Control of Heterologous Simian Immunodeficiency Virus SIVsmE660 Infection by DNA and Protein Coimmunization Regimens Combined with Different Toll-Like-Receptor-4-Based Adjuvants in Macaques

Abstract

We developed a method of simultaneous vaccination with DNA and protein resulting in robust and durable cellular and humoral immune responses with efficient dissemination to mucosal sites and protection against simian immunodeficiency virus (SIV) infection. To further optimize the DNA-protein coimmunization regimen, we tested a SIV_mac251-based vaccine formulated with either of two Toll-like receptor 4 (TLR4) ligand-based liposomal adjuvant formulations (TLR4 plus TLR7 [TLR4+7] or TLR4 plus QS21 [TLR4+QS21]) in macaques. Although both vaccines induced humoral responses of similar magnitudes, they differed in their functional quality, including broader neutralizing activity and effector functions in the TLR4+7 group. Upon repeated heterologous SIV_smE660 challenge, a trend of delayed viral acquisition was found in vaccinees compared to controls, which reached statistical significance in animals with the TRIM-5α-resistant (TRIM-5α R) allele. Vaccinees were preferentially infected by an SIV_smE660 transmitted/founder virus carrying neutralization-resistant A/K mutations at residues 45 and 47 in Env, demonstrating a strong vaccine-induced sieve effect. In addition, the delay in virus acquisition directly correlated with SIV_smE660-specific neutralizing antibodies. The presence of mucosal V1V2 IgG binding antibodies correlated with a significantly decreased risk of virus acquisition in both TRIM-5α R and TRIM-5α-moderate/sensitive (TRIM-5α M/S) animals, although this vaccine effect was more prominent in animals with the TRIM-5α R allele. These data support the combined contribution of immune responses and genetic background to vaccine efficacy. Humoral responses targeting V2 and SIV-specific T cell responses correlated with viremia control. In conclusion, the combination of DNA and gp120 Env protein vaccine regimens using two different adjuvants induced durable and potent cellular and humoral responses contributing to a lower risk of infection by heterologous SIV challenge.IMPORTANCE An effective AIDS vaccine continues to be of paramount importance for the control of the pandemic, and it has been proven to be an elusive target. Vaccine efficacy trials and macaque challenge studies indicate that protection may be the result of combinations of many parameters. We show that a combination of DNA and protein vaccinations applied at the same time provides rapid and robust cellular and humoral immune responses and evidence for a reduced risk of infection. Vaccine-induced neutralizing antibodies and Env V2-specific antibodies at mucosal sites contribute to the delay of SIV_smE660 acquisition, and genetic makeup (TRIM-5α) affects the effectiveness of the vaccine. These data are important for the design of better vaccines and may also affect other vaccine platforms.

Keywords: A/K variant; ADCC; ADCD; ADNP; Ab glycosylation structures; DNA; HIV; QS21; SIVmac251; SIVsmE660; SIVsmE660 T/F; T cell responses; TLR4; TLR7; TRIM-5α; V2 responses; acquisition delay; adjuvant; binding antibody; correlate of viremia control; cyclic V2; humoral responses; immunization; linear peptide; mucosal responses; neutralizing antibody; protein; reduced risk of infection; repeated low-dose rectal challenge; rhesus macaque; scaffolded gp70-V1V2; systems serology; vaccination; vaccine; viremia control.

FIG 1

Vaccination of macaques with SIV_mac251 DNA-protein coimmunization vaccine regimens. (A) Indian rhesus macaques were vaccinated three times (0, 2, and 6 months) with SIV_mac251-derived env plasmids (SIV_mac239 and T/F M766) coadministered with monomeric M766 gp120 protein adjuvanted with TLR4+7 (n = 12) or TLR4+QS21 (n = 12). The DNA mixtures also contained SIV_mac239 gag DNA and rmIL-12 DNA. Five months after the 3rd vaccination, the animals were subjected to weekly intrarectal exposures using a titrated dose of the heterologous SIV_smE660 virus, and the infected animals were monitored for 6 months. (B) Schematic representation of the DNA-protein vaccine comprising RNA/DNA-optimized expression vectors producing SIV_mac251-derived membrane-bound gp120e-TM and the soluble trimeric gp140 Env proteins. The vaccine contained monomeric M766 gp120 Env. Amino acid positions follow SIV_mac239 numbering. (C) HEK293T cells were transfected with SIV M766 env plasmid DNAs (pDNAs) expressing gp150 (lane 1), gp140 (lane 2), gp120e-TM (lane 3), and gp120-TM (lane 4). Proteins from the cell-associated and extracellular (1/200 of each sample) fractions were analyzed by Western immunoblotting and detected using a mouse anti-gp120 Ab. Equal loading of the blot with the cell-associated fractions was controlled by probing the membrane with an antiactin antibody. (D) Histogram overlay showing the membrane localization of mac239 and M766 gp120e-TM proteins on transfected HEK293 cells using a mouse anti-gp120 Ab followed by an APC-conjugated goat anti-mouse Ab. The mock-transfected cells are shown (gray histogram).

