High-Throughput Repurposing Screen Reveals Compounds with Activity against Toxoplasma gondii Bradyzoites

Abstract

Toxoplasma gondii causes widespread chronic infections that are not cured by current treatments due to the inability to affect semidormant bradyzoite stages within tissue cysts. To identify compounds to eliminate chronic infection, we developed an HTS using a recently characterized strain of T. gondii that undergoes efficient conversion to bradyzoites in vitro. Stage-specific expression of luciferase was used to selectively monitor the growth inhibition of bradyzoites by the Library of Pharmacological Active Compounds, consisting of 1280 drug-like compounds. We identified 44 compounds with >50% inhibitory effects against bradyzoites, including new highly potent compounds, several of which have precedent for antimicrobial activity. Subsequent characterization of the compound sanguinarine sulfate revealed potent and rapid killing against in vitro-produced bradyzoites and bradyzoites harvested from chronically infected mice, including potent activity against intact cysts. These findings provide a platform for expanded screening and identify promising compounds for further preclinical development against T. gondii bradyzoites that are responsible for chronic infection.

Toxoplasma gondii is a widespread parasite of animals that causes zoonotic infections in humans.¹ Serological studies suggest that ∼1/3 of humans are chronically infected with T. gondii worldwide, although prevalence rates vary widely by geographic region.² Human infections are caused by ingestion of undercooked meat, harboring tissue cysts, or ingestion of oocysts shed by infected cats.³ The acute stage of infection is predominated by fast-growing tachyzoites that disseminate widely, including to sites of immune privilege like the brain, followed by differentiation into cysts harboring semidormant bradyzoites, which divide slowly and asynchronously.⁴ Although most infections are only mildly symptomatic and controlled by the immune system, they persist chronically and hence predispose individuals to subsequent reactivation if they become immunocompromised.⁵ The current standard of care treatment does not eradicate chronic infection; hence, infected individuals remain at risk of reactivation for life.

Long known as a cause of congenital infection,⁶ recent studies highlight toxoplasmosis as a cause of ocular disease due to newly acquired infections in otherwise healthy adults.⁷ Frequent and severe outbreaks of ocular toxoplasmosis have been described in South America⁸ and India,^9,10 and may occur in other localities. Notably, infection of healthy individuals in South America often leads to severe, recurrent ocular toxoplasmosis,¹¹ with an estimated disease burden of 30 million individuals requiring treatment annually in Brazil alone.¹² Additionally, prior T. gondii infection has been linked to increased cognitive decline in Alzheimer’s Disease, although not all studies have confirmed this association, likely due to the complexity of the risk factors involved.¹³

Current therapies for the treatment of toxoplasmosis rely on the inhibition of the folate pathway in the parasite.¹⁴ The standard of care therapy (i.e., sulfadiazine and pyrimethamine) inhibits the rapidly growing tachyzoite stage but has minimal activity on bradyzoites within tissue cysts and consequently does not eliminate chronic infection.¹⁵ Unfortunately, there are also significant adverse effects of this treatment regimen due to intolerance or allergic reactions^16,17 and contraindication during the first two trimesters of pregnancy.¹⁸ Although drug resistance is not frequently encountered in the treatment of toxoplasmosis, some isolates are naturally resistant to sulfonamides due to natural variants in dihydropteroate synthetase, or other molecular mechanisms, thus complicating treatment in some cases.¹⁷ Clindamycin and several macrolide antibiotics have also been shown to inhibit growth in vitro and in animal models^19,20 and such antibiotics have been used to treat toxoplasmosis in humans.²¹ However, these compounds are not specific to the parasite and disrupt the normal endogenous microbiota leading to the possible emergence of pathobionts like Clostridium difficile.²²

A review of the literature of compounds that are clinically approved in humans or animals identified several compounds that inhibit parasite growth in vitro and/or in murine models of toxoplasmosis.²¹ For example, repurposing of guanabenz, an FDA-approved drug that interferes with translation, showed activity against acute and chronic toxoplasmosis in mice,²³ although this effect was dependent on the strain of the mouse,²⁴ and treatment did not eliminate cysts, resulting in rebound after discontinuation.²⁵ A repurposing screen for inhibition of tachyzoite growth in vitro using the Tocriscreen Total Library, which consists of 1,280 biologically active small molecules with confirmed molecular targets, identified multiple compounds that affect dopaminergic and estrogen signaling, including tamoxifen that was shown to act by upregulating xenophagy to restrict parasite growth and cause clearance.²⁶ Additionally, a number of new investigational compounds have shown an ability to block tachyzoite growth in vitro and during acute infection in animal models, including several that have activity against bradyzoite growth or chronic infection.¹⁷

