Identification of highly-protective combinations of Plasmodium vivax recombinant proteins for vaccine development

Abstract

The study of antigenic targets of naturally-acquired immunity is essential to identify and prioritize antigens for further functional characterization. We measured total IgG antibodies to 38 P. vivax antigens, investigating their relationship with prospective risk of malaria in a cohort of 1-3 years old Papua New Guinean children. Using simulated annealing algorithms, the potential protective efficacy of antibodies to multiple antigen-combinations, and the antibody thresholds associated with protection were investigated for the first time. High antibody levels to multiple known and newly identified proteins were strongly associated with protection (IRR 0.44-0.74, p<0.001-0.041). Among five-antigen combinations with the strongest protective effect (>90%), EBP, DBPII, RBP1a, CyRPA, and PVX_081550 were most frequently identified; several of them requiring very low antibody levels to show a protective association. These data identify individual antigens that should be prioritized for further functional testing and establish a clear path to testing a multicomponent P. vivax vaccine.

Keywords: IgG antibody; Plasmodium vivax; clinical malaria; epidemiology; global health; natural immunity; protection; vaccine.

Figure 1.. Breadth of IgG antibodies to 38 P. vivax proteins in Papua New Guinean children aged 1–3 years.

For each protein, antibody levels were stratified into tertiles and scored as 0, 1 or 2 for the low, medium, and high tertiles, respectively. Scores were then added up to reflect the breadth of antibodies per child. (a) Antibody breadth by age group. Age is shown as median (interquartile range). (b) Antibody breadth by lifetime exposure group. For each child, exposure was defined as the total number of P. vivax blood-stage clones acquired per time-at-risk (molFOB), and lifetime exposure as a product of age and molFOB. P values are from negative binomial regression and were deemed significant if <0.05.

Figure 2.. Association between cumulative IgG levels to 38 P. vivax proteins and exposure to P. vivax infections in Papua New Guinean children aged 1–3 years.

(a) Association with age. (b) Association with lifetime exposure. For each child, exposure was defined as the total number of P. vivax blood-stage clones acquired per time-at-risk (molFOB), and lifetime exposure as a product of age and molFOB. n = 225. P values < 0.05 were deemed significant.

Figure 3.. Association between high IgG levels to 38 P. vivax proteins and protection against clinical malaria (density >500/μL) in Papua New Guinean children aged 1–3 years old.

Data are plotted as incidence rate ratios and 95% confidence intervals adjusted for exposure (molFOB), age, season, and village of residency. Incidence rate ratios are for high versus low tertiles of responders, 95% confidence intervals and P values are from GEE models. P values < 0.05 were deemed significant.

Figure 4.. Correlations between IgG to 38 P. vivax proteins in Papua New Guinean children aged 1–3 years.

Correlation coefficients between antibody levels to every pair of antigens were calculated using Spearman’s rank correlation tests.

Figure 5.. Association between antibodies to combinations of P. vivax proteins and malaria risk in Papua New Guinean children aged 1–3 years.

(a) Potential protective efficacy (PPE) for combinations of antigens with the maximum number of antigens indicated on the x-axis. Dashed lines represent the range of PPE for all possible combinations. Solid lines represent the range of PPE from 1000 implementations of the simulated annealing algorithm. (b) Sum of residuals (as a measure of model goodness of fit). (c) The heatmap shows the frequency of including an antigen (x-axis) in a multi-component vaccine with a fixed number of antigens (y-axis). (d) Predicted PPE of a single antigen. (e) Immunogenicity of each antigen represented as seroprevalence with a cut-off as 10% of the antibody levels of fully-immune PNG adults.

Figure 6.. Estimated dose-response curves for the associations between antibody responses specific to P. vivax antigens and protection from clinical malaria.

Solid black lines depict exposure-adjusted dose-response curves, and the grey shaded regions depict the 95% credible intervals. Histograms show the observed distribution of antibody levels (relative to the PNG immune pool) colored per tertiles (low = blue; medium = green; high = red), and the darkly-colored portions of the histograms show the proportion of individuals with that antibody level who had a P. vivax episode (>500 parasites/μL). (a–c) Examples of antigens that need low antibody levels to provide 50% of protection. (d–f) Examples of antigens that need medium antibody levels to provide 50% of protection. (g–i) Examples of antigens that need high antibody levels to provide 50% of protection.

Figure 6—figure supplement 1.. Estimated dose-response curves for the associations between antibody levels to P. vivax antigens and protection from clinical malaria.

Solid black lines depict exposure-adjusted dose-response curves, and the grey shaded regions depict the 95% credible intervals. Histograms show the observed distribution of antibody levels (arbitrary antibody units relative to standard curves made of immune pooled serum), and the darkly-colored portions of the histograms show the proportion of individuals with that antibody levels who had a P. vivax episode (>500 parasites/μL). (a) Antigens that need low antibody levels to show an association with 50% of protection against clinical malaria. (b) Antigens that need medium antibody levels to show an association with 50% of protection against clinical malaria. (c) Antigens that need high/very high antibody levels to show an association with 50% of protection against clinical malaria.