Safe drugs with high potential to block malaria transmission revealed by a spleen-mimetic screening

Abstract

Malaria parasites like Plasmodium falciparum multiply in red blood cells (RBC), which are cleared from the bloodstream by the spleen when their deformability is altered. Drug-induced stiffening of Plasmodium falciparum-infected RBC should therefore induce their elimination from the bloodstream. Here, based on this original mechanical approach, we identify safe drugs with strong potential to block the malaria transmission. By screening 13 555 compounds with spleen-mimetic microfilters, we identified 82 that target circulating transmissible form of P. falciparum. NITD609, an orally administered PfATPase inhibitor with known effects on P. falciparum, killed and stiffened transmission stages in vitro at nanomolar concentrations. Short exposures to TD-6450, an orally-administered NS5A hepatitis C virus inhibitor, stiffened transmission parasite stages and killed asexual stages in vitro at high nanomolar concentrations. A Phase 1 study in humans with a primary safety outcome and a secondary pharmacokinetics outcome ( https://clinicaltrials.gov , ID: NCT02022306) showed no severe adverse events either with single or multiple doses. Pharmacokinetic modelling showed that these concentrations can be reached in the plasma of subjects receiving short courses of TD-6450. This physiologically relevant screen identified multiple mechanisms of action, and safe drugs with strong potential as malaria transmission-blocking agents which could be rapidly tested in clinical trials.

Fig. 1. High-throughput screening based on mitochondrial staining and cell deformability identifies compounds with both killing effect and stiffening activity on P. falciparum late gametocytes.

Screening progression cascade of three different libraries: Malaria Pathogen Box (A), Kinase Inhibitors Box (B), and ReFrame library (C). A 3 hits from primary screening were submitted to dose-response analysis along with 12 compounds found active in some but not all screening replicates. The 3 hits were confirmed but none of them was selected for further post-screening validation. B Four hits from primary screening along with 5 compounds found active in some but not all screening replicates were submitted to dose-response analysis raising 3 confirmed hits. None of them was selected for further post-screening validation. C 112 hits from primary screening were submitted to dose-response analysis, raising 74 confirmed hits. 63 compounds with uninterpretable results during primary screening were added to the hits for dose-response analysis raising additional two confirmed hits. The 76 confirmed hits were allocated to seven groups (panels on the right), based on their activity and molecular target. For each group, one representative hit has been selected for illustration. Hit scoring based on route of administration, safety in human subjects, and pharmacokinetics resulted in the selection of 3 drugs submitted to final confirmation experiments (dark blue).

Fig. 2. Compiled screening results and representative plate analysis.

Representative screening output scatterplots for the specific (A) and ReFrame (B) libraries, and for 2 single plates from the ReFrame primary screening, with NITD609 (C) and TD-6450 (D) as finally selected hits. In panel A, “K” refers to Kinase Inhibitors Box and “P” refers to Pathogen Box. Killing and retention rates are on the x and y axes, respectively. Negative controls (red squares), and positive controls including calyculin A 75 nM for stiffening activity (green triangles), Gentian violet 50 µM for killing effect (green circles) and NITD609 0.5 µM for combined stiffening and killing effect (green squares), fell inside and outside of the main cloud of inactive test compounds (empty gray circles) defined by linear regression +/−95% prediction bands SD (full line and dotted lines, respectively). Hits were either confirmed by DRA (Blue X), or not (blue empty triangles). Source data are provided as a Source Data file.

Fig. 3. Dose-response curves of selected hits.

Dose-response curves and chemical structures of the 3 selected hits: NITD609 (A), TD-6450 (B) and L-THP (C). TD-6450 was a pure enantiomer. Stereocenters are: (S)−1-((S)−2-(5-(4’-(6-((2 R,5 S). Data are presented as mean values +/− SEM. For NITD609 and L-THP, n = 2 independent experiments, for TD-6450, n = 4 independent experiments. Source data are provided as a Source Data file.

Fig. 4. Activity of NITD609 and TD-6450 upon post-screening analysis.