The majority of efforts to identify new compounds with activity against T. gondii have focused on in vitro assays using tachyzoite growth as a readout.^17,21 Differentiation of bradyzoites in vitro can be achieved by treatment with stress, such as high pH;²⁷ however, the stages that develop under stress often continue to express tachyzoite traits,^28,29 thus complicating screening efforts. Nonetheless, combining dual promoters to drive firefly luciferase (FLUC) in the cytosol of bradyzoites and Nanoluc luciferase (nLuc) that was engineered to be secreted into the cyst matrix, allowed evaluation of compounds for selective activity against chronic stages.³⁰ Another promising development is the recent development of an in vitro system for the development of bradyzoites using a specialized KD3 muscle cell line where spontaneous differentiation occurs.³¹ Consistent with the lack of available treatments for tissue cysts in chronically infected mice, treatment with pyrimethamine and/or sulfadiazine was not effective in restricting the growth of mature bradyzoites formed in vitro in this system.³¹ Hence, this system could provide a useful platform for testing compounds for their activity against bradyzoites. However, this system requires the use of a specialized culture system, and the treatment and recovery phases needed for evaluation take ∼50 days, complicating its use for screening.

As an alternative, we recently described a type II strain called Tg68 that has a high propensity to differentiate into bradyzoite in vitro under conditions of stress that include high pH or cultivation in high glutamine, low glucose, that forces metabolism based on glutaminolysis.³² Unlike other type II strains that undergo partial differentiation, Tg68 forms fully mature bradyzoites without associated breakthroughs of tachyzoites following stress induction in vitro.³² Here, we engineered this strain to express Fluc under the control of a constitutive promoter and separately generated a line expressing nLuc under the control of a bradyzoite promoter. We used these reporter lines to develop a high throughput screen (HTS) and used it to evaluate the Library of Pharmacologically Active Compounds (LOPAC). Several compounds with activity against both stages were identified, providing a proof of concept for further HTS projects designed to find new treatments for chronic toxoplasmosis.

Results and Discussion

To facilitate HTS using the Tg68 strain, we generated a clonal line of Tg68 constitutively expressing Fluc under the pTUB1 promoter (Figure 1A). The pTub1:Fluc plasmid also contained a resistant DHFR cassette that was integrated into the genome after electroporation followed by selection with pyrimethamine. Following passage in HFF cells, a cloned line constitutively expressing Fluc was isolated and is referred to as Tg68-pTub1:Fluc. Confluent HFF cells grown in 384-well plates were infected with Tg68-pTub1:Fluc tachyzoites at 2 × 10⁴ parasites per well, and the infection was allowed to proceed for 72 h under 5% CO₂ at 37 °C (Figure 1B). Comparison of Fluc activity between day 3 and day 0 (4 h postinfection) increased significantly, representing the growth of the parasite (Figure 1C). In addition, we analyzed Fluc expression from replicate 14 wells, across 5 plates, and determined the Coefficient of Variation (CV) for positive wells was 7 ± 3% (Table S1). The addition of BRD7929, which targets parasite phenylalanine tRNA synthetase,³³ completely inhibited parasite growth at a concentration of 10 μM (Figure 1D). A comparison of the luciferase signal from untreated and treated wells from 5 plates demonstrated an average Z′ value of 0.77 ± 0.11 (Table S1). Additionally, the Tg68-pTub1:Fluc line demonstrated sensitivity comparable to that of BRD7929 and atovaquone compared to the reference type II strain TgMe49-Fluc (Figure 1D). Taken together, these findings indicate that Tg68-pTub1:Fluc provides a robust readout for HTS of compounds against tachyzoite growth.

Figure 1.

Development of a High Throughput Screening assay for growth inhibition of Tg68 tachyzoites. (A) A tachyzoite-specific firefly luciferase (Fluc) reporter strain of Tg68 was generated using the pTUB1 promoter. (B) Confluent HFF cells in 384-well plates were infected with Tg68 tachyzoites, compounds were added immediately in a dilution series, and luciferase activity was measured on day 3. (C) Replicates of 12 wells from six independent plates were analyzed for luciferase activity. Fluc expression increased more than 1,000-fold and was found significantly higher at day 3 (72 h) compared to day 0 (4 h postinfection). Data accumulated from 12 wells across 6 plates. Mann–Whitney test, P < 0.02. (D) EC₅₀ values were determined for Tg68-pTub1:Fluc and ME49-Fluc treated with serial dilutions of BRD7929 or Atovaquone. All EC₅₀ values are presented as the mean of three biological replicates (n = 3).

