Stock solutions of individual test compounds in 100% DMSO (10 µL of 10 mM solution per compound) were provided by the Medicines for Malaria Venture (MMV; ¹ ). The compounds originated from the MMV Pathogen Box, an open-access compound collection of 400 structurally diverse, drug-like molecules with activity against pathogens responsible for neglected tropical diseases. The pathogen box is distributed across five 96-well plates, each containing 80 compounds.

In this study, two plates (plates A and E) comprising 160 compounds were selected for initial screening based on assay throughput considerations, providing a representative subset for preliminary hit identification. These plates contain compounds spanning multiple chemical classes and disease indications, providing a manageable yet chemically diverse representative subset. Compound handling and dilution were performed according to MMV guidelines. ² Briefly, 1 mM parent stock solutions were prepared in 100% DMSO from the 10 mM master stocks supplied by MMV, aliquoted into 96-well plates, and stored at −20 °C. Working solutions were freshly prepared by diluting the intermediate stocks into autoclaved distilled water and subsequently into the assay medium to achieve a final compound concentration of 10 µM (1% DMSO). The two compounds selected for downstream mode-of-action studies, MMV667494 and MMV028694, were selected based on their sub-micromolar activity in secondary screening assays (Supplementary Material 1). Fresh supplies of these compounds were subsequently obtained from MMV, and 10 mM stock solutions were prepared in DMSO for all confirmatory and mechanistic experiments.

Ethical approval and informed content were not required for this study, as all experiments were conducted using established cell lines and did not involve human participants or animal subjects.

Wild-type bloodstream-form Trypanosoma brucei brucei GUTat 3.1 cells in logarithmic growth phase were cultured at 37 °C and 5% CO₂ in HMI-9 medium [35], supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin.

Antitrypanosomal activity was assessed using a resazurin-based viability assay adapted from a previously described method [36]. This assay measures metabolic activity as a proxy for parasite viability. Screening was conducted in two phases: (i) a single-concentration primary screen to identify active compounds, and (ii) a dose-response secondary screen to determine IC₅₀ values of the compounds. For the primary screen, a single-concentration assay (1 µM) was used to screen all compounds. Compounds were incubated with T. b. brucei cells (2 × 10⁴ cells/well) for 24 h, followed by the addition of alamarBlue dye solution (500 µM of resazurin sodium salt in PBS pH 7.3) and further incubation for 24 h. Fluorescence was measured at an excitation wavelength of 530 nm and an emission wavelength of 500 nm using a Varioskan™ multimode plate reader (Thermo Fisher Scientific). Antitrypanosomal activity in this context was defined as a reduction in resazurin fluorescence relative to untreated controls, indicating reduced metabolic activity. The assay was performed in four biological replicates. Untreated cells served as negative controls, while cells treated with 1 µM diminazene aceturate, a standard antitrypanosomal drug, served as positive controls. For secondary screening, selected compounds were subjected to two-fold serial dilution (maximum concentration: 1 µM) to determine the IC₅₀ values. The upper concentration limit was capped at 1 µM to prioritize the identification of high-potency compounds with sub-micromolar activity. The assay was carried out as described for the primary screening. IC₅₀ values were determined using a non-linear regression (four-parameter logistic model) in GraphPad Prism version 9. The experiments were conducted in three biological replicates, each with three technical replicates.

Mammalian cytotoxicity was assessed using RAW 264.7 murine macrophages and HEK 293 human embryonic kidney cells in a resazurin-based cell viability assay. These cell lines were selected to provide a preliminary assessment of compound toxicity across immune and non-immune mammalian cell types. Cells were seeded at 1.5 × 10⁴ cells/well and incubated for 24 h in Dulbecco’s Modified Eagle’s Media (DMEM) before compound exposure. After 24-h treatment, alamarBlue dye solution (500 µM of resazurin sodium salt in PBS) was added, and fluorescence was measured after an additional 24 h. As with the antitrypanosomal assay, this assay reflects metabolic activity as a proxy for cell viability. The half-maximal cytotoxic concentration (CC₅₀) values were determined using a non-linear regression (four-parameter logistic model) in GraphPad Prism version 9. The assay was conducted in three biological replicates, each with three technical replicates. Cells treated with phenylarsine oxide (PAO; Sigma-Aldrich) served as the positive control. Selectivity indices (SI = CC₅₀/IC₅₀) were calculated.

