KAI407, a potent non-8-aminoquinoline compound that kills Plasmodium cynomolgi early dormant liver stage parasites in vitro

Abstract

Preventing relapses of Plasmodium vivax malaria through a radical cure depends on use of the 8-aminoquinoline primaquine, which is associated with safety and compliance issues. For future malaria eradication strategies, new, safer radical curative compounds that efficiently kill dormant liver stages (hypnozoites) will be essential. A new compound with potential radical cure activity was identified using a low-throughput assay of in vitro-cultured hypnozoite forms of Plasmodium cynomolgi (an excellent and accessible model for Plasmodium vivax). In this assay, primary rhesus hepatocytes are infected with P. cynomolgi sporozoites, and exoerythrocytic development is monitored in the presence of compounds. Liver stage cultures are fixed after 6 days and stained with anti-Hsp70 antibodies, and the relative proportions of small (hypnozoite) and large (schizont) forms relative to the untreated controls are determined. This assay was used to screen a series of 18 known antimalarials and 14 new non-8-aminoquinolines (preselected for blood and/or liver stage activity) in three-point 10-fold dilutions (0.1, 1, and 10 μM final concentrations). A novel compound, designated KAI407 showed an activity profile similar to that of primaquine (PQ), efficiently killing the earliest stages of the parasites that become either primary hepatic schizonts or hypnozoites (50% inhibitory concentration [IC50] for hypnozoites, KAI407, 0.69 μM, and PQ, 0.84 μM; for developing liver stages, KAI407, 0.64 μM, and PQ, 0.37 μM). When given as causal prophylaxis, a single oral dose of 100 mg/kg of body weight prevented blood stage parasitemia in mice. From these results, we conclude that KAI407 may represent a new compound class for P. vivax malaria prophylaxis and potentially a radical cure.

FIG 1

In vitro liver stage assay, intracellular development of P. cynomolgi, and PVM staining of small EEFs. (A) Schematic representation of the P. cynomolgi in vitro liver stage assay workflow. The experiment starts with the infection of a rhesus monkey with blood stage parasites from thawed stock. Mosquitoes are fed on the parasitized rhesus blood around peak parasitemia. Salivary gland sporozoites are harvested ∼14 days after the infected blood meal to infect primary rhesus hepatocytes. After sporozoite invasion, compounds are added in a concentration series, and the assay mixtures are incubated for 6 days. Parasites are visualized by staining Hsp70 with fluorescent antibodies. Small and large liver stage parasites are differentially counted using a high-content imager, resulting in a dose-response curve. (B) Sporozoite-infected primary rhesus hepatocytes were fixed 1, 3, 5, 8, or 10 days after inoculation with P. cynomolgi sporozoites. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) (left column). Parasites were stained with anti-Hsp70 antibodies and fluorescein isothiocyanate (FITC)-labeled secondary antibodies (middle column). An overlay is shown in the right column. Within 24 h after infection, intracellular parasites are visible (images A to C), containing one nucleus (image A). The growth rates of all intracellular parasites appear to be similar until day 3 (images D to F). After 5 (images G to I) and 8 (images J to L) days of intracellular development, schizonts and hypnozoite forms (arrowhead) can be distinguished. Approximately 53% of the EEFs develop into mature schizonts, and 47% of the parasites remain small for the duration of the assay. Ten days after sporozoite inoculation, in the same well, free merozoites were observed (images M to O), as well as persistent small forms with a single nucleus (P to R). The developing liver stage parasites grow out to full maturity in a similar time frame, as reported in vivo (51). (C) Etramp staining of P. cynomolgi liver stage parasites. A liver schizont and a hypnozoite form (arrowheads) are seen in this field. Images: A, nuclei of P. cynomolgi-infected primary rhesus hepatocytes stained with DAPI; B, mouse anti-Hsp70/anti-mouse FITC staining of P. cynomolgi liver stage parasites; C, PVM stained with rabbit-anti-P. cynomolgi Etramp/anti-rabbit-tetramethyl rhodamine isocyanate (TRITC), showing the presence of a PVM in both parasites; D, merge of images A, B, and C. The presence of a PVM indicates that small EEFs are not the product of an aborted in vitro invasion but, rather, that the parasites have successfully invaded the hepatocyte.

FIG 2

P. cynomolgi liver stage drug assay validation. (A) Atovaquone sensitivity profile of large EEFs and hypnozoite forms. Three independent atovaquone IC₅₀ assays were averaged and are shown here (± standard deviations [SD]) with representative images of assay wells. Developing EEFs are more sensitive to the compound than hypnozoite forms. At higher atovaquone concentrations, only small parasites are visible. (B) Reproducibility of the primaquine IC₅₀ assay. Ten independent P. cynomolgi primaquine in vitro liver stage drug assays were performed. IC₅₀s were determined for individual assays and plotted. Indicated are the mean values and the 95% confidence intervals. IC₅₀s from the same experiment for schizonts and hypnozoite forms are indicated with identical symbols. Note that the IC₅₀ for hypnozoite forms is always higher than for schizonts within a single assay.

FIG 3

P. cynomolgi drug assays with selected malaria box compounds. (A) Structure formulas of KAE609, KAF156, KAI407, and primaquine. (B) Three-point 10-fold dilution series (0.1, 1, and 10 μM) on P. cynomolgi liver stage cultures. The percentage of untreated control is shown as a function of the compound concentration. Differential counting of schizonts and hypnozoite forms was performed based on size and number of parasite nuclei. The results of one representative assay are shown; assays were performed at least twice. KAE609 was not active against liver stage parasites and was therefore not taken forward for IC₅₀ determination. (C) IC₅₀ determination for KAF156, KAI407, and primaquine. Drugs were tested in a 2-fold dilution series. The percentage of untreated control is shown as a function of the compound concentration. Differential counting of schizonts and hypnozoite forms was performed based on size and number of parasite nuclei. IC₅₀s were calculated based on fitted curves (nonlinear fit). The results of one representative assay are shown, and at least two independent assays were performed. The activity of KAF156 is limited to schizonts, while KAI407 and primaquine also kill hypnozoite forms. The error bars indicate standard deviations.