Selected HIV-1 Env trimeric formulations act as potent immunogens in a rabbit vaccination model

Abstract

Background: Ten to 30% of HIV-1 infected subjects develop broadly neutralizing antibodies (bNAbs) during chronic infection. We hypothesized that immunizing rabbits with viral envelope glycoproteins (Envs) from these patients may induce bNAbs, when formulated as a trimeric protein and in the presence of an adjuvant.

Methods: Based on in vitro neutralizing activity in serum, patients with bNAbs were selected for cloning of their HIV-1 Env. Seven stable soluble trimeric gp140 proteins were generated from sequences derived from four adults and two children infected with either clade A or B HIV-1. From one of the clade A Envs both the monomeric and trimeric Env were produced for comparison. Rabbits were immunized with soluble gp120 or trimeric gp140 proteins in combination with the adjuvant dimethyl dioctadecyl ammonium/trehalose dibehenate (CAF01). Env binding in rabbit immune serum was determined using ELISAs based on gp120-IIIB protein. Neutralizing activity of IgG purified from rabbit immune sera was measured with the pseudovirus-TZMbl assay and a PBMC-based neutralization assay for selected experiments.

Results: It was initially established that gp140 trimers induce better antibody responses over gp120 monomers and that the adjuvant CAF01 was necessary for such strong responses. Gp140 trimers, based on HIV-1 variants from patients with bNAbs, were able to elicit both gp120IIIB specific IgG and NAbs to Tier 1 viruses of different subtypes. Potency of NAbs closely correlated with titers, and an gp120-binding IgG titer above a threshold of 100,000 was predictive of neutralization capability. Finally, peptide inhibition experiments showed that a large fraction of the neutralizing IgG was directed against the gp120 V3 region.

Conclusions: Our results indicate that the strategy of reverse immunology based on selected Env sequences is promising when immunogens are delivered as stabilized trimers in CAF01 adjuvant and that the rabbit is a valuable model for HIV vaccine studies.

Figure 1. Analysis of purified gp140 proteins.

Reduced SDS-PAGE gel (10–12%) analysis of four gp140 proteins (lane 1 to 4) after purification and staining with Coomassie Blue. Lane 1 ITM1_anc; lane 2 ITM1_4; lane 3 ACS19642; lane 4 ACS19554. Proteins were run next to a molecular weight marker (Novex Sharp Pre-stained protein standard, Invitrogen). Proteins were typically >95% pure prior to immunisation.

Figure 2. Comparison of trimeric gp140 versus monomeric gp120.

(A) End-point binding titers of rabbits immunized with either trimeric gp140 (filled squares) or monomeric gp120 (open squares) of UG_A HIV-1 Env. Each dot represents one rabbit. Horizontal lines indicate mean titers. Statistical analysis was done using Mann-Whitney test. (B) Neutralization data using IgG (at a final concentration of 250 µg/ml) isolated from week 12 sera of rabbits immunized, in the presence of CAF01, with either trimeric gp140 or monomeric gp120 of UG_A. Neutralization of two clade B viruses, SF162 (left, squares) and Bx08 (right, diamonds) is depicted. Each dot represents one rabbit. Horizontal lines indicate the mean percent neutralization. Statistical analysis was done using Mann-Whitney test.

Figure 3. Immunization in the presence or absence of CAF01.

(A) End-point binding titer of rabbits immunized with ITM1_4, in the presence (filled triangles) or absence (open triangles) of CAF01 respectively. Results are given for weeks 8, 12 and 14. Each dot represents one rabbit. Horizontal lines indicate mean titers. Statistical analysis was done using Mann-Whitney test. (B) Neutralization using IgG (at a final concentration of 250 µg/ml) isolated from week 12 (filled squares) and 14 (open triangles) sera of rabbits immunized with ITM1_4 in the presence or absence of CAF01. Two clade B viruses, SF162 (left) and Bx08 (right) were used. Horizontal lines indicate the mean percent neutralization. The dotted (green) line represents 50% neutralization. Statistical analysis was done using Mann-Whitney test.

Figure 4. Comparison of Env immunogens.

(A) Kinetics and comparison of end-point binding titers of all rabbits immunized with Env gp140 trimers (100 µg/dose) in the presence of CAF01. The mean of four rabbits/group with standard deviation is given for weeks 4, 8, 12 and 14. For UG_A the data for week 4 and 8 are lacking. The green dotted line represents an end-point titer of 10⁵. No statistical differences between the various trimeric gp140 Envs were observed (p = 0.18 and 0.61 at week 12 and 14 respectively; Kruskal-Wallis). (B) Neutralization of SF162 virus using IgG (at a final concentration of 250 µg/ml) isolated from week 12 sera of rabbits immunized with trimeric gp140, all in the presence of CAF01. Each dot represents one rabbit. Horizontal lines indicate mean percent neutralization. The dotted (green) line represents 50% neutralization. Overall, no statistical differences were found between the trimeric gp140 Envs (p = 0.092, Kruskal-Wallis).

Figure 5. Correlation between end-point binding titers and neutralization responses against SF162.

(A) Correlation using week 12 sera (ELISA) or IgG (neutralization responses at a final concentration of 250 µg/ml) of all groups, including the monomers (44 data points). Spearman correlation r = 0.7243; p <0.0001. (B) Correlation using week 12 sera (ELISA) or IgG (neutralization responses at a final concentration of 250 µg/ml). Only data from rabbits immunized with trimers ITM1_4, UG_A and ACS19554, giving the highest neutralization responses (12 data points) were used. Spearman correlation r = 0.8616; p = 0.0003.

Figure 6. Kinetics of neutralizing responses against SF162 pseudovirus.

Kinetics using three 2-fold dilutions of IgG isolated from weeks 4, 8, 12 and 14 sera of rabbits immunized with trimer ITM1_4 (Figure 6a), ACS19554 (Figure 6b) and ACS19642 (Figure 6c). Horizontal lines indicate the mean percent neutralization. Final IgG concentrations were 250 µg/ml (filled dots), 125 µg/ml (semi filled dots) and 62.5 µg/ml (open circles) respectively. The dotted (green) line represents 50% neutralization.

Figure 7. Peptide inhibition experiments.

Four serial 2-fold dilutions of IgG purified from rabbits immunized with ITM1_4 (ITM1_4-1; -2; -3 and -4) and UG_A (UG_A-1; -2; -3 and -4) were used in a modified pseudovirus neutralization assay using SF162. Two linear peptides 797 and 7012.1, recognizing the V3-crown motif of clade B (GPGR) and clade A (GPGQ) respectively and 1 cyclic peptide 7019 recognizing GPGR (clade B) were used at a final concentration of 16 µg/ml. Each graph represents a single rabbit.