To assess effects on parasite proliferation, bloodstream-form T. b. brucei cells were treated with compounds at different concentrations (1 × IC₅₀, 2 × IC₅₀, and 4 × IC₅₀). Cells were seeded at 2 × 10⁵ cells/mL, and viable parasites were counted every 24 h using a haemocytometer over a period of up to 4 days. Cells were subcultured daily to maintain densities at 2 × 10⁵ cells/mL. Cumulative cell numbers were calculated to account for daily dilution during subculturing, and doubling times were obtained directly from non-linear regression analysis (exponential growth model) in GraphPad Prism version 9, using the best-fit parameter output. Statistical comparison between the treated and untreated cells was made using an ordinary one-way ANOVA with Dunnett’s multiple comparison test. Diminazene aceturate-treated cells served as positive controls and untreated cells served as the negative control. Experiments were conducted in three biological replicates, each with three technical replicates.

The mode of T. b. brucei cell death following treatment with MMV667494 and MMV028694 was investigated using the Annexin V-FITC apoptosis detection assay (Sigma-Aldrich). The assay was used to identify apoptosis-like phenotypes based on phosphatidylserine exposure and membrane integrity; however, it does not confirm classical apoptosis in T. b. brucei. The experiment was carried out following the manufacturer’s instructions. Briefly, cells (2 × 10⁵) were treated for 24 h at 1 × IC₅₀, 2 × IC₅₀, and 4 × IC₅₀, except the untreated cells which served as the controls. Propidium iodide was added immediately before acquisition to discriminate membrane-compromised cells. Data were acquired using a BD FACS LSR Fortessa X-20 flow cytometer with BD FACSDiva software version 8. Data analysis was carried out using FlowJo software 10.7.1. Statistical analysis was performed using two-way ANOVA with Dunnett’s multiple comparison test in GraphPad Prism version 9. Three independent experiments (biological replicates) were conducted.

Cell-cycle distribution was assessed using the Guava cell cycle reagent (EMD Millipore, Cat No 4500-0220). Cells were treated for 24 h at 1× IC₅₀, 2 × IC₅₀, and 4 × IC₅₀, and fluorescence acquisition was performed using a BD FACS LSR Fortessa X-20 flow cytometer with BD FACSDiva software version 8. Data were analysed using FlowJo software 10.7.1, and statistical analysis was performed using two-way ANOVA with Dunnett’s multiple comparison test in GraphPad Prism version 9. Two independent experiments (biological replicates) were conducted.

Mitochondrial membrane potential (Δψm) was measured using tetramethylrhodamine ethyl ester (TMRE; Thermo Fisher Scientific, T669). This assay provided a population-level assessment of mitochondrial membrane potential. Cells (2 × 10⁵ cells/mL) were treated for 24 h at 1× IC_50. Valinomycin (100 nM) and troglitazone (10 µM) were used as depolarization and hyperpolarization controls, respectively [37]. Fluorescence was acquired using a BD FACSCalibur (4-colour) cytometer with CellQuest Pro software. Data were analyzed using FlowJo software 10.7.1, and statistical analysis was performed using two-way ANOVA with Dunnett’s multiple comparison test in GraphPad Prism version 9. Two independent experiments (biological replicates) were carried out.

Mitochondrial reactive oxygen species (ROS) production was quantified using the MitoSOX Red assay as a specific indicator of mitochondrial oxidative stress associated with compound-induced cytotoxicity. Bloodstream-form T. b. brucei cells were treated with MMV667494 and MMV028694 at 1 × IC₅₀, 2 × IC₅₀ and 4 × IC₅₀ for 1h and 24 h to distinguish early versus delayed mitochondrial ROS responses. Following treatment, cells were incubated with MitoSOX Red according to the manufacturer’s protocol, and fluorescence intensity was measured by flow cytometry using a BD FACS LSR Fortessa X-20 flow cytometer equipped with BD FACSDiva software version 8.0.1. A minimum of 10,000 events per sample were acquired, and debris and doublets were excluded by forward- and side-scatter gating. Mean fluorescence intensity (MFI) values were calculated for each condition and normalized to untreated controls at the corresponding time point. Data analysis was performed using FlowJo software version 10.7.1, The assay was conducted in three biological replicates. Statistical analysis was carried out using repeated measures two-way ANOVA with Dunnett’s post hoc test. Fluoresence intensity was used as a measure of mitochondrial superoxide levels, indicative of mitochondrial oxidative stress.

