Introduction

Dermatophytosis is a highly prevalent superficial fungal infection that can affect skin, nails and hair. It affects an estimated 20–25% of the world’s population [1], and the worldwide expenditure on antifungal agents to treat dermatophytosis is estimated at 2.5 billion USD. Risk factors include participation in sports activities, type 2 diabetes and ageing [2], and because of the latter two, the prevalence of dermatophytosis is expected increase further. The disease is generally mild, but it does impact the quality of life. However, in diabetics it contributes to the severity of the diabetic foot [3] and in immunocompromised patients the disease can be more severe and even life-threatening [4].

Worldwide, the most common causative agent of dermatophytosis is the anthropophilic fungus Trichophyton rubrum [5]. This organism and other dermatophytes are able to invade keratinised tissues where they obtain nutrients by degrading the keratin, a main component of hair, the stratum corneum of the skin and the nail plate. The disease is normally transmitted by contact with infected dead skin, nail or hair particles containing hyphae or spores called arthroconidia [6]. In vitro, Trichophyton spp. often form microconidia, which are small pyriform spores that form along the sides of hyphae. These spores are not observed in vivo, but both arthro- and microconidia from related Trichophyton spp. (T. interdigitale and T. mentagrophytes) are infective and show similar levels of adherence to corneocytes [7, 8].

To study the infection process of skin, a number of infection models have been developed. These include the use of isolated corneocytes [7, 8], explanted human skin [9,10,11,12], cell cultures using Chinese hamster ovary (CHO) cells [13] or immortal human keratinocytes (HaCaT cells) [14], and reconstituted human epidermis (RHE) [15]. With some of these models, such as those based on cell cultures, it is unclear how relevant they are to the initial stages of infection as there is no skin structure and a lack of the stratum corneum. This problem is partially solved by using RHE grown on polycarbonate culture inserts, but RHE still lacks an integrated dermis and has a significantly higher permeability compared to human skin [16, 17]. Explanted human skin does not have these issues, but sourcing human tissue can be challenging and costly and is complicated by ethical issues. Our aim in this study was to evaluate the feasibility of using an explanted porcine skin model, which is relatively simple and affordable. We have previously used a porcine skin model for bacterial infections [18], and an important advantage of this model is that it is ethically neutral as the tissue can be obtained from the abattoir as surplus material from pigs that go into the food chain. Also, porcine skin is a well-established model for transdermal and topical delivery of drugs and has proved to be very useful in preclinical stages of research as it is very similar to human skin with similar thickness and hair density [19]. Here, we show that porcine skin is useful to study the early stages of infection and can be used to evaluate the treatment with antifungals.

Materials and Methods

Cultures and Growth Conditions

Trichophyton rubrum ATCC 28188, obtained from Fisher Scientific (Loughborough, UK) was routinely grown on potato dextrose agar (Sigma-Aldrich, St. Louis, MO, USA) for 15 days at 30 °C to induce full sporulation. Microconidia were harvested with sterile water containing 1% Tween-20 (Fisher Scientific), washed and resuspended in sterile water and stored at − 20 °C until use. The freezing process does not lead to a significant reduction in viability of conidia but longer storage does [20], and to ensure consistent viability of inocula, aliquots were used within 2 weeks of freezing.

Extracellular Protease Activity

Keratin was isolated from human hair (obtained from a local hairdresser) as described [21] and after dialysis against water was adjusted to a concentration of 20 mg/mL keratin. This was diluted tenfold in potato dextrose broth (PDB) or minimal salts medium (50 mM NH4H2PO4, 5.75 mM K2HPO4, 3.4 mM KH2PO4, 2 mM CaCl2 and 0.25% glucose; pH 5.2), which was then inoculated with T. rubrum conidia to a final concentration of 5 × 102 conidia/mL. Cultures (1.5 mL each) were then grown for 10 days in 24-well plates at 30 °C. Mycelium was removed by centrifugation, and the supernatant retained for protease activity assays using azocasein as described [22].

