Identification of two potential inhibitors of Sporothrix brasiliensis and Sporothrix schenckii in the Pathogen Box collection

Sporotrichosis is a neglected endemic mycosis with a high incidence in Latin America, mainly in Brazil. Sporothrix schenckii is the most frequent species in Latin America, whereas Sporothrix brasiliensis is the predominant species observed in Brazil and is associated with both human and animal sporotrichosis. Sporotrichosis treatment remains restricted to a few options, itraconazole being the first choice for human and animal therapy. In this work, we screened the molecular library Pathogen Box (Medicines for Malaria Venture [MMV], Switzerland) in search of compounds with anti-Sporothrix activity. Our initial screen of the 400 compounds identified five compounds that inhibited more than 80% of S. brasiliensis and S. schenkii growth. Among those, three compounds (MMV675968, MMV102872, and MMV002817 (known as iodoquinol)) not previously described as antifungals or agrochemicals, were selected for further evaluation. MMV102872 and iodoquinol showed the most promising combination of antifungal activity (lower inhibitory concentration) and fungal selectivity (lower cytotoxicity in LLC-MK2 cells). Scanning electron microscopy and flow cytometry analyses revealed that MMV102872 and iodoquinol induced changes in cell morphology, membrane integrity, and the presence of neutral lipids, impairing fungal survival. Our results indicate that MMV102872 and iodoquinol are promising molecules for use as scaffolds for the development of new antifungal agents.


Introduction
Sporotrichosis is a neglected endemic mycosis caused by thermodimorphic fungi of the genus Sporothrix that affects humans and animals [1]. In Brazil, sporothrichosis is an endemic disease and an important zoonosis, with thousands of humans and cats being infected in the last two decades [2]. Between 1992 and 2015, 782 hospitalizations and 65 deaths due to human sporotrichosis were reported in Brazil [3], and outbreaks of the disease are currently present in areas of the southeast, south, and northeast of the country [4].
Sporothrix schenckii and Sporothrix brasiliensis are the most virulent species in the Sporothrix genus and the main etiological agents of sporotrichosis in Latin America and Brazil, respectively [4,5]. Infection with S. schenckii occurs by traumatic inoculation of fungus present in contaminated organic materials, while S. brasiliensis transmission is associated with the zoonotic route involving cats as intermediate hosts [6]. Importantly, S. brasiliensis can induce a pronounced inflammatory response, leading to more severe and/or atypical clinical presentations of sporotrichosis [7,8].
Although sporotrichosis is endemic in several countries, treatment remains restricted to a few options, relying mainly on itraconazole [4]. The emergence of clinical and feline-borne isolates with low in vitro susceptibility to itraconazole and the constantly reported therapeutic failure highlights the need to seek new options for sporotrichosis treatment [9][10][11].
High-content screening of libraries containing hundreds of drug-like compounds is a powerful strategy to search for novel antifungal drugs and to explore new targets [12]. Screening of large chemical libraries aim to identify compounds able inhibit the growth of pathogenic fungi and have been the most successful approach in the discovery of chemical scaffolds with promising antifungal action [12]. However, good quality of compounds and the establishment of standardized assays for susceptibility tests are required for this approach to succeed [12].
The Pathogen Box library, developed by the Medicines for Malaria Venture organization (MMV), contains 400 compounds with repurposing potential and also new molecules (https:// www.mmv.org/mmv-open/pathogen-box), selected on the basis of their promising activity against pathogens associated with neglected tropical diseases [13]. In an effort to identify new molecules with antifungal activity, the Pathogen Box library was previously screened for its antifungal activity against Candida species, Cryptococcus species, and chromoblastomycosis agents [14][15][16][17]. Thus, the purpose of this study was to screen this collection in search of compounds with potential anti-S. brasiliensis and S. schenckii activity.

