In vitro modeling of Batrachochytrium dendrobatidis infection of the amphibian skin

The largest current disease-induced loss of vertebrate biodiversity is due to chytridiomycosis and despite the increasing understanding of the pathogenesis, knowledge unravelling the early host-pathogen interactions remains limited. Batrachochytrium dendrobatidis (Bd) zoospores attach to and invade the amphibian epidermis, with subsequent invasive growth in the host skin. Availability of an in vitro assay would facilitate in depth study of this interaction while reducing the number of experimental animals needed. We describe a fluorescent cell-based in vitro infection model that reproduces host-Bd interactions. Using primary keratinocytes from Litoria caerulea and the epithelial cell line A6 from Xenopus laevis, we reproduced different stages of host cell infection and intracellular growth of Bd, resulting in host cell death, a key event in chytridiomycosis. The presented in vitro models may facilitate future mechanistic studies of host susceptibility and pathogen virulence.

Bd-induced chytridiomycosis was first described 20 years ago [4] and several studies have documented Bd growth and development at morphological and ultrastructural levels [19][20][21]. PLOS  The general Bd-infection steps have been described as attraction to a suitable host, attachment of zoospores to the host skin, zoospore germination, germ tube development and penetration into the skin cells, leading to endobiotic growth of this pathogen inside host cells which eventually results in the loss of host cytoplasm [20]. Despite recent advances in understanding the pathogenesis, fundamental knowledge about the early infection process at a cellular level, crucial in understanding disease pathogenesis, is however still limited [6][7][8][9][10][11][21][22]. Infectious diseases are commonly studied in vitro by assessing the interaction of a pathogen with host cells. This is a reductionist approach, but one that can advance the understanding of mechanisms that underlie infection and disease. After two decades of chytrid research, a cellbased assay is lacking and the focus still remains on in vivo experimentation. To date, infectivity and the pathogenicity of Bd have mostly been studied using light microscopy (LM), scanning electron microscopy (SEM) and transmission electron microscopy (TEM) on in vivo-infected skin tissues or ex vivo-infected skin explants [20,23]. The main objective of the current study was to establish a cell-based assay that mimics the colonization stages of Bd in vitro, allowing rapid and efficient screening of host-Bd interactions. We first optimized an early-infection model showing attachment of Bd to primary amphibian keratinocytes (PAK), followed by internalization of Bd in these host cells. Secondly, we developed an invasion model using the Xenopus laevis kidney epithelial cell line A6 mimicking the complete Bd colonization cycle in vitro.

Experimental animals
We isolated PAK from adult Litoria caerulea (captive bred). Upon arrival and before the startup of the experiments we examined kin swabs for the presence of Bd by quantitative PCR (qPCR) [24]. Husbandry and euthanasia methods were in accordance with the guidelines of the Ethical committee of the Faculty of Veterinary Medicine (Ghent University). Animals were euthanized by intracoelomic injection of sodium pentobarbital (Annex IV of the EU directive 2010/63). For the isolation of primary keratinocytes, ethical permission by the ethical committee of the Faculty of Veterinary Medicine (Ghent University) was not required under Belgian and European legislation (EU directive 2010/63/EU).

Batrachochytrium dendrobatidis growth conditions
We carried out the inoculations with Bd strain JEL 423. This strain was isolated from an infected Phyllomedusa lemur frog in Panama and is a representative of the Bd global panzootic lineage [25]. The Bd strain was routinely cultured in TGhL broth (1.6% tryptone, 0.4% gelatin hydrolysate and 0.2% lactose in H 2 O) in 75 cm 2 cell culture flasks at 20˚C for 5 days. We collected the Bd spores from a full-grown culture containing mature sporangia. Once the zoospores were released, the medium containing the zoospores was collected and passed over a sterile mesh filter with pore size 10 μm (PluriSelect, Leipzig, Germany). We used the flowthrough as the zoospore fraction (> 90% purity).

Cell culture: Isolation of PAK
Isolation of PAK from Litoria caerulea frogs was performed as previously described [20,23], with minor modifications. In brief, after euthanizing the frogs, we washed them in plastic containers containing respectively 70% ethanol, 70% Leibovitz L-15 medium without phenol red (3 times) (Fisher Scientific, Aalst, Belgium), Ca 2+ /Mg 2+ -free Barth's solution (CMFB; Bilaney Consultants GmbH, Düsseldorf, Germany), 1.25 mM ethylenediaminetetraacetic acid (EDTA; Sigma-Aldrich, Overijse, Belgium) in CMFB for 5 min and 70% L-15 medium (twice) at 4˚C. Next, we excised ventral skin, which we rinsed at apical and basal side with 70% L-15 medium. From each donor animal a skin sample was taken, fixed in 70% EtOH and tested for the presence of Bd by qPCR [24]. We then cut the skin into 10 x 20 mm wide strips, which were incubated overnight in MatriSperse TM Cell Recovery Solution (BD Biosciences, Massachusetts, USA) at 4˚C. Subsequently, we peeled off the the cornified skin layers using sterile needles and forceps. To obtain single cell suspension, we incubated the cornified skin in 10 U/ml dispase solution (Fisher scientific) in 70% L-15 medium at 20˚C, 5% CO 2 . Finally the cells were suspended by repetitive pipetting, washed in 70% L-15 medium and resuspended in the appropriate cell culture medium for invasion assays.

