Compounds from the Medicines for Malaria Venture Box Inhibit In Vitro Growth of Babesia divergens, a Blood-Borne Parasite of Veterinary and Zoonotic Importance

Babesiosis is an infectious disease with an empty drug pipeline. A search inside chemical libraries for novel potent antibabesial candidates may help fill such an empty drug pipeline. A total of 400 compounds (200 drug-like and 200 probe-like) from the Malaria Box were evaluated in the current study against the in vitro growth of Babesia divergens (B. divergens), a parasite of veterinary and zoonotic importance. Novel and more effective anti-B. divergens drugs than the traditionally used ones were identified. Seven compounds (four drug-like and three probe-like) revealed a highly inhibitory effect against the in vitro growth of B. divergens, with IC50s ≤ 10 nanomolar. Among these hits, MMV006913 exhibited an IC50 value of 1 nM IC50 and the highest selectivity index of 32,000. The atom pair fingerprint (APfp) analysis revealed that MMV006913 and MMV019124 showed maximum structural similarity (MSS) with atovaquone and diminazene aceturate (DA), and with DA and imidocarb dipropionate (ID), respectively. MMV665807 and MMV665850 showed MMS with each other and with ID. Of note, a high concentration (0.75 IC50) of MMV006913 caused additive inhibition of B. divergens growth when combined with DA at 0.75 or 0.50 IC50. The Medicines for Malaria Venture box is a treasure trove of anti-B. divergens candidates according to the obtained results.


Introduction
Babesiosis is a tick-borne parasitic disease causing serious economic losses in the livestock industry worldwide [1]. In general, babesiosis is manifested by fever, hemoglobinuria, hemolytic anemia, jaundice, and death [2,3].
Babesia divergens (B. divergens) is considered one of the main causes of veterinary and human babesiosis in Europe [3]. The currently available drugs used to control Babesia infection in animals, imidocarb dipropionate (ID) and diminazene aceturate (DA), have proven their limitations regarding toxicity to hosts and parasite resistance [4]. Safety and residue problems have led to the withdrawal of these anti-B. divergens drugs from the market [5]. Severe cases of human babesiosis have not responded to treatment with the most commonly used anti-B. divergens drugs, such as clindamycin, azithromycin, quinine,