To determine whether compound-induced cell death was associated with intracellular oxidative stress, intracellular ROS levels were measured using a fluorometric ROS detection kit (Sigma-Aldrich, Cat No MAK145). Live bloodstream-form T. b. brucei cells (2 × 10⁴ cells per well) were treated with MMV667494 or MMV028694 at 1 × IC₅₀, 2 × IC₅₀ and 4 × IC₅₀ for 1h and 24 h, respectively. Following treatment, the cells were washed with PBS and incubated with 100 µL of the master reaction mix for 1 h at 37 °C and 5% CO₂, protected from light. Fluorescence was measured using a Varioskan™ multimode plate reader at excitation 520 nm and emission 605 nm. Fluorescence values were background-subtracted and normalized to untreated controls at each time point prior to statistical analysis to account for time-dependent changes in cell density and metabolic activity. The assay was conducted in three biological replicates. Statistical analysis was performed using repeated-measures two-way ANOVA with Dunnett’s and Sidak’s post hoc test.

To assess mitochondrial morphology and kinetoplast–nuclear organization, cells were stained with MitoTracker™ Red CMXRos (Invitrogen, Thermo Fisher Scientific) before fixation. The Mitotracker stain served as a qualitative assessment of mitochondrial morphology. Following compound treatment at 1 × IC₅₀, cells were incubated with MitoTracker dye according to the manufacturer’s recommendations and subsequently fixed in 4% formaldehyde, as previously described [31, 38]. Fixed cells were mounted onto poly-L-lysine-coated slides (Sigma-Aldrich) and allowed to adhere for 1 h, washed with PBS, and blocked with 100 mM of Tris-HCl in PBS (pH 7.5) for 10 min and washed again with PBS to reduce background fluorescence. Nuclear DNA was counterstained with Hoechst 33342 (Invitrogen, Cat No H3570) before the final PBS wash and mounting with Vectashield antifade mounting medium (Vector laboratories, Cat No H-1000).

Fluorescence imaging was performed using a Zeiss Axio Observer Z1 microscope equipped with an LSM 800 confocal unit using a 63× oil-immersion objective (Plan-Apochromat 63x/1.4 Oil DIC M27). Image acquisition was carried out using Zen Blue 2.3 software, with fluorescence channels configured using the “smart setup” function for two-colour imaging. Hoechst 33342 and Mitotracker Red CMXRos were excited using the 405 nm (5 mW) and 561 nm (10 mW) laser lines, respectively, with emission collected at 410–546 nm and 590–700 nm. Detector saturation was avoided using the range-indicator function, and Z-stacks were acquired at 1 µm intervals with a pinhole size of 1 Airy unit. Images were acquired in CZI format and exported as TIFF files for analysis in ImageJ (Fiji). Linear adjustments to contrast and brightness were applied uniformly across conditions, and Z-stacks were visualized using maximum-intensity projections. Kinetoplast/nucleus (K/N) counts as well as cell and mitochondrial morphology profiling were carried out manually using cell counts. At least 500 cells per condition were analyzed across biological replicates. Statistical analysis was carried out using one-way ANOVA with Sidak’s post hoc test.

A total of 160 compounds from the MMV Pathogen Box (plates A and E compounds of theathogen box) were subjected to single-concentration in vitro antitrypanosomal phenotypic screening against bloodstream-form T. b. brucei. The primary screen was designed to be a one-point phenotypic assay to identify compounds that reduced parasite metabolic activity at 1 µM rather than to determine IC₅₀ values. After this initial screen, 10 compounds demonstrating activity were identified for secondary screening. Secondary screening was performed using eleven two-fold serial dilutions (maximum concentration 1 µM), allowing the determination of IC₅₀ values (Table 1). Two compounds, MMV667494 and MMV028694 (Figure 1), exhibited sub-micromolar potency (IC₅₀ < 0.5 µM were prioritized for further analysis based on their potency. Both compounds are classified by MMV as antimalarial hits showed the most promising in vitro antitrypanosomal activity among the compounds tested (Table 1). To assess parasite selectivity, mammalian cytotoxicity was evaluated using RAW 264.7 murine macrophages and HEK 293 human kidney cells. Both MMV667494 and MMV028694 showed good selectivity toward T. b. brucei, with selectivity indices greater than 10 (calculated using 24 h CC₅₀ and IC₅₀ values), indicating their preferential toxicity toward trypanosomes relative to mammalian cells (Tables 1, 2; Supplementary Material 2).

