Host responses in an ex-vivo human skin model challenged with Malassezia sympodialis

Malassezia species are a major part of the normal mycobiota and colonise mainly sebum-rich skin regions of the body. This group of fungi cause a variety of infections such as pityriasis versicolour, folliculitis and fungaemia. In particular, Malassezia sympodialis and its allergens have been associated with non-infective inflammatory diseases such as seborrheic dermatitis and atopic eczema. The aim of this study was to investigate the host response to M. sympodialis on oily skin (supplemented with oleic acid) and non-oily skin using an ex-vivo human skin model. Host-pathogen interactions were analysed by SEM, histology, gene expression, immunoassays and dual species proteomics. The skin response to M. sympodialis was characterised by increased expression of the genes encoding β-defensin 3 and RNase7, and by high levels of S100 proteins in tissue. Supplementation of oleic acid onto skin was associated with direct contact of yeasts with keratinocytes and epidermal damage. In oily conditions, skin response to M. sympodialis showed no gene expression of AMPs, but increased expression of IL18. In supernatants from inoculated skin plus oleic acid, TNFα levels were decreased and IL-18 levels were significantly increased.


INTRODUCTION
The genus Malassezia, previously known as Pityrosporum, is a group of lipophilic yeasts. Malassezia species are part of the normal mycobiota and colonise several regions of the body, mainly sebum-rich skin areas such as the scalp and thorax (Marcon & Powell, 1992). To date, 17 Malassezia species have been proposed (Sparber & LeibundGut-Landmann, 2017). Malassezia spp. are the most abundant genus on the skin of individuals with psoriasis and atopic eczema (Zhang et al, 2011;Paulino et al, 2008;Paulino LC et al, 2006) and are highly increased in seborrheic dermatitis and dandruff (Park et al, 2012;DeAngelis et al 2005).
Most Malassezia species are unable to synthesise fatty acids and degrade carbohydrates and are dependent upon the acquisition of exogenous fatty acids.
Malassezia species have a large repertoire of lipolytic enzymes such as lipases, phospholipases and esterases (Sparber & LeibundGut-Landmann, 2017;Saunders et al, 2012). Lipases hydrolyse sebum triglycerides from the host skin to release fatty acids (oleic acid and arachidonic acid). Lipase activity is significantly higher in M. globosa and M. pachydermatis than in M. sympodialis and M. slooffiae, although normal phospholipase activity is found in these species (Juntachai et al, 2009). The fatty acids oleic acid (OA) and arachidonic acid are released by phospholipase and lipase activity and have an irritating/inflammatory effect on skin (Ashbee & Evans, 2002). OA is also increased in dandruff scalps compared with non-dandruff scalps (Jourdain et al, 2016) and increases the severity of flaking (or stratum corneum desquamation) facilitating better penetration of M. sympodialis so that the fungus directly interacts with cells of the inner skin layers (DeAngelis et al, 2005).
Exogenously acquired fatty acids contribute to the formation of a thick cell wall in Malassezia sp. characterised by a unique lipid-rich outer layer, which contributes to triggering the immune response against this group of fungi (Thomas et al, 2008;Gioti et al, 2013b). Human studies have shown that cytokine levels  in the skin of individuals with seborrheic dermatitis and Malassezia folliculitis were higher than levels in the skin of healthy volunteers (Faergemann et al, 2001). However, this pattern of cytokine induction varied depending on the fungal cell wall structure. Higher levels of IL-8 and lower levels of IL-10 were produced by keratinocytes in vitro when they were stimulated with M. sympodialis lacking the lipid-rich outer layer compared to the same yeasts with the outer layer (Thomas et al, 2008).
Antimicrobial peptides (AMPs) are key in the innate immune response to Malassezia spp. Individuals with pityriasis versicolour have significantly higher AMP levels (βdefensin 2, β -defensin 3, S100A7 and RNase7) in their skin (Brasch et al, 2014). The β -defensins are specifically increased in the stratum corneum, RNase7 in the stratum granulosum, and S100A7 in the stratum corneum, granulosum and spinosum (Brasch et al, 2014). Malassezia sp. also play a role in NLRP3 inflammasome activation when yeast cells are sensed by dendritic cells but not by keratinocytes (Kistowska et al, 2014). Activation of NLRP3 depends on Dectin-1 and leads to high expression of caspase-1-dependent IL-1β in dendritic cells of patients with seborrheic dermatitis (Kistowska et al, 2014). The adaptive immune response against Malassezia is characterised by the production of specific IgG and IgM antibodies in healthy individuals and specific IgE antibodies in atopic eczema (Glatz et al, 2015).
Atopic eczema (AE) is a chronic inflammatory disease affecting up to 20% of children and 3% of adults (Nutten, 2015). Multiple factors have been associated with AE, such as impairment of skin barrier function due to physical (scratching, skin dryness) or chemical (pH changes due to soaps) damage, genetic factors (mutations in FLG and SPINK5 genes encoding filaggrin and serine protease inhibitor Kazal-type 5 protein, respectively), and environmental factors (cold climate, no breastfeeding, pollution) (Nutten, 2015;Sääf et al, 2008). Malassezia sp. has been linked with AE pathogenesis as Malassezia sp. allergens induce specific IgE antibodies and autoreactive T-cells that can cross-react with skin cells (Glatz et al, 2015;Zargari et al, 2001).
The aim of this study was to investigate the host skin response to Malassezia sympodialis. An ex-vivo human skin model was used either directly for non-oily skin or supplemented with oleic acid to represent oily skin. Host-pathogen interactions were analysed by SEM, histology, gene expression, immunoassays and proteomics.
The skin response to M. sympodialis was characterised by increased expression of the genes encoding β -defensin 3, and RNase7 and high levels of S100 proteins.
Supplementation of the skin with oleic acid resulted in epidermal damage, direct contact of yeasts with keratinocytes, and AMP gene expression was not detected.
TNFα levels were decreased in supernatants from inoculated skin plus oleic acid. IL-18 levels were significantly increased in the same supernatants and IL18 gene expression was increased in tissue.

