A high throughput screen for next-generation leads targeting malaria parasite transmission

Abstract

Spread of parasite resistance to artemisinin threatens current frontline antimalarial therapies, highlighting the need for new drugs with alternative modes of action. Since only 0.2-1% of asexual parasites differentiate into sexual, transmission-competent forms, targeting this natural bottleneck provides a tangible route to interrupt disease transmission and mitigate resistance selection. Here we present a high-throughput screen of gametogenesis against a ~70,000 compound diversity library, identifying seventeen drug-like molecules that target transmission. Hit molecules possess varied activity profiles including male-specific, dual acting male-female and dual-asexual-sexual, with one promising N-((4-hydroxychroman-4-yl)methyl)-sulphonamide scaffold found to have sub-micromolar activity in vitro and in vivo efficacy. Development of leads with modes of action focussed on the sexual stages of malaria parasite development provide a previously unexplored base from which future therapeutics can be developed, capable of preventing parasite transmission through the population.

Fig. 1

Screening progression cascade. The Global Health Chemical Diversity Library (GHCDL) was screened in parallel against Plasmodium falciparum asexual stage and male and female gametocyte. Actives were reconfirmed and cytotoxic compounds removed. Asexual hits were further filtered based upon potency and 26 compounds selected for further profiling based upon potency and commercial availability. Five compounds with different transmission-blocking properties were further investigated for activity phenotype against male gametogenesis and their physiochemical properties (DMPK) investigated. Activity of three molecules was confirmed by standard membrane-feeding assay (SMFA)

Fig. 2

Profiling the transmission blocking properties of selected compounds. a 26 compounds were profiled in seven assays that interrogate different ranges of parasite cell biology (represented by position and length of coloured bars). Ability to prevent asexual replication was tested by asexual growth assay. Ability to inhibit the metabolic viability of stage IV/V gametocytes (Pf GCT) was tested by late stage gametocyte ATP depletion assay (ATP) and stage V gametocyte by mitotracker assay (MITO). By varying compound incubation period, the carry-over/wash-out (WO)/add-in Pf DGFA assays permit discrimination between gametocyte- and gamete-targeted activity. The P. berghei ookinete development assay tests activity against early mosquito stage parasites and cross-reactivity to P. berghei. b Table summarising outcome of compound profiling. Compounds fell into four categories: gamete-targeted (purple); transmission-specific (gametocyte-targeted, green); dual asexual/gametocyte targeted (blue); asexual-specific (orange). Red blocks indicate compound was active in respective assay (IC₅₀ > highest assay concentration: 12.5 µM asexual; 10 µM ATP; 25 µM MITO; 12.5 µM WO; 10 µM Add-in; >12.5 µM Pb Ookinete), grey indicates inactive and white indicates not tested. The exception being the Pf DGFA columns: compounds that were irreversible (IRR), e.g. active in carry-over and wash-out (WO) were designated gametocyte-active; compounds that lost activity in wash-out format were designated gamete-active

Fig. 3

Chemical structure of GHCDL compounds selected for further study. Five compounds showing a range of transmission-blocking activity at the profiling stage were selected for detailed study. DDD01027599 (found to actually be BPCA—see Supplementary Fig. 1-2) and DDD01245291 are active against gametocytes and asexuals; DDD01255968 and DDD01249504 are active specifically against male gametocytes; and DDD01035881 specifically targets male gamete formation

Fig. 4

Monitoring the effect of compound treatment on male gamete formation by IFA. Gametocytes were treated for 24 h with DMSO or 10 µM test compound and then gamete formation triggered by the addition of ookinete medium and reduction to room temperature. Aliquots were sampled at different time points and immediately fixed before being stained directly for α-tubulin (green), glycophorin A (red) and DNA via DAPI labelling (blue). Images show representative non-induced cells (0 min), representative cells during gamete formation progression (2–10 min), and representative cells showing the most advanced phenotype at 20 min. a Upon activation, male gametocytes initially round up before undergoing three endomitotic genome replications, egressing from the surrounding erythrocyte membrane and releasing up to eight flagellated gametes. b Images from left to right show DMSO-treated cells undergoing normal male gamete formation over 20 min. Male gametocytes initially have an elongated morphology with a faint nucleus, relatively homogenous tubulin distribution and are surrounded by an erythrocyte membrane. Upon induction of gametogenesis, cells round up and begin three rounds of endomitosis, with ordered microtubule spindle fibres and increasingly intense DNA staining. Towards the end of gametogenesis, the surrounding erythrocyte membrane is shed and up to eight microgametes emerge from the cell possessing a microtubule-rich flagellum and one replicated genome. c DDD599/BPCA-treated cells; d DDD504-treated cells; e DDD968-treated cells; f DDD291-treated cells; g DDD881-treated cells. Bars = 5 µm

