Sequential and Simultaneous Immunization of Rabbits with HIV-1 Envelope Glycoprotein SOSIP.664 Trimers from Clades A, B and C

Abstract

We have investigated the immunogenicity in rabbits of native-like, soluble, recombinant SOSIP.664 trimers based on the env genes of four isolates of human immunodeficiency virus type 1 (HIV-1); specifically BG505 (clade A), B41 (clade B), CZA97 (clade C) and DU422 (clade C). The various trimers were delivered either simultaneously (as a mixture of clade A + B trimers) or sequentially over a 73-week period. Autologous, Tier-2 neutralizing antibody (NAb) responses were generated to the clade A and clade B trimers in the bivalent mixture. When delivered as boosting immunogens to rabbits immunized with the clade A and/or clade B trimers, the clade C trimers also generated autologous Tier-2 NAb responses, the CZA97 trimers doing so more strongly and consistently than the DU422 trimers. The clade C trimers also cross-boosted the pre-existing NAb responses to clade A and B trimers. We observed heterologous Tier-2 NAb responses albeit inconsistently, and with limited overall breath. However, cross-neutralization of the clade A BG505.T332N virus was consistently observed in rabbits immunized only with clade B trimers and then boosted with clade C trimers. The autologous NAbs induced by the BG505, B41 and CZA97 trimers predominantly recognized specific holes in the glycan shields of the cognate virus. The shared location of some of these holes may account for the observed cross-boosting effects and the heterologous neutralization of the BG505.T332N virus. These findings will guide the design of further experiments to determine whether and how multiple Env trimers can together induce more broadly neutralizing antibody responses.

Fig 1. Schematic of the immunization schedule.

The longitudinal course of the various sub-studies, in weeks, is depicted, with each immunization indicated by an arrow. Every animal in each group was immunized as indicated at the time-points, except at week 73, when only a subset of animals was immunized (specifically #5715–1, #5716–1, #5718–2, #5719–2, #5721–2, #5726–3, #5727–3, #5734–5, #5735–5, #5736–5, #5749–8). The rabbits were bled to obtain sera for NAb and serology assays immediately before and 2 weeks after each immunization. A, The color of the arrows denotes the identity of the trimer, indicated in the key to the right. The amount of trimer given at each time was 30 μg. B, The schematic is similar to the one in panel A. The size of the arrows is proportional to the dose of the trimer. For groups 4 and 7, the total trimer dose in the week 0–24 period was 30 μg, whereas in group 8 it was 90 μg, with clade A (BG505)- and clade B (B41)-based trimers present in ~1/1 ratios (see text). For group 2, the clade A (BG505) content of the mixture (total 30 μg) was ~20% (~6 μg). Clade C (CZA97) trimers (30 μg) were used to boost groups 4, 8 and 2 in the period from week 36 to 60, as indicated. Group 7 was terminated at week 26, group 4 at week 62. C, The schematic is similar to panel-A, with the color code to the right. The sizes of the symbols approximately correspond to the absolute and relative doses of the trimer immunogens (see text).

Fig 2. Autologous neutralization responses to clade A, B and C SOSIP.664 trimers.

The neutralization titers (IC₅₀) against Tier-2 Env-pseudotyped viruses, derived at DUMC, are plotted on the y-axis as a function of time after the first immunization on the x-axis (weeks). The scales are kept constant within each panel to facilitate comparisons among groups. The schedule is shown in Fig 1 and involved immunizations at weeks 0, 4, 20, 24, 36, 48, 60 and 73, with blood taken for analysis immediately before and 2 weeks after each immunization. Note that only a subset of animals was immunized at week 73 (see Fig 1); for the other animals, the final immunization was at week 60. The test viruses were as follows: A, BG505.T332N; B, B41; C, DU422; D, CZA97. The curves connecting the reciprocal neutralization titers are color-coded for each rabbit as per the key associated with each group. Thus, changes in titers can be monitored over time both for individual animals and on a group-wide basis.

Fig 3. Autologous NAb responses to sequentially delivered clade A and clade B trimers.

The neutralization titers (IC₅₀) against Tier-2 Env-pseudotyped viruses are plotted on the y-axis as a function of time after the first immunization on the x-axis (weeks). The scales are kept constant within each panel to facilitate comparisons among groups. The schedule is shown in Fig 1C with immunizations at weeks 0, 4, 20, 24, 36, 48 and 60, and for rabbits #5734–5, #5735–5 and #5736–5 also at week 73. Blood was taken for analysis immediately before and 2 weeks after each immunization. The test viruses were as follows: A and C, BG505.T332N; B and D, B41. The curves connecting the reciprocal neutralization titers are color-coded for each rabbit as per the key associated with each group. Thus, changes in titers can be monitored over time both for individual animals and on a group-wide basis.

Fig 4. Cross-boosting of clade A and clade B autologous NAb responses by clade C trimers.

NAb titers at the indicated time points for all five animals in the listed groups are compared. Bars show arithmetic means + s.e.m. The clade C trimer boosts occurred at weeks 36 and 48. A and C, BG505.T332N titers; B and D, B41 titers.