An ex-vivo human skin model
The ex-vivo human skin model was set up as previously described with modifications (Corzo-León et al, 2019). Human skin tissue (no stretch marks), without an adipose layer (adipose layer removed by the surgeon during surgery), from abdominal or breast surgeries was supplied by Tissue Solutions ® Ltd. (Glasgow, UK). Human tissue from four different donors was obtained according to the legal and ethical requirements of the country of collection, with ethical approval and anonymous consent of the donor or nearest relative. Tissue Solutions ® comply with the UK Human Tissue Authority (HTA) on the importation of tissues. Explants were transported at 4 °C with cooling packs and maintained at 4 °C until processed, which occurred within 36 h of surgery.
Skin was washed with DMEM (supplemented with 1% v/v antibiotics (penicillin and streptomycin) and 10% heat inactivated FBS, Thermo-Fisher Scientific, Loughborough, UK) and kept moist in a Petri dish in the same medium. The explant was cut into 1 cm 2 pieces. The surface of each piece was gently wounded using a needle, without penetrating the entire skin thickness, by pricking with a needle several times (6-10) to a depth of approximately 2-3 mm.
After wounding, each piece of skin was placed into an individual well of a 6-well plate. An air-liquid interphase was maintained by adding 1 ml of supplemented DMEM. A well containing only DMEM growth medium was added as a negative control and served as a contamination control. The medium was changed every 24 h and the spent medium was stored in 2 ml tubes for subsequent analysis at the same time points as the rest of the samples. Recovered medium was stored at -80 o C until analysed. In a previous study (Corzo-León et al, 2019) the skins samples were demonstrated to remain viable for 14 days under these conditions using the TUNEL system. Skin explants were inoculated by applying 10 µl of fungal suspension (1x10 6 yeasts) directly on to the epidermis. Yeast suspension was prepared as described below and resuspended in PBS. Uninfected/non-inoculated skin controls were included in all experiments. Two additional experimental conditions were added: 1) uninfected skin with 10 µl 100% oleic acid (Thermo-Fisher Scientific) applied to the surface of the skin explant and 2) infected skin with same M. sympodialis inoculum, followed by application of 10 µl 100% oleic acid on to the surface of the skin explant.
Skin samples were incubated for six days at 37 °C and 5% CO 2 before being recovered in a Petri dish. Prior to processing, the macroscopic appearance of explants was evaluated by eye and images captured with a Stemi 2000-c Stereo Microscope (Carl Zeiss, Oberkochen, Germany). Samples were then processed depending for further analyses.
Tissue samples for histology were placed into moulds, embedded in OCT compound (Cellpath Ltd. Newtown, UK) and flash-frozen with dry ice and isopentane. These samples were stored at -20 o C for immediate analysis or at -80 o C for longer term storage. For scanning electron microscopy (SEM), tissue samples were fixed in glutaraldehyde buffer (2.5% glutaraldehyde in 0.1 M cacodylate) overnight at 4 o C and sent to the Microscopy and Histology technology hub, University of Aberdeen, for further sample preparation. Tissue samples for RNA extraction were cut into smaller pieces and placed in a microcentrifuge tube with RNAlater ® (Sigma, Dorset UK) for subsequent RNA extraction. These samples were stored at -20 o C for immediate analysis or at -80 o C for longer term storage. Experiments were replicated at least three times using skin from different human donors.