Microorganisms
The reference isolates S. brasiliensis CBS 133006 and S. schenckii ATCC 32286 were used in this study. In addition, eight clinical isolates, six from human sporotrichosis patients (S. brasiliensis B972 and HE06; and S. schenckii Ss22, Ss42, Ss73, and Ss110) and two from felines (S. brasiliensis: Ss53 and Ss245) were used [18]. All isolates were stored in saline solution containing 10% glycerol and 10% glucose at -20˚C. The filamentous form was cultivated for 7 days in Sabouraud broth (Difco, United States), with orbital shaking (at 150 rpm), at 36˚C. Then, an aliquot containing 10 5 CFU/ml was inoculated into brain heart infusion broth (Difco, USA) supplemented with 2% glucose (pH 7.8) and cultivated at 36˚C, with orbital shaking for 7 days, in order to obtain yeast cells. All assays were carried out using the yeast form as it is the pathogenic form.

Compounds
The Pathogen Box library was obtained from Medicines for Malaria Venture (MMV, Switzerland) and contained 400 compounds diluted in dimethyl sulfoxide (DMSO) at 10 mM. A stock solution of all compounds was prepared in DMSO at 1 mM and stored at -20˚C. Itraconazole (Sigma Chemical Co., USA) was used as a reference antifungal, and a stock solution (1 mM in DMSO) was kept at -20˚C.

Screening the compound library for inhibition of Sporothrix spp. growth
The reference isolates S. brasiliensis CBS 133006 and S. schenckii ATCC 32286 were used to screen The Pathogen Box compounds for putative inhibitors of fungal growth according to the following protocol: dx.doi.org/10.17504/protocols.io.bje9kjh6. The final concentration of compounds was 1 μM following the MMV recommended guidelines for screening (https:// www.mmv.org/mmv-open/pathogen-box/pathogen-box-supporting-information), while the final concentration of fungal cells was 1 x 10 5 CFU/ml. Itraconazole (reference antifungal) was also tested at 1 μM. After 48 h (at 35˚C in a 5% CO 2 atmosphere), fungal growth was analyzed by visual inspection in an inverted light microscope and quantified by spectrophotometric readings. Inhibitions of more than 80% were defined as the cut-off to select for the most promising compounds, corresponding to clearly visible prevention of growth when samples were initially analyzed by visual inspection. Experiments were performed in quadruplicate.

Determination of minimum inhibitory concentration values (MICs) against S. brasiliensis and S. schenckii
The MIC of the most active compounds were determined using the in vitro broth microdilution technique described by the CLSI [19] with modifications according to the following protocol: dx.doi.org/10.17504/protocols.io.bjgxkjxn. MIC values were compared with that of the reference drug itraconazole. Yeasts (1 x 10 5 CFU/ml) were treated with different concentrations of compounds (0.002-1 μM) for 48 h, at 35˚C, in a 5% CO 2 atmosphere. Fungal growth was analyzed by visual inspection in an inverted light microscope and quantified by spectrophotometric readings. Concentrations that inhibit 50% and 80% of fungal growth (IC 50 and IC 80 , respectively) were estimated. Results are presented as the mean of two independent experiments, performed in duplicate.

Cytotoxicity assays
Cytotoxicity assays with MMV102872 and iodoquinol (selected as the most promising compounds) were performed using a mammalian epithelial cell line (LLC-MK 2 ; ATCC CCL-7) according to the following protocol: dx.doi.org/10.17504/protocols.io.bjgbkjsn. Several concentrations of compounds (0.1-10 μM) were tested, concentrations that elicited 50% cytotoxicity (CC 50 ) were estimated, and the selectivity towards Sporothrix spp. was determined using the IC 80 values obtained previously. The selectivity index (SI) of compounds was calculated using the following equation: SI = CC 50 / IC 80 median. Results are representative of two independent experiments, performed in triplicate.

Treatment of fungal cells
The effect of MMV102872 and iodoquinol (MMV002817) on the morphology and metabolism of S. brasiliensis yeasts (CBS133006) was evaluated by scanning electron microscopy and flow cytometry. Standardized yeast suspensions (1 x 10 5 CFU/ml) were treated with the IC 50 concentrations (0.25 μM) of MMV102872 or iodoquinol (MMV002817) in supplemented RPMI, for 48 h, at 36˚C, with orbital shaking. Untreated controls were grown in the absence of drugs in parallel. Next, samples were centrifuged at 927 G for 3 min, washed three times in sterile PBS and processed for scanning electron microscopy or flow cytometry as described below.

