Successful treatment of Batrachochytrium salamandrivorans infections in salamanders requires synergy between voriconazole, polymyxin E and temperature

Chytridiomycosis caused by the chytrid fungus Batrachochytrium salamandrivorans (Bsal) poses a serious threat to urodelan diversity worldwide. Antimycotic treatment of this disease using protocols developed for the related fungus Batrachochytrium dendrobatidis (Bd), results in therapeutic failure. Here, we reveal that this therapeutic failure is partly due to different minimum inhibitory concentrations (MICs) of antimycotics against Bsal and Bd. In vitro growth inhibition of Bsal occurs after exposure to voriconazole, polymyxin E, itraconazole and terbinafine but not to florfenicol. Synergistic effects between polymyxin E and voriconazole or itraconazole significantly decreased the combined MICs necessary to inhibit Bsal growth. Topical treatment of infected fire salamanders (Salamandra salamandra), with voriconazole or itraconazole alone (12.5 μg/ml and 0.6 μg/ml respectively) or in combination with polymyxin E (2000 IU/ml) at an ambient temperature of 15 °C during 10 days decreased fungal loads but did not clear Bsal infections. However, topical treatment of Bsal infected animals with a combination of polymyxin E (2000 IU/ml) and voriconazole (12.5 μg/ml) at an ambient temperature of 20 °C resulted in clearance of Bsal infections. This treatment protocol was validated in 12 fire salamanders infected with Bsal during a field outbreak and resulted in clearance of infection in all animals.

The rate at which amphibian populations have been declining the past decades is alarming 1,2 . One of the factors in part responsible for these declines is the infectious disease chytridiomycosis caused by the chytrid fungus Batrachochytrium dendrobatidis (Bd) [3][4][5] . Recently, the related chytrid fungus Batrachochytrium salamandrivorans (Bsal) has been identified as a novel threat to amphibian populations, with a potentially major impact on salamander diversity worldwide 6,7 . For amphibian chytridiomycosis caused by Bd, topical antimycotic treatment using voriconazole at a concentration of 1.25 μ g/ml during 7 days has proven highly successful and safe 8 . However, applying this treatment to Bsal infected salamanders is unable to clear infections (see Case report section). Thermal treatment consisting of exposure to the critical thermal maximum for Bsal (25 °C) for 10 days was shown to be able to clear Bsal infections from infected salamanders 9 . However, this temperature approaches the critical thermal