To expand the use of Tg68 for HTS of compounds with activity against bradyzoites, we generated a clonal line expressing nLuc under the control of the pBAG1 promoter, which is strongly upregulated upon differentiation.³⁴ The pBAG1:nLuc plasmid also contained a resistant DHFR cassette that was integrated into the genome after electroporation, followed by selection with pyrimethamine. Following serial passage in HFF cells, a positive clone was isolated and referred to as Tg68-pBAG1:nLuc (Figure 2A). In previous studies, we have shown that the Tg68 strain undergoes highly efficient differentiation to bradyzoites in alkaline media at low CO₂ in vitro and that it retains a mature bradyzoite profile even at relatively high multiplicities of infection.³² Tg68 also undergoes efficient differentiation when grown in a glucose-free medium supplemented with 10 mM glutamine, a process that forces glutaminolysis for energy production via the mitochondrion.³² Confluent HFF cells in 384-well plates were infected with pBAG1:nLuc parasites (3 × 10³ parasites per well) for 2 h, followed by washing and shifting to alkaline or glutamine media. Cultures were maintained for 10 days under CO₂-free conditions (ambient air), with media changes on days 3 and 6. On day 6, compounds were added, and at day 10 luciferase assays were performed (Figure 2B). Expression of nLuc was very low in cultures of tachyzoites and in the initial culture conditions at day 0 (4 h postinfection) in normal or differentiation media (Figure 2C). Expression of nLuc increased dramatically by over 4 logs by day 3 and continued to increase significantly on days 6 and 10 (Figure 2C). On day 10, analysis of luciferase values from 14 wells showed an optimal Coefficient of Variation (CV) of 10 ± 2% for alkaline media and 7 ± 2% for glutamine media, indicating suitability for high-throughput screening (HTS) (Table S1). Treatment with BRD7929 at a concentration of 10 μM completely inhibited the growth of the parasites, and comparison of nLuc activity from treated and untreated wells revealed a Z′ value of 0.67 ± 0.05 for alkaline media and 0.76 ± 0.06 for glutamine media (Table S1). We also compared the potency of BRD7929 and atovaquone under both conditions that induced bradyzoite differentiation (Figure 2D). Both compounds showed a reduction in EC₅₀ values under glutamine differentiation when compared to alkaline conditions, although this was much more dramatic for atovaquone which showed a > 30-fold shift (Figure 2D). The greater potency of compounds in glutamine medium may reflect a decreased ability of the parasite to generate energy stores from glutaminolysis vs glycolysis as suggested previously by the modest growth defects in knockouts of T. gondii hexokinase³⁵ and glucose transporter 1.³⁶ Consequently, the use of glutamine medium for bradyzoite development may preferentially reveal compounds that act on the mitochondrion, as is the case for atovaquone which inhibits the bc1 complex.³⁷

Figure 2.

Development of a High Throughput Screening assay for the growth inhibition of Tg68 bradyzoites. (A) A bradyzoite-specific Nanoluc luciferase (nLuc) reporter strain of Tg68 was generated using the pBAG1 promoter. Parasites were grown in alkaline medium (pH 8.2), CO₂ free, or in the absence of glucose supplemented with glutamine, which stimulates in vitro development of bradyzoites. (B) Confluent HFFs in 384-well plates were infected with Tg68 tachyzoites for two h, washed, and then cultured either in D10 under normal conditions (Tz) or switched to alkaline or glutamine conditions to induce bradyzoites. The cultures were then maintained under CO₂-free conditions for 10 days with media changes on days 3 and 6, with a compound treatment beginning at day 6 and readout at day 10. (C) Luciferase signals from tachyzoites (Tz) harvested at day 0 (4 h postinfection) or bradyzoites induced for different times (day 0 (4 h) to day 10) by culture in alkaline or glutamine media. Comparisons between sequential time points using the Mann–Whitney test, ****P < 0.0001. (D) Determination of EC₅₀ values for BRD7929 and atovaquone treatment of in vitro induced bradyzoites culture in alkaline or glutamine media. All EC₅₀ values are presented as the mean of four biological replicates (n = 4).

The LOPAC library (1157 of 1280 total compounds) was plated in a 384-well format and tested for parasite growth inhibition in duplicate using a single concentration of 10 μM in the tachyzoite and both alkaline and glutamine-induced bradyzoite assays. Growth inhibition was averaged from the two replicates and plotted as a Venn diagram summarizing the outcome of each of the three assays (Figure 3A). A total of 27 compounds showed selective inhibition of tachyzoite growth without affecting growth in the other assays (Figure 3A). The largest number of compounds was identified in the glutamine-induced bradyzoite assay, perhaps reflecting the metabolic liability of this growth condition (Figure 3A). In total, 21 compounds that inhibited parasite growth in all three assays by 50% or more were identified as Primary Hits (Figure 3A). Furthermore, we identified 9 compounds that inhibited bradyzoite growth in both assays by 50% or more but not tachyzoite growth (Figure 3A). Finally, we identified 14 compounds that specifically inhibited glutamine-induced bradyzoite growth by ≥80% but were not effective in growth inhibition in the other two assays (Figure 3A). Robust validation of the within plate and between replicates was performed through collective calculation of the coefficient of variation (CV) and Z′ analysis of assay plates (Table S2). Of the 44 primary hit compounds defined above (Figure 3B), 36 compounds were available in quantity at Calibr and were chosen for further testing (Table S3). These 36 compounds were tested in a dilution series and 9 compounds were found to have EC₅₀ values of ≤2 μM in either tachyzoite and/or bradyzoite assays, thus defining a set of top hits (Table S3, Figure 3B).

