Antifungal Testing and High-Throughput Screening of Compound Library against Geomyces destructans, the Etiologic Agent of Geomycosis (WNS) in Bats

Bats in the northeastern U.S. are affected by geomycosis caused by the fungus Geomyces destructans (Gd). This infection is commonly referred to as White Nose Syndrome (WNS). Over a million hibernating bats have died since the fungus was first discovered in 2006 in a cave near Albany, New York. A population viability analysis conducted on little brown bats (Myotis lucifugus), one of six bat species infected with Gd, suggests regional extinction of this species within 20 years. The fungus Gd is a psychrophile (“cold loving”), but nothing is known about how it thrives at low temperatures and what pathogenic attributes allow it to infect bats. This study aimed to determine if currently available antifungal drugs and biocides are effective against Gd. We tested five Gd strains for their susceptibility to antifungal drugs and high-throughput screened (HTS) one representative strain with SpectrumPlus compound library containing 1,920 compounds. The results indicated that Gd is susceptible to a number of antifungal drugs at concentrations similar to the susceptibility range of human pathogenic fungi. Strains of Gd were susceptible to amphotericin B, fluconazole, itraconazole, ketoconazole and voriconazole. In contrast, very high MICs (minimum inhibitory concentrations) of flucytosine and echinocandins were needed for growth inhibition, which were suggestive of fungal resistance to these drugs. Of the1,920 compounds in the library, a few caused 50% - to greater than 90% inhibition of Gd growth. A number of azole antifungals, a fungicide, and some biocides caused prominent growth inhibition. Our results could provide a theoretical basis for future strategies aimed at the rehabilitation of most affected bat species and for decontamination of Gd in the cave environment.


Introduction
Geomyces destructans (Gd) causes geomycosis, a devastating disease of hibernating bats that has resulted in the death of tens of thousands if not more than a million individuals of six bat species in the northeastern U.S. The fungus is the only etiologic agent thus far identified that is associated with White Nose Syndrome (WNS) [1,2]. To date, Gd has most seriously affected an insectivorous species, Myotis lucifugus (''little brown bat''), but recent studies indicate that it causes high mortality in other bat species and in geographical areas beyond the northeastern U.S. [3]. The projected regional extinction of M. lucifugus could lead to a serious imbalance in the ecosystem with unforeseen consequences [3]. There is an urgent need to understand the pathogenesis of geomycosis and to devise control measures to halt and reverse this potential mass extinction.
Any scheme for the treatment of infected bats with antifungal drugs faces a number of challenges. At the outset, there is no information about the effectiveness of the currently available drugs for a fungus that grows at very low temperatures; these drugs are usually effective at the physiological range. An equally challenging task will be the delivery of drugs to an infected population in the wild. Recent studies on the precipitous decline in amphibians due to infection by the chytrid fungus Batrachochytrium dendrobatidis have revealed some important insights including the need for the judicious selection of a therapeutic agent [4]. Treatment attempts directed at tadpoles have shown that itraconazole was effective against the fungus, but it caused severe damage to the host [5]. Conversely, the same treatment worked well in a majority of anuran species [6]. Other recent studies revealed that the treatment of fungal infections in amphibians would be best targeted by selecting drugs that function first in the laboratory [7,8].
The natural habitat of Gd is unknown although caves are obviously high on the list because it is believed that hibernating bats become infected in caves that contain Gd. Caves harbor unique fungal communities, which are greatly affected by human activity [9,10]. It is not known if human activities played a role in the initial outbreak and subsequent spread of geomycosis. However, the United States Fish and Wildlife Service has temporarily blocked public access to many caves and mines as a preventative measure aimed at stopping further spread of the disease by humans (http://www.fws.gov/whitenosesyndrome/ pdf/NWRS_WNS_Guidance_Final1.pdf). Many of the affected caves are important for tourism and recreational activities. Thus, it might become necessary to decontaminate heavily infested caves and other natural sites using safe and effective chemicals that selectively eradicate Gd. This study aimed to identify antifungal drugs and antiseptic compounds for potential treatment of geomycosis and decontamination of Gd.