Porcine Skin Sterilisation

Porcine skin was sourced from a local abattoir. The skin was obtained from 3-month-old gilts (female, approximately 60 kg) of the Large White breed. Skin from two animals was used; results on skin obtained from one animal were similar to those from the second animal. Full thickness skin was obtained, and the skin was then stored at 4 °C. The next day, the skin was dermatomed to a thickness of 750 µm and stored frozen until use. To avoid damage to the skin, it was not submitted to any cleansing process at the abattoir such as scalding with hot water to remove hair, neither before nor following harvesting it from the animal, and was simply rinsed with water at the laboratory before being dermatomed. On the day of the experiment, dorsal skin samples were defrosted at room temperature and cut in approximately 1 cm2 square pieces. The hair on the skin was removed with scissors, and the skin was washed three times in sterile phosphate-buffered saline (PBS). Next, the porcine skin was sterilised with chlorine gas as described [23] with minor modifications. Briefly, the skin pieces were placed in a Petri dish, which was then put in a chamber with chlorine gas that was generated by mixing 20 mL glacial acetic acid (VWR, Fontenay-sous-Bois, France) and 10 mL sodium hypochlorite solution (10–15%, Sigma-Aldrich) in a beaker. The skin was sterilised with chlorine gas at room temperature for 15 min on an orbital platform shaker at 15 rpm, after which the skin was turned around and incubated for a further 15 min in the chlorine gas. After the treatment, the skin was rinsed twice with sterile PBS, followed by washing by being placed in a tube with 10 mL sterile PBS for 30 min on a shaker at 200 rpm. The skin pieces were then transferred onto minimal salts agar plates (50 mM NH4H2PO4, 5.75 mM K2HPO4, 3.4 mM KH2PO4, 2 mM CaCl2 and 0.25% glucose; pH 5.2) for 30 min before experiments were continued.

Adherence to Porcine Skin

Time-dependent adherence of T. rubrum to the stratum corneum (SC) of porcine skin was determined by measuring the number of conidia that did or did not adhere. To this purpose, 10 µL of T. rubrum conidia (1 × 104 colony-forming units (CFU)/mL) was pipetted on the top of sterilised porcine skin that was placed on minimal salts agar, followed by incubation for 0–24 h at 30 °C. After incubation, the skin pieces were washed with 1 mL sterile PBS and the number of conidia that did not adhere to the skin was measured by a viable plate count of the wash on Sabouraud dextrose agar (SDA). To measure the number of conidia that adhered to the skin, washed skin pieces were incubated with Trypsin–EDTA (0.25%) (Gibco, Paisley, UK) for 15 min at 37 °C on shaker at 220 rpm, followed by viable plate counting.

Adherence to HaCaT Cells

HaCaT cells [24] were plated at a density of 1 × 105 cells per well in 24-well tissue culture plates, cultured in 1 mL Dulbecco’s Modified Eagle Medium (DMEM; Sigma-Aldrich)) supplemented with 10% foetal bovine serum and penicillin–streptomycin (10,000 U/mL). The HaCaT cells were incubated at 37 °C and 5% CO2 for 3 days to reach confluence. The medium was replaced by DMEM, without foetal bovine serum and antibiotics, the night before inoculation. 5 µL of T. rubrum conidia (1 × 105 CFU/mL) was added into each well and incubated for 0–24 h at 37 °C and 5% CO2.

To measure the number of conidia not adhered to the HaCaT cells, the supernatant was collected and the CFU of T. rubrum were determined on SDA. To measure the number of conidia adhered to the HaCaT cells, the DMEM solution in the well was removed and the wells were washed three times in PBS. 0.5 mL of Trypsin–EDTA (0.25%) was added followed by incubation for 5 to 10 min at 37 °C. Once the cells detached, 0.5 mL of DMEM solution was added to the well and cells were resuspended. The solution was transferred to SDA plates for a viable cell count.

Histology Staining

To study the effect of chlorine gas to the structure of the skin after sterilising, specimens were fixed overnight at 4 °C in 10% neutral-buffered formalin. To cryoprotect the tissues, specimens were immersed in 30% sucrose (Sigma-Aldrich) in PBS until the tissue sank, followed by freezing the sample in optimal cutting temperature compound (OCT; CellPath, Powys, UK) using dry ice methanol slurry. Samples were then cryosectioned and stained using haematoxylin (Gill No. 3; Sigma-Aldrich) and eosin (BDH, Poole, UK), then mounted with DPX Mountant (Sigma-Aldrich) and imaged using light microscopy (Zeiss Axioscope).

