naive BCRs isolated by prior human B cell sorting studies (21, 43) (week 0, far left). Thick lines are medians; boxes show 25% and 75% quantiles; whiskers approximate 10% and 90% quantiles. For median K_D, “−” indicates median ≥ 100 μM. For the 1321 K_Ds in (A), 1171 (88.6%) were measured once, 48 (3.6%) were measured twice, and 102 (8.7%) were measured three times. (B) K_D for eOD-GT8 versus K_D for eOD-GT6 and variants, with estimated regression lines shown as solid lines and 95% confidence prediction intervals indicated by shading. Regressions were restricted to K_Ds <100 μM. K_Ds ≥100 μM for eOD-GT6 and variants are shown in grey. S, slope of regression line; P, p-value for slope; ns, not significant (p>0.05).