Figure 3.

Summary of LOPAC screening for the growth inhibition of T. gondii. (A) Venn diagram showing the number of compounds with ≥50% growth inhibition at 10 μM in each of three growth assays. Tz = tachyzoite growth assay, Alk = alkaline induced bradyzoite growth assay, Gln = glutamine induced bradyzoite growth assay. Red circled numbers indicate the selection criteria for Primary Hits. (B). Summary of LOPAC screen and prioritization of Hits for follow up. Of 44 Primary Hits, 36 were available for dilution series to determine EC₅₀ values. Top 9 Hits were available for further biological testing.

We were able to source the 9 Top HIT compounds from commercial sources. The 9 available compounds were tested for cytotoxicity against HepG2 and differentiated THP-1 cell lines, along with a control drug, atovaquone (Figure 3B). Most compounds exhibited a favorable selectivity index (SI), which is a measure of the compound’s EC₅₀ against the parasite compared to the CC₅₀ against host cells (Table 1). However, a few compounds, namely, Idarubicin hydrochloride and MS012, showed high toxicity, while Auranofin and JFD00244 showed modest toxicity toward the cell lines. Auranofin is a gold-containing compound that has been approved for the treatment of rheumatoid arthritis and it has previously been reported to have activity against several parasites.³⁸ Consistent with this profile, auranofin has previously been shown to inhibit replication of the type I RH strain in vitro, reduce infection in a chicken embryo model,³⁹ and reduce the burden of cysts in chronically infected mice.⁴⁰ However, the high level of growth inhibition for HepG2 and differentiated THP-1 cells treated with auranofin in the present study would appear to limit the potential of this compound for further clinical development. Additionally, Brefeldin A, which blocks ER to Golgi transport, showed a favorable SI in HepG2 cells but not in differentiated THP-1 cells (Table 1), suggesting that block of protein export has very different consequences on host cell growth in different lineages. Although this level of toxicity may seem surprising, the compounds included in LOPAC have been chosen for being “pharmacologically active and drug-like” but most have not been screened for safety or toxicity. Additionally, most have not been screened for inhibition of mammalian cell growth, although a number of the compounds in Table 1 have entries in PubChem that are consistent with toxicity (https://pubchem.ncbi.nlm.nih.gov/). As a result, these compounds were excluded from further assays due to their undesirable cytotoxic effects.

Table 1. Potency and Selectivity of Top Hits^a.

compound	Tg68Fluc EC50(tachyzoite)	Me49Fluc EC50(tachyzoite)	Tg68nLuc EC50(alkaline)	Tg68nLuc EC50 (glutamine)	HepG2 CC50	SIb	THP-1c CC50	SI
diphenyleneiodonium sulfate	0.01	0.01	0.27	0.06	1.28	96	4.85	365
emetine dihydrochloride hydrate	0.03	0.03	2.50	1.08	1.92	64	0.23	8
sanguinarine sulfate	0.3	0.1	4.00	1.25	9.44	32	4.34	15
brefeldin A	0.06	0.05	0.40	0.16	60.00	987	0.07	1
idarubicin hydrochloride	0.19	0.67	0.30	0.82	0.14	1	0.59	3
T0070907	0.75	0.87	0.64	0.60	18.12	24	21.98	29
MS012	2.46	2.68	0.85	2.23	2.03	1	1.19	0.5
auranofin	2.75	1.12	0.68	0.38	9.81	4	2.83	1
JFD00244	3.89	3.39	2.88	4.48	16.07	4	2.95	1
atovaquone	0.16	0.10	2.07	0.06	9.67	62	14.26	91
BRD7929	0.11	0.05	0.22	0.08	6.08	54	2.87	25

^aValues in μM, average of two or more biological replicates.

^bSI = selectivity index (Host cell CC₅₀/EC₅₀ Tg68Fluc).

^cTHP-1 cells were differentiated as described in the Methods.