Results
All five strains of Gd had similar antifungal susceptibility profiles in our tests irrespective of the test devices used ( Figure 1A-1C). Amphotericin B and four azoles (itraconazole, ketoconazole, posaconazole and voriconazole) were effective at low MICs against Gd strains while fluconazole MICs were suggestive of Susceptible-Dose Dependent (S-DD) pattern. Flucytosine and all three echinocandins (caspofungin, anidulafungin and micafungin) exhibited relatively high MICs suggestive of resistance ( Figure 1C). Unlike pathogenic fungi, susceptibility testing of Gd strains required prolonged incubations for 10-21 days at 15uC and 6uC, respectively. The antifungal susceptibility testing proceeded relatively faster at 15uC than at 6uC as expected due to the better growth of the fungus seen at 15uC [2]. It is noteworthy that YeastOneH microdilution test yielded results on an average of 14 days while Etest was positive in 10 days. This difference in turnaround time was accentuated when the tests were done at 6uC ( Figure S1).
The results of HTS identified 27 compounds (1.4% of 1,920 tested), which caused 50% to-greater than 90% inhibition of Gd growth at 15uC while other compounds tested had no inhibitory effect (Figure 2A, 2B and Table S1). Of the 27 compounds, ten were randomly selected for further analysis using 10 mM original concentration and ten-fold (1.0 mM) and hundred-fold (0.1 mM) lower concentrations in the microtiter plates. These compounds and Gd were incubated both at 15uC and 6uC. This secondary screen confirmed excellent growth inhibitory profiles of selected compounds at both temperatures ( Figure 3 and Tables 1, S2, S3). Not surprisingly, antifungals (econazole, sulconazole, chloroacetoxyquinoline, pyrithione zinc, ciclopirox olamine, and chloroxine), fungicide (phenylmercuric acetate), and biocide (benzalkonium chloride) caused prominent growth inhibition. Notably, pyrimidine analog (fluorouracil) and a biomembrane disruptor (digitonin) were also highly effective.

Discussion
A remarkable observation in this study is that the antifungal drugs are active at the psychrophilic range, a fact not previously Suggestive resistance to flucytosine and caspofungin was found in testing using E-test and YeastOneH, and suggestive resistance to anidulafungin and micafungin was found in testing using YeastOneH. doi:10.1371/journal.pone.0017032.g001 documented in the published literature. Geomyces destructans was susceptible to amphotericin B and azoles at MICs that are similar to the values reported for pathogenic yeasts and filamentous fungi [11,12,13]. Azoles and amphotericin B represent some of the most effective antifungal drugs that act by inhibition of ergosterol biosynthesis or by binding to ergosterol in the plasma membrane. Our results expand their broad-spectrum of activity to a psychrophilic pathogen. Thus, we have identified many antifungal drugs that work well in the laboratory against Gd at temperature ranges likely to be encountered in bat hibernacula. This observation could be the basis for future trials of these drugs in efforts aimed at the rehabilitation of most affected bat species.
The observed high flucytosine MIC for Gd was suggestive of resistance. Such resistance to flucytosine is fairly widespread among filamentous fungi and the resistance mechanisms are manifold. In studies with Aspergillus species, it was found that resistant strains were not defective in the drug uptake nor in its conversion to 5-flourouracil, but they utilized cytosine as a nitrogen source to circumvent the drug effect [14]. Another noteworthy finding of the present study is the high echinocandins MICs for Gd; this suggests intrinsic resistance similar to other filamentous pathogens such as Fusarium species, Scedosporium prolificans, and zygomycetes [13,15]. This finding suggests that Gd shares an important characteristic with fungal pathogens, which cause serious and intractable infections in humans. Echinocandinsresistance in the aforementioned fungi is due to specific mutations in the membrane protein Fks1, which encodes b-1,3-glucan synthase catalytic subunit crucial for cell wall synthesis [16]. It would be interesting to investigate if a similar or a novel mechanism of echinocandins-resistance exists in Gd.
The yield of effective compounds in HTS relative to the number of compounds tested is similar to other HTS for antifungal activities [17,18]. We believe that the psychrophilic conditions needed for this screen and the innate ability of Gd to grow in the presence of a diverse group of chemicals limited the outcome to a few well-known fungal inhibitors. The list of the most effective compounds included antifungals, fungicide and biocides. In contrast to these known inhibitors, the high susceptibility of Gd to fluorouracil was notable and unexpected since Gd was found to be resistant to flucytosine. This observation suggested that the  [14,19,20]. By the same logic, it is more likely that the observed flucytosine-resistance in Gd involves either defects in the uptake of this drug or its inefficient conversion to 5-fluorouracil (5-FU) and 5-fluorodeoxyuridine monophosphate (5-Fd-UMP).
The effective biocides and fungicides identified in this study are currently in use, and therefore, could potentially be useful for the decontamination of Gd infested sites. This is a positive step as earlier publications on decontamination of sites harboring pathogenic fungi such as Histoplasma capsulatum used formalin, which is not an option at present due to the associated toxicity [21]. However, any proposal for the decontamination of infested sites requires a careful balance considering the likely deleterious effects on the environment. It is worth repeating that caves support a rich diversity of life including many species of fungi and bacteria [10,22]. So far, extensive published data is only available for benzalkonium chloride. This chemical was found to be an effective disinfectant for household items contaminated with the dermatophytic pathogen Microsporum canis [23]. Benzalkonium chloride has been used for many years in the famous Lascaux cave (Montignac, France) to control Fusarium solani infestation. Although effective, this treatment has resulted in the replacement of native bacterial communities with more resistant bacteria [24]. The long-term environmental effects of such biocide-induced changes remain unknown. The effective inhibition of Gd growth by phenylmercury acetate suggests that organomercury fungicides could also be effective agents for the decontamination of natural sites. Similar to the concerns with biocides, any future remedial plans involving fungicides would require further refinements to ensure that the non-target fungi and other living forms are not harmed [25,26]. In conclusion, a number of antifungal drugs and inhibitory compounds were systematically identified in this study; these results could serve as a starting point for the selection of appropriate agents to treat Gd-infected bats in rehabilitation centers and to decontaminate heavily infested sites.

