Passive Transfer of Vaccine-Elicited Antibodies Protects against SIV in Rhesus Macaques

Abstract

Several HIV-1 and SIV vaccine candidates have shown partial protection against viral challenges in rhesus macaques. However, the protective efficacy of vaccine-elicited polyclonal antibodies has not previously been demonstrated in adoptive transfer studies in nonhuman primates. In this study, we show that passive transfer of purified antibodies from vaccinated macaques can protect naive animals against SIVmac251 challenges. We vaccinated 30 rhesus macaques with Ad26-SIV Env/Gag/Pol and SIV Env gp140 protein vaccines and assessed the induction of antibody responses and a putative protective signature. This signature included multiple antibody functions and correlated with upregulation of interferon pathways in vaccinated animals. Adoptive transfer of purified immunoglobulin G (IgG) from the vaccinated animals with the most robust protective signatures provided partial protection against SIVmac251 challenges in naive recipient rhesus macaques. These data demonstrate the protective efficacy of purified vaccine-elicited antiviral antibodies in this model, even in the absence of virus neutralization.

Keywords: ADCC; Ad26; HIV; SIV; adoptive transfer; antibodies; systems biology; transcriptomics; vaccines.

Figure 1.. Common protective signature for SIV and SHIV.

(A) Immunologic features correlated with protective efficacy in two previously published SIVmac251 and SHIV-SF162P3 challenge studies (Barouch et al., 2015; Barouch et al., 2018). Features most strongly correlated with protection in both studies are ranked from left to right, and the 12 features most strongly correlated with protection are shown to the left of the vertical line. Red to blue shading highlights a positive or negative correlations, respectively, with protective efficacy. The heat map shows the selected 12 immune features across all animals ranked by the number of challenges required for infection in the (B) SIVmac251 and (C) SHIV-SF162P3 studies. Black to white shading on the right of each heatmap depicts the number of challenges required for infection for each animal. Yellow to blue shading represents the magnitude (positive to negative) of each response in the animals. Receiver operative characteristic (ROC) curves showing predictive and cross-predictive capacity of the minimal correlates within study (red) or across study (blue) are shown in the (D) SIVmac251 and (E) SHIV-SF162P3 studies. To estimate the statistical significance of ROC curve generated by the model, we employed two types of permutation tests: shuffling the outcome labels across the samples and through the use of a random dataset. Each dataset was then used to test the likelihood of obtaining a model prediction accuracy by chance. Each permutation test performed 10⁴ times to generate an empirical null distribution and an exact p-value. See also Figs. S1, S2.

Figure 2.. Antibody responses following Ad26/Env vaccination.

(A) ELISA titers to SIVmac32H Env gp140 following vaccination. (B) Neutralizing antibody responses to primary and T cell line-adapted clones of SIVmac251 and SIVsmE660 at week 54. (C) Functional antibody responses at week 54, including antibody-dependent NK cell activation (CD107a, IFN-γ, MIP-1β), antibody-dependent complement deposition (ADCD), antibody-dependent cellular phagocytosis (ADCP), and antibody-dependent neutrophil phagocytosis (ADNP). The dashed lines represent limits of quantitation of the assays. (D) Heterogeneity of polyfunctional antibody responses in individual animals reflected in pie charts. Pie charts represent the level of vaccine-induced NK cell degranulation (CD107a, orange), vaccine-induced NK cell cytokine secretion (IFN-g, black), vaccine induced NK cell chemokine secretion (MIP-1b, green), antibody dependent cellular phagocytosis (ADCP, blue), antibody dependent neutrophil phagocytosis (ADNP, reds), and antibody dependent complement deposition (ADCD). Data were normalized and individuals over the median were scored as 1 and individuals below the median were scored as 0 to create the pie charts. See also Fig. S3.

Figure 3.. Cellular immune responses in vaccinated macaques.

IFN-γ ELISPOT responses to Gag, Env1, Env2, Pol1, and Pol2 peptide pools in macaques at baseline and at week 4 following the boost immunization. Spot-forming cells (SFC) per million peripheral blood mononuclear cells (PBMC) are shown. The dotted lines represent limits of quantitation of the assay. Red bars indicate median responses.

Figure 4.. Ranking of vaccinated animals by the protective signature.

(A) Principal component analysis (PCA) of vaccinated animals stratified by the enrichment of the protective signature defined in Fig. 1 where each dot represents a vaccinated animal. The dot intensity depicts the degree of the protective antibody signature, where dark dots are highly enriched for the protective signature and light dots have lower levels of the protective signature. (B) PCA of individual protective features in the signature in the same multivariate space. (C) Heatmap showing the ranking of each vaccinated animal by enrichment of the protective antibody signature. Black to white shading shows the strength of the protective signature. Yellow to blue shading depicts the magnitude (positive to negative) of each of the signature features in the vaccinated animals.

Figure 5.. Innate immune correlates of the protective signature.

(A) Heatmap representing the top positively correlated genes on day 1 post-boost associated with the protection rank score. Rows represent genes and columns represent samples. Gene expression is represented as a gene-wise standardized expression (Z-score). Red and blue correspond to up- and down-regulated genes respectively. Green to white correspond to high to low protection rank score. (B) An over-representation test of gene expression pathways on day 1 post-boost correlated with the predicted protection rank score, demonstrating IL-4/IL-13 signaling and reactive oxygen species (ROS) and mitochondrial metabolism pathways. Scatter plots highlight correlations between (C) PINK1 and (D) ALOX15 gene expression on day 1 and protection rank score. (E) Heatmap representing the top positively correlated genes on day 7 post-boost associated with the protection rank score as in (A). (F) An over-representation test of gene expression pathways on day 7 post-boost correlated with the predicted protection rank score, demonstrating interferon signaling and B-cell receptor (BCR) pathways. Scatter plots highlight correlations between (G) CD22 and (H) CD19 gene expression and protection rank score. The Cluego plug-in in Cytoscape was used to plot the transcriptomic cascades (B, F). P-values represent Spearman rank correlation tests (C, D, G, H). See also Figs. S4–S6.

Figure 6.. Protective efficacy in vaccinated animals.

(A) Kaplan-Meier curve depicting the number of challenges required for infection in sham control macaques and in vaccinated macaques in quartiles I-IV defined by the predicted protective signature (Fig. 4). (B) Statistical analyses of protective efficacy by log-rank tests, Cox proportional hazard models, and Weibull survival models. (C) Correlation of predicted protection rank score and the number of challenges required for protection. P-value represents Spearman rank-correlation test.

Figure 7.. Protective efficacy following adoptive transfer of purified IgG.

(A, B) ELISA titers to SIVmac32H Env gp140 in vaccinated macaques (Donors) and in recipient macaques (Recipients) following adoptive transfer of purified IgG from animals in quartiles I-IV. ELISA titers in animals that received (C) sham IgG or (D) IgG purified from vaccinated macaques in quartiles I-IV following the initial SIVmac251 challenge. (E) Kaplan-Meier curve depicting the number of challenges required for infection in sham control macaques and in macaques that received IgG from animals in quartiles I+II and quartiles III+IV. (F) Statistical analyses of protective efficacy by log-rank tests, Cox proportional hazard models, and Weibull survival models. See also Fig. S7.