To visualise the infection process, sterilised porcine skin was placed on minimal salts agar and inoculated with 10 µL of T. rubrum conidia (1 × 105 CFU/mL), followed by incubation for 0–72 h at 30 °C. The skin was fixed with 10% neutral-buffered formalin overnight at 4 °C; then, the fixed specimens were dehydrated, embedded in paraffin and sectioned. The sections were deparaffinised and then stained using a periodic acid–Schiff–diastase (PAS-D) staining kit (Sigma-Aldrich) following the manufacturer’s instructions, and the samples were observed using light microscopy.

Field-Emission Scanning Electron Microscopy

All specimens were treated with the same procedures as outlined in histopathology section. After the treatment, the specimens were fixed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.3) overnight, followed by postfixing with 1% osmium tetroxide for 2 h. This was followed by rinsing three times in sodium cacodylate buffer (pH 7.3) for 10 min. The specimens were dehydrated in a graded acetone series (10, 30, 50, 70, 80, 90, 95 and 100%) for 15 min per change. After that, the specimens were mounted onto a SEM plug and coated with chromium. The samples were then imaged using FESEM (JSM 6301F, JEOL, Tokyo, Japan).

Antifungal Susceptibility Testing

Explanted porcine skin was sterilised and infected with conidia as above, with the porcine skin placed on mineral salts agar. This was incubated for 3 days at 30 °C to induce fungal skin infection. The infected skin was then treated topically using clinically used antifungal creams, being 1% clotrimazole (Bayer, Reading, UK) or 1% terbinafine hydrochloride (GlaxoSmithKline, Brentford, UK). Both were applied as directed in the patient information leaflets. In brief, the creams were applied thinly on the skin (using a sterile swab) once per day (terbinafine) or twice per day (clotrimazole), for a period of 7 days. The explanted skin was kept at 30 °C for the duration of the experiment and was rinsed once per day with sterile PBS to simulate washing. Control specimens were treated the same but were wiped with a sterile swab dipped in sterile water instead of with antifungal cream. After 1 week, all specimens were washed with sterile PBS and gently dried with sterile tissue paper. The specimens were then transferred to fresh mineral salts agar and incubated for a further 7 days at 30 °C.

Statistical Analysis

The results were represented as the mean ± standard deviation (SD). The mean differences between two groups of data were analysed by a repeated measures analysis of the variance (ANOVA). The data were considered statistically significant with a p value of less than 0.05.

Results

Keratinases are important virulence factors for dermatophytes, as these degrade the keratin in skin, nail or hair that these organisms thrive on [25]. We firstly sought to identify conditions in which these enzymes are secreted by T. rubrum. In a nutritionally rich medium (PDB), production of extracellular protease activity was very low, also when it was supplemented with keratin. In contrast, significantly more proteolytic activity was found in culture supernatant when T. rubrum was grown in minimal salts medium containing keratin (Supplementary Figure 1). There was no fungal growth in the absence of keratin (not shown), demonstrating that in these conditions the growth of T. rubrum is dependent on the keratin. This medium was therefore used in the porcine skin infection model.

When testing fungal growth on porcine skin, it became clear that it was important to sterilise the skin first, as otherwise there was a lot of bacterial growth. Initially, we tested surface disinfection with 70% ethanol, but that did not stop bacterial growth (not shown). Also, ethanol can be damaging to skin as it selectively extracts lipids from the stratum corneum [26, 27]. However, sterilisation with chlorine gas did not damage skin as described before [23] and observed by histology staining (Supplementary Figure 2), while this treatment was sufficient to prevent bacterial growth on the porcine skin.

A schematic of the infection model, in which dermatomed porcine skin, on a polycarbonate membrane, is placed on top of minimal salts agar, is shown in Fig. 1a. When conidia were inoculated on the porcine skin, growth of T. rubrum was observed after several days of growth at 30 °C (Fig. 1b). Similar to the growth in liquid medium, this was dependent on the presence of skin, as the fungi were unable to grow on minimal salts agar without porcine skin (Fig. 1c).