Fig. 5

Computer-aided quantification of male gamete formation from IFA images. Individual cells were extracted from IFA images stained for α-tubulin, glycophorin A and DNA from samples fixed between 0 and 20 min post induction of gamete formation (n = 195–668 cells per compound treatment). Cell shape (elongated/irregular vs. round) and haploid chromosome number (1nto 8n) were estimated from α-tubulin and DAPI staining (DNA content) respectively (arbitrary units). a Schematic diagram illustrating how cell shape and DNA content parameters vary over the course of male gamete formation. Initially gametocytes are elongated with 1n haploid chromosome number. Upon induction, they round up then undergo three rounds of DNA replication whilst maintaining a round shape. Finally, male gamete release (exflagellation) changes the shape of the cell whilst maintaining a high DNA content. b Genome copy number during gamete formation was estimated by benchmarking mean DAPI intensity with respect to the non-activated gametocytes at t = 0 min (1n) and exflagellating cells at t = 10–20 min (8n). DNA replication was compared between the DMSO control and five test compounds. c–h Graphs plotting the cell shape and DAPI intensity data for each individual cell analysed. Red dots show the initial distribution of cells before induction (t = 0 min). Grey dots show induced cells (t = 2–20 min)

Fig. 6

Structure–activity relationship between BPCA/DDD599 and analogues. Activity in the Pf DGFA male readout and parasite selectivity compared to Tox₅₀ in HepG2 cells shown. Red and blue groups indicate modifications to respective parts of the parent molecule. Black molecules are modified in both groups. Analogue IDs correspond to PubChem ID (https://pubchem.ncbi.nlm.nih.gov/)

Fig. 7

Pf DGFA structure activity relationship between DDD01035881 and analogues. Red and blue groups indicate modifications to respective parts of the parent molecule. Analogue IDs correspond to molecules within the GHCDL with the exception of OJF-034 which was synthetised in-house

Fig. 8

Transmission-blocking efficacy of leads in P. falciparum SMFA. The predictive power of the Pf DGFA to identify transmission-blocking compounds was assessed by treating GFP/luciferase-expressing P. falciparum gametocytes for 24 h with DMSO, DDD291 or DDD599/BPCA and then feeding them to A. stephensi mosquitoes in indirect mode in which test compound is absent from the blood meal. As DDD881 targets gamete formation rather than gametocytes, membrane feeds with this compound were performed in direct mode, added directly to the blood meal with no prior incubation. a Normalised luciferase intensity from individual mosquitoes (n = 56–72 per condition per replicate) was compared to DMSO and unfed (uninfected) controls. Data combined from three independent biological replicate feeds. Grey dots indicate individual luciferase readout for individual mosquitoes and black bar shows the mean. b Calculated infection prevalence of mosquito feeds normalised to respective DMSO control. Error bars denote the standard error of the mean

Fig. 9

In vivo transmission-blocking efficacy in P. berghei direct feeding assays. Eight phenylhydrazine-treated mice were infected with P. berghei constitutively expressing GFP. On day 3 of infection, they were assigned into four groups of two mice (DMSO A + B; DDD291 A + B; DDD881 A + B and C + D). Three groups received either 50 mg kg⁻¹ DDD291, 50 mg kg⁻¹ DDD881 or DMSO + vehicle control by IP injection. Prior to dosing and 24 h later, immediately before feeding to A. stephensi mosquitoes, a asexual parasitemia and b gametocytaemia were recorded. The fourth group received a 50 mg kg⁻¹ DDD881 dose on day 4, 30 min prior to mosquito feed. c Seven days after feeding, mosquitoes were dissected (n = 17–52 per condition) and oocysts recorded by fluorescent microscopy and automated image analysis. Horizontal bars indicate the mean. d PK analysis of DDD881. BALB/c mice were dosed with 50 mg kg⁻¹ DDD881 (n = 3) and plasma concentrations were monitored over time. After an initial distribution phase, DDD881 was eliminated with a half-life of ~90 min. Error bars denote the standard deviation