Fig 5. De novo cross-neutralization of BG505.T332N (clade A) induced by clade C trimers after priming with clade B trimers.

BG505.T332N NAb titers at the indicated time points for the five animals in group 1, which received only clade B and C trimers are shown. Bars show arithmetic means + s.e.m. The NAb titers were compared with the cut-off value of 20 by the Wilcoxon signed rank test (see Results).

Fig 6. Neutralization of heterologous Tier-1 viruses.

The arithmetic mean neutralization titers (IC₅₀) against the Env-pseudotyped Tier-1 viruses MN.3 (clade B) and MW965.26 (clade C), derived at DUMC, are plotted over time for the rabbit groups denoted by the key to the right.

Fig 7. Mapping autologous NAb specificities using BG505.T332N, B41 and CZA97 virus mutants.

A, Neutralization of BG505.T332N virus mutants. The values recorded for the various mutant viruses are the percentage neutralization at a dilution of 1/50, relative to the BG505.T332N parental virus (labeled WT and defined as 100%), and are the averages of 2 replicates ± s.e.m. Red boxes highlight fully or substantially resistant viruses (<25% neutralization); yellow, moderately resistant viruses (25–75% neutralization); green, sensitive viruses (>75% neutralization). The Q130N, S241N and P291T changes introduce N-linked glycans at positions 130, 241 and 289, respectively. The S241N+P291T double mutant contains glycans at both positions 241 and 289. The MG505 cl.A2 and cl.H3 viruses differ from BG505.T332N at several positions (see S1 Fig). Of note is that MG505 cl.A2 has a lysine residue at position-241 (i.e., as per the S241K mutant), whereas cl.H3 has a glycan site (i.e., as per the S241N mutant). A glycan is present at position-289 in both MG505 clones. The K241S change in MG505 cl.A2 restores the Ser residue that is present at position-241 in the BG505.T332N virus. Full titration curves for the sera showed only a single example of reduced neutralization of the glycan knock-in mutants, compared with BG505.T332N, that was not evident at the standard test dilution of 1/50 (rabbit #5739 at week 62; S4 Fig). Overall, the extent of neutralization of the glycan knock-in mutants did not correlate well with the titers against the wild-type BG505.T332N virus (Spearman rank correlation for the 241-glycan knock-in: r = 0.30, p = 0.099; for the 289-glycan knock-in: r = 0.21, p = 0.25; for the double mutant: r = 0.27, p = 0.15). B, Neutralization of B41 virus mutants. The organization is the same as in panel-A. The B41 parental virus is labeled WT and its neutralization defined as 100%. The N132T and A291T changes into this virus introduce N-linked glycans at positions 130 and 289, respectively. The CH01 and VRC01 bNAbs were used as control reagents for assessing overall neutralization sensitivity (only shown for B41 mutants for which a partial effect on CH01 neutralization was observed). C, Neutralization of CZA97 virus mutants. The organization is the same as in panel-A. The serum dilution used was 1/60. The CZA97 cl.12 parental virus is labeled cl.12 and its neutralization defined as 100%. The CZA97 cl.29 virus differs from cl.12 at several positions (see S1 Fig) but of note is that it contains a glycan at position-411. The CZA97 cl.12-D411N and cl.29-N411D mutants contain and lack glycans at position-411, respectively.

Fig 8. A glycan hole in the BG505 trimer.

Left panel: The BG505 SOSIP.664 trimer and the BG505.T332N virus lack glycans at positions 241 and 289, which results in a large exposed peptide surface on the side of gp120 near the gp120/gp41 interface. For clarity one protomer of the BG505 SOSIP.664 trimer has been colored, with the polypeptide chain in blue ribbons and glycans denoted by green spheres. Right panel: The restored glycans at positions 241 and 289 (as in the BG505 S241N+P291T double mutant virus) now mask the underlying peptide surface. The individual glycans at positions 241 and 289 both have a substantial masking effect, accounting for the similar phenotypes of the single and double mutant knock-in viruses (S3 Table).

Fig 9. Trimer-binding antibody titers induced by trimer immunizations.

The serum dilutions giving half-maximal binding (EC₅₀) are plotted on the y-axes and the time of sampling on the x-axes for sera from individual rabbits (color-coded) in groups 1, 2, 3, 4, 7 and 8. The respective trimers used in the binding assays are A, BG505; B, B41; C, DU422 (group 1) and CZA97 (groups 2, 3, 4, 7 and 8). D, The corresponding plots for the sera from sequentially immunized rabbits (groups 5 and 6) show antibody binding to BG505 and B41 trimers.

Fig 10. Correlations between trimer-binding and neutralizing antibody titers.

The scatterplots show neutralization titers on the y-axes and the trimer-binding antibody titers on the x-axes. Within each plot the symbols corresponding to neutralization of the BG505.T332N and B41 viruses are color-coded as indicated on the figure panels. Spearman correlation coefficients (r-values) and the corresponding significances (p-values) are color-coded analogously. Correlation analyses were performed for both BG505 and B41 NAb and binding antibody titers for sera from the early monovalent immunogen regimens during weeks 4–36 (top panel, group 1; lower panel, group 3).