Results and Discussion
The Malaria Box is a valuable source of commercially available compounds, representing families of structures identified in phenotypic screens of pharmaceutical and academic libraries against P. falciparum, and was made available to identify novel drugs to combat other apicomplexan parasites [8]. Malaria Box, a collection of 400 compounds, is divided into 200 diverse drug-like compounds and 200 diverse probe-like ones [6]. The compounds that are absorbed by the oral route and toxophores are defined as drug-like compounds. Other hits are assigned to the probe-like category [2,8]. The in vitro inhibitory effects of Malaria Box compounds against 16 protozoa, 7 helminths, 9 bacterial and mycobacterial species, and the NCI60 human cancer cell line were published within a large, complete dataset [6]. The promising candidates identified from such wide screening of Malaria Box compounds, together with the zoonotic importance of B. divergens, encouraged us to perform, in the present study, a large-scale in vitro screening of MMV compounds (n = 400) against B. divergens.
In fact, a previous study [9] measured the susceptibilities of multiple blood-stage Plasmodium and Babesia parasite species to the Malaria Box compounds to investigate the possible sharing of the erythrocyte-specific apicomplexans to chemical sensitivities during the clinically relevant stages of parasitic infection. In this previous study, the authors evaluated the inhibitory effects of MMV compounds from Malaria Box against B. divergens only without either determination of the parasite viability after stopping the treatment, evaluating the possible synergistic or antagonistic interaction between MMV compounds and other antibabesial drugs, or preforming any bioinformatic analysis to the identified potent MMV hits. In our study, we have evaluated the inhibitory effects of MMV compounds from the Malaria Box on B. divergens growth. Then, we have evaluated the parasite viability after four days of treatments by the identified potent MMV hits. Of note, we have calculated the structural similarity between the drug-like MMV compounds and drugs commonly used to treat babesiosis using APfp. Additionally, combination therapies of MMV compounds with the lowest IC 50 s and the highest SIs with DA against the in vitro growth of B. divergens parasite were evaluated in our study. In the study performed by Paul et al. [9], the authors evaluated the inhibitory effects of MMV compounds from Malaria Box against B. divergens using a [ 3 H]-hypoxanthine uptake assay. Indeed, radiolabeled precursors of phospholipids, such as the sources of phospholipid polar head groups, [ 3 H]-hypoxanthine, are widely used radioisotopes for the antimalarial in vitro drug-sensitivity assay [10][11][12]. This assay is based on the inability of either platelets or uninfected erythrocytes to synthesize their RNA, DNA, and proteins, or their membranes or leukocytes to multiply during the in vitro malaria parasite culture. Thus, only malaria parasites have actively dividing cells during the in vitro culture process, and the addition of radioactive substances into the culture enables the parasites to integrate the radioactive precursors themselves, which is considered an indirect measurement of the parasite metabolic activity [10,11]. Generally, Plasmodium spp. are unable to synthesize purines de novo [13], which force Plasmodium parasites to use exogenous purines. Therefore, the incorporation of a radioisotope, such as [ 3 H]hypoxanthine, which is the main purine base needed by P. falciparum, is directly related to the count of P. falciparum-infected erythrocytes [11,12]. For Babesia parasites, there is paucity on the occurrence or the absence of de novo synthetic pathways of purines. Studies on the uptake of purines by Babesia-infected erythrocytes indicate that hypoxanthine is likely to be the major purine source in vivo [14]. For B. divergens, the findings indicate that the parasite can synthesize purine nucleotides by two routes, one involving PRTases and the other employing nucleoside kinases [15]. Thus, the parasite does not need to use exogenous purines. Subsequently, the incorporation of a radioisotope, such as [ 3 H]-hypoxanthine, with the parasite does not have a significant value for counting of Babesia-infected erythrocytes. Therefore, in the current study, we used a recently established, accurate, and automatic fluorescence assay using SGI stain to screen the huge library-like Malaria Box against the in vitro growth of B. divergens.