The effect of both compounds on the proliferation of T. b. brucei cells was assessed by longitudinal growth profiling following treatment with increasing concentrations of MMV667494 (Figures 2A,B) and MMV028694 (Figures 2C,D). At 24 h, both compounds reduced cell numbers across all the tested concentrations (1 × IC₅₀, 2 × IC₅₀, and 4 × IC₅₀), although statistical significance was reached only at higher concentrations - specifically MMV667494 at 2 × IC₅₀ (p = 0.041), and 4 × IC₅₀ (p = 0.014) and MMV028694 at 4 × IC₅₀ (p = 0.026) (Figure 2). At 48 h and 72 h, both compounds produced dose-dependent reductions in parasite numbers, with statistically significant inhibition observed at all concentrations tested (p ≤ 0.0001) (Figures 2A,C). Notably, a net decline in parasite numbers was observed at 4 × IC₅₀ after 72 h, indicating that both MMV667494 and MMV028694 killed a proportion of the trypanosome population under these conditions. At lower concentrations, parasite growth was inhibited without net loss of viable cells, consistent with a predominantly trypanostatic effect that transitions to trypanocidal activity at higher exposure, like previously reported for a series of adenosine analogues [39]. Consistent with these observations, cell doubling times were significantly prolonged following treatment. At 1 × IC₅₀, doubling times increased to 17.7 h (MMV667494) and 18.7 h (MMV028694), while at 2 × IC₅₀, doubling times further increased to 20.0 h and 27.6 h, respectively, compared with 7.4 h for untreated cells (Figure 2E). These findings confirm that both compounds markedly impair parasite proliferation in a concentration-dependent manner.

The annexin V-FITC/PI assay was used to investigate whether the observed trypanocidal cell death was associated with apoptosis-like or necrotic phenotypes. Phosphatidylserine is usually found in the inner leaflet of the plasma membrane of healthy cells. Upon initiation of apoptosis, phosphatidylserine is translocated to the extracellular membrane leaflet, making it easily accessible for annexin V-FITC to bind to it. Propidium iodide (PI), a nucleic acid binding dye, is a fluorescent indicator that is excluded from intact cells but enters when the plasma membrane becomes compromised. Annexin V-FITC/propidium iodide (PI) staining therefore allows to distinguish between viable (FITC-/PI-), early apoptotic-like (FITC+/PI-), late apoptotic-like (FITC+/PI+), and necrotic (FITC-/PI+) populations. The term “apoptotic-like” is used because T. b. brucei lacks canonical caspases, which are the classical apoptotic cell death signalling machinery, despite displaying hallmark features such as phosphatidylserine externalisation [40]. Treatment with MMV667494 for 24 h resulted in a significant increase in early apoptotic-like cells defined as increased Annexin V fluorescence but unchanged cell permeability measured with PI, at 2 × IC₅₀ (10.9%, p = 0.0004) and 4 × IC₅₀ (12.0%, p < 0.0001) compared with untreated controls 0.49%) (Figure 3). Late apoptotic-like cells were also significantly increased at 4 × IC₅₀ (15.6%, p < 0.0001) with increased fluorescence on both channels, relative to untreated cells (0.52%) (Figures 3D,E). A similar pattern was observed for MMV028694, although significant increases in early apoptotic-like cells were already evident at 1 × IC₅₀ (5.3%. p = 0.0193) (Figure 4). Despite these increases, the overall proportion of Annexin V-positive cells remained relatively modest indicating that apoptosis-like features were restricted to a subset of the population at the time points examined.

The mitochondrial membrane potential (ΔΨm) of T. b. brucei cells treated with MMV667494 and MMV028694 was determined by flow cytometry in TMRE-stained live cells exposed to a concentration of 1 × IC₅₀ of the compounds for 24 h (Figures 5A–F). Both MMV667494 and MMV028694 treatments resulted in a statistically significant reduction in mean ΔΨm (p = 0.029 and p = 0.001 respectively). However, analysis of fluorescence distributions revealed that depolarization was confined to a distinct subpopulation of cells, while the majority retained mitochondrial membrane potential. This is evident from the observed low-fluorescence peak in treated samples (Figures 5D,E), suggesting that mitochondrial depolarization is associated with cells undergoing death or severe stress rather than a uniform effect across the population. Consistent with this interpretation, confocal microscopy of MitoTracker-stained cells showed largely preserved mitochondrial morphology in most treated cells, particularly following MMV667494 exposure (Figure 5F; Supplementary Figure S1). Control treatments behaved as expected: troglitazone induced mitochondrial hyperpolarization, while valinomycin caused complete depolarization (both p < 0.0001) (Figures 5A–C), in agreement with previous reports [37, 41].

