Protective CD8+ T-cell immunity to human malaria induced by chimpanzee adenovirus-MVA immunisation

Abstract

Induction of antigen-specific CD8(+) T cells offers the prospect of immunization against many infectious diseases, but no subunit vaccine has induced CD8(+) T cells that correlate with efficacy in humans. Here we demonstrate that a replication-deficient chimpanzee adenovirus vector followed by a modified vaccinia virus Ankara booster induces exceptionally high frequency T-cell responses (median >2400 SFC/10(6) peripheral blood mononuclear cells) to the liver-stage Plasmodium falciparum malaria antigen ME-TRAP. It induces sterile protective efficacy against heterologous strain sporozoites in three vaccinees (3/14, 21%), and delays time to patency through substantial reduction of liver-stage parasite burden in five more (5/14, 36%), P=0.008 compared with controls. The frequency of monofunctional interferon-γ-producing CD8(+) T cells, but not antibodies, correlates with sterile protection and delay in time to patency (P(corrected)=0.005). Vaccine-induced CD8(+) T cells provide protection against human malaria, suggesting that a major limitation of previous vaccination approaches has been the insufficient magnitude of induced T cells.

Trial registration: ClinicalTrials.gov NCT00890760.

Figure 1. Immunogenicity of prime-boost vaccination with ChAd63-MVA ME-TRAP.

(a) Median ME-TRAP IFNγ ex vivo ELISPOT responses to the T9/96 strain for each group, P=0.04 for peak immunogenicity (D28 for A (n=10) and D63 for AM (n=14)), *P=0.012 for difference in immunogenicity at D166, both two-tailed Mann–Whitney. (b) Breadth of TRAP-specific ELISPOT responses before (D14, D56) and after boosting (D63) with MVA ME-TRAP, using T9/96 strain peptides, showing mean (±s.e.m.) number of pools recognized. (c) Correlation of peak ELISPOT responses to T9/96 TRAP post-prime and at time of sporozoite challenge. (d) Individual responses to TRAP and ME at time of sporozoite challenge (data points in red represent sterilely protected volunteers). (e) Median with IQR IFNγ ELISPOT responses to TRAP peptides from the heterologous vaccine and challenge strains of P. falciparum. (f) IgG antibodies to TRAP boosted by MVA compared with ChAd63 alone (medians); differences not statistically significant. (g) Spearman’s correlation of pre-existing antibodies to ChAd63 with peak vaccine-induced ELISPOT responses to TRAP (T9/96 strain).

Figure 2. Vaccine efficacy assessments by time to microscopic patency and PCR measures.

Kaplan–Meier log-rank comparison of days to positive blood film (a) and >20 parasites per ml by PCR (b) between adenovirus-MVA prime-boost (Ad-M) group (n=14), adenovirus-only (Ad) group (n=10) and controls (n=12). Mean days to positivity by blood film: Ad-M 14.6 days (95% CI: 12.3–16.8), adenovirus-alone group (Ad) 11.3 days (95% CI: 10.2–12.5), control group 11.8 days (95% CI: 10.8–12.7); and to 20 parasites per ml by PCR: Ad-M 11.6 days (95% CI: 8.5–14.7), Ad 7.8 days (95% CI: 7.0–8.6), control group 8.1 days (95% CI: 4–8.8). (c) Group mean log-transformed PCR data (error bars represent s.e.m.). Area under curve analysis of parasite densities comparing controls to vaccinees (either including (d) or excluding (e) volunteers that were sterilely protected) at days 6.5–8 (first cycle post hepatocyte release), days 8.5–10 (second cycle) and 10.5–12 (third cycle) post challenge. Over the second and third cycles, there is a significant reduction in vaccinees’ parasite densities compared with controls; the lack of significance at cycle one probably reflects low power due to very low parasite densities. Comparison of areas under curves for all three cycles combined also shows a significant reduction in parasite densities between vaccinees and controls (P=0.003 when including sterilely protected vaccinees, P=0.01 when excluding sterilely protected vaccinees, (two-tailed t-test).

Figure 3. Functionality of TRAP-specific CD4⁺ and CD8⁺ T cells.

(a) Peak immunogenicity (day 63=7 days after boosting) and (b) time of challenge (day 76). N=10 for adenovirus-only group and n=14 for Ad-M group. Frequencies of cytokine-secreting CD4⁺ and CD8⁺ T cells. *Staining for CD107a⁺ expression performed at challenge B only. Mean (with s.e.m.) responses are shown. (c,d) T cell responses shown are grouped according to number of functions at peak (c) and challenge (d). Pie charts summarize the fractions of the total response that are positive for three, two or one functions. All possible combinations of functions are shown in the bar chart stratified by vaccination regimen, with the y axis showing the percentage of CD4⁺ or CD8⁺ cells. Data points represent individual volunteers.

Figure 4. Correlates of protective efficacy.

(a) Correlation of time to parasitaemia with ex vivo ELISPOT responses for Ad-MVA vaccinees, P=0.97. (b) Correlation of time to parasitaemia with frequency of CD8⁺ IFNγ⁺/IL-2⁻ /TNFα⁻ for Ad-MVA vaccinees, P=0.0005. For vaccinees receiving Ad alone or Ad-MVA (n=24), correlation of time to parasitaemia with frequency of CD8⁺ IFNγ⁺/IL-2⁻ /TNFα⁻, r_s=0.61, P=0.002. Both (a) and (b) were assessed at the time of sporozoite challenge. (c) Correlation of time to parasitaemia with frequency of CD107a⁺/ IFNγ⁻/IL-2⁻/TNFα⁻ CD8⁺ T cells at day 150 post challenge in challenge A, P=0.02. (d) Correlation of parasite density in second replication cycle with frequency CD8⁺ IFNγ⁺/IL-2⁻ /TNFα⁻ for Ad-MVA vaccinees. All correlations were performed using two-tailed Spearman’s correlation.