The in vitro inhibitory effects of MMV Malaria Box compounds against the growth of the B. divergens parasite were varied: potent (IC 50 < 1µM, n = 156), moderate (IC 50 1-10 µM, n = 186), or weak (IC 50 > 10µM, n = 58) (Tables S1-S20). Subsequently, Malaria Box compounds exhibited a 39% hit rate against the in vitro growth of the screened parasite ( Figure S1). This hit rate is higher than those caused by Malaria Box compounds against the in vitro growth of Babesia bovis, Babesia bigemina, Theileria equi, and Babesia caballi, respectively ( Figure S1) [2]. Such high in vitro susceptibility of B. divergens to compounds derived from Malaria Box may be attributed to the fact that B. divergens and P. falciparum use common receptors, glycophorins A and B, to invade the infected red blood cell [16]. Future studies are required to explore whether glycophorin receptors located on the RBCs' surface are drug targets where the identified potent MMV compounds produce their inhibitory effects in B. divergens infection. Indeed, previous studies [17][18][19][20][21] have demonstrated that certain compounds other than MMV compounds could inhibit the parasite invasion of RBCs through targeting the parasite secretory proteins on RBCs in both B. divergens and P. falciparum. Boyle, et al. [19] reported that heparin inhibits the invasion process of P. falciparum through inhibiting the earliest step in the invasion of merozoite to RBCs, initial RBC attachment, and binding merozoite surface protein 1 (MSP1). Additionally, heparin inhibits the merozoite invasion of RBCs in P. falciparum by binding to rhoptry and microneme proteins that are involved in reorientation and signaling steps of the invasion process [18,20]. For Babesia spp. (B. bovis, B. bigemina, B. equi, and B. caballi in vitro and B. microti in vivo), Bork, et al. [17] established that heparin covers the surfaces of babesial merozoites and inhibits their subsequent invasion of erythrocytes.
The hit rate of MMV compounds against the in vitro growth of B. divergens is considered higher than that observed by the in vitro screening of Malaria Box compounds against either Toxoplasma gondii (T. gondii) and Entamoeba histolytica [22], Schistosoma mansoni [23], or Cryptosporidium parvum [24] ( Figure S1).
A total of 99 MMV drug-like and 57 MMV probe-like compounds exhibited potent anti-B. divergens activity in vitro, with nanomole levels of IC 50 (Figure 1). From these compounds, 38 compounds (19 drug-like and 19 probe-like) were more effective than DA (Figures 2 and 3, Figure S2, and Tables S1-S20). Treatment with 100 nM of these potent 38 compounds significantly inhibited (p < 0.05) the in vitro growth of B. divergens (Figures 2 and 3).
Interestingly, seven compounds (four drug-like and three probe-like) exhibited a highly inhibitory effect against the in vitro growth of B. divergens, with IC 50 s less than 10 nanomolar (Table 1 and Figure 4). MMV019124, MMV006913, MMV019995, MMV000699, and MMV666054 exhibited IC 50 s lower than those observed for P. falciparum (Table 1). In general, several characteristics associated with the screening parasite, such as parasite species, strain, and size, have an impact on the therapeutic efficacy of the tested drugs [2,4]. In vitro culture parameters, such as the utilized medium, HCT, and the presence or absence of serum, also influence the calculated IC 50 s of the tested medication [25,26]. As a result, differences in parasite species or culture conditions between Babesia and Plasmodium might explain the inconsistencies in MMV compound IC 50 values between two parasites.     Interestingly, seven compounds (four drug-like and three probe-like) exhibited a highly inhibitory effect against the in vitro growth of B. divergens, with IC50s less than 10 nanomolar (Table 1 and Figure 4). MMV019124, MMV006913, MMV019995, MMV000699, and MMV666054 exhibited IC50s lower than those observed for P. falciparum (Table 1). In general, several characteristics associated with the screening parasite, such as parasite species, strain, and size, have an impact on the therapeutic efficacy of the tested drugs [2,4]. In vitro culture parameters, such as the utilized medium, HCT, and the presence or absence of serum, also influence the calculated IC50s of the tested medication [25,26]. As a result, differences in parasite species or culture conditions between Babesia and Plasmodium might explain the inconsistencies in MMV compound IC50 values between two parasites.

