A malaria vaccine protects Aotus monkeys against virulent Plasmodium falciparum infection

Abstract

The Plasmodium falciparum protein, apical membrane antigen 1 forms a complex with another parasite protein, rhoptry neck protein 2, to initiate junction formation with the erythrocyte and is essential for merozoite invasion during the blood stage of infection. Consequently, apical membrane antigen 1 has been a target of vaccine development but vaccination with apical membrane antigen 1 alone in controlled human malaria infections failed to protect and showed limited efficacy in field trials. Here we show that vaccination with AMA1-RON2L complex in Freund's adjuvant protects Aotus monkeys against a virulent Plasmodium falciparum infection. Vaccination with AMA1 alone gave only partial protection, delaying infection in one of eight animals. However, the AMA1-RON2L complex vaccine completely protected four of eight monkeys and substantially delayed infection (>25 days) in three of the other four animals. Interestingly, antibodies from monkeys vaccinated with the AMA1-RON2L complex had significantly higher neutralizing activity than antibodies from monkeys vaccinated with AMA1 alone. Importantly, we show that antibodies from animals vaccinated with the complex have significantly higher neutralization activity against non-vaccine type parasites. We suggest that vaccination with the AMA1-RON2L complex induces functional antibodies that better recognize AMA1 as it appears complexed with RON2 during merozoite invasion. These data justify progression of this next generation AMA1 vaccine towards human trials.

Fig. 1

PfAMA1–RON2L complex protects against virulent P. falciparum challenge. a–c Time course of parasitemia after challenge with 10⁴ Pf FVO strain infected RBCs for animals in the control, adjuvant only (a) Group 1), AMA1 (b) Group 2) and AMA1–RON2L complex (c) Group 3). The dashed line at the top indicates limit of parasitemia (200,000 parasites per μL) at which time animals were treated with the antimalarial drug, mefloquine. The dotted line at the bottom indicates the absence of thin smear detectable parasites. T, treatment due to high parasitemia; A, treatment due to anemia; +, a single animal found dead possibly due to anemia; SC, self-cured; SP: animals that remained thin-smear negative until day 41 after challenge when the study was terminated. (d) Kaplan–Meier plot of time-to-patency of animals in Groups 2 and 3. Log-rank was performed to compare time-to-patency (parasite positive by thin blood smear) of animals between Group 2 and Group 3 by the Mantel-Cox test (P = 0.0003). (e) The cumulative parasitemia up to day 14, the day on which the first animal in Group 2 was treated, between Groups 2 and Group 3 by the Mann–Whitney test (P = 0.012)

Fig. 2

Vaccination with AMA1–RON2L complex induces a shift in the quality of blocking antibody. a FVO AMA1-specific antibody levels in plasma and purified IgG. ELISA was performed on individual samples collected before the day of challenge. Arbitrary ELISA units based on standard curves were generated to compare anti-FVO AMA1 antibody levels in plasma and purified IgG of Group 2 and Group 3 animals by the Mann–Whitney test (plasma: P = 0.854; IgG: P = 0.045). Data are shown for individual animals and represented as mean ± SEM. (b) In vitro GIA using purified IgG against the homologous FVO strain. Total IgG from each animal was tested at 2.5 mg mL⁻¹ final concentration and inhibitory activity was compared between Group 2 and 3 by Mann–Whitney test (P = 0.0006). Data shown are from two independent experiments represented as mean ± SEM. Antibody data are from plasma samples collected 4 weeks after last vaccination (before parasite challenge)

Fig. 3

Complex-induced enhancement in antibody quality is associated with protection. a The relationship between time-to-patency and in vitro growth inhibitory activity of purified IgG for the 16 animals in Groups 2 (blue) and 3 (red). Spearman’s rank correlation coefficient (r _s) and P value are shown for the relationship of time-to-patency and in vitro growth inhibitory activity of purified IgG from the same animals. The horizontal dotted line represents the 50% GIA and the vertical dotted line separates the animals that either had a significant delay in patency (>15 days) or were SP until the end of the study. b Anti-PfAMA1(FVO) antibody levels do not correlate with in vitro growth inhibitory activity (r _s = −0.082, P = 0.76). c The relationship between anti-PfAMA1(FVO) antibody levels and time-to-patency shows no correlation (r _s = −0.008, P = 0.94). GIAs were performed at 2.5 mg mL⁻¹ total IgG from each immunized animal

Fig. 4

Levels of AMA1-RON2 blocking antibodies correlate with protection. a The relationship between levels of AMA1-RON2 blocking antibodies (IC₅₀ Log10[EU]) and in vitro neutralization activity of the corresponding purified IgG. b The relationship between levels of AMA1-RON2 blocking antibodies (IC₅₀ Log10[EU]) in the plasma and time to infection after challenge in these animals. Spearman’s rank correlation coefficient (r _s) and P value are shown for each comparison. c Growth inhibitory antibodies largely target conformational epitopes in AMA1. The conformational and allele-specific dependency of antibodies to block invasion was assessed by measuring the ability of recombinant AMA1 to block the inhibitory activity of IgG. Saturating concentrations (2 μM) of recombinant FVO AMA1 (rFVO), reduced and alkylated FVO AMA1 (RArFVO) or 3D7 AMA1 (r3D7) were pre-incubated with IgG before assessing their GIA activity against FVO strain parasites. The amount of IgG from each animal was chosen such that they had 40–50% GIA before recombinant proteins were added and four or more animals each from Groups 2 and 3 were tested. ns, not significant

Fig. 5

AMA1–RON2L complex induces an increase in the proportion of antibodies targeting conserved epitopes. a 3D7 AMA1-specific antibody levels in purified IgG were measured by ELISA. Arbitrary ELISA units based on standard curves were generated to compare anti-3D7 AMA1 antibody levels in purified IgG of Group 2 and Group 3 animals by Mann–Whitney test (P = 0.003). b In vitro GIA using purified IgG against the heterologous 3D7 strain. Total IgG from each animal were tested at 2.5 mg mL⁻¹ final concentration and inhibitory activity was compared between Group 2 and 3 by the Mann–Whitney test (P = 0.02). c Correlation of growth inhibitory activity of IgG from Group 3 (red squares) and Group 2 (blue circles) between homologous FVO parasites and heterologous 3D7 parasites (blue circles) (r _s = 0.89, P < 0.0001). d Structural representation of AMA1 (brown) bound to RON2L (cyan) (PDB ID: 3ZWZ). The polymorphic residues that are conserved between FVO, 3D7 and GB4 parasites are shown in dark blue and residues that differ are shown in red. The black lines indicate the respective loops in AMA1 (Ib, Ic, Id, Ie and If) surrounding the RON2 binding site