The Liver-Stage Plasmodium Infection Is a Critical Checkpoint for Development of Experimental Cerebral Malaria

Abstract

Cerebral malaria is a life-threatening complication of malaria in humans, and the underlying pathogenic mechanisms are widely analyzed in a murine model of experimental cerebral malaria (ECM). Here, we show abrogation of ECM by hemocoel sporozoite-induced infection of a transgenic Plasmodium berghei line that overexpresses profilin, whereas these parasites remain fully virulent in transfusion-mediated blood infection. We, thus, demonstrate the importance of the clinically silent liver-stage infection for modulating the onset of ECM. Even though both parasites triggered comparable splenic immune cell expansion and accumulation of antigen-experienced CD8⁺ T cells in the brain, infection with transgenic sporozoites did not lead to cerebral vascular damages and suppressed the recruitment of overall lymphocyte populations. Strikingly, infection with the transgenic strain led to maintenance of CD115⁺Ly6C⁺ monocytes, which disappear in infected animals prone to ECM. An early induction of IL-10, IL-12p70, IL-6, and TNF at the time when parasites emerge from the liver might lead to a diminished induction of hepatic immunity. Collectively, our study reveals the essential role of early host interactions in the liver that may dampen the subsequent pro-inflammatory immune responses and influence the occurrence of ECM, highlighting a novel checkpoint in this fatal pathology.

Keywords: Plasmodium; cerebral malaria; experimental cerebral malaria; liver-stage; malaria; pre-erythrocytic stage; sporozoites.

Figure 1

Improved pre-erythrocytic development of PRF parasites. (A) Prepatent period of sporozoite-induced infections. Appearance of blood-stage parasites was monitored by daily microscopic examination of Giemsa-stained blood films. C57BL/6 mice were infected by intravenous injection of 5,000 WT or PRF hemocoel sporozoites (n = 22 each). Shown is a Kaplan-Maier analysis of time to first detection of blood infection, ^***P < 0.001 (Mantel-Cox test). (B) In vivo quantification of parasite loads in the liver of infected mice. Livers were harvested 42 h after infection of C57BL/6 mice by intravenous injection of 5,000 WT or PRF hemocoel sporozoites. Expression levels of P. berghei 18S rRNA were quantified by real-time RT-PCR and normalized to mouse GAPDH. Results represent mean values (± SEM) (n = 8 each for infected mice; n = 3 for naïve mice). Differences between WT- and PRF-infected livers were non-significant (Mann-Whitney test). (C) Sporozoite cell traversal. Hepatoma cells were incubated for 30 min or 2 h with medium (white; Con), FITC-dextran only (dotted line), and FITC-dextran together with either WT (blue; WT), or PRF (red; PRF) hemocoel sporozoites. Cells were analyzed by flow cytometry to enumerate the percentage of dextran-positive cells indicative of sporozoite traversal. Results represent mean values (±SD) of at least three independent experiments with duplicates each. ^*P < 0.05 (Mann-Whitney test).

Figure 2

Full virulence after PRF blood infection, but very low ECM pathology after sporozoite infection. (A) Time course of blood infection after sporozoite inoculation or transfusion of iRBCs. C57BL/6 mice were infected by intravenous injection of 5,000 hemocoel sporozoites (solid lines) of either wild type (WT; blue) or PRF (red) parasites (n = 22 each) or 5,000 iRBCs (dashed lines) of WT (blue) or PRF (red) parasites (n = 15 each). Parasitemia was monitored daily by microscopic examination of Giemsa stained blood-films. Differences between asexual blood-stage propagation were non-significant (Mantel-Cox test). (B) Kaplan-Meier analysis of time to development of signature symptoms of experimental cerebral malaria (ECM). ^***P < 0.001 (Mantel-Cox test). (C) Quantification of parasite loads in the brain (square), BM (polygon) and liver (triangle) was done by qPCR. Organs were harvested 6 days after infection of C57BL/6 mice with 5,000 WT or PRF sporozoites. Animals were perfused via the left heart ventricle to remove non-sequestering infected red blood cells from the circulation of blood. Relative expression levels of P. berghei 18S rRNA were normalized to mouse GAPDH. Results represent mean values (±SD). ^*P < 0.05 (Mann-Whitney test). (D) Visualization of integrity of blood-brain barrier in infected mice. Infected mice were intravenously injected with 2% Evans blue dye 8 days after infection by 5,000 WT sporozoites (upper left), PRF sporozoites (upper right), WT- iRBCs (bottom left), and PRF-iRBCs (bottom right). (E) Vascular leakage also occurs in the spinal cords of mice with ECM symptoms. Isolated spinal cords are indicated by arrows and show leakage of the dye in mice with signature ECM symptoms.