B. Probe-like compounds
Of note, MMV006913 was the most promising hit for further medicinal chemistry/biological screenings, with SI 32000 and IC50 1 nM (Table 1). To the best of our knowledge, the inhibitory effect of this hit in an animal model of babesiosis remains unknown. Moreover, the mechanism by which MMV006913 inhibits either Plasmodium or Babesia needs more deep investigation. Some theories suggested cGMP-dependent or mitogen-activated protein kinases (PK) as targets of this MMV hit in the Plasmodium parasite [6]. Generally     Of note, MMV006913 was the most promising hit for further medicinal chemistry/biological screenings, with SI 32000 and IC 50 1 nM (Table 1). To the best of our knowledge, the inhibitory effect of this hit in an animal model of babesiosis remains unknown. Moreover, the mechanism by which MMV006913 inhibits either Plasmodium or Babesia needs more deep investigation. Some theories suggested cGMP-dependent or mitogen-activated protein kinases (PK) as targets of this MMV hit in the Plasmodium parasite [6]. Generally speaking, PK has important physiological functions in the mammalian system, e.g., in vascular and gastrointestinal smooth muscles, platelets, kidneys, bones, and the central nervous system [27]. Among protozoan parasites, PKs are found in P. falciparum, T. gondii, and Babesia spp. [27]. The PK similarity between Babesia and P. falciparum is over 70% [27], possibly explaining why antimalarial compounds that target the PK gene are effective against Babesia parasites.
The inhibitory assay results obtained in the current study showed some differences from those previously published by Paul et al. [9]. Several factors affect the parasite susceptibility to the screened chemicals, including parasite strain and species, the inhibitory assay used for chemical screening, and the in vitro culture conditions [4,26]. Subsequently, the differences in the used parasite strain (Rouen 1987 vs. German bovine strain), the used inhibitory assay ( 3 H-hypoxanthine uptake assay vs. fluorescence assay using SGI stain assay), and the in vitro culture conditions (standard hypoxic conditions (e.g., 1% O 2 , 5% CO 2 , 94% N2), 2%-hematocrit, RPMI 1640 medium supplemented with 25 mM HEPES, 50 mg/l hypoxanthine, 2.42 mM sodium bicarbonate, and 4.31 mg/mL AlbuMAX II, 6.75 pH of the media vs. RPMI 1640 medium containing 40% normal bovine serum; 60 units per milliliter of penicillin G, 60 µg/mL of streptomycin, and 0.15 µg/mL of amphotericin B; cultures of parasitized RBCs (pRBCs) incubated at 37 • C in an atmosphere of 5% CO 2 , 5% O 2 , and 90% N 2 ), explain the discrepancies in the results between our results and the results obtained from the previous study that screened Malaria Box compounds against B. divergens parasite [9].
To determine the ability of potent identified MMV hits to suppress the regrowth of B. divergens after four days of in vitro treatment, viability testing was performed. Results showed that there was no regrowth of the screened parasite after treatment by all MMV compounds, even at the lowest concentration (0.1 µM) ( Table S21).
The HCA analysis revealed that MMV006913 (CID: 780973) showed the maximum structural similarity with either atovaquone or DA ( Figure 5). In the same way, the MMV019124 compound (CID: 1347726) exhibited the maximum structural similarity with either DA or ID ( Figure 5). MMV665807 (CID: 5065884) and MMV665850 (787413) showed the maximum structural similarity with each other or with ID ( Figure 5). Compounds falling under different clusters indicate structural dissimilarity among them and in comparison with other commonly used antibabesial drugs, indicating a probable different antibabesial mechanism. Subsequently, MMV hits that showed the lowest IC 50 s against the growth of B. divergens and the highest SIs (MMV006913 and MMV666054) were selected to further evaluate their inhibitory effects when used in combination with the commonly used antibabesial drug, DA. All combination ratios of MMV006913, when used with DA, revealed additive and indifference inhibitory effects against the in vitro growth B. divergens (Table 2). Of note, a very high concentration (0.75 IC 50 ) of MMV006913 caused an additive interaction on the growth of B. divergens when used in combination, with either 0.75 IC 50 or 0.50 IC 50 of DA (Table 2). On the contrary, the combined use of MMV666054 and DA revealed indifference and antagonistic interactions regarding the in vitro growth of B. divergens (Table 2). Further in vivo experiments are required to confirm the in vitro enhancement in the inhibitory effect of MMV006913 when used in combination with DA.
Of the hits with an IC 50 less than 10 nM, five compounds (two drug-like and three probe-like) exhibited SI ranging from 1195.64 to 7785.47 (Table 1 and Figure 4). Future in vivo experiments are required to evaluate the inhibitory effects of these MMV compounds using B. microti as the infectious agent.    Figure 5. Hierarchical clustering analysis. The hierarchical clustering analysis was performed using ChemmineR software and highlights the structural similarities between the MMV compounds with potent anti-B. divergens activity and the commonly used antibabesial drugs (diminazene aceturate, imidocarb dipropionate, and atovaquone). Dissimilarity is reported using a scale unit of 0.05.