Cell culture: Continuous A6 cell line
The Xenopus laevis kidney epithelial cell line A6 (ATCC-CCL 102) was grown in 75 cm 2 cell culture flasks and maintained in complete growth medium (74% NCTC 109 medium (Fisher Scientific), 15% distilled water, 10% fetal bovine serum (FBS) and 1% of a 10 000 U/ml penicillinstreptomycin solution (Fisher Scientific)) and the cells were incubated at 26˚C and 5% CO 2 until they reached confluence [26]. Using trypsin, we detached the cells, washed them with 70% Hanks' Balanced Salt Solution without Ca 2+ , Mg 2+ (HBSS-; Fisher Scientific) by centrifugation for 5 min at 1500 rpm and resuspended them in the appropriate cell culture medium for invasion assays.

Fluorescent in vitro model to assess adhesion and invasion of Bd in PAK
PAK are only usable for 1 to 4 days and the lifecycle of Bd takes approximately 4 to 5 days [19]. As such, these cells are not appropriate to examine the complete maturation process of this pathogen, but they can be used to investigate the early steps in Bd-host cell interaction, including adhesion and invasion. To visualize these early pathogen interactions (4 hours: adhesion and 24 hours: invasion), we stained the PAK with 3 μM CellTracker TM Green CMFDA (Fisher Scientific) according to the manufacturer's guidelines. After centrifugation for 5 min at 1500 rpm, we suspended the cells in cell medium A (70% L-15 medium, 20% distilled water and 10% FBS) and seeded 10 5 cells per well in 24-well tissue culture plates containing collagen-coated glass coverslips. PAK were allowed to attach for 1 hour at 20˚C and 5% CO 2 after which they were washed with 70% Hanks' Balanced Salt Solution with Ca 2+ , Mg 2+ (HBSS+; Fisher Scientific). Next, we inoculated the cells with Bd zoospores stained with 3 μM CellTracker TM Red CMTPX (Fisher Scientific) [27] in cell medium B (40% L-15 medium, 55% distilled water and 5% FBS), to ensure the mobility of the zoospores, at a multiplicity of infection (MOI) of 1:10. After 2 hours, we gently washed the infected cells three times with 70% HBSS+ to remove non-adherent spores and we replaced cell medium B by cell medium A. To asses early Bd-PAK interactions, the infected cells were washed three times and fixed in 0.5 ml 3.0% paraformaldehyde for 10 min at 4 hours and 24 hours post infection (p.i.). Finally, we used Hoechst (Fisher Scientific) for nuclear staining and we mounted the coverslips using ProLong Gold antifade mountant (Fisher Scientific). For visual confirmation of Bd-PAK interactions, we studied the cells using fluorescence microscopy and confocal laser scanning microscopy (CLSM), using appropriate filter sets. Detailed protocols are available at protocols.io (dx.doi.org/10.17504/protocols.io.8ihhub6).

Fluorescent in vitro model to assess adhesion of Bd to A6 cells
To study Bd adhesion (less than 24 hours p.i.) in the epithelial cell line A6, the protocol described in PAK cells was slightly modified. After staining the A6 cells with 3 μM CellTracker TM Green CMFDA, we seeded 10 5 cells per well in 24-well tissue culture plates containing collagen-coated glass coverslips and the cells were allowed to attach for 2 hours at 20˚C and 5% CO 2 . We then washed the cells three times with 70% HBSS+, after which we inoculated them with Bd zoospores in cell medium B, at a MOI of 1:10. Two hours p.i., the cells were washed three times with 70% HBSS+ and we replaced cell medium B by cell medium A. Four hours p.i. the infected cells were washed three times with HBSS+ and we incubated them with Calcofluor White stain (10 μg/ml in 70% HBSS+; Sigma-Aldrich) for 10 min. Next, we washed the cells three times with 70% HBSS+, followed by fixation. Finally the cells were mounted and we analyzed them using fluorescence microscopy. Detailed protocols are available at protocols. io (dx.doi.org/10.17504/protocols.io.8thhwj6).