Fungal strains and culture conditions
Modified Dixon (mDixon) broth and agar (3.6% w/v malt extract, 2% w/v desiccated ox-bile, 0.6% w/v bacto tryptone, 0.2% v/v oleic acid, 1% v/v Tween 40, 0.2% v/v glycerol, 2% w/v bacto agar for agar plates) was used to grow Malassezia sympodialis (ATCC 42132) for skin experiments. Yeast cells were grown on mDixon agar at 35 °C for 4-5 days in a static incubator, one colony was selected from an agar plate and inoculated into 10 ml of mDixon broth for 4 days at 37 °C in a shaking incubator at 200 rpm. Yeast cells were recovered and a final inoculum of 1x10 6 yeasts in 10 µl of PBS (1x10 8 /ml) was prepared and applied to the skin surface.

Scanning Electron Microscopy (SEM) and Histopathology
Several microscopic analyses were performed on recovered skin tissue for histological confirmation of fungal infection. Sections (0.6 µm) were cut from the frozen OCT blocks for histological analysis and stained with fluorescent dyes (1 µg/ml calcofluor white (CFW), and propidium iodide (1 µg/ml). Images were captured with a Deltavision™ confocal microscope (GE Healthcare, Buckinghamshire UK). SEM samples were observed using a Zeiss™ EVO MA10 Scanning Electron Microscope. Images were generated by detection of secondary electrons (SE1) and backscatter electron detection (NTS BSD) and captured at 10 kV resolution and at different magnifications.