Chemical Reagents
SYBR Green I (SGI) nucleic acid stain (Lonza, Rockland, NY, USA; 10,000×) was stored at -20 • C and thawed before use. A lysis buffer consisting of Tris (130 mM; pH 7.5), ethylenediaminetetraacetic acid (EDTA) (10 mM), saponin (0.016%; w/v), and TritonX-100 (1.6%; v/v) was prepared in advance and stored at 4 • C. DA (Novartis, Tokyo, Japan) and MMV compounds (MolPort, Riga, Latvia) were prepared as a 100 mM stock solution and stored at -30 • C until use. The Malaria Box compounds were supplied in V-shaped 96-well plates in 20 µL of a 10 mM DMSO (Sigma-Aldrich, Tokyo, Japan) solution and shipped frozen. Plate mapping and full data on the Malaria Box with original GSK/St. Jude/Novartis compound number, ID, set, molecular weight, canonical SMILES, source, and biological data are provided as supporting information for the Malaria Box (in pone.0062906.s002-supporting information). A list of vendors used to supply compounds for the Malaria Box, including the vendor's web address and the number of compounds from each vendor in the Malaria Box, is also available (in pone.0062906.s002-supporting information).

In Vitro Cultivation of B. divergens Parasite
Babesia divergens (German bovine strain of catalog number 201708-4) used in this study were cultivated in purified bovine red blood cells (RBCs) using a microaerophilic, stationary-phase culture system [26,28]. Briefly, B. divergens [26,29] was cultivated in RPMI 1640 medium (Sigma-Aldrich) containing 40% normal bovine serum. Sixty units per milliliter of penicillin G, 60 µg/mL of streptomycin, and 0.15 µg/mL of amphotericin B (all from Sigma-Aldrich) were added to all of the culture medium used. Antibiotics were added to the parasite culture at a concentration that did not harm the RBCs to prevent bacterial contamination and, subsequently, minimize the loss of valuable cells, reagents, time, and effort due to contamination. Cultures of parasitized RBCs (pRBCs) were incubated at 37 • C in an atmosphere of 5% CO 2 , 5% O 2 , and 90% N 2 .

Malaria Box Compound Screening against B. divergens
The inhibitory effects of 400 compounds (200 drug-like and 200 probe-like) from the Malaria Box on B. divergens growth were tested using the fluorescence-based highthroughput screening (HTS) assay [26]. Briefly, B. divergens pRBCs were cultivated in 96-well plates at 1% parasitemia with nonparasitized RBCs to 5% hematocrit (HCT). RPMI 1640 medium alone or with indicated concentrations of 400 MMV compounds (0.1, 1, 10, and 25 µM) in 100 µL final volume was added to the culture. The medium containing 0.1% dimethyl sulfoxide (DMSO) with infected RBCs (iRBCs) and uninfected RBCs was used as positive and negative controls, respectively. DA was used as the positive control and tested at a final concentration of 10 µM. One plate was used in each drug experiment for 50% inhibitory concentration (IC 50 ) value calculation on the fourth day of culture. A lysis buffer containing SGI was added to each well, as reported above. The mean fluorescence values were then plotted against the logarithm of drug concentrations. Selectivity indices (SIs) for the identified potent compounds were calculated as the ratio of the cytotoxic concentration 50 (CC 50 ) to the IC 50 for each MMV compound. The cytotoxicity was evaluated using the human fibroblast cell line MRC-5 and retrieved from ChEMBL.

Viability Test
MMV hits that exhibited potent anti-B. divergens efficacy with IC 50 ≤ 10 nM were selected to evaluate the B. divergens parasite viability after four days of treatments by these compounds, as previously described by [26]. Briefly, experiments similar to those described in Section 3.3 were performed. On the fourth day of treatment, 3.5 µL of the DMSO-or MMV-treated infected RBCs were removed and diluted in a fresh growth medium. Then, the plates were incubated at 37 • C for the next four days. The amount of parasite DNA was measured using a fluorescence spectrophotometer and used as an indicator of parasite recrudescence. Each experiment was repeated three times.