Fig. 1
figure 1

Explanted porcine skin infection model. a Schematic of the set-up used for the infection model. b Fungal growth on explanted porcine skin, 7 days after inoculation with conidia. No growth is observed in the absence of porcine skin (c)

Full size image

The adherence of conidia observed in this explanted porcine skin was compared to a previously established in vitro model using HaCaT cells, an immortal line of keratinocytes [24]. Using these two approaches, we measured the number of conidia that adhered to skin (Fig. 2a). With porcine skin, maximum adherence was reached within 2 h, while that took ~ 3 h with HaCaT cells. A repeated measures ANOVA on those first 3 h showed that these differences were not significant. The adherence of conidia was mirrored by the number of non-adherent conidia (Fig. 2b). However, the number of non-adherent conidia continued to drop to less than 5%, while the number of adherent conidia did not increase beyond 60% of the number of conidia initially applied.

Fig. 2
figure 2

Measurement of T. rubrum conidia bound (a) and unbound (b) to porcine skin and HaCaT cells between 0 and 24 h. There is a reverse relationship between bound and unbound conidia in both models, as they showed an increase in conidial adherence with a decrease in non-adherent conidia (data derived from 2 to 3 independent experiments, with 3–6 repeats per experiment)

Full size image

To visualise the infection process, we performed histology analysis of infected porcine skin. These are shown in Fig. 3, with examples of the features described below indicated with arrows. Also, the position of the stratum corneum, epidermis and dermis is indicated in Fig. 3a. After 3-h incubation (Fig. 3a), conidia can be observed above the stratum corneum, while after 6 h (Fig. 3b) some conidia appear to be embedded in the stratum corneum. Development of germ tubes, the first step in the formation of a mycelium, can be observed after 9 h (Fig. 3c). Upon longer incubation of the infected skin samples, mycelium can be observed after 24 h, and hyphae can be seen in the stratum corneum (Fig. 3d), while after 48 h the stratum corneum was damaged and sometimes absent (Fig. 3e). After 72 h (Fig. 3e), the epidermis also appeared damaged and degraded, and there was clearly invasion of the dermis by T. rubrum.

Fig. 3
figure 3

Histopathological analysis of infected porcine skin. Explanted porcine skin was infected with T. rubrum conidia, incubated for 3 h (a), 6 h (b), 9 h (c), 24 h (d), 48 h (e) and 72 h (f). After incubation, samples were prepared for PAS-D staining and visualised using light microscopy. Features described in the main text are indicated with arrows. In a, the position of the stratum corneum (SC), epidermis (e) and dermis (d) is indicated. Scale bar = 20 µm

Full size image

To get a more detailed view at the infection process, samples were prepared for SEM imaging. In these images, microconidia were easily recognised with their characteristic piriform shape. Sometimes after 3 h (Fig. 4a), but more frequently after 6 h (Fig. 4b), fibril-like material that appeared to connect the conidia was observed. Longer incubation, at 9–12 h (Fig. 4c, d), showed conidia also connecting to the skin. After 24 h, conidia could be observed that appeared completely embedded in the skin with a fibrous film (Fig. 4e), while hyphae were also present (Fig. 4f, g). After 48 h (Fig. 4h), the skin was mostly covered in mycelium. It should be noted that some of the images show that the conidial preparation does contain some hyphal fragments as shown in Fig. 4a and b, but in all experiments the percentage of microconidia was at least 90%.