Although Tg68 undergoes efficient conversion to bradyzoites in vitro, with a transcriptional profile that resembles in vivo bradyzoites,³² it may still lack some of the features of mature tissue cysts. Hence, we tested select Top Hits, along with several reference compounds, against tissue cysts that were harvested from chronically infected mice (Figure 4A). For these assays, we used the ME49 EW strain, which produces high numbers of cysts in vivo.⁴¹ As shown in Table 1, the ME49 strain has a sensitivity very similar to that of the Top Hits and reference compounds, thus validating the choice of strain. Compounds were tested using continuous exposure of ex vivo bradyzoites to 3XEC₉₀ as measured on tachyzoites. Alternatively, ex vivo bradyzoites were treated at 3XEC₉₀ for only 4 h followed by washout to ascertain how irreversibly they might act (Figure 4A). All three of the reference compounds BRD7929, atovaquone, and pyrimethamine showed potent inhibition of bradyzoites when they were used continuously (Figure 4B). Although atovaquone and BRD7929 are active on bradyzoites (Table 1), the activity of pyrimethamine in this continuous treatment assay is likely because they inhibit the outgrowth of tachyzoites. Consistent with this specificity, pyrimethamine was largely ineffective when used for only 4 h, indicating it has minimal effects on bradyzoites present at the start of the assay (Figure 4B).

Figure 4.

Testing of compounds for activity against ex vivo bradyzoites. (A) Schematic for the isolation of bradyzoites from chronically infected mice and testing in vitro during continuous treatment or after 4 h treatment and washout. (B) Testing of sanguinarine and reference compounds in continuous treatment vs 4 h treatment and washout. Values are determined from plaque counts after 10–14 days of outgrowth and are normalized to the DMSO control for each condition. (C) Schematic for 4 or 24 h treatment of intact cysts followed by in vitro plaque formation in media with or without compounds. (D) Testing of sanguinarine and reference compounds on intact cysts for 4 h treatment. (E) Testing of sanguinarine and reference compounds on intact cyst for 24 h treatment. Values are determined from plaque counts after 10–14 days of outgrowth and are normalized to the DMSO control for each condition. Compounds were used at 3XEC₉₀ based on the sensitivity of Tg68Fluc tachyzoites (Table S3). Data from 3 to 6 biological replicates, bar graph represents the average of percentage of number of plaques in the treatment group normalized to DMSO control.

Similar to pyrimethamine, testing of diphenyleneiodonium and T0070907 in the ex vivo bradyzoite assay revealed that they only work when used in continuous treatment, suggesting their activity is static rather than cidal (Figure S1). Diphenyleneiodonium is an inhibitor of NADPH oxidase that separately induces oxidative stress.⁴² Previous studies have shown diphenyleneiodonium inhibits the growth of T. gondii tachyzoites in ARPE-19 cells through the production of ROS.⁴³ In a separate study, it also showed activity against P. falciparum with an EC₅₀ of 0.06 nM,⁴⁴ and it has also been shown to have broad-spectrum antibacterial activity.⁴⁵ T0070907 is an inhibitor of peroxisome proliferator activator receptor γ that induces G2/M arrest and thus has activity against cancer cells.⁴⁶ This nuclear hormone pathway is not conserved in T. gondii, and this compound has not been described to have antimicrobial activity previously, so the potential mechanism of action is uncertain. Emetine is a natural product that binds to ribosomes and blocks translation, and studies of the structure of Giardia ribosomes have recently refined the binding mode and mechanism of action.⁴⁷ Emetine has also been shown to be potent against Plasmodium falciparum in vitro⁴⁸ and against Trypanosoma brucei and Trypanosoma cruzi, although it displayed poor selectivity relative to host cells.⁴⁹ Although we did not observe toxicity in short-term CC₅₀ assays, it was toxic to monolayers in a longer-term culture, preventing us from testing its effects against ex vivo bradyzoites. Advancing emetine as a treatment for parasitic infections would likely be compromised by its cardiotoxicity.⁵⁰

Atovaquone showed partial inhibition in the washout ex vivo bradyzoite assay, consistent with previous studies showing it is partially active in reducing cyst numbers during chronic infection in vivo.⁵¹ BRD7929 was highly effective in preventing the outgrowth of bradyzoites even when removed after 4 h (Figure 4B), consistent with previous findings.³³ Testing of the Top Hit Sanguinarine sulfate revealed potent activity in the ex vivo bradyzoite washout assay, thus confirming that it has activity against bradyzoites as well as tachyzoites (Figure 4B). In an extended assay, intact cysts were treated with 3xEC₉₀ concentrations of compounds for 4 or 24 h before liberating the bradyzoites, followed by a plaquing assay in media ± compounds to determine the effects on intact cysts (Figure 4C). Although BRD7929 was highly effective against liberated bradyzoites (Figure 4A), it was only minimally effective against intact cysts after 4 h (Figure 4D) and moderately effective after 24 h of treatment (Figure 4E). Similarly, treatment with atovaquone or pyrimethamine for 4 or 24 h was only minimally active against intact cysts, although extended culture with these compounds was effective at preventing outgrowth (Figure 4D,E). In contrast, treatment with sanguinarine sulfate showed stronger activity against intact cysts compared to liberated bradyzoites treated for 4 h (Figure 4D). Moreover, sanguinarine was highly effective on intact cysts after 24 h of treatment (Figure 4E). The greater potency of sanguinarine on intact cysts vs liberated bradyzoites suggests that the compound may be concentrated within the intact cyst, thus reaching a higher effective concentration. The cyst wall is comprised of glycoproteins that surround a dense matrix that harbors the individual bradyzoites^52,53 and the permeability of this barrier to small molecules is uncertain. Hence further studies are needed to define the features that mediate the uptake and concentration of small molecules across this interface.

