Propensity of selecting mutant parasites for the antimalarial drug cabamiquine

Abstract

We report an analysis of the propensity of the antimalarial agent cabamiquine, a Plasmodium-specific eukaryotic elongation factor 2 inhibitor, to select for resistant Plasmodium falciparum parasites. Through in vitro studies of laboratory strains and clinical isolates, a humanized mouse model, and volunteer infection studies, we identified resistance-associated mutations at 11 amino acid positions. Of these, six (55%) were present in more than one infection model, indicating translatability across models. Mathematical modelling suggested that resistant mutants were likely pre-existent at the time of drug exposure across studies. Here, we estimated a wide range of frequencies of resistant mutants across the different infection models, much of which can be attributed to stochastic differences resulting from experimental design choices. Structural modelling implicates binding of cabamiquine to a shallow mRNA binding site adjacent to two of the most frequently identified resistance mutations.

Fig. 1. Numbers of parasites across pre-clinical and clinical assays.

(1) In vitro P. falciparum minimum inoculum for resistance; (2) P. falciparum-infected human red blood cells engrafted in NSG mice; (3) Volunteer infection study with humans infected with blood-stage P. falciparum parasites; and (4) Field settings, i.e. endemic regions. NSG, NOD/SCID/IL2rγnull.

Fig. 2. PfeEF2 mutations mediating varying levels of cabamiquine resistance across infection models.

a Venn diagram of PfeEF2 amino acids and their locations subject to mutations under cabamiquine exposure in vitro, in NSG mouse, and in P. falciparum VIS. b Table with biomarkers of mutants selected in vitro in MIR studies (3D7, 7G8, and Dd2), in culture-adapted field isolates, in P. falciparum-infected NSG mice (3D7), or in P. falciparum VIS (3D7). Wild-type EC₅₀ values were 0.28 nM (3D7), 0.47 nM (Pf3D7^0087/N9 in serum), 0.24 nM (7G8), 0.19 nM (Dd2), 0.57 nM (3D7_MM), 0.46 nM (3D7_FS), 0.64 nM (EEF192), and 0.41 nM (EEF209). Low: EC₅₀ = 1–10 nM (blue); Medium: EC₅₀ > 10–100 nM (orange); High: EC₅₀ > 100 nM (red). *EC₅₀ not determined; ** field isolates (Mali, 2021); *** obtained from a free concentration ratio of the human C_av0–24h (corrected for human plasma protein binding of 83%) and P. falciparum 3D7 EC₅₀ (corrected for Albumax binding of 45.3%); **** obtained from a free concentration ratio of the C_av0–24h (249 nM) in NSG mice treated with a single dose of cabamiquine at 12 mg/kg (p.o.) and P. falciparum 3D7 EC₅₀. ^§ Mixed population. MIR minimum inoculum for resistance, VIS volunteer infection studies, NSG NOD/SCID/IL2rγnull, ND not defined, WT wild type.

Fig. 3. Characterization of P. falciparum culture-adapted field isolates under cabamiquine drug pressure.

a In vitro P. falciparum asexual blood-stage parasite growth of two recently culture-adapted field isolates (EEF192 and EEF209) and two laboratory lines of 3D7 parasites (3D7_FS and 3D7_MM). Parasites were cultured for over 50 days with cabamiquine drug pressure (15× EC₅₀) applied for three consecutive intra-erythrocytic developmental cycles starting on day 10. b Table summarizing the EC₅₀ values pre- and post cabamiquine drug pressure for 52 field isolates, cultured in parallel with 52 sets of laboratory 3D7 parasites, along with sequencing data. c Electropherograms of the recrudescent field isolates (EEF192 and EEF209) and 3D7 strains (3D7_FS and 3D7_MM). WT wild-type, AA amino acid. Source data are provided as a Source Data file.

Fig. 4. Mathematical modelling of resistant mutants.

a Estimated frequency of resistant mutants in different infection model systems and in the stochastic model simulations of these different experimental settings. For each model system, the dot represents the estimated frequency of resistant mutants, and the horizontal line represents the 95% confidence interval. The frequency of resistant mutants was estimated using a limiting dilution assay (* for the 3D7 in vitro regrowth, we assumed that the parasite number at the time of treatment for each culture is the mean parasite number from the cultures for which the parasitemia at treatment was known, see Supplementary Material and Methods). The data differ between in vivo and in vitro experiments and between the parasite strain and RBCs used. The P-values on the right-hand side indicate the comparison of the respective estimates from the data using a likelihood ratio test. For the model estimates, we simulated the stochastic model 100,000 times for each experiment-specific setting (i.e. using the inoculation size, pre-treatment parasite multiplication rate (PMR), and time from inoculation to treatment for each experiment; Supplementary Table 8). For each simulated experiment, we then estimated the frequency of resistant parasites in the same way as for the data. Supplementary Table 9 lists the median (square) and the 2.5th and 97.5th percentiles (dashed line) of the estimated frequency of resistant parasites from 100,000 simulated experiments. We assumed a fitness cost of a 7% reduction in the PMR and a PMR of four per replicative cycle. We also assumed that 11 different mutations are able to mediate resistance and that the mutation rate is 1.05 × 10⁻⁹ per base pair per generation. For estimates with different fitness costs, resistance mutation numbers, mutation rate, and PMR, see Supplementary Fig. 1. b The probability of the emergence of resistant mutants after treatment compared with the probability that resistant mutants were pre-existent at treatment for different parasite reduction ratios. The model predicts that the probability of resistant mutants already being present at the time of treatment is larger than that of resistant mutants emerging only after treatment for a parasite reduction ratio of ≥5. These quantities were computed using the deterministic model (Supplementary Materials and Methods). Source data are provided as a Source Data file.

Fig. 5. Potential binding site and top-ranked docking pose of cabamiquine.

a Surface and known mutants in the predicted protein structure of PfeEF2 and potential binding of cabamiquine near residues E134 and Y186. b The ligand–receptor interaction map highlights amino acid residues of PfeEF2 within a 5 Å radius around the potential binding pose of cabamiquine. Amino acids are coloured based on their properties (olive, hydrophobic; blue, polar; purple, positively charged; and orange, negatively charged) and interactions are shown by arrows (purple, hydrogen bonding; dark blue, salt bridge; and green, π–π stacking). Single-letter code for amino acid residues: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.