Results and Discussion
In vitro susceptibility of Bsal to antimycotic compounds. The results of the experiments to determine the MICs of the tested antimicrobials for Bsal are summarized in Table 1. Florfenicol was the only compound tested that was not able to limit growth or kill Bsal at the concentrations tested. In contrast, florfenicol is capable of limiting growth of Bd at concentration of 0.5-1.0 μ g/ml (Muijsers et al. 2012). Interestingly, the inhibitory concentrations of the other compounds against Bsal differed noticeably from those against Bd. The mechanism underlying this difference remains unknown. Whereas polymyxin E did not show any inhibitory potential in Bd MIC tests at the concentrations used 12 , Bsal was inhibited by polymyxin E at a concentration of 8000 IE/ml (Table 1). Terbinafine limited Bsal growth at a concentration of 0.2 μ g/ml, which is in accordance with its activity against Bd at 0.063 μ g/ml 13 . Itraconazole, which is frequently used to treat amphibians infected with Bd, had a MIC against Bsal 2.5-5 times lower (0.006 μ g/ml) compared to its MIC against Bd (0.016-0.032 μ g/ml) 13 . Finally, the MIC of voriconazole for inhibiting Bsal growth was 10 times higher (0.125 μ g/ml) than the MIC for inhibiting Bd (0.0125 μ g/ml) 8 . This result at least partly explains the failed initial treatment of the wild fire salamanders using the voriconazole dosage for treating chytridiomycosis in amphibians infected with Bd (1.25 μ g/ml sprays, twice a day for 7 days) 8  Synergy between polymyxin E and voriconazole or itraconazole in inhibiting Bsal growth. The three main techniques used for testing interactions between compounds in antifungal activity are Etest, time-kill methods and checkerboard dilution methods [14][15][16][17] . In synergy testing for bacterial pathogens, the biggest disadvantage is that no two methods will produce comparable results, and therefore clinical applicability of results is under debate 18 . These limitations also apply for antifungal synergy testing 17 . Furthermore, a vast amount of studies exist that describe in vitro synergy without linking (or being able to link) these results to a beneficial treatment outcome of combined treatment 18 . The goal of this study was to evaluate potential synergy between antimycotic compounds in inhibiting Bsal growth to allow development of an experimental treatment protocol using antifungal concentrations below toxicity levels. In this study, a checkerboard dilution method adopted from the method used to evaluate minimum inhibitory concentrations for Bd 8,12 was used, which in comparison to the time-kill method, is easier to carry out and interpret 17 . The combinations of compounds tested both included an azole antifungal (voriconazole or itraconazole) and polymyxin E, which were already shown to be able to inhibit Bsal growth (Table 1). Apart from polymyxins exerting antifungal activity on their own, combinations of polymyxins and azole antifungals showed synergistic antifungal activity against infections with Aspergillus spp., Candida spp. and Cryptococcus spp [19][20][21] . The bactericidal activity of polymyxin E against Gram-negative bacteria is an added advantage for treating Bsal associated lesions, since histological preparations of skin samples of salamanders infected with Bsal often revealed severe bacterial overgrowth of the skin in concordance with Bsal infection 6 . Secondary bacterial infections in immunocompromised amphibians are often caused by opportunistic Gram-negative bacteria 22 [24][25][26] , so studies evaluating the efficacy of reduced concentrations of itraconazole alone 27 or in combination therapies to successfully treat chytridiomycosis could be a major advantage. The results of the experiments to determine the FICs of polymyxin E combined with voriconazole or itraconazole are graphically depicted in isobolograms ( Fig. 1). Two of the tested combinations of polymyxin E with voriconazole (2000 IE/ml + 0.02 μ g/ml and 1000 IE/ml + 0.03 μ g/ml) and two combinations of polymyxin E with itraconazole (2000 IE/ml + 0.0016 μ g/ml and 1000 IE/ml + 0.0016 μ g/ml) resulted in a FICI that demonstrates synergism (FICI ≤ 0.5, Fig. 1). All combinations that inhibited Bsal growth also killed Bsal completely after 10 days of incubation.
Effective treatment of Bsal infections in fire salamanders based on synergy between polymyxin E, voriconazole and temperature. All initially tested treatment conditions, composed out of itraconazole or voriconazole alone and in combination with polymyxin E were unable to clear Bsal infections from infected amphibians (Fig. 2, panels A-E ). Although the combination therapies of itraconazole or voriconazole with polymyxin E did reduce Bsal infection loads to undetectable levels in 3 out of the 5 animals and 5 out of 5 animals respectively, recrudescence of infection did occur in all animals ( Fig. 2, panels C and E). The difference between in vitro and in vivo effects of the combination therapy at 15 °C might lie in the exposure to the compounds; in the in vitro experiments, Bsal was exposed continuously to both compounds as opposed to the periodical exposure in the in vivo experiments. At least for polymyxin E (2000 IU/ml) longer exposure times are unusable due to occurrence of toxicity (personal observations). The conditions of the additional treatment (Fig. 2, panel F) instituted after failure of the initial conditions to clear Bsal infections, were based on the in vitro synergy between voriconazole and polymyxin E in inhibiting Bsal growth, the increased but suboptimal inhibition of Bsal in vivo by combined exposure to voriconazole and polymyxin E (Fig. 2, panel C) and the previously determined temperature dependent infection dynamics of Bsal 9 . Using the same concentrations of voriconazole and polymyxin E, but raising the temperature to 20 °C did result in successful elimination of Bsal in all infected animals (Fig. 2, panel F). The results of this study underline the key influence temperature plays in Bsal infection dynamics, which already allowed development of a Bsal temperature treatment protocol for amphibian species able to endure a continuous ambient temperature of 25 °C for 10 days 9 . Ethical considerations allowed only one additional treatment condition to be tested. Therefore, the positive treatment effect could theoretically be attributed to the sole influence of either one of the compounds at 20 °C, as experimental treatments with individual compounds were only tested in vivo at a temperature of 15 °C. The results of this study show that synergy between voriconazole and polymyxin E together with the temperature dependent infection dynamics of Bsal allow Bsal infections to be eliminated in amphibian species with critical thermal maxima lower than that of Bsal. The efficacy of the treatment protocol was validated by successful treatment of fire salamanders naturally infected with Bsal during a field outbreak (Fig. 3). In conclusion, in vitro synergy between antimycotic compounds in inhibiting Bsal, together with the temperature dependent infection dynamics of Bsal allowed development of a treatment protocol successful in  eliminating Bsal from experimentally and naturally infected amphibians. Although exploiting synergism between temperature and chemical compounds was effective in this study, in case of therapeutic failure, acclimatisation of Bsal to higher temperatures and/or the emergence of Bsal strains with higher thermal preferences should be considered 28,29 .