A Giemsa-stained erythrocytes infected by a mature gametocyte of P. falciparum exposed for 24 hours to DMSO (negative control, green border), TD-6450 5 µM (blue border) and NITD609 1 µM (red border). Circularity of gametocytes by imaging flow cytometry was compared measuring the aspect ratio showing a significant difference between DMSO and NITD609. The experiment was performed once. B Cumulative dot-plot of 5 microsphiltration experiments where stage V gametocytes of P. falciparum (3D7 pULG8-GFP strain) were exposed for 24 hours to TD-6450 (blue), NITD609 (red) or the combination of both (purple). C Cumulative dot-plot of 4 microsphiltration experiments where gametocytes were exposed to the drugs for 24 h and kept in culture for an additional 24 hours after removing the drug. Dots indicate the retention rate of single wells of 96-well microsphiltration plates. For B, C, each experiment was performed with a single gametocyte induction and a single 8-well column for each condition was loaded, unless the cultured gametocyte population was not large enough to fill the entire column. Numbers of repetitions are, from left to right: 40 (8 for each experiment), 40 (8 for each experiment), 32 (8 for each of 4 experiments), 32 (8 for each of 4 experiments). Boxplots indicate medians and IQRs. Exposure to DMSO (green) was the negative control. P and F-values after the one-way ANOVA test were lower than 0.0001 and equal to 110 (B) and equal to 0.0005 and 6.365 (C), respectively. Individual p-values legend: *p = 0.05–0.01, **p = 0.01–0.001, ***p = 0.001–0.0001, ****p < 0.0001. Individual p-values, when higher than 0.0001, are: panel B, NITD609 vs combination, 0.0058. Panel C, DMSO vs NITD609, 0.0096. Source data are provided as a Source Data file.

Fig. 5. Activity of NITD609 and TD-6450 on asexual parasites.

A Giemsa-stained erythrocytes infected by a ring-stage of P. falciparum exposed to DMSO (negative control, green border), TD-6450 5 µM (blue border) and NITD609 1 µM (red border). The optical microscopic observations were repeated 5 times. B Dose-response of synchronized ring-stage P. falciparum (NF54 and 3D7 pULG8-GFP strains), exposed for 48 h to TD-6450 (blue) and NITD609 (red). C Dose-response of F32-TEM strain either before (dark blue), or after 5 (blue) or 10 (light blue) drug pulses with TD-6450. Each point shows the mean (±SD) of 3 replicates. Source data are provided as a Source Data file.

Fig. 6. TD-6450 modeling based on phase I results.

Goodness-of-fit (A) of the final pharmacokinetic model. Simulation-based (n = 2000 independent simulations) diagnostics (pvcVPC), displaying the predictive performance of the final PK model, stratified on single (B) vs multiple (C) dosing. PK data used to realize the modeling were made available by Theravance Biopharma. Source data are provided as a Source Data file.

Fig. 7. TD-6450 showed an activity at concentrations that can be reached in healthy subjects receiving a single dose.

Simulated population pharmacokinetic profiles of a single oral dose of TD-6450 when administered in a fasting state (A) and with food (B), in healthy volunteers. The black solid lines are the population mean profiles and the shaded gray area show the 90% prediction interval. The red dashed lines indicate a concentration of 200 nM, associated with significant stiffening of mature gametocytes. PK data used to realize the modeling were made available by Theravance Biopharma. C Cumulative dot-plot of 3 microsphiltration experiments where gametocytes were exposed to the drugs for 8 h, diluted by a factor of 10 and then filtrated after 24 h. Dots indicate the retention rate of single wells of 96-well microsphiltration plates. Each experiment was performed with a single gametocyte induction and a single 8-well column for each condition was loaded, unless the cultured gametocyte population was not large enough to fill the entire column. Numbers of repetitions are, from left to right: 23 (8 for the first 2 experiments, 7 for the 3rd), 16 (8 for each of the first 2 experiments), 23 (8 for the first 2 experiments, 7 for the 3rd), 7 (3rd experiment only) and 7 (3rd experiment only). Boxplots indicate medians and IQRs. Exposure to DMSO (green) was the negative control. P and F-values after the one-way ANOVA test were lower than 0.0001 and equal to 12.09, respectively. Individual p-values legend: *p = 0.05–0.01, **p = 0.01–0.001, ***p = 0.001–0.0001, ****p < 0.0001. Individual p-values, when higher than 0.0001, are: DMSO vs TD-6450 200 nM, 0.0284. Source data are provided as a Source Data file.