Fig. 4
figure 4

Electron microscopy analysis of infected porcine skin. Explanted porcine skin was infected with T. rubrum conidia, incubated for 3 h (a), 6 h (b), 9 h (c), 12 h (d), 24 h (e–g) and 48 h (h), followed by visualisation using FESEM. Different magnifications were used depending on the features that were visualised. Scale bars shown are 2 µm, except for b (5 µm), f (10 µm) and g (5 µm)

Full size image

The porcine skin model was further tested using antifungal creams that are commonly used for the topical treatment of tinea infections. These were 1% terbinafine hydrochloride and 1% clotrimazole, which were both used as directed in the patient information leaflet on infected porcine skin. The skin was rinsed once per day—to mimic washing—and kept at 30 °C for the duration of the experiment, a temperature that closely matches that of skin in feet and hands [28, 29]. After treatment for a week, the explanted skin was washed, dried and incubated for a further 7 days at 30 °C. As shown in Fig. 5, the control (without antifungal) was fully covered with mycelium, whereas no T. rubrum was visible on the skin that was treated with the antifungal creams, showing successful treatment of the infected skin with the topical products.

Fig. 5
figure 5

Topical treatment of infected porcine skin. Explanted porcine skin was infected with T. rubrum conidia and incubated for 3 days at 30 °C. After this, the skin, which was kept at 30 °C on minimal salts agar, was treated daily for a period of 7 days with water (control, a), 1% clotrimazole (b), or 1% terbinafine hydrochloride (c). Treatment was as directed in the patient information leaflet, and the skin was also rinsed daily to simulate washing. After treatment, the skin was washed and incubated for a further 7 days at 30 °C

Full size image

Discussion

As we have shown here, explanted porcine skin is a useful model system to monitor the infection process by dermatophytes such as T. rubrum. Porcine skin was chosen, as its structure is very similar to that of human skin, including a similar thickness of the stratum corneum, epidermis and dermis, and a comparable hair density [19, 30]. A difference is a thicker layer of subcutaneous layer of fat in pigs and possibly differences in the immune system [30], but here only dermatomed skin was used. Because of these similarities, pig skin is considered a suitable model for, for example, wound healing, toxicology and dermal delivery of drugs, also since the flux of drugs through porcine skin is within the same order of magnitude as that for human skin [17]. In contrast, skin from, for example, rodents is very different, with significantly higher hair density, a much thinner epidermis and dermis, and higher permeability of drugs [17, 30]. Domestic pigs do not suffer often from dermatophytosis, but they are sometimes infected with the zoonotic species Trichophyton mentagrophytes [31]. Furthermore, while the related species T. rubrum is considered anthropophilic, it occasionally causes infections in animals such as dogs [32], while the species has also been used in rodent models of infection [33]. Thus, even though there are no reported cases of pigs being infected with T. rubrum, the above does suggest that porcine skin is a suitable model to study infections with this species.

In dermatophytes, extracellular proteases have important roles in nutrient acquisition, invasion [25] and, for some specific proteases, adherence [11]. While we have not tested which specific proteases are expressed, we have shown that proteolytic activity can be detected in vitro in minimal salts medium containing keratin, while this activity is absent in a rich medium containing keratin. Similarly, we showed that fungal growth on plates with minimal salts agar only occurs when porcine skin is present. This minimal salts medium does contain a carbon source (glucose), but no nitrogen source, and fungal growth on porcine skin is therefore likely to be dependent on the release of nitrogen-containing compounds that are obtained by proteolytic degradation of the keratin in skin. Previous studies have also shown that these proteases are expressed when nutrients are limiting and proteins such as keratin are present [34].

Maximum adherence to the porcine skin was achieved within 2 h, which compared well to a previously published in vitro model using keratinocytes [14], for which we found maximum adherence within 3 h. We did not make a comparison with other models as it was not our aim to comprehensively contrast and compare these with the porcine skin model. Moreover, the keratinocyte model using HaCaT cells is a very simple epidermal model, but it was nevertheless useful to make this comparison as, firstly, this model has been used previously as a model for T. rubrum infection [14] and, secondly, HaCaT cells are very frequently used in skin research.

Adherence has also been measured using corneocytes, which are terminally differentiated keratinocytes that form most of the stratum corneum. Zurita and Hay [8] showed that maximum adherence of Trichophyton interdigitale to corneocytes occurred within 3–4 h. Similarly, Aljabre et al. [7] showed time-dependent increase in adherence of conidia to corneocytes of both T. interdigitale and T. mentagrophytes over a 6-hour period. It should, however, be noted that in the studies with keratinocytes or isolated corneocytes, adherence was measured with cells immersed in liquid, rather than on intact skin exposed to air. Furthermore, those models lack a stratum corneum and other skin structures, although it is unclear whether that would affect the adherence.