Fluorescent in vitro model to assess invasion and intracellular maturation of Bd in A6 cells
To assess Bd-A6 cell interactions starting from 24 hours p.i., we seeded unstained A6 cells which we inoculated with unstained Bd zoospores as described above. At different time points (from 1 to 6 days p.i.), the Bd-A6 cell interactions were visualised as follows: infected cells were stained with 3 μM CellTracker TM Green CMFDA, washed three times with 70% HBSS+ and they were incubated with Calcofluor White stain (10 μg/ml in 70% HBSS+) for 10 min. After washing 3 times with HBSS+, we fixed the infected cells, permeabilized them for 2 min with 0.1% Triton 1 X-100 and incubated them for 60 min with a polyclonal antibody against Bd raised in rabbit (1/1000) [28]. After washing three times with 70% HBSS+, we incubated the samples with a monoclonal goat anti-rabbit Alexa Fluor 568 (1/500) antibody (Fisher Scientific; A11011). After an incubation of 1 hour, we washed the samples three times with 70% HBSS+, mounted them and finally analyzed them using fluorescence microscopy and CLSM. The Alexa Fluor 568 targeting Bd and Calcofluor White stainings were used in concert to assess the ability of Bd to penetrate the host cell. Calcofluor White is not internalized by A6 cells, whereas the Alexa Fluor 568 staining (targeting Bd) was applied after permeabilization of the host cells. As such, intracellular Bd will only be targeted by the Alexa Fluor 568 whereas extracellular Bd bodies will be bound by both the Alexa Fluor 568 and Calcofluor White stain. We included shaminfected cells as a negative control to check the cell morphology over different time points (S1 File). Detailed protocols are available at protocols.io (dx.doi.org/10.17504/protocols.io.8ishuee).

Fluorescent caspase-3 staining to assess induction of apoptosis in A6 cells
To visualize the induction of cell death in Bd-infected A6 cells, we performed a fluorescent caspase-3 staining. Therefore, we seeded unstained A6 cells which were inoculated with unstained Bd zoospores as described above. At different time points (day 4 to 6 p.i.), the infected cells were fixed, permeabilized and incubated for 60 min with anti-caspase-3 primary antibody raised in rabbit (Sigma-Aldrich; C8487) 1/1000 diluted. After washing 3 times with 70% HBSS +, we treated the samples with goat anti-rabbit Alexa Fluor 568 (1/500) for 1 hour. We then washed the cells and treated them with Hoechst for 15 min. Finally, the cells were washed three times with 70% HBSS+, mounted and analyzed using fluorescence microscopy. We included sham-infected cells as a negative control and staurosporin-treated A6 cells (1 μM, 24 hours; Sigma-Aldrich) as a positive control (S2 File). Detailed protocols are available at protocols.io (dx.doi.org/10.17504/protocols.io.8tihwke)

PAK can be used to reproduce the early infection stages of Bd
We first optimized an in vitro model using PAK, that could be used in research unravelling factors that underpin early pathogenesis in primary cells. Therefore, an invasion experiment was performed with fluorescently-labelled PAK cells of Litoria caerulea and fluorescently-labelled Bd spores that were incubated for 4 and 24 hours (Fig 1). After a 4-hour invasion period, clear contact between the spores and host cells was observed (Fig 1A-1D). When increasing the contact time to 24 hours, host cells seemed to be invaded by Bd spores as intracellular chytrid thalli were observed (Fig 1E-1H). To confirm this, confocal microscopy was used to determine the exact position of the chytrid thalli, showing a clear intracellular localization (Fig 2A-2C).

The entire Bd colonization cycle can be mimicked using A6 cells
Fluorescent microscopy of PAK was shown to be useful to visualize the early host-pathogen steps, including attachment and invasion of Bd. We next tested whether similar results could be obtained working with the Xenopus laevis kidney epithelial cell line A6 (Figs 3 and 4). After an incubation of 4 hours, we observed the formation and growth of tubular structures, called germ tubes [20] (Fig 3A). Using a combination of Alexa Fluor 568 targeting Bd and a Calcofluor White staining allowed us to discriminate between the intracellular and extracellular localization of Bd. From 1 day p.i. on, the germ tubes penetrated the A6 cells (Fig 3B). After germ tube protrusion into the cells, both epibiotic and endobiotic Bd growth were observed. Epibiotic Bd growth was limited to Bd development outside the host cells, whereas endobiotic growth was characterized by intracellular Bd colonization (Fig 4). Both at 1 and 2 days p.i., an intracellular swelling was formed at the end of the germ tube, giving rise to a new Bd thallus ( Fig 3C). As shown in Fig 3D, 2 to 3 days p.i. intracellular colonization was observed in A6 cells, which was also confirmed using CLMS (Fig 2D-2F). At day 3 and 4 p.i., maturation of the intracellular thalli was observed with the formation of large intracellular zoosporangia ( Fig  3E). The formation of a discharge tube could be seen at day 4 to 5 p.i. (Fig 3F) through which Bd contents was released into the cell (Fig 3G), eventually leading to the induction of host cell death. From day 5 p.i. onward, marked caspase-3 activation was observed in Bd-associated A6 cells, a key event in apoptosis induction of host cells (Fig 3H).