Sanguinarine sulfate is a natural product produced by the opium poppy and several other plants and it consists of a benzoquinoline alkaloid.⁵⁴ Sanguinarine has anti-inflammatory, antitumor, and antimicrobial activities and it is thought to act on numerous signaling pathways in human cells.⁵⁵ The antitumor effects of sanguinarine are related to the induction of mitochondrial reactive oxygen species and the triggering of apoptotic cell death pathways,^56,57 although other mechanisms have also been described.^54,55 Similarly, the induction of reactive oxygen intermediates has been linked with antiparasitic effects against Eimeria in chickens.⁵⁸ The toxicity of sanguinarine to host cells varies by type⁵⁵ and it was only moderately toxic to HEPG2 and differentiated THP-1 cells tested here and based on its lower potency against T. gondii, it exhibited a reasonable therapeutic index. However, the LD₅₀ in mice is 18 mg/kg by i.p. injection,⁵⁹ and it has a relatively short half-life in vivo before being metabolized into various derivatives,⁶⁰ which may limit its use in efficacy testing for toxoplasmosis. Nonetheless, various chemical derivatives have been synthesized and tested for anticancer activities,⁶¹ and testing these derivatives against T. gondii could establish further SAR of this series and potentially identify analogs with greater potency and selectivity.

Conclusions

The newly discovered Tg68 strain is permissive for bradyzoite differentiation and forms the basis for the screening protocol in this study. We have developed protocols for HTS of compound libraries that target tachyzoites and bradyzoites and further validated hits in assays against ex-vivo-produced bradyzoites. Several of the hits identified in this screen have precedent for being antimicrobial; however, they pose challenges for selectively leading to considerable host toxicity. The top hits identified here were potent, and at least one compound showed potent activity against intact tissue cysts; therefore, potential medicinal chemistry efforts could address the selectivity issue in future studies. Additionally, the methodology developed here offers promise for future HTS to identify potent and selective leads.

Experimental Section

Compounds and Liquid Handling

The LOPAC compound library was plated in a 384-well plate format for primary screening and determination of the half-maximal effective concentration (EC₅₀) of primary hits by Calibr at Scripps Research (La Jolla, CA). The plates were stored at −80 °C prior to use. All liquid handling steps (host and parasite cell seeding, media exchange, compound transfer to assay plates, and addition of luciferase reagents) were carried out in a semiautomated facility to ensure efficient and consistent execution of assays across all replicates (High-Throughput Screening Center, Washington University School of Medicine). Nine top-hit compounds and two control compounds, atovaquone (Sigma#A7986) and pyrimethamine (Sigma#46706), were obtained from commercial sources, and follow-up assays were manually performed in-house.

Construction of Transgenic Parasites and Parasite Culture

Tg68 pTub1:Fluc,DHFR parasite lines were generated by utilizing a pre-existing pTub:Fluc plasmid,³³ which provides constitutive Fluc expression from the alpha-tubulin promoter and a pyrimethamine-resistant DHFR selectable marker. Tg68 tachyzoites were electroporated with 50 μg of the plasmid and subsequently selected using pyrimethamine (3 μM).

The plasmid pSAG1:EGFP-DHFR-BAG1:nLuc was constructed by assembling fragments encoding EGFP driven by the SAG1 promoter, nLuc driven by the BAG1 promoter, DHFR pyrimethamine resistance marker, and pNJ-26 vector using the NEBuilder HiFi DNA Assembly Master Mix (NEB). Tg68 pBAG1:nLuc, DHFR tachyzoites (Tg68pSAG1:EGFP-DHFR-BAG1:nLuc) were electroporated with 50 μg of this plasmid and selected with pyrimethamine (3 μM) in order to establish stable parasite lines. Detailed primer information can be found in Table S4.