FIG 2

Systemic and mucosal humoral immune responses induced by the DNA-protein coimmunization vaccine. Humoral responses were measured 2 weeks after the 3rd vaccination (bars indicate median values). (A) bAb to SIV_mac251 and SIV_smE660 in plasma determined by SIV-BAMA are shown as the area under the curve of the binding magnitude (AUC). (B) gp140_smE660-specific binding activity measured by SIV-BAMA in rectal and vaginal (samples from 4 of the 5 females could be analyzed) mucosal samples are shown as Env-specific binding antibody (MFI × dilution/total IgG). (C) Responses in plasma to cyclic V2 were measured to SIV_smE543 V2 peptide (aa 168 to 206) by a SPR assay. (D and E) bAb recognizing gp70-scaffolded SIV_smE660-specific V1V2 measured in plasma (D) and in mucosal (rectal and vaginal) samples (E) using SIV-BAMA. (F) NAb to SIV_mac251.6, SIV_smE660 (BR-CG7G.IR1 and 2A5-VTRN), and SIV_mac251 M766 in plasma. Titers are calculated as 50% infectious doses (ID₅₀) (dilution) in TZM-bl cells with a threshold of 300. (G) Serially diluted serum samples were used to determine the ADCC titers (left) and peak granzyme B activity (right) using SIV_mac251 gp120-coated CEM-NKR target cells. (H to K) Systems serology shows distinct function and Ab glycoforms. (H) Violin plot illustrating the performance of the actual model and 2 negative-control null models (random features and permuted data) for comparison of the two vaccine groups. The violin plot illustrates the distribution of the classification accuracies of the actual and null models, as measured across 100 independent 5-fold cross-validation replicates. (I) Scores plot of a LASSO/PLS model illustrating separation between animals from the 2 vaccine groups. (K) Variable importance in projection (VIP) plot showing the variables that were identified by the model distinguishing the animals in each vaccine groups. The length of the bar corresponds to the relative importance of the variable, and the color of the bar corresponds to which arm the variable is higher.

FIG 3

Delay in SIV_smE660 acquisition in vaccinees. (A and B) Kaplan-Meier curves of the number of SIV_smE660 challenges for infection of the two vaccine groups (n = 12 each) and the control (n = 12) (A) and the combined group of animals with the TRIM-5α R genotype (vaccinees, n = 10) and controls (n = 5) and of animals with the TRIM-5α M/S genotype (vaccinees, n = 14) and controls (n = 7) (B). P values comparing vaccinees and controls with the TRIM-5α R genotype are from a Gehan-Breslow-Wilcoxon test. (C) Vaccinees with positive rectal V1V2 responses and carrying the TRIM-5α R or TRIM-5α M/S allele are compared to V1V2-negative animals. P values are from an exact log rank test.

FIG 4

Sieve effect with selection of SIV_smE660 neutralization-resistant virus variants. (A) Single-genome amplification and direct sequencing of the T/F env genes from the plasma samples (Table 1) were used to determine the number of T/F variants in the vaccinees and controls. (B) Genetic analysis of T/F Env sequences. Fifty-five informative sites were plotted as a proportion of each amino acid (AA) compared to the consensus sequence (see Fig. S6 in the supplemental material), and the plot shown excludes the 25 sites with a <10% difference in relative proportions between any of the groups for clarity. Residue 45 (T) and residue 47 (R), associated with neutralization resistance (54), showed the most prominent changes in the vaccinees compared to controls. (C) The percentage of Env sequences with consensus T/R (neutralization-sensitive), A/K (neutralization-resistant), and A/R (neutralization-intermediate) sequences are shown for the SIV_smE660 challenge stock and the infected controls and vaccine groups. No viruses with T/K changes were found. (D to G) Two-tailed nonparametric Spearman correlation plots show direct correlations of SIV_smE660 bAb in plasma of the TLR4+7 group (D), NAb to pseudotyped T/F SIV_smE660-CG7G (E), the neutralization-sensitive SIV_smE660 (SIVsmE660/2A5-VTRN) (F), and the number of SIV_smE660 exposures to infection (G). Associations with bAb and NAb were measured 2 weeks after the 3rd vaccination (D to F) and 2 weeks before challenge (F). Spearman r and P values are shown.

FIG 5

Control of viremia. (A) The dot plot shows the peak VL of each animal with the TRIM-5α M/S genotype, and the median is indicated. The P value is from multiple comparisons to controls using ANOVA (Kruskal-Wallis test). (B) Geometric means of virus loads monitored for 25 weeks are shown for TLR4+7 (n = 7) and controls (n = 7) (top) and TLR4+QS21 (n = 7) and controls (n = 7) (bottom). (C and D) Inverse correlation of SIV_smE660-specific V1V2 responses in plasma (C) and rectal mucosa (D) and peak virus loads. gp70-V1V2 responses were measured by SIV-BAMA. (E and F) Inverse correlation of SIV_smE660-specific cyclic V2 responses in plasma and virus loads (AUC) during the acute phase (weeks 2 to 4) (E) and chronic phase (weeks 6 to 25) (F) of infection.

FIG 6

Antigen-specific T cell responses and control of viremia. (A) Bars show frequencies of Env-specific and Gag-specific T cell responses measured in PBMC samples collected 2 weeks after the 3rd vaccination. (B) Dot plots show SIV-specific IFN-γ⁺ CD4⁺ (top), CD8⁺ (middle), and granzyme B-positive (GrzB⁺) (bottom) SIV-specific T cells. (C to E) Inverse correlations of Gag-specific IFN-γ⁺ T cells (C) and Env-specific IFN-γ⁺ T cells (D) measured 2 weeks before challenge start and virus load at peak and between Gag-specific IFN-γ⁺ T cells measured 18 weeks after the 3rd vaccination and VL during the chronic phase (weeks 6 to 25 postinfection) (E). Spearman r and P values are shown.

FIG 7

Association between Env-specific CD4⁺ T cells and NAb. Shown are direct correlations of Env-specific CD4⁺ T cells measured 2 weeks after the 3rd vaccination and NAb to SIV_smE660 (SIV_smE660-BR-CG7G [left] and SIV_smE660/2A5-VTRN [right]) in plasma. Spearman r and P values are shown.