A second mechanism employed by artemisinins to suppress Plasmodium falciparum hinges on inhibition of hematin crystallization

Abstract

Malaria is a pervasive disease that affects millions of lives each year in equatorial regions of the world. During the erythrocytic phase of the parasite life cycle, Plasmodium falciparum invades red blood cells, where it catabolizes hemoglobin and sequesters the released toxic heme as innocuous hemozoin crystals. Artemisinin (ART)-class drugs are activated in vivo by newly released heme, which creates a carbon-centered radical that markedly reduces parasite density. Radical damage to parasite lipids and proteins is perceived to be ARTs' dominant mechanism of action. By contrast, quinoline-class antimalarials inhibit the formation of hemozoin and in this way suppress heme detoxification. Here, we combine malaria parasite assays and scanning probe microscopy of growing β-hematin crystals to elucidate an unexpected mechanism employed by two widely administered antimalarials, ART, and artesunate to subdue the erythrocytic phase of the parasite life cycle. We demonstrate that heme-drug adducts, produced after the radical activation of ARTs and largely believed to be benign bystanders, potently kills P. falciparum at low exogenous concentrations. We show that these adducts inhibit β-hematin crystallization and heme detoxification, a pathway which complements the deleterious effect of radicals generated via parent drug activation. Our findings reveal an irreversible mechanism of heme-ART adduct inhibition of heme crystallization, unique among antimalarials and common crystal growth inhibitors, that opens new avenues for evaluating drug dosing regimens and understanding growing resistance of P. falciparum to ART.

Keywords: atomic force microscopy; drug adducts; hemozoin crystals; malaria; parasites.

Figure 1

In vitro activation of artemisinins to heme–drug adducts.A–E, molecular structures of (A) ART, (B) the radical ART∗ generated from the reaction between ART and heme (II) (Fig. S1), and (C) the adduct H-ART formed between ART∗ and heme(III). Carbons-labeled α, β, δ, and γ indicate sites for isomorphic ART∗ substitutions. The chemical bond is arbitrarily shown at the δ position (noting that all positions can be alkylated). D and E, molecular structures of ARS, in (D), and its adduct (H-ARS), in (E), generated by a similar reaction mechanism (Fig. S2). F, mass spectrum of isolated H-ART with characteristic peaks at m/z = 838 and 898. Inset: absence of mass ion peaks associated with the parent drug ART. G, mass spectrum of isolated H-ARS with its characteristic peak at m/z = 1000.1. Inset: absence of mass ion peaks associated with the parent drug ARS. ARS, artesunate; ART, artemisinin; H-ART, heme–artemisinin adduct; H-ARS, heme–artesunate adduct.

Figure 2

P. falciparum inhibition by heme–drug adducts.A and B, percent inhibition of three P. falciparum strains, NF54, C580Y, and CamWT, after 72 h of continuous drug (parent or adduct) exposure as a function of drug concentration: (A) ARS and H-ARS; (B) ART and H-ART. C, IC₅₀ assays were performed with each P. falciparum isolate for the parent drugs (number of biological replicates, n = 4) and adducts H-ART (n = 3) and H-ARS (n = 5). D, results of ring stage survival assays with 500 nM of drugs pulsed for 6 h, then washed, and incubated for an additional 66 h compared with assays without drug (n = 3). E and F, isobolograms for combination drugs with an outcome of IC₅₀ and corresponding fractional inhibitory concentrations (FICs): (E) CQ and ARS; (F) H-ARS and ARS. G and H, MS spectra of extracted hemozoin crystals after a 6-h incubation of trophozoites with ARS in (G) and H-ARS in (H). The samples were frozen and hemozoin was pelleted and decrystallized (for details of the extraction procedure and expanded MS analysis, see the Methods in the Supporting Information and Fig. S5). ART, artemisinin; ARS, artesunate; CQ, chloroquine; H-ART, heme–artemisinin adduct; H-ARS, heme–artesunate adduct; MS, mass spectrometry.

Figure 3

AFM analysis of parent drug and heme–drug adduct inhibition of β-hematin growth.A, AFM deflection mode image of a β-hematin (100) surface in supersaturated solution with heme concentration C_H = 0.28 mM and at 28 °C, the temperature of the AFM liquid cell. B and C, images of an identical surface in the presence of 10 μM growth inhibitors: (B) H-ART and (C) H-ARS. D, measurements of the rate of two-dimensional nucleation, J_2D, relative to that in the absence of any inhibitor, J_2D,o, as a function of drug concentration. E, measurements of step velocity, v, in the [001] direction with increasing drug concentration relative to the value measured in the absence of drug, v_o. F, linearized step velocity coordinates based on a reported equation (28) used to confirm crystal growth inhibition by a kink blocking mode of action. G, illustration of a crystal surface highlighting different sites for solute/drug binding (kinks, steps, and terraces) where both H-ART and H-ARS (orange spheres) act as kink blockers. AFM, atomic force microscopy; ART, artemisinin; ARS, artesunate; H-ART, heme–artemisinin adduct; H-ARS, heme–artesunate adduct.

Figure 4

Dual action mechanism of artemisinins.A, illustrated erythrocytic life cycle of P. falciparum. Optical micrographs represent P. falciparum parasites at ring (I), trophozoite (II), and schizont (III) stages. B, idealized scheme of hemozoin accumulation and free heme concentration variations during the parasite life cycle. Concentration profiles are adapted from data presented by Heller and Roepe (34). The shaded red regions correspond to the approximate times reported by Klonis et al. (10), where P. falciparum is most sensitive to 4-h pulses of ART. C–E, staged in situ AFM measurements of β-hematin surface growth at different combinations of heme and drug concentrations. Step velocities are measured in the absence of drug (stage 1); in the presence of 10 μM (C) H-ART or (E) CQ (stages 2 and 3); at increased heme concentration, 0.50 mM, in the presence (stage 3) and absence (stage 4) of drug; and again at the initial heme concentration, 0.28 mM, in the absence of drug (stage 5). D, AFM deflection mode image of a β-hematin (100) surface during stage 3 analysis showing the presence of deposited nanocrystals (arrow). AFM, atomic force microscopy; ART, artemisinin; CQ, chloroquine; H-ART, heme–artemisinin adduct.