Structural Similarity Measurements
To calculate the structural similarity between the drug-like MMV compounds (n = 4) and drugs commonly used to treat babesiosis (DA, ID, and atovaquone), we used atom pair fingerprints (APfp) [30]. APfp was calculated using the chemical ID obtained from the PubChem for each compound. The CIDs were then loaded into the ChemMine tools software for calculating APfp of all compounds [31]. The APfp was submitted to ChemmineR software for hierarchical clustering analysis (HCA) [32].

In Vitro Drug Combination Test
Combination therapies of MMV compounds (MMV006913 and MMV666054) that exhibited the lowest IC 50s (1 nM) and the highest SIs with the commonly used antibabesial drug DA against the in vitro growth of B. divergens parasite were evaluated. The combination ratios ranged from 0.75 IC 50 MMV compounds:0.75 IC 50 DA to 0.50 IC 50 MMV compounds:0.50 IC 50 DA (Table 2), as previously described [3], with some modifications. Each combination was loaded in triplicate wells in 96-well plates and incubated with the parasite at 1% parasitemia and 5% HCT for 4 consecutive days. RBCs collected from healthy cattle were loaded into each well in triplicate with the medium and DMSO and used as blank controls. On the fourth day of culture, 100 µL of a lysis buffer containing a 2 × SGI nucleic acid stain was added directly to each well on the plates, and then the plates were incubated for 6 h in a dark place at room temperature and the fluorescence values were determined as described above. Three independent experiments were performed for each combination therapy.

Statistical Analyses
The IC 50 values of MMV compounds were calculated using the nonlinear regression curve fit in GraphPad Prism 5.0 (GraphPad Software, Inc., San Diego, CA, USA). The differences in the emitted fluorescence signals for the in vitro cultures and drug combination test were analyzed with a commercial statistical software program (GraphPad Prism) using the independent unpaired Student's t-test [33]. A p value of < 0.05 was considered statistically significant for all tests. The regrowth of the parasite in the viability test was determined based on the statistically significant differences between the MMV-hits-treated and the positive control group (Babesia treated with RPMI 1640 medium containing the vehicle only, DMSO) [26].

Conclusions
Screening the Malaria Box compounds against the in vitro growth of B. divergens assisted in the discovery of new, more effective drugs than the traditionally used ones, either at the animal level, DA and ID, or human level, clindamycin, azithromycin, and quinine. Attractively, seven novel potent anti-B. divergens, including MMV665807, MMV665850, MMV019124, MMV006913, MMV019995, MMV000699, and MMV666054, were identified. One hit, MMV006913, exhibited the lowest IC 50 and the highest selectivity indices, highlighting the possible promising in vivo inhibitory effect of this hit when administrated in either a mouse infected with B. microti or gerbil infected by B. divergens. Seven other MMV hits showed excellent SIs. Although the present study helped in the discovery of novel MMV with potential anti-B. divergens efficacy, the inhibitory effects of these hits may be enhanced when used in combination with each other. Future studies are required to analyze the two-drug interactions of MMV hits when used in combination with each other for further clinical application. The obtained results indicate that the Malaria Box is a treasure trove of anti-B. divergens candidates. Knowledge of the targets and action mechanisms of these compounds might provide new insights into Babesia biology. Therefore, further researches are essential to calculate the distance matrix between the identified MMV compounds with anti-B. divergens activity and the commonly used antibabesial drugs, including highlighting the structural differences between these compounds using PubChem fingerprint for similarity workbench. Furthermore, future studies are required to establish a simple mathematical relationship between structural or property structural molecular features (descriptors) of the identified potent MMV compounds and their physicochemical activities using quantitative structure-activity relationships (QSAR). Such physicochemical descriptors, which include parameters to account for hydrophobicity, topology, electronic properties, and steric effects, will be determined empirically or, more recently, by computational methods. Three main steps are required for accomplishing QSAR modeling: (1) collect or, if possible, design a training set of chemicals; (2) choose descriptors that can properly relate chemical structure to biological activity; and (3) apply statistical methods that correlate changes in structure with changes in biological activity.

Informed Consent Statement: Not applicable.
Data Availability Statement: The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.