An image-based high-content screening for compounds targeting Toxoplasma gondii repurposed inhibitors effective against the malaria parasite Plasmodium falciparum

Abstract

Apicomplexa phylum includes numerous obligate intracellular protozoan parasites that are life threatening for humans and animals. In this context, Plasmodium falciparum and Toxoplasma gondii are of particular interest, as they are responsible for malaria and toxoplasmosis, respectively, for which efficient vaccines are presently lacking and therapies need to be improved. Apicomplexan parasites have a highly polarized morphology, with their apical end containing specific secretory organelles named rhoptries and micronemes, which depend on the unique receptor and transporter sortilin TgSORT for their biogenesis. In the present study, we took advantage of the subcellular polarity of the parasite to engineer a clonal transgenic Toxoplasma line that expresses simultaneously the green fluorescent protein TgSORT-GFP in the post-Golgi-endosome-like compartment and the red fluorescent protein rhoptry ROP1-mCherry near the apical end. We utilized this fluorescent transgenic T. gondii to develop a miniaturized image-based phenotype assay coupled to an automated image analysis. By applying this methodology to 1,120 compounds, we identified 12 that are capable of disrupting the T. gondii morphology and inhibiting intracellular replication. Analysis of the selected compounds confirmed that all 12 are kinase inhibitors and intramembrane pumps, with some exhibiting potent activity against Plasmodium falciparum. Our findings highlight the advantage of comparative and targeted phenotypic analysis involving two related parasite species as a means of identifying molecules with a conserved mode of action.

Keywords: ROP1-mCherry; TgSORT-GFP; Toxoplasma gondii; high-content screening; image-based analysis; malaria parasite; repurposing drugs.

Figure 1

Generation of transgenic T. gondii expressing TgSORT-GFP/ROP1-mCherry. (A) corresponds to the vector that allows the knock-in of ROP1-mCherry in the parasites. One clone was selected and the red color indicates the ROP1 signal located at the apical end of four intracellular dividing tachyzoites (B). Thus, the vector expressing the TgSORT-GFP was transfected (C) and one positive clone (green signal), which replicates into eight intracellular tachyzoites (D) for the expression of this receptor in the post-Golgi and endosome-like compartment (ELC) is shown. Bar scale = 3 µm.

Figure 2

Assay workflow showing all steps of the library screening protocol, namely cell culture, parasite infection, sample processing and labeling, and image acquisition followed by automated image analysis. MOI, multiplicity of infection.

Figure 3

Typical images and image analysis workflow (described in detail in the Columbus analysis script): (1) Raw image acquisition from the reader; (2) Segmentation on the blue channel to detect the HFF nuclei (stained with DAPI reagent), followed by false detection removal based on the intensity/morphology properties in order to retain true nuclei only; (3) HFF cytosol delimitation around nuclei using predefined algorithm based on the use of the DAPI channel, which also stains cytosol (producing a signal of lower intensity); 4) Parasite detection within the cytosol region allowing the DAPI-labeled plots to be segmented into parasite nuclei; (5) Vacuole determination by the total proximal parasites nuclei area (as parasites grow in the vacuole the nucleus area can be mathematically expanded around them); (6) mCherry intensity, enabling determination of disrupted vacuoles. In order for a vacuole to be considered as positive, the mean of the mCherry fluorescence intensity should not be too high and the ratio of mCherry intensity between the vacuole center and the periphery must exceed 1.4.

Figure 4

Pre-screen optimization using pyrimethamine, the well-known anti-Toxoplasma drug as control. (A) Checker-board layout; (B) Determination of the parasite number per HFF cells for DMSO and Pyr at 10-µM concentration. VAL1 and VAL2 correspond to the two analyzed plates.

Figure 5

Plate heat map for each of the three screening plates (P37, P43 and P49) for each well, where each square corresponds to a well. DMSO was dispensed in the 1^st eight rows of the 1^st column and the last eight rows of the 21^st column, while the two compounds—SB 239063 (Cpd2) and SB 208 (Cpd3)—were placed in the last eight rows of the 1^st and the first eight rows of the 21^st column, respectively. For each well, the Z-score determined on the percentage of vacuoles displaying disruption is given, with Z-score > 3 indicating a compound impacting vacuole disruption.

Figure 6

Data dot plot for each of the three screening plates—P37 (circles), P43 (diamonds) and P49 (triangles). The percentage of positive vacuoles is shown before normalization on the Y-axis. Red color corresponds to the negative control (DMSO), and blue and pink denote the positive controls—SB 239063 (Cpd2) and SB 208 (Cpd3)—in each plate, while yellow color is used for the tested compounds and black for non-infected wells.

Figure 7

Screening hits. Five hits were obtained using our automatic image-based screening protocol and these fluorescence images (featuring necrotized intracellular parasites) illustrate the disruption of classical apical location of ROP1 signal compared to the negative DMSO control that shows intracellular dividing tachyzoites with normal morphology.

Figure 8

Chemical structures of all 12 hits with their corresponding names in the Tocris library, along with the nomenclature used in the present study (Cpd) and with the number corresponding to the order of identification.

Figure 9

Dose-dependent inhibitory activities on the 12 compounds using β-galactosidase assays. Panel (A) shows the assays performed on the first six compounds (including Cpd5, which probably represents a false positive). Panel (B) illustrates the six remaining compounds. These two panels correspond to single-dose assays and all p values are below 0.0001. Panel (C) pertains to a mix of three compounds (p value = 0.0024), while Panel (D) shows a dose-dependent inhibition of plaque formation by T. gondii after nine days of drug exposure. As two independent experiments yielded identical results, one is illustrated here. Panel (E) showing confocal microscopy of intracellular tachyzoites treated with 25 µM of drugs indicated on the images for 24h followed by staining with different antibodies specific to rhoptries, micronemes, dense granules and the inner complex membrane. The control corresponds to intracellular parasites treated with DMSO alone. Nuclei are stained with DAPI. Bar, 3 µm.

Figure 10

Dose-dependent inhibitory effects of the studied drugs on P. falciparum growth in red blood cells (RBCs). Panels (A–C) respectively showing ring stage, trophozoite and schizonte counts after drug treatments based on flow cytometry, with chloroquine as positive control (p < 0.0001). Panel (D) shows a blood smear used to visualize ring stages of the DMSO control versus selected drugs in all used concentrations.