Infection-Induced Resistance to Experimental Cerebral Malaria Is Dependent Upon Secreted Antibody-Mediated Inhibition of Pathogenic CD8+ T Cell Responses

Abstract

Cerebral malaria (CM) is one of the most severe complications of Plasmodium falciparum infection. There is evidence that repeated parasite exposure promotes resistance against CM. However, the immunological basis of this infection-induced resistance remains poorly understood. Here, utilizing the Plasmodium berghei ANKA (PbA) model of experimental cerebral malaria (ECM), we show that three rounds of infection and drug-cure protects against the development of ECM during a subsequent fourth (4X) infection. Exposure-induced resistance was associated with specific suppression of CD8⁺ T cell activation and CTL-related pathways, which corresponded with the development of heterogeneous atypical B cell populations as well as the gradual infection-induced generation and maintenance of high levels of anti-parasite IgG. Mechanistically, transfer of high-titer anti-parasite IgG did not protect 1X infected mice against ECM and depletion of atypical and regulatory B cells during 4X infection failed to abrogate infection-induced resistance to ECM. However, IgMi mice that were unable to produce secreted antibody, or undergo class switching, during the repeated rounds of infection failed to develop resistance against ECM. The failure of infection-induced protection in IgMi mice was associated with impaired development of atypical B cell populations and the inability to suppress pathogenic CD8⁺ T cell responses. Our results, therefore, suggest the importance of anti-parasite antibody responses, gradually acquired, and maintained through repeated Plasmodium infections, for modulating the B cell compartment and eventually suppressing memory CD8⁺ T cell reactivation to establish infection-induced resistance to ECM.

Keywords: B cells; T cells; antibody; brain; cerebral malaria; spleen.

Figure 1

Three rounds of infection-drug cure promote resistance to ECM in susceptible C57BL/6 mice. (A) Schematic of the experimental design. C57BL/6 mice were infected with PbA (10⁴ pRBCs i.v.) or left uninfected. Mice were treated (i.p.) with chloroquine and artesunate on stated days post each infection, and re-infections were performed after a minimum interval of 30 days following cessation of drug treatment. (B) Kinetics of ECM development shown as percentage of survival of mice. N = 7–10 per group, pooled from two independent experiments. (C) Peripheral parasitaemia (% of pRBCs) ± SD in different infection groups. N = 5–9 per group. Results are pooled from two experiments for the 1X, 2X, and 3X infection and from 3 experiments for the 4X infection. Kruskal-Wallis test with Dunn's multiple comparisons test was used for statistical analysis p < 0.05 denoted by #, for 1X v 3X, ‡ for 1X v 4X, and ~ for 2X v 4X. (D) Representative H & E stained cortical brain sections demonstrating presence (in 1X infected mice) and absence (in 4X infected mice) of cerebrovascular pRBCs (filled arrow), hemorrhage (unfilled arrow) and leukocyte packed vessels (arrow head) in 1 and 4X infected mice on day 8 post infection. Scale bar 100 μm.

Figure 2

4X PbA infected mice have a distinct whole brain transcriptional signature (A–D) Perfused whole brains were removed from 4X infected and age-matched 1X infected C57BL/6 mice on day 8 p.i. (when 1X developed ECM), and age-matched naïve mice. (A) Representative flow cytometric plots showing identification of and (B) numbers of CD8⁺ T cells in the brain (mean ± S.D.). (C) Representative flow cytometric plots showing granzyme B expression by and (D) frequencies of granzyme B expressing CD8⁺ T cells in the brain (mean ± S.D.). (A,C) Numbers denote the percentage of cells within the gate. (B,D) N = 4–8 per group, pooled from two independent experiments. Statistical analyses were performed with Kruskal-Wallis test with Dunn's multiple comparisons test or with one-way ANOVA with Tukey's test, depending on normality of data (^*p ≤ 0.05 and ^***p ≤ 0.001). (E–G) Microarray analysis was performed on perfused whole brains from 4X infected and age-matched 1X infected C57BL/6 mice on day 8 p.i. (when 1X developed ECM), and age-matched naïve mice. (E) Principal components analysis of whole-brain transcriptomes. N = 6 per group. Results are generated from the pooled array data from brains taken from two independent experiments. (F) Venn diagrams defining unique and overlapping genes differentially expressed between 4X infected vs. uninfected mice and 1X infected mice vs. uninfected mice. (G) Genes in (F) and Supplementary Figure 2 were filtered to identify genes differentially expressed in brains of 1 and 4X infected mice on day 8 of infection. Heat maps showing differentially expressed genes grouped by enriched biological processes identified within DAVID bioinformatics database. Results (n = 6 per group) are generated from the pooled array data from brains taken from two independent experiments. (H,I) Perfused whole brains were removed from pre-4X infected mice (minimum 30 days post clearance of 3X infection) and age-matched uninfected mice for microarray analysis. (H) Principal components analysis of whole-brain transcriptomes (n = 6 per group). Results are generated from the pooled array data from brains taken from two independent experiments. (I) Scatter plot comparing gene expression between pre-4X infected mice and uninfected mice. Each data point represents the mean expression level of a gene. Results (n = 6 per group) are generated from the pooled array data from brains taken from two independent experiments.