Stable clones were isolated through limiting dilution, and the expression of the transgene was confirmed by luciferase expression. Clonal transgenic Tg68 tachyzoites were maintained by serial passage in T25 flasks with confluent HFFs in D10 medium (Dulbecco’s modified Eagle’s medium, DMEM; Thermo Fisher, 10% fetal bovine serum (FBS), 2 mM glutamine (Sigma), and 10 μg/mL gentamicin (Thermo Fisher)) at 37 °C and 5% CO₂.

In Vitro Assays for Bradyzoite and Tachyzoite Growth Inhibition

Host Cell Culture and Seeding

Human foreskin fibroblast (HFF-1 SCRC-1041, ATCC) cells were cultured in T175 flasks using a D10 medium. Prior to seeding the cells to the assay plates, a cell suspension was prepared by treating them with trypsin (0.5 g/L porcine trypsin and 0.2 g/L EDTA in Hank’s balanced salt solution with phenol red) for 5 min. EC₅₀ assays (half maximal effective concentration) were conducted in a 384-well plate format. Confluent HFF cells were seeded 4 days in advance using a Multidrop Combi dispensing 80 uL/well, while stirring the cell suspension at 350 rpm. The plates were stacked in a Cytomat rack on a level surface at room temperature until all plates were completed, and then, they were placed in a 37 °C incubator with 5% CO₂. To minimize edge effects, only the inner 240 wells of each plate were utilized.

Bradyzoite Primary Screening and EC₅₀ Assay

The medium from the HFF cell plates was aspirated using a Biotek ELx405CW, leaving 20 μL/well before infection with parasites. Freshly harvested Tg68pBAG1:nLuc,DHFR parasites (3 × 10³) in a 40 μL volume were then added to each well (total volume of 60 μL/well, containing 0.1% DMSO in D10) using a Fluotics BXi-50F while stirring the parasite suspension at 350 rpm. The plates were incubated for 2 h at 37 °C in a 5% CO₂ incubator to allow parasite invasion. Afterward, the culture medium was removed and switched to either 80 μL/well of Alkaline (RPMI 1640 (Sigma, #R6504) containing 1% FBS and 50 mM HEPES (Sigma), adjusted to pH 8.2), or Glutamine (glucose-free RPMI 1640 (Sigma, #R1383) containing 1% FBS and 50 mM HEPES (Sigma), 10 mM glutamine, adjusted to pH 7.2) medium. The culture was then incubated at 37 °C in ambient CO₂ and maintained for 10 days, with media changes occurring on days 3 and 6 using a Biotek ELx405CW and Multidrop Combi inside a biohood to minimize contamination. On day 6, before the compound was added, the plates were aspirated, and 40 μL/well of fresh media was added, followed by the transfer of 40 μL/well of 2X compound using a Fluotics BXi-50F. On day 10, the plates were equilibrated to room temperature for 30 min prior to the assay readout. To read the plates, the medium was aspirated, 20 uL of Promega Nano-Glo reagent (Promega N1150) was dispensed into each well using a Multidrop Combi, and the plates were covered with a black lid. The plates were then incubated for 10 min at room temperature, and the readout was performed using Envision.

Tachyzoite Primary Screening and EC₅₀ Assay

Before being infected with parasites, the medium was removed from the HFF cell plates using a Biotek ELx405CW, leaving 20 μL/well. Freshly harvested Tg68pTub1:Fluc parasites (2 × 10⁴) in a 20 μL volume were then added to each well while stirring the parasite suspension at 350 rpm, and 40 μL of 2X compound solutions were transferred (80 μL/well total volume, containing 0.1% DMSO in D10) using a Fluotics BXi-50F. The plates were incubated for 72 h at 37 °C in a 5% CO₂ incubator. Before the assay readout, the plates were equilibrated to room temperature for 30 min. To read the plates, the medium was aspirated to 20 μL using a Biotek ELx405CW, 20 μL of Promega Bright-Glo reagent (Promega E2650) was dispensed into each well, and the plates were covered with a black lid. The plates were then incubated for 10 min at room temperature using a Multidrop Combi instrument, and the readout was performed using Envision.