Scanning electron microscopy
Untreated and treated cells (obtained as described in "Treatment of fungal cells") were fixed in 2.5% glutaraldehyde and 4% formaldehyde, in 0.1 M cacodylate buffer, for 1 h and then washed three times in 0.1 M cacodylate buffer. Samples were adhered to poly-L-lysine-coated glass coverslips, dehydrated in a graded ethanol series, critical-point-dried in CO 2 , and coated with gold. Images were obtained in a ZEISS EVO 10 scanning electron microscope (ZEISS Company, Germany) and processed using Photoshop software (Adobe, USA). The area distribution of 100 cells was measured and the count of damaged cells was performed using ImageJ software (National Institute of Health, USA).

Flow cytometry
Neutral lipid accumulation was evaluated by flow cytometry using the fluorescent marker BODIPY 493/503 (Thermo Fisher Scientific, USA). Plasma membrane integrity was evaluated using SYTOX Green (Thermo Fisher Scientific, USA), which is impermeable to intact membranes, but penetrates compromised membranes and stains nucleic acids. Untreated and treated cells (obtained as described in "Treatment of fungal cells") were incubated with 200 μl containing 20 μM of each fluorochrome for 30 min at room temperature in the dark. Cells were washed in PBS, lightly fixed in 2% formaldehyde, and washed again in PBS. The fluorescence intensity of stained cell populations was quantified in a BD Accuri C6 flow cytometer (BD Biosciences, USA) that counted 10,000 or 2000 events per sample for BODIPY 493/503 or SYTOX Green, respectively. Data were analyzed using BD Accuri C6 software. Experiments were repeated three times. Data depicted in the results section shows the mean and standard error of the mean of one representative experiment.

Statistical analyses
Statistical analysis was performed using Prism 6.01 software (GraphPad Software, USA), by one-way ANOVA (with Dunnett's post hoc test). Statistical significance was considered when p < 0.05.
Among the five compounds that are active for both species, compounds MMV688754 and MMV688934 were excluded from subsequent experiments as they are commercial antifungal drugs (MMV688754-posaconazol) and an agrochemical insecticide (MMV688934-tolfenpyrad). Therefore, MMV675968, MMV102872, and MMV002817 (iodoquinol) were selected for further studies exploring their potential antifungal activity and targets.

MMV102872 and iodoquinol showed low cytotoxicity against human cell lines in vitro
The CC 50 of MMV102872 and iodoquinol using the mammalian cell line LLC-MK 2 were 8.7 μM and 5 μM, respectively, resulting in high selectivity index of each compound to both S. brasiliensis and S. schenckii ( Table 2).

MMV102872 and iodoquinol altered fungal cell morphology and lipid metabolism and disrupted cell membrane integrity
The effects of MMV102872 and iodoquinol were evaluated against S. brasiliensis as a representant of genus Sporothrix. S. brasiliensis was selected based on its epidemiological importance in Brazil, where its associated with endemic transmission and also unusual clinical presentations of human sporotrichosis. Scanning electron microscopy analysis of S. brasiliensis (CBS 133006) revealed the presence of disrupted cells and overflow of intracellular contents after exposure to iodoquinol (Fig 3C, arrow), which was not observed for untreated or MMV102872 treated cells (Table 3). Moreover, treatments with MMV102872 and iodoquinol induced an increase in cell size of 7% and 17%, respectively ( Table 3). Flow cytometry results revealed that MMV102872 and Iodoquinol induced significant accumulation of neutral lipids and disruption of the plasma membrane integrity; being such cellular alterations more pronounced after treatment with iodoquinol (Fig 4).  Table. https://doi.org/10.1371/journal.pone.0240658.g001