Figure 3

4X infected mice develop significantly altered splenic CD8⁺ and CD4⁺ T cell responses. (A–H) Spleens and plasma were removed from 4X infected and age-matched 1X infected C57BL/6 mice on day 8 post-infection (when 1X infected mice developed ECM), from pre-4X infected C57BL/6 mice (minimum 30 days post clearance of 3X infection), and age-matched uninfected mice. (A) Representative flow cytometric plots and the frequencies of splenic CD8⁺ T cells and CD4⁺ T cells, gated on live CD45⁺ CD3⁺ cells. Total numbers of splenic (B) CD8⁺ T cells, (C) CD4⁺ T cells and (D) CD45+ cells. (E,F) Representative flow cytometric plots and the frequencies of splenic (E) CD8⁺ T cells and (F) CD4⁺ T cells expressing markers of activation, function and differentiation. (G) Nanostring analysis of gene expression in whole spleen tissue from 4X and 1X infected mice (on day 8 of infection), expressed relative to gene expression in age-matched uninfected mice (fold change of 1 defines level of gene expression in uninfected mice). (H) Cytokine bead array of plasma cytokine levels in 4X, 1X infected mice and aged matched uninfected C57BL/6 mice. (A,E,F) Numbers denote the percentage of cells within the gate. (B–G) Results are the mean ± SD of the group, representative of two independent experiments with (B–D) n = 3–6 per group, (E,F) n = 3–10 per group and (G) n = 3 per group. (H) N = 4–7 per group, pooled from two independent experiments. (B–F,H) Statistical analyses were performed with Kruskal-Wallis test with Dunn's multiple comparisons test or with one-way ANOVA with Tukey's test, depending on normality of data (^*p ≤ 0.05, ^**p ≤ 0.01, ^***p ≤ 0.001, ^****p ≤ 0.0001). (G) Statistical analysis by Student's t-test, with Welch's correction (^*p ≤ 0.05).

Figure 4

Anti-parasite IgG titres and atypical B cell populations are significantly increased in 4X infected mice. (A–G) Spleens and plasma were obtained from 4X infected and age-matched 1X infected C57BL/6 mice on day 8 post-infection (when 1X infected mice developed ECM), from pre 4X infected mice (minimum 30 days post-clearance of 3X infection) and from age-matched uninfected mice. (A) Representative flow cytometric plots and the frequencies of splenic CD19⁺ B cells, gated on live CD45⁺ CD3⁻ cells. Numbers denote the percentage of cells within the gate. (B) Total numbers of splenic CD19⁺ B cells. (C) Expression of CD138, CD80, CD11c, PDCA1, PDL-1, and MHC II by CD19⁺ splenic B cells from uninfected, 1X infected and 4X infected mice. (D) Left, Representative flow cytometric plots (gated on live CD45⁺CD3⁻CD19⁺ cells), right, the frequencies and numbers of non-switched (IgM⁺IgD⁺ and IgM⁻IgD⁺) and switched (IgM⁺IgD⁻, and IgM⁻IgD⁻) splenic CD19⁺ B cells. The end point plasma titres of (E) anti- PbMSP1₁₉ and (F) anti-PbA IgG, with background titer from naïve mice subtracted. (G) Graphical representation of the fold change in Ab titer between 1X and 2X infected mice and between 2X and 4X infected mice. (B,D) N = 2–6 per group, representative of two independent experiments (mean ± SD). Statistical analyses were performed with Kruskal-Wallis test with Dunn's multiple comparisons test ^*p ≤ 0.05, ^**p ≤ 0.01, and p < 0.05 denoted by #, ~, ^†for indicated groups. (E,F) N = 6 per group, pooled from two independent experiments (mean ± SD). Statistical analyses were performed by one-way ANOVA with Tukey's multiple comparisons test, shown for 4X infected mice compared to each other group (^****p ≤ 0.0001).