The effect of MMV667494 and MMV028694 on mitochondrial levels of reactive oxygen species was determined using the MitoSOX red assay at 1 h and 24 h of treatment (Figure 6). MMV667494 did not significantly alter mitochondrial ROS levels at 1 h, whereas MMV028694 caused a significant increase only at 4 × IC₅₀ (p < 0.0001) but not at lower concentrations (Figure 6A). After 24 h, both compounds significantly increased mitochondrial ROS levels at all concentrations tested, relative to untreated controls (Figure 6A). Time-course analysis revealed a significant increase (p < 0.0001) in mitochondrial ROS between 1 h and 24 h for MMV667494 at all concentrations (Figures 6A,B). For MMV028694, significant increases between time points were observed at 1 × IC₅₀ and 2 × IC₅₀ (p < 0.0001 for both), but not at 4 × IC₅₀ (p = 0.76), likely reflecting advanced cell damage or loss of viable mitochondria at higher exposure (Figure 6B). Overall, the magnitude of ROS increases was modest, and the delayed onset suggests that mitochondrial oxidative stress is more likely a downstream consequence of compound-induced cellular effects rather than the primary cause of cell death.

The intracellular levels of ROS in MMV667494-treated, MMV028694-treated and untreated cells were determined using a fluorescence-based commercially available kit as earlier described. The results obtained showed that neither MMV667494 nor MMV028694 significantly altered intracellular ROS levels at 1 h or 24 h across all tested concentrations, except for a just-significantly modest decrease observed at 1 h for MMV028694 at 4 × IC₅₀ (p = 0.029) (Supplementary Material 4). As this effect was not sustained at 24 h, it is unlikely to be biologically meaningful. A similar increase in intracellular ROS was observed over time in both treated and untreated cultures, consistent with increased cell density and prolonged culture, leading to media changes, rather than compound-specific effects. These findings suggest that intracellular oxidative stress is not the primary mechanism underlying MMV667494- or MMV028694-induced parasite death.

Cell cycle effects were analyzed using DNA-content flow cytometry (Figure 7) and confocal microscopy (Supplementary Material 3, 5). Following MMV667494 treatment, a significant reduction in G1-phase cells was observed at 4 × IC₅₀ (p < 0.0001), accompanied by a concomitant increase in S-phase cells (Figure 7A). G2-phase cells were significantly reduced at all concentrations tested, while cells with sub-G1 DNA content (<G1) increased in a dose-dependent manner, consistent with DNA fragmentation or aberrant division. Similarly, MMV028694 treatment resulted in the accumulation of <G1 cells and depletion of G2-phase cells i.e., with a completely replicated diploid set of chromosomes, at all concentrations (Figure 7B). A modest reduction in G1-phase cells was also observed, reaching statistical significance at 2 × IC₅₀ (p = 0.0326). These profiles are consistent with disrupted cell cycle progression and accumulation of cells with abnormal DNA content. Overall, both compounds produced cell-cycle profiles consistent with slowed S-phase progression and impaired cell division, in agreement with the observed growth defects. The observations suggest that the compounds exerts a direct or indirect inhibitory action on DNA synthesis, likely not only prolonging S-phase but causing incomplete replicate of daughter chromosomes, leading to the highly significant drop in cells with fully replicated nuclei and associated increase in cells that contain less DNA than the normal diploid set of chromosomes (<G1 DNA content) after cell division, which would not be viable.

Confocal microscopy was used to validate the results obtained from the flow cytometry-based cell cycle progression assay. The confocal microscopy corroborated these findings at 1 × IC₅₀ and 24 h of exposure. Both compounds increased the proportion of 0K1N cells (∼6% vs. 1.9% in controls) and 2K1N cells, while 1K2N and other aberrant K/N configurations remained rare (Supplementary Material 5A). MMV028694 uniquely increased the proportion of 1K0N cells, and rare events such as cojoined nuclei and cytokinesis failure (<1%) were observed (Supplementary Material 5B). Gross morphological distortions were limited to a very small fraction of treated cells and were not widespread (Supplementary Material 5C), indicating that cell death and growth inhibition primarily result from functional disruption of cell-cycle progression rather than widespread structural damage. Together, the most prominent changes in the N/K configuration therefore concerned the lack of an intact (visible) nucleus or kinetoplast at the point of cell division or thereafter. This is consistent with an impact of the treatments on DNA synthesis both of kinetoplast and nuclear DNA.