Antifungal susceptibility testing
Two different commercial devices were used for antifungal testing. These devices yield reproducible results and are reported to be good alternatives to standard methods, which might be too cumbersome for many laboratories working with Gd [12,27,28,29]. Five strains of Gd characterized in an earlier study were used in the susceptibility tests [2]. In brief, Gd strains were grown on potato dextrose agar for 10 days at 15uC. Spore suspensions were prepared in sterile water, adjusted to OD 530 = 0.1. For testing with Etest strips, diluted spore suspension was streaked on to RPMI/MOPS agar and the Etest strips were placed per the manufacturer's instruction (bioMerieux, Inc., Durham, NC). For 96-microwell dilution tests, the YeastOneH with Anidulafungin, Caspofungin and Micafungin panels were used (Part #YO-9, TREK Diagnostic Systems, Inc., Cleveland, OH). The susceptibility patterns of antifungals were assessed 10-14 days post-incubation at 15uC and 21-28 days post-incubation at 6uC. We have previously reported that Gd grows faster at 15uC than at 6uC without any apparent changes in its morphology or secretory enzymes [2].

High-throughput screening (HTS)
The method used for HTS was done as previously described by Li et al. [30]. A SpectrumPlus (MicroSource) library representing Figure 3. Summary of secondary screening for G. destructans growth inhibition. The results shown were obtained at 15uC and 6uC with 10 promising compounds tested ten-fold and hundred-fold lower concentrations than used in the initial screen. Note that two topical azole antifungals and a fungicide were most effective at all concentrations tested. (Abbreviation: ECO econazole nitrate, SCO sulconazole nitrate, PMA phenylmercuric acetate, CAQ chloroacetoxyquinoline, 5-FU fluorouracil, ZnP pyrithione zinc, CPX ciclopirox olamine, DIG digitonin, CHL chloroxine, BZCL benzalkonium chloride). doi:10.1371/journal.pone.0017032.g003 1920 structurally diverse compounds including known drugs, experimental actives, and pure natural products, was used. The compounds were distributed in sterile 96-microwell plates with a Sciclone ALH 3000 workstation (Caliper Life Sciences, Hopkinton, MA). The control well of the 96-microwell plate contained 100 ml of 0.1% DMSO in sterile water+100 ml of Gd spore suspension (OD 530 = 0.1) in 26 yeast nitrogen glucose broth (2% yeast nitrogen base+4% glucose; YNGB). All other wells contained 100 ml of different drugs (10 mM concentration in 0.1% DMSO)+100 ml of spore suspension in 26 YNGB. The growth inhibition was assessed spectrophotometrically after 10 days of incubation at 15uC. Inhibition of the compounds was calculated as percentages from the ratio of absorbance for Gd with compound wells compared to no compound in 0.1% DMSO alone. Secondary screening of selected compounds proceeded similar to the primary screen except that the 10-and 100-fold lower concentrations of the effective compounds identified in the initial screens were tested both at 15uC and 6uC.