In Vitro Assays for Host Cell Cytotoxicity

Cytotoxicity against host cells was tested using human hepatocellular carcinoma cells (HepG2, ATCC-HB-8065) and a human monocytic tumor line (THP-1, ATCC-TIB-202), which were maintained according to ATCC recommendations. Host cell lines were tested negative for mycoplasma using an e-Myco plus kit (Intron Biotechnology). Compounds were diluted to 2X concentration, and a 10-dose serial dilution series was prepared through stepwise, 3-fold dilutions in recommended media with 0.1% DMSO. Host cells were seeded at a density of 10⁴ cells/well (100 μL vol) to achieve subconfluent monolayers for a 72 h growth assay. Prior to compound addition, THP-1 cells were treated with 10 ng/mL phorbol 12-myristate 13 acetate for 24 h to facilitate differentiation into macrophages. HepG2 cells were treated with compounds 6 h postseeding (200 μL final volume, 0.05% DMSO) and incubated under 37 °C, 5% CO₂ culture conditions. At 72 h post compound addition, culture media were aspirated to 80 μL, and an equal volume of CellTiter-Glo Luminescent Cell Viability Assay reagent (Promega G7571) was added. Luciferase activity was measured using a BioTek Cytation 3 instrument equipped with Gen5 software (v3.08). Each assay was repeated with two technical replicates within two independent biological replicates. Statistical analyses were performed using Prism 10 (GraphPad Software, Inc.). Dose–response inhibition curves for host cell toxicity (CC₅₀ values) were generated using the “Log(inhibitor) vs normalized response variable slope” function. The reported values represent averages from three biological replicates.

Ex Vivo Bradyzoite Assay

CBA/CaJ mice (Strain #000654) from Jackson Laboratory were housed in an approved facility at Washington University School of Medicine, and all animal studies followed ethical guidelines approved by the Institutional Animal Care and Use Committee. CBA/CaJ mice were infected by oral gavage with 5–10 tissue cysts from the brain homogenate of previously infected mice. The brains of CBA/CaJ mice infected with the TgME49-EW strain⁴¹ were collected at 1–2 mos postinfection, homogenized, and tissue cysts were isolated using Percoll gradients, as described previously.⁶² To release the bradyzoites from the tissue cysts, purified tissue cysts were treated with an acid-pepsin solution (170 mM NaCl and 60 mM HCl) and a freshly prepared pepsin solution (0.1 mg/mL in 1xPBS) for 10 min at 37 °C. The reaction was stopped by adding a neutralization buffer (94 mM Na₂CO₃). The liberated bradyzoites were then evenly distributed into a duplicate set of 6-well plates (technical replicates), each containing 5 mL of culture media. Each plate included a negative control (media with 0.1% DMSO), the previously characterized PheRS inhibitor BRD7929 (0.5 μM),³³ and pyrimethamine (2.0 μM). Compounds were tested at EC₉₀ and 3XEC₉₀ concentrations based on their corresponding in vitro tachyzoite growth inhibition (TgEC₅₀) assays (Table S3). After a 4 h treatment, the compounds were removed from one set of plates by washing them 3 times in PBS and replaced with compound-free D10 medium. The plates were then incubated at 37 °C in 5% CO₂ undisturbed for 12–14 days to form plaques. Plaque quantification was performed by fixing the plates in 100% ethanol for 5 min at room temperature, followed by staining with a 0.1% crystal violet solution for 10 min. After rinsing with water and air drying, plaque quantification was conducted using a Nikon Eclipse ts2 microscope equipped with a 4× objective. The number of plaques from two biological replicates was normalized to DMSO control as a percentage. To test the effect of the compound on intact cysts, replicate samples of Percoll purified cysts (n = 10 cysts per sample) were treated with 3XEC₉₀ concentrations in a 1 mL Eppendorf tube at 37 °C for 4 or 24 h. Compound concentrations were based on the respective in vitro tachyzoite growth inhibition (TgEC₅₀) assays (Table S3). Following treatment, cysts were either washed (replicate 3) or not washed (replicates 1 and 2) three times with PBS, bradyzoites were released using an acid-pepsin solution, and the reaction was stopped with a neutralization buffer. The bradyzoites were then distributed equally into parallel 6-well plates in media with 3XEC₉₀ concentration compounds or media without compounds. Plates were incubated at 37 °C and 5% CO₂ for 12–14 days to form plaques. Plaques were quantified as described above.

Statistical Analysis

Statistical comparisons were performed in Prism (GraphPad). Data were first analyzed for normal distribution and according to the outcome, nonparametric tests were then applied. The statistical test and P values are given in the figure legends.

Glossary

Abbreviations

DMSO

dimethylsulfoxide

DHFR

dihydrofolate reductase

EDTA

ethylenediaminetetraacetic acid

HFF

human foreskin fibroblasts

HepG2

human hepatocellular carcinoma cell

THP-1

human monocytic tumor cell

HTS

high-throughput screening

nLuc

nano luciferase

Fluc

firefly luciferase

PheRS

phenylalanine tRNA synthetase

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsinfecdis.4c00689.

• Testing of additional TOP HIT compounds for activity against ex vivo bradyzoites (PDF)

  HTS parameters for lines expressing luciferase (XLSX)

• HTS parameters for the LOPAC screen (XLSX)

• Activity of primary hits in growth inhibition assays (XLSX)

• Primers used in the study (XLSX)

The authors declare no competing financial interest.