Figure 5

Ablation of B cell populations or transitory boosting of anti-parasite antibody does not alter dynamics of ECM in 1X or 4X infected mice. (A–G) C57BL/6 mice were injected (i.p) one day prior to 4X infection and on days 2, 5, 8, 11 of infection, with either (250 μg) anti-CD20 mAb or (250 μg) control anti-ragweed mAb. (A) Total numbers of splenic CD19⁺ B cells. (B) Left, Representative flow cytometric plots (gated on live CD45⁺CD3⁻CD19⁺ cells), right, the frequencies of non-switched (IgM⁺IgD⁺ and IgM⁻IgD⁺) and switched (IgM⁺IgD⁻, and IgM⁻IgD⁻) splenic CD19⁺ B cells. (C) Total numbers of splenic CD19⁻ CD138⁺ plasma B cells and splenic CD19⁺ B cells expressing CD138, CD80, CD11c, PDCA1, PDL-1, and MHC II in 4X infected mice that received anti-CD20 mAb or anti-ragweed mAb. (D,E) The frequencies of Granzyme B expressing CD8⁺ T cells in the (D) spleen and (E) brain in 4X infected mice that received anti-CD20 mAb or anti-ragweed mAb. (F) Survival of 4X infected mice given anti-CD20mAb or control mAb. (G) Parasitaemia (% of pRBCs) of 4X infected mice given anti-CD20mAb or control mAb. (H–K) C57BL/6 mice received 500 μl of heat-inactivated plasma (obtained from 4X infected mice or, as a control, from age-matched naïve mice) prior to and during 1X infection (4 i.p. injections between day −1 and day 5 of infection). (H) Peripheral parasitaemia (% of pRBCs) in 4X infected mice that received plasma from 4X or uninfected mice (I) Kinetics of ECM development shown as percentage survival of 1X infected mice that received plasma from 4X or uninfected mice. (J,K) The frequencies of granzyme B expressing CD8⁺ T cells in the (J) spleen and (K) brain in 1X infected mice that received plasma from 4X or naïve mice. (A–K) Results are the mean ± SD of the group. (A–C) N = 3–4 per group, representative of two independent experiments. (D,E) N = 7–8 per group, pooled from two independent experiments. (F, G) N = 10–13 per group, pooled from two independent experiments. (H, I) N = 6–7 per group, pooled from two independent experiments. (J,K) N = 4, representative of two independent experiments. Statistical analyses were performed with by Student's t-test, with Welch's correction or Mann-Whitney test depending on normality of data (^*p ≤ 0.05, ^**p ≤ 0.01 ^***p ≤ 0.001 and p < 0.05 denoted by #, ~, ‡ for indicated groups).

Figure 6

IgMi mice do not develop infection-induced resistance to ECM. (A–F) IgMi mice and WT littermate controls were infected with PbA (10⁴ pRBCs i.v.) or left uninfected. Mice were treated (i.p.) with chloroquine and artesunate from day 5 or 6 post each infection, and re-infections were performed after a minimum interval of 30 days following cessation of drug treatment. (A) Kinetics of ECM development, shown as percentage of survival of mice. (B) Peripheral parasitaemia (% of pRBCs). (C) Left, Representative flow cytometric plots (gated on live CD45⁺CD3⁻CD19⁺ cells), right, the frequencies of non-switched (IgM⁺IgD⁺ and IgM⁻IgD⁺) and switched (IgM⁺IgD⁻, and IgM⁻IgD⁻) splenic CD19⁺ B cells. (D) Expression of CD138, CD80, CD11c, PDCA1, PDL-1 and MHC II by CD19⁺ splenic B cells from 1X infected and 4X infected WT and IgMi mice. (E,F) Activation phenotype of (E) splenic and (F) brain accumulating CD8⁺ T cells. (B,C,E,F) Results are the mean ± SD of the group. (A,B) N = 7–8 per group, pooled from two independent experiments. (C) N = 3–4 per group, representative of two independent experiments. (E,F) N = 5–8 per group, pooled from two independent experiments. (B,E,F) Statistical analyses were performed with Kruskal-Wallis test with Dunn's multiple comparisons test or with one-way ANOVA with Tukey's test depending on normality of data (^*p ≤ 0.05, ^**p ≤ 0.01, ^***p ≤ 0.001, ^****p ≤ 0.0001, and p < 0.05 denoted as follows ^*, WT 1X v IgMi 1X, ~, WT 4X v IgMi 1X, #, WT 4X v IgMi 4X, †, IgMi 1X v IgMi 4X). (C) Statistical analysis by Student's t-test, with Welch's correction for indicated groups (p < 0.05 denoted by #, ~, ‡ for the indicated groups).