Platelets kill circulating parasites of all major Plasmodium species in human malaria

Abstract

Platelets are understood to assist host innate immune responses against infection, although direct evidence of this function in any human disease, including malaria, is unknown. Here we characterized platelet-erythrocyte interactions by microscopy and flow cytometry in patients with malaria naturally infected with Plasmodium falciparum, Plasmodium vivax, Plasmodium malariae, or Plasmodium knowlesi Blood samples from 376 participants were collected from malaria-endemic areas of Papua, Indonesia, and Sabah, Malaysia. Platelets were observed binding directly with and killing intraerythrocytic parasites of each of the Plasmodium species studied, particularly mature stages, and was greatest in P vivax patients. Platelets preferentially bound to the infected more than to the uninfected erythrocytes in the bloodstream. Analysis of intraerythrocytic parasites indicated the frequent occurrence of platelet-associated parasite killing, characterized by the intraerythrocytic accumulation of platelet factor-4 and terminal deoxynucleotidyl transferase deoxyuridine triphosphate nick-end labeling of parasite nuclei (PF4⁺TUNEL⁺ parasites). These PF4⁺TUNEL⁺ parasites were not associated with measures of systemic platelet activation. Importantly, patient platelet counts, infected erythrocyte-platelet complexes, and platelet-associated parasite killing correlated inversely with patient parasite loads. These relationships, taken together with the frequency of platelet-associated parasite killing observed among the different patients and Plasmodium species, suggest that platelets may control the growth of between 5% and 60% of circulating parasites. Platelet-erythrocyte complexes made up a major proportion of the total platelet pool in patients with malaria and may therefore contribute considerably to malarial thrombocytopenia. Parasite killing was demonstrated to be platelet factor-4-mediated in P knowlesi culture. Collectively, our results indicate that platelets directly contribute to innate control of Plasmodium infection in human malaria.

Graphical abstract

Figure 1.

Platelet binding in clinical malaria blood samples. (A) Photos of platelet-bound iRBCs from patient Giemsa smears (black arrowhead = platelet). Images were taken at 1000× magnification using a Samsung Note-3 camera attached to an Olympus CX31 microscope. Scale bar, 5 µm. (B) Representative flow cytometry gating strategy to measure platelet binding in a P falciparum patient and healthy control. (C) Frequency of platelet-bound uRBCs by flow cytometry in patients with malaria compared with control patients in samples from (Ci) Papua (control patients, n = 17; Pf, n = 23; Pv, n = 26; Pm, n = 9; mixed, n = 7) and (Cii) Sabah (control patients, n = 28; Pf, n = 14; Pv, n = 85; Pk, n = 106; Kruskal-Wallis, *significantly different to all other groups). (D) Frequency of platelet-bound iRBCs and uRBCs by flow cytometry in samples from (Di) Papua (n as per Ci) and (Dii) Sabah (n as per Cii; Wilcoxon test). (E) Inverse correlation of platelet-bound iRBCs with parasitemia in samples from (Ei) Papua and (Eii) Sabah (Spearman). Scatterplots indicate median ± interquartile range for each group. Parasitemia values are log transformed. Data presented in supplemental Table 1. Pf, P falciparum; Pv, P vivax; Pm, P malariae; Pk, P knowlesi.

Figure 2.

PF4-associated parasite killing in clinical malaria samples. (A) Representative immunofluorescent images from Pf, Pv, Pk, and Pm patient blood smears illustrating PF4-associated parasite killing (PF4⁺TUNEL⁺ iRBCs). Scale bars, 5 µm. Arrows and arrowheads indicate platelets and parasites, respectively. Images were taken at 630× magnification on an Axio Scope A1 fluorescent microscope coupled to an Axiocam ICm-1 CCD camera, or an Axio Observer inverted fluorescence microscope coupled to an Axiocam 503 monochrome camera. ZEN 2 software was used for image acquisition and processing (all from Carl Zeiss, Germany). (B) Percentage of PF4⁺TUNEL⁺ parasites in clinical samples with Pf (Papua, n = 50; Sabah, n = 14), Pv (Papua, n = 32; Sabah, n = 13), Pm (n = 11), Pk (n = 15), and mixed species infection (n = 6). (C) Comparison of intraerythrocytic PF4 (PF4⁺) parasites as a percentage of dying (TUNEL⁺) parasites in Pf, Pv, Pm, Pk, and mixed species infection from Papua and Sabah (n as per B). (D) Inverse correlation of PF4⁺TUNEL⁺ iRBCs with parasitemia in Pf and Pv clinical samples (Spearman). (E) Proportions of PF4⁺TUNEL⁺ rings vs mature stages in Pv (rings, n = 15; mature, n = 16), Pm (rings, n = 6; mature, n = 9), and Pk (rings, n = 8; mature, n = 5). (F) Proportions of rings that were PF4⁺TUNEL⁺ and (G) PF4⁺ in Pf (Papua, n = 50; Sabah, n = 14), Pv (n = 15), Pm (n = 6), and Pk clinical samples (n = 8). Scatterplots indicate median ± interquartile range for each group. Parasitemia values are log transformed. Kruskal-Wallis or Mann-Whitney test used for statistical comparisons. Data presented in Table 4.

Figure 3.

In vitro cultures of P knowlesi are sensitive to platelets and PF4. The growth of (A) P knowlesi (n = 4) and (B) P falciparum (n = 3) treated with different platelet concentrations or Tyrodes buffer for 48 hours. (C) The growth of P knowlesi treated with platelet lysate, with and without anti-PF4 antibodies or immunoglobulin G isotype control (n = 2). (D) P knowlesi PF4 dose-response curve (n = 2). (E) The growth of P knowlesi treated with platelets (60 million/mL), platelet lysate or PF4 (0.5 µM), and cocultured in standard wells or Transwells (n = 2). (F) Micrographs showing platelets bound to uninfected and P knowlesi-infected cells. (G) Percentage platelet binding to uninfected, P knowlesi (n = 4) or P falciparum iRBCs (n = 3), determined by flow cytometry. (H) Percentage TUNEL-labeled (TUNEL⁺) P knowlesi parasites costained for PF4 (PF4⁺) or not PF4-stained (PF4⁻; n = 3). (I) Micrographs showing a PF4⁺TUNEL⁺ P knowlesi infected cell after platelet treatment. Scale bars, 5 µm. Images were taken at 630× magnification on an Axio Observer inverted fluorescence microscope coupled to an Axiocam 503 monochrome camera with ZEN 2 software (Carl Zeiss, Germany). Bars indicate means of replicate data points. Kruskal-Wallis test or 1-way ANOVA used for statistical comparisons, *P < .05 and **P < .01. DIC, differential interference contrast.