Discussion
Chytridiomycosis is increasingly recognized as a challenge for wildlife conservation. The power of a single disease to affect an entire vertebrate class and the fact that mitigation of Bd and Bsal in nature is still in its infancy [29], makes it one of the most destructive diseases ever recorded [1]. To date, a lot of research has focused on ecology and epidemiology of this fungal disease [1,[30][31] and although fundamental knowledge of the disease's pathogenesis is increasing, still knowledge gaps remain [32]. We here present in vitro infection models, intended to study Bd-host interactions in order to further explore the gaps in our understanding of chytridiomycosis. To date, in vivo experimentation still is the golden standard in Bd and Bsal research. To understand host-pathogen interactions in natural systems, researchers often turn to laboratory infection experiments. Although in vivo research has tremendous value for understanding disease processes, the availability of in vitro infection models could provide a first line tool to gain insight into host-pathogen interactions which will reduce the number of animals used in infection trials [33].
We showed that primary keratinocytes could be useful to mimic and examine the early Bdhost interactions, which until now have only been described using light microscopy and TEM of Bd-infected skin explants [20]. Previously, it has been stated that these cells cannot be used to study host-chytrid interactions because of the incompatibility of commonly-used culture media and the motility of Bd zoospores [23]. This obstacle was circumvented by diluting the cell culture medium during the first two hours of contact between Bd and the cells, guaranteeing the motility of the Bd spores during the adhesion process. Not all amphibian species are equally sensitive to chytridiomycosis and factors contributing to susceptibility of amphibians to this disease are not completely known [34][35][36]. However, specific attachment to a suitable host, induction of encystment and invasion of host cells are crucial and underexplored processes for successful colonization of this fungus. The described in vitro model may for example be used to look into adhesion factors or adherence mediators, that are possibly linked to host susceptibility.
Although working with primary cell cultures is more closely linked to the in vivo situation, cell lines provide the major advantage that they are standardized, immortalized and that no animals are needed. By using the epithelial cell line A6 from Xenopus laevis we were able to mimic the complete infection cycle of Bd and we showed that this model can be used to assess adhesion, invasion and maturation interactions, reflecting endobiotic development which is observed in susceptible amphibians [20]. Besides the intracellular colonization, we also observed epibiotic development of Bd, a type of growth which previously has been described in infection trials with ex vivo skin explants of Xenopus laevis [20]. Up to date there is however no histological evidence of epibiotic growth of Bd occurring in nature. Therefore it could be suggested that the reported epibiotic growth is linked to the in vitro/ex vivo conditions, including the extracellular presence of nutrients from the cell culture medium, the lack of mucus and fungicidal skin secretions [37][38][39][40][41][42] and the lack of a normal skin microbiome [43][44]. In our in vitro model, apoptosis of A6 cells was observed when the zoospores were discharged into the cell by intracellular zoosporangia. As a result of this cell death the zoospores were released into the extracellular environment, ready to colonize new host cells, which (partly) deviates from the in vivo situation. In susceptible animals, germ tube-mediated invasion, establishment of intracellular thalli and spread of Bd to the deeper skin layers have been described, but this is followed by an upward migration by differentiating epidermal cells resulting in the releasement of the zoospores at the skin surface [19][20][21]. During Bd-induced chytridiomycosis, apoptosis has been reported as a key event, but the exact mechanism and role remains to be elucidated [45]. However, since epidermal cell death is positively associated with infection loads and morbidity [45], it is likely that cell death originates by colonization of many zoosporangia rather than the intracellular releasement of zoospores as observed in this in vitro model. Caution should always be exercised when extrapolating in vitro data to the in vivo situation, but in vitro cell culture models allow an experimental flexibility making them highly suitable to study host-pathogen interactions. Interestingly, the whole genome sequences of Xenopus laevis and Bd are known, permitting the study of transcriptional responses in host and pathogen during different infection stages. To date, different Bd lineages have been detected, all with their own virulence properties [25,[46][47]. The availability of an in vitro model using a continuous cell line may be used to analyze the differences in host-pathogen interactions between different Bd strains.
Summarized, for the first time we describe in vitro cell infection models that mimic Bd interactions with the amphibian skin ranging from adhesion, germ tube development, penetration into skin cells and invasive growth to the induction of host cell death. These in vitro models provide an import tool that may help understanding Bd-host interactions.