Figure 3

Cerebral vessels from mice infected by PRF sporozoites maintain a healthy status. (A) Hematoxilin and eosin (H&E) stain of a brain tissue section from an animal displaying signature ECM symptoms. An arrow indicates RBCs that leaked into the cerebral parenchyma. Scale bar, 10 μm. (B) Quantification of hemorrhage sites based on H&E stains of horizontal cross-sections of the brain (n = 10 each). Results represent mean values (±SD) from at least two independent experiments. ^***P < 0.001 (Mann-Whitney test). (C) Visualization of brain vessel integrity. Shown are H&E stains of representative brain histological cross-sections. Comparison of a healthy vessel (left) from a mouse infected with PRF sporozoites and a collapsed vessel (right) from a mouse infected with WT sporozoites. The latter vessel also contains accumulated leukocytes. Scale bar, 10 μm. (D) Morphological scoring of cerebral vessels based on histological sections. Quantification of mice infected with 5,000 WT sporozoites, PRF sporozoites, WT iRBCs or PRF iRBCs (n = 10 animals for infected groups, n = 3 mice for naïve group; >20 cerebral vessels were quantified for each animal).

Figure 4

Quantification of total and antigen-experienced CD8⁺ T cells in the brains of infected mice. (A) Time course of total cerebral lymphocyte counts during the late stages of blood infection. Mice were infected with WT (blue) or PRF (red) sporozoites by intravenous injection (n = 6 each), and cerebral lymphocytes were isolated and analyzed at indicated time points. Shown are total CD8⁺ T cell numbers. The dotted line represents the mean cerebral CD8⁺ T cell number in all naïve animals (n = 16). The scatter dot plots represent mean values (±SD). ^**P < 0.01 (Mann-Whitney test). (B) Gating strategy for detection of intracellular IFN-γ expression in CD3⁺CD8⁺ cells after re-stimulation with PbGAP50₄₀₋₄₈ peptide. Shown are representative plots for naïve animals, as well as mice infected with WT or PRF sporozoites. (C) Quantification of cerebral CD8⁺ T cells expressing intracellular IFN-γ after re-stimulation with PbGAP50₄₀₋₄₈ peptide. Shown are numbers of this population in the brain. The scatter dot plots represent mean values (±SD) from samples isolated on day 8 from three independent experiments (n = 10 each for infected mice; n = 4 for naïve mice). IFN-γ-secreting antigen-specific CD8⁺ T cells number from WT vs. PRF infected animals is non-significant. ^**P < 0.01 (Mann-Whitney test).

Figure 5

Quantification of myeloid cells after infection. (A) Representative contour plots from day 8 after infection showing CD45 vs. CD11b expression on live cells obtained from naïve mice (left), and mice infected with WT (center) and PRF (right) sporozoites. Indicated are gated populations of lymphocytes, CD45^hiCD11b⁻ cells (gray circles) and myeloid cells, CD45^med/loCD11b⁺ cells (red circles). (B) CD45 vs. CD11b expression on the Ly6G⁻ myeloid population. CD45^loCD11b⁺ cells (orange circles) represent microglia and CD45^medCD11b⁺ cells (green circles) activated microglia, monocytes and macrophages, respectively. (C) Quantification of CD45^lo microglia. Shown are percentage (left) and absolute numbers (right) of CD45^loLy6G⁻CD11b⁺ microglia. (D) Quantification of CD45^medLy6G⁻CD11b⁺ cells, which correspond to CD45^med microglia, monocytes and macrophages. Shown are percentage (left) and absolute numbers (right) of CD45^med microglia/monocytes/macrophages in the brain. The scatter dot plots in (C,D) represent mean values (±SD) from samples (n = 4–7) isolated 8 days after infection from two independent experiments. n.s, non-significant; ^*P < 0.05; ^**P < 0.01 (Mann-Whitney test). (E–G) Representative images from the IBA-1⁺ microglial cells of brain histological cross-sections. Cerebral vessels are indicated by arrows. The IBA-1 staining reaction was visualized with diaminobenzidine (DAB), highlighting some microglial cells with thin processes in naïve mice, while they are more prominent in WT sporozoite-infected mice and they also tend to cluster around vessels. This feature is also pronounced in PRF sporozoite-infected mice, dark brown. Scale bar, 50 μm.