Discussion
Over the last 50 years, several fungal infections have emerged as important public health problems [20]. In the present work, we screened the Pathogen Box library in search of promising molecules with anti-Sporothrix activity, showing that 3% (12/400) and 1.25% (5/400) of them were able to inhibit S. brasiliensis and S. schenckii, respectively.
S. brasiliensis and S. schenckii can exhibit distinct in vitro susceptibility profiles to commercial antifungals [21]. Besides, higher rates of in vivo resistance, as wells as prolonged treatment

PLOS ONE
Identification of Sporothrix spp. inhibitors in the Pathogen Box in sporotrichosis human cases, are reported in S. schenckii infections [8]. We observed that S. brasiliensis was susceptible to a higher number of compounds from the library, compared with S. schenckii. The lower sensitivity of S. schenckii to the compounds could be in part related to differences in the structure and composition of the cell wall of this species compared to S. brasiliensis [22]. Three compounds that showed activity against S. brasiliensis in the initial screening are usually used as agrochemicals: difenoconazole (MMV688943), azoxystrobin (MMV021057), and tolfenpyrad (MMV688934), the last one also being effective against S. schenckii. The expanding use of antifungal-based agrochemicals has been correlated with the emergence of antifungal resistance among pathogenic fungi [23], and Sporothrix spp. are most often found in the environment, which increases the probability of exposure to these compounds, ultimately reducing its susceptibility to them. Therefore, as our goal is to treat humans and animals, we decided to exclude those agrochemical molecules from our pool of candidates for the development of new human and/or animal antifungal agents. However, it is important to emphasize that the use of some of those compounds (mainly tolfenpyrad) could be explored for decontamination of environmental areas with high fungal loads that may be a focus of animal/human infection [6].
The initial screening also showed that posaconazole (MMV688754) and three other commercial drugs were able to inhibit S. brasiliensis growth: pentamidine (MMV000062), buparvaquone (MMV689480), and iodoquinol (MMV002817); however, only posaconazole (MMV688754) and iodoquinol (MMV002817) were also effective against S. schenckii (Fig 1). The in vitro and in vivo activity of posaconazole against Sporothrix spp. is well known [21,24]; however, its high costs compromises usage in lower-income countries of South America, like Brazil, where sporotrichosis also figures as a higly prevalent mycosis.
Among the new molecules, five compounds (MMV675968, MMV102872, MMV676477, MMV687807, and MMV658988) exhibited high activity against S. brasiliensis; however, only MMV675968 and MMV102872 were also effective against S. schenckii (Fig 1). The new compounds that showed anti-Sporothrix activity are diverse, and the previously described mechanisms of action for several cell types and microorganisms are summarized in Table 4. The identification of five new molecules with possibly distinct cellular targets opens perspectives for future studies, because these molecules have a different mechanism of action than antifungals currently available to treat fungal infections, a relevant discovery due to the increase in fungus resistance [23].
According to the above-cited results from the initial susceptibility assays, iodoquinol (MMV002817) and the new molecules MMV675968 and MMV102872 were selected for further evaluation as anti-Sporothrix agents. Iodoquinol (MMV002817) is a quinoline derivative used as an antiparasitic drug for intestinal worms (also known as diiodohydroxyquinoline) (Fig 2) [25]. MIC values showed that iodoquinol is highly active against Sporothrix spp. at concentrations as low as 1 μM, inducing almost complete inibition of proliferation (IC 80 ) of all 10 clinical isolates tested ( Table 1). Iodoquinol has also been reported to inhibit in vitro growth

PLOS ONE
of Candida species and chromoblastomycosis agents [16,17]. The IC 50 for planktonic cells of Candida spp. was similar to what we described here for Sporothrix spp. (0.25 μg/ml-corresponding to 0.63 μM), and growth inhibition of 100% of chromoblastomycosis agents was