It was noted that the number of conidia that adhered did not rise above ~ 50–60% of the number of conidia applied with both models, while the number of non-adherent conidia continued to decrease over at least a 6-h period. This suggests that either the conidia lose their viability, or that the conidia adhere so strongly that they can no longer be removed. The latter may indeed be the case, as SEM revealed formation of fibril-like material between adjacent conidia within 6 h and between conidia and the stratum corneum in 9 h. After 24 h, some conidia appeared completely embedded in the stratum corneum, and it is conceivable that this leads to a strong interaction, making them difficult to remove. Similar fibrils have also been observed with T. mentagrophytes microconidia [10] and arthroconidia [12] when grown on explanted human skin, and both of these studies found indeed a strong adherence of the conidia to the skin. Our results for T. rubrum are consistent with these previous findings. The fibril-like material could consist of the same material as the extracellular matrix (ECM) of T. rubrum biofilms, which was shown to be rich in polysaccharides [35], but at present the nature of the fibril-like material is unknown and their role in attachment is purely speculative.

Adherence of conidia is followed by invasion of the skin, in which formation of germ tubes plays an important role. We observed germ tubes within 9 h, which is similar to what has been observed in vitro [36]. Penetration of hyphae into the skin was observed with histology staining and light microscopy after 24 h. This penetration was not observed with SEM, but this technique will only visualise the surface which will make it more difficult to identify hyphae that penetrate into the skin. Development of mycelium occurred after 24 h, while longer incubation (48–72 h) resulted in the skin being completely covered by a fungal mat, damage of the epidermis and invasion of the dermis. Such luxuriant growth and invasion are not observed in vivo, but deeper tissues can be affected in immunocompromised patients [37,38,39]. Indeed, the immune system has an important role in restricting the growth of dermatophytes to dead keratinised tissue only [40]. However, the skin used in this study is not kept alive after harvesting and is simply stored dry at 4 °C, dermatomed the next day and then stored at − 20 °C. Although we have not tested this, it is unlikely to contain an active immune system, thus explaining the invasion of the epidermis and dermis. We also do not observe lesions as observed in vivo or with a human skin infection model in which the skin is kept alive [41], which again may be a consequence of using porcine skin that is not kept viable.

In summary, we have tested the feasibility of a porcine skin infection model to study dermatophytosis and treatment strategies. The successful treatment with creams commonly used for fungal skin infections demonstrates that the model is also suitable to test novel strategies for prevention and treatment of fungal skin infections. The infection model is relatively simple, cheap and ethically neutral as the porcine skin is obtained from slaughterhouse material. When compared to other models such as cultured keratinocytes, the model also has the advantage that it contains a complete skin structure comprising stratum corneum, epidermis and dermis. Reconstituted human epidermis does contain a stratum corneum but lacks a dermis and is more permeable to drugs as compared to intact skin [16]. This might suggest that also fungal invasion would be influenced, but there is no evidence for this at present. Furthermore, we have only made a comparison with the keratinocyte model and have not determined which model is a better reflection of a fungal skin infection. While explanted human skin has been used in other studies [9,10,11,12, 41], sourcing human skin can be costly and complicated by ethical issues. It is important to note that porcine skin is very similar to human skin, with a similar thickness and hair density, and even when stored frozen the barrier function of the stratum corneum is well preserved [17, 19]. Importantly, it behaves similarly to human skin with respect to penetration of drugs, and porcine skin is therefore an important model to study dermal delivery or drugs. Indeed, as shown here the porcine model can also be used to analyse topical treatment with antifungals, albeit that in this case the data are only qualitative as we did not test different dosages. A limitation of our model is the lack of an immune system, resulting in invasion of the dermis and more luxuriant growth than would be observed in vivo. Therefore, late stages of infection as well as chronic infections cannot be modelled well with this approach. However, we believe the model is very useful to study in particular the early stages of infection, and we are currently planning to identify virulence factors involved. Moreover, the model can be used to test novel regimes to prevent or treat fungal skin infections.