Figure 6

Reduced CD115 expressions on splenic Ly6C⁺CD11b⁺ monocytes in mice infected with WT sporozoites during the development of ECM. (A) Representative contour plots showing gating of Ly6C vs. CD115 expression on CD45⁺Ly6G⁻CD11b⁺ myeloid cells from naïve mice (left), and mice infected with WT (center) and PRF (right) sporozoites 8 days after infection. An arrow indicates the disappearance of CD115⁺Ly6C⁺ monocytes (blue circles) in spleens of mice infected with WT sporozoites. CD115⁻Ly6C⁺ monocytes are gated in gray circles. (B) Quantification of splenic CD115⁺Ly6C⁺ monocytes. Shown are percentage and numbers of CD115⁺Ly6C⁺ monocytes in the spleen. (C) Quantification of splenic CD115⁻Ly6C⁺ cells. Shown are percentage and numbers of CD115⁻Ly6C⁺ myeloid cells in the spleen. The scatter dot plots in (B,C) represent mean values (±SD) from samples (n = 8–11) isolated 8 days after infection from two independent experiments. n.s, non-significant; ^*P < 0.05; ^**P < 0.01; ^***P < 0.001 (Mann-Whitney test).

Figure 7

Serum cytokine levels are up-regulated 3 days after infection with PRF sporozoites, at the onset of blood infection. (A–F) Serum from peripheral blood was collected on days indicated after infection with 5,000 WT (blue triangles) or PRF (red circles) sporozoites. Systemic cytokines were captured using a Cytokine Bead Array and analyzed by flow cytometry. Day 0 results represent data from uninfected mice. Shown are the steady-state levels of the systemic cytokines (A) IL-10, (B) IL-12p70, (C) IL-6, (D) TNF, (E) IFN-γ, and (F) MCP-1. The results represent mean values (±SD) (n = 5 for WT, n = 4 each for PRF). ^*P < 0.05 (Mann-Whitney test).

Figure 8

PRF sporozoites do not induce more liver-stage antigen specific CD8⁺ T cells compared to WT sporozoites upon immunization. (A) Immunization and challenge protocol. C57BL/6 mice were immunized twice at weekly intervals with 10,000 irradiated hemocoel sporozoites (10 k γ-spz). Animals were challenged with 10,000 WT salivary gland sporozoites (10 k spz) 14 days after the last immunization. (B) Kaplan-Meier analysis of time to blood infection. Naïve mice (n = 9), WT and PRF sporozoite-immunized mice (n = 12 each). Blood parasitemia was determined by daily microscopic examination of Giemsa-stained blood films. The statistics for WT vs. PRF sporozoite-immunized mice were non-significant for Log-rank (Mantel-Cox) test. (C) Enumeration of total liver CD8⁺ T cell numbers after 14 days from the last immunization. Livers were isolated for flow cytometric analysis from naïve and WT and PRF sporozoite-immunized mice (n = 5). The results represent mean values (±SD). (D) Quantification of IFN-γ-secreting antigen-specific liver CD8⁺ T after 14 days from the last immunization. Hepatic leukocytes were stained for intracellular IFN-γ production without (left) or after re-stimulation with PbTRAP_130−138 peptide (right). The results represent mean values (±SD). ^*P < 0.05; ^**P < 0.01 (Mann-Whitney test).