PLOS ONE
Identification of Sporothrix spp. inhibitors in the Pathogen Box observed with 1.25 μM iodoquinol [16,17]. Therefore, a comprehensive study including a panel of diverse fungi would be of interest to elucidate whether iodoquinol has a broad spectrum of antifungal activity. Compound MMV675968 (5-chloro-6-[(2,5-dimethoxyanilino) methyl] quinazoline-2,4-diamine) was designed as a dihydrofolate reductase inhibitor, an enzyme conserved in prokaryotes and eukaryotes that catalyzes the NADPH-dependent reduction of dihydrofolate into tetrahydrofolate [13]. Previous studies demonstrated that MMV675968 also exhibited an inhibitory effect against some protozoa (Toxoplasma gondii and Cryptosporidium parvum) and fungi (Pneumocystis jrovecii and Candida albicans) [13,14] (Table 4). Our results showed that although MMV675968 presented high activity against S. brasiliensis, it was less effective at preventing proliferation of S. schenckii at concentrations lower than 1 μM ( Table 1). The IC 50 of MMV675968 against planktonic cells of C. albicans was also higher than 1 μM (3.12 μM), and inhibition of biofilms was only achieved at 50 μM [14].
Despite its demonstrated antimicrobial activities, no mechanism of action has been reported for MMV102872 in the literature, while the amoebicidal activity of iodoquinol has been associated with its ability to chelate iron [13]. MMV102872 is an azole-derivative, and previous reports have shown that inhibition of ergosterol biosynthesis by azoles leads to accumulation of neutral lipids; therefore, the accumulation of neutral lipids was quantified in Sporothrix-treated and non-treated cells, as well as cell membrane integrity effects given its central role in maintaining cellular homeostasis [26]. Disruption in the membrane was observed to a lesser extent after treatment with MMV102872, as well as neutral lipid accumulation, when compared to iodoquinol (Fig 4). Since this compound is an azole-derivative, we speculate that, in Sporothrix spp., MMV102872 may be interfering with ergosterol biosynthesis [26]. Alterations in iron absorption could disrupt enzymatic activity and metabolic pathways, culminating in irreversible cell damage and rupture, as observed by SEM in Sporothrix cells after treatment with iodoquinol (Fig 3C). However, further studies are needed to elucidate the mechanism(-s) involved in the antifungal activity exerted by MMV102872 and iodoquinol.
Importantly, the cytotoxicity results reported here demonstrate that MMV102872 and iodoquinol are more selective towards fungi than mammalian cells ( Table 2). Previous reports using HepG2 cells showed a CC 50 for MMV102872 of 9 μM; and a CC 20 for iodoquinol of 2.5 μM (reported in Pathogen Box documents: https://www.mmv.org/mmv-open/pathogen-

PLOS ONE
Identification of Sporothrix spp. inhibitors in the Pathogen Box box/about-pathogen-box#composition). Together, these data suggest that both compounds have reduced toxicity and are potential candidates to advance to in vivo and clinical studies. Iodoquinol has been used for the treatment of protozoal infections of the gastrointestinal tract; however, it is not able to penetrate into the intestinal epithelium [25]. Topical formulations containing 1% iodoquinol are available to treat dermatoses and exhibit antifungal and antibacterial properties [27] and could be repurposed as an adjuvant topical treatment for sporotrichosis in association with oral itraconazole therapy. Thus, future in vivo and clinical studies are required to evaluate the effectiveness of iodoquinol in the treatment of sporotrichosis.

Conclusion
In this work we screened the Pathogen Box library and identified two compounds, MMV102872 and iodoquinol, with promising anti-Sporothrix activity. These compounds were effective against both Sporothrix species endemic in Latin America: S. schenkii and S. brasiliensis and demonstrated reduced toxicity to mammalian cell lines in vitro. Mechanistic studies indicated that both compounds induced neutral lipid accumulation and disruptions in cell membrane homeostasis. Finally, based on the described results we suggest that MMV102872 and iodoquinol are promising molecules and could be used as a scaffold for the development of new antifungal agents.
Supporting information S1

PLOS ONE
Identification of Sporothrix spp. inhibitors in the Pathogen Box