Methods
All experiments were performed in accordance with the relevant guidelines and regulations. All experiments with experimental animals were carried out with approval of the ethical committee of the Faculty of Veterinary Medicine, Ghent University.
Chytrid strain & culture conditions. The Bsal type strain (AMFP13/1) 6 was grown in TGhL broth (16 g tryptone, 4 g gelatin hydrolysate, 2 g lactose per liter of distilled water) in 25 cm 3 cell culture flasks and incubated at 15 °C. To obtain a suspension containing a mixture of zoosporangia and zoospores, the walls of a cell culture flask containing 5-day-old culture were scraped with a sterile cell scraper and the suspension subsequently collected.
In short, two-fold dilutions series of the antimicrobial agents were prepared in TGhL broth, and 200 μ l of these prepared dilutions were added to wells of 24 well cell culture plates (Table 1). Two hundred μ l of a suspension containing a mixture of Bsal sporangia and zoospores (approximately 10 5 Bsal organisms per ml) were added to all wells. Finally, 1600 μ l of TGhL broth were added to all wells resulting in a final volume of 2 ml per well. Plates were incubated at 15 °C (optimum growth temperature of Bsal 6 ) and checked for viability and growth daily for 10 days with an inverted light microscope. Wells containing TGhL broth with viable Bsal sporangia and zoospores and wells containing heat treated (85 °C, 10 minutes) Bsal sporangia and zoospores served as positive and negative growth controls respectively. The MIC value was determined as the lowest concentration of the antimicrobial agent at which no growth could be observed after 10 days of incubation. To test which concentrations of the antimicrobial agents were lethal for Bsal after 10 days of exposure, we removed the medium and replenished all wells with fresh TGhL broth without antimicrobial agents. Plates were incubated at 15 °C and checked for viability and growth daily for an additional 14 days with an inverted light microscope. A concentration was considered to be lethal to Bsal when no signs of growth could be observed after this incubation period of 14 days (Table 1). All conditions were tested in triplicate.

Determination of the fractional inhibitory concentrations of antimicrobial agents against
Bsal. To test for synergy in combinations of polymyxin E with voriconazole or with itraconazole in inhibiting Bsal growth, fractional inhibitory concentrations (FICs) 30 were determined using a  18,31,32 . Isobolograms were used to graphically depict the FIC and FICI values of all tested combinations of antimicrobial agents ( Fig. 1) 33 . To test which combination of concentrations of the antimicrobial agents was lethal for Bsal after 10 days of exposure, we replaced the broth containing the compound(s) with fresh TGhL broth without antimicrobial agents. Plates were incubated at 15 °C and checked for viability and growth daily for an additional 14 days with an inverted light microscope. A combination of concentrations was considered to be lethal to Bsal when no signs of growth could be observed after this incubation period of 14 days. All conditions were tested in quadruplicate.

