Experimental medicine study with stabilised native-like HIV-1 Env immunogens drives long-term antibody responses, but lacks neutralising breadth

Abstract

Background: We report findings from an experimental medicine study of rationally designed prefusion stabilised native-like HIV envelope glycoprotein (Env) immunogens, representative of global circulating strains, delivered by sequential intramuscular injection.

Methods: Healthy adult volunteers were enrolled into one of five groups (A to E) each receiving a different schedule of one of two consensus Env immunogens (ConM SOSIP, ConS UFO, either unmodified or stabilised by chemical cross-linking, followed by a boost with two mosaic Env immunogens (Mos3.1 and Mos3.2). All immunogens were co-formulated with liposomal Monophosphoryl-Lipid A (MPLA) adjuvant, and volunteers were followed up for 28 days post final Mosaic booster injection. Participants gave written informed consent to join the study. The study is registered on ClinicalTrials.gov ID NCT03816137.

Findings: Fifty-one participants (men n = 23 and women n = 28) aged 18-55 were enrolled. The seroconversion rate against Env was 100% with all participants having measurable anti-Env IgG antibodies after their second injection and throughout the study. Neutralisation was detected against the ConM pseudovirus in sera of those who had received both ConM and ConS immunogens. However, this activity was limited in breadth and was neither boosted nor broadened in those receiving the Mos3.1 and Mos3.2 immunogens. Neutralising antibody function correlated with binding to V1/V3 and V5 epitopes and peaked after the third injection.

Interpretation: Rationally designed prefusion-stabilised native-like Env trimers are robustly immunogenic in a prime-boost schedule. When given alone they are insufficient to induce neutralising antibody titres of significant breadth, but they represent potentially valuable polishing immunogens after germline-targeting.

Funding: European Aids Vaccine initiative (EAVI2020) received funding from EU Horizon 2020, grant number 681137. Structural studies were supported by the Bill and Melinda Gates Foundation (INV-002916).

Keywords: Experimental medicine; HIV envelope; HIV-1; Prefusion; Trimer; Vaccine.

Fig. 1

CONSORT diagram of EAVI_01. Diagram showing the flow of the experimental medicine study from screening through to completion of the final visits.

Fig. 2

Antigen-specific IgG responses in serum samples from participants of the EAVI2020_01 experimental medicine study. The concentration of IgG specific for (a) ConM, (b) ConM-EDC, (c) ConS or (d) ConS-EDC were assessed. In part 1 (circles) Group A (Grey) received three injections with 100 μg ConM, Group B (Light Blue) received three injections with 100 μg ConM-EDC, Group C (Green) received three injections with 100 μg ConS, Group D (Pink) received three injections with 100 μg ConS-EDC and Group E (Yellow) received two injections with 100 μg ConS followed by one injection with 100 μg ConM. Injections 1–3 were given at visits 1, 6 and 11. All groups were boosted with a cocktail of 50 μg Mosaic 3.1 and 50 μg Mosaic 3.2 during the fourth injection (visit 17; part 2 (squares)). All injections were adjuvanted with 500 μg MPLA. The arrows indicate the timepoints after injection at which blood was taken for immunogenicity analysis. Median values with IQR are shown. The Kruskal–Wallis with Dunn’s multiple correction test was performed to compare the statistical differences between groups at day 0, 7, 14, 28 and 84 post injection. ∗ <0.05; ∗∗<0.01; ∗∗∗ <0.001; ∗∗∗∗ <0.0001.

Fig. 3

Memory B cell ELISpot responses to the study-specific Env in participants of the EAVI2020_01 experimental medicine study. % antigen-specific memory B cells are shown when PBMC from participants of the study were stimulated with either (a) ConM or (b) ConS. The timepoints assessed were V1 (time of first injection), V15 (28 days post-third IM injection), V17 (time of Mosaic Boost Injection) and V21 (28 days post-Mosaic Boost). In part 1 (circles), Group A (Grey) received three injections with 100 μg ConM, Group B (Light Blue) received three injections with 100 μg ConM-EDC, Group C (Green) received three injections with 100 μg ConS, Group D (Pink) received three injections with 100 μg ConS-EDC and Group E (Yellow) received two injections with 100 μg ConS followed by one injection with 100 μg ConM. All groups were boosted with a cocktail of 50 μg Mosaic 3.1 and 50 μg Mosaic 3.2 (total = 100 μg) during the fourth injection (part 2 (squares)). All injections were adjuvanted with 500 μg MPLA. All data have been background subtracted. Median values with IQR are shown. Kruskal–Wallis Test with Dunn’s multiple correction test was performed to compare the statistical differences between V1 the remaining timepoints within each group, as well as a comparison between groups at V15, V17 and V21. ∗ <0.05; ∗∗<0.01; ∗∗∗ <0.001.

Fig. 4

Virus neutralisation assay responses in serum samples from participants of the EAVI2020_01 experimental medicine study. Serum samples were tested against (a) ConM HIV-1 virus. The timepoints assessed were, Part 1; V1 (time of first injection), V14 (14 days post-third IM injection), Part 2; V17 (time of Mosaic Boost injection) and V20 (14 days post-Mosaic Boost). Serum sample responses shown longitudinally (b). These responses were then correlated with ConM-specific IgG serum concentrations (c) at V20 (2 weeks post-third injection). In part 1 (circles), Group A (Grey) received three doses of 100 μg ConM, Group B (Light Blue) received three doses of 100 μg ConM-EDC, Group C (Green) received three doses of 100 μg ConS, Group D (Pink) received three doses of 100 μg ConS-EDC and Group E (Yellow) received two doses of 100 μg ConS followed by one dose of 100 μg ConM. All groups were boosted with a cocktail of 50 μg Mosaic 3.1 and 50 μg Mosaic 3.2 during the fourth injection (part 2 (squares)). All injections were adjuvanted with 500 μg MPLA. For (a) median values with IQR are shown. Kruskal–Wallis with Dunn’s multiple correction test was performed to compare the statistical differences between V1 the remaining timepoints within each group, as well as a comparison between groups at V14, V17 and V20. ∗ <0.05; ∗∗<0.01. For (c), Spearman’s correlation was performed, and the r and p values are shown within the graphs.

Fig. 5

Electron microscopy-based polyclonal epitope mapping (EMPEM) of plasma samples from participants of the EAVI2020_01 experimental medicine study. (a) Composite map of HIV-1 trimer with Fabs binding to potential epitopes against V1/V3 (green), V5 (blue), N611 (yellow/beige) and base (purple) regions. (b) Summary of ID50 neutralisation results against ConM and ConS HIV-1 trimer. Numbers highlighted in (red) show highest neutralisation response. (c) Overview of epitope responses by participants. Dashed circles represent epitopes clearly visible in 2D classes while solid circles represent same epitopes also seen in 3D maps after 3D classification. Some participants had Fab bound to HIV-1 protomers visible in the 2D classes shown in (grey).