Treatment of experimentally infected fire salamanders. Fire salamanders (Salamandra salamandra)
were inoculated with Bsal in order to study in vivo efficacy of different antimicrobial treatment protocols. The animal experiment was performed with the approval of the ethical committee of the Faculty of Veterinary Medicine (Ghent University, EC2013/87 and EC2014/65). Thirty captive bred fire salamanders were housed individually in plastic containers in a climatized room with an ambient temperature of 15 °C. The animals were kept on a moist tissue, with access to a hiding place and water container. Crickets powdered with mineral and vitamin supplement were provided ad libitum as food source. All animals were clinically healthy and free of Bd and Bsal, as determined with duplex real-time PCR examination of skin swabs 34 . An acclimatization period of 2 weeks was admitted before the start of the experiment. The experimental animals were randomly assigned to one of the 6 experimental treatment groups (5 animals per treatment group, kept individually). All salamanders were inoculated with Bsal by topically applying one mL of inoculum containing 10 5 zoospores per ml on the skin. Animals were kept at 15 °C (except for the animals in group F; details below) and skin swabs for Bsal real-time PCR analysis 34 were collected every 7 days. Individual treatment commenced when Bsal infection was established (determined as an increase in Bsal infection load between 2 consecutive samplings). The different groups were untreated negative control (group A), voriconazole treatment (12.5 μ g/ml) alone (group B), voriconazole and polymyxin E treatment (concentrations of 12.5 μ g/ml and 2000 IU/ml respectively, group C), itraconazole treatment (0.6 μ g/ml) alone (group D) and itraconazole and polymyxin E treatment (0.6 μ g/ml and 2000 IU/ml respectively, group E). After initial failure to clear Bsal infections in these first experimental groups, we trialed another treatment condition composed of voriconazole and polymyxin E treatment (concentrations of 12.5 μ g/ml and 2000 IU/ml respectively, identical to the treatment described for group C) but with an ambient temperature of 20 °C instead of 15 °C (group F). Due to ethical considerations only one additional condition was performed. All experimental treatments were carried out twice a day for 10 days. Polymyxin E was administered through submersion baths (10 minutes) and voriconazole and itraconazole were administered through spraying the animals and tissue in their housing (after polymyxin E baths if applicable). After the treatment period, all animals were kept at 15 °C. Skin swabs for Bsal real-time PCR analysis were collected immediately after the treatment period and subsequently every 7 days for another 3 weeks. An animal was considered negative for Bsal after 3 consecutive negative real-time PCR results. Development/progression of symptoms associated with Bsal infections together with presence of Bsal as determined with real-time PCR analysis in an animal was determined as experimental endpoint and resulted in withdrawal of the animal from the experiment. If an animal tested positive for the presence of Bsal in the post-treatment follow-up phase (starting from day 10 in Fig. 2), the treatment was considered as failed. Remaining Bsal infections in animals that were removed from the experiment due to reaching the described endpoint, and animals still positive for Bsal Scientific RepoRts | 5:11788 | DOi: 10.1038/srep11788 at the last sampling time point were exposed to an ambient temperature of 25 °C during 10 days to clear the Bsal infection 9 .
Treatment of naturally infected fire salamanders. Thirty-five fire salamanders from the population from which Bsal (strain AMFP13/1) was originally isolated (Bunderbos, Netherlands, N50°54′ 51″ , E5°44′ 59″ ) were transferred to our research facility for treatment. Upon arrival, 12 of the translocated animals tested positive for presence of Bsal DNA as tested with the Bsal real-time PCR 34 . Based on the results of the treatments of experimental infections, the animals were treated with polymyxin E submersion baths (2000 IU/ml, 10 minutes) followed by spraying voriconazole (12.5 μ g/ml) twice a day for 10 days at an ambient temperature of 20 °C. Housing conditions of the animals were identical to the conditions described for the experimental animals. After the treatment period all animals were put back at 15 °C. Skin swabs for Bsal real-time PCR analysis were collected directly after the treatment period and subsequently every 7 days for another 3 weeks. An animal was considered negative for Bsal after 3 consecutive negative real-time PCR results.

Cost of treatment.
A step-by-step treatment protocol can be found as supplementary file. Based on the products, volumes and estimated pricing in this protocol, the treatment cost of a single animal would be 6 €. It should be noted however, that this estimated cost applies for individual treatment of animals. Treatment of groups of amphibians with the same volumes is possible (although this has not yet been validated), resulting in a reduction of the cost of treatment per animal.