Gene expression and proteomics analyses
Tissue samples for RNA extraction were thawed and the RNAlater ® discarded before processing the samples. Samples were washed twice with PBS. RNA and proteins were extracted from the same recovered samples in a single two-day sequential process based on previously published methods (Chomczynski & Sacchi, 1987;Berglund et al, 2007;Corzo-León et al, 2019).
The RNA yield and purity were evaluated by Nanodrop™ spectrophotometry (Thermo-Fisher Scientific) and samples were stored at -80 o C until further use. All samples had an initial yield between 200 and 800 ng/µl and a 260/280 ratio between 1.8 and 2.0. To produce cDNA, RNA samples (1 µg) were treated with DNase I (1 U/µl per 1 µg of RNA sample) (Thermo-Fisher Scientific), then reverse transcription carried out using the SuperScript™ IV first-strand synthesis system (Thermo-Fisher Scientific) with Oligo dT primer, following the manufacturer's instructions.
Intron spanning primers and qRT-PCR assays were designed using Roche's Universal Probe Library Assay Design Centre (lifescience.roche.com/en_gb/brands/universal-probe-library.html#assay-designcenter) for different target genes known to be expressed in skin. Genes, accession numbers, Roche probes paired with target primers and primer sequences are shown in Table 1. The Roche probes were hydrolysis probes labelled at the 5 end with fluorescein (FAM) and at the 3 end with a dark quencher dye. A reference gene (B2M encoding β 2-microglobulin) primer pair and probe was also designed (Lossos et al, 2003) using the Eurogentec web tool (secure.eurogentec.com/life-science.html) ( Table 1). The probe was modified at the 5 end with Cy5 and at the 3 end with quencher QXL 670.
qRT-PCR reactions (10 µl) were set up in Light cycler 480 plates using the LightCycler 480 probe master mix (Roche, Welwyn Garden City UK) according to the manufacturer's instructions (Tables 2). Dual hydrolysis probe assays were analysed in the same well, FAM probe was used for target gene primer pairs and Cy5 probe for the reference gene. For each cDNA, assays were performed in triplicate.
Reactions were run in a LightCycler 480 (Roche). Following manufacturer's recommendations, reaction settings were as follows: one cycle at 95 o C for 10 min Facility (www.abdn.ac.uk/ims/facilities/proteomics/)). Four biological replicates were analysed per condition (infected and uninfected skin). Only proteins having two or more identified peptides and two or more Peptide Spectrum Matches (PSM) were selected. Finally, proteins found in at least two out of four analysed samples per condition were included for further Gene Ontology (GO) analysis using the GO consortium online tool (geneontology.org). Area under the curve (AUC) values for each protein were averaged and compared between conditions, inoculated and noninoculated skin, non-inoculated skin with and without oleic acid and, finally analysed using the Student's t-test or Mann-Whitney test depending on the distribution of the data, with a value of p<0.05 considered statistically significant (Prism 8 software).
Proteins identified as significant were compared to the CRAPome database (www.crapome.org/). The CRAPome web tool is a Contaminant Repository for Affinity Purification and contains lists of proteins identified in negative control samples, collected using affinity purification followed by mass spectrometry (AP-MS).
Proteins found in the CRAPome database in >10% of the cases were not considered significant as they are probably the result of carryover contamination during mass spectrometry experiments.

Immunoassays
Recovered supernatants from inoculated skin and non-inoculated skin controls were analysed for different cytokines (TGFβ1, TNFα, IL-1β, IL-6, IL-8, IL-18, IFNγ and, TNFRI) at 2-3 days of incubation. Four biological experiments were analysed in duplicate. Multiplex immunoassays were carried out following the manufacturer's instructions (Milliplex ® Map kits. EMD Millipore Corporation, Livingston UK). Data were analysed using either one-way ANOVA or Kruskal-Wallis test depending on the homogeneity of variance (tested by Bartlett's test). A p value <0.05 was considered statistically significant (Bonferroni correction) and post-hoc analysis done by Dunnett's or Dunn's test (Graphpad Prism 8).

M. sympodialis invaded ex vivo skin supplemented with oleic acid and interacted directly with the inner epidermal layer
Host-pathogen interactions between human skin and M. sympodialis were Intact epidermis was also seen in the uninfected skin control without oleic acid. In uninfected skin receiving OA, thinner epidermis was observed but no detachment of epidermal layers was observed (Figure 3).
Local skin response to M. sympodialis is characterised by higher expression of genes encoding β -defensin 3, ribonuclease 7 and higher levels of S100 proteins RNA was extracted from skin tissue samples, at day 6 post-inoculation under the four experimental conditions described above, for gene expression analysis. Due to the importance of AMPs and cytokines in the innate immune response, the expression of eight human genes encoding different AMPs and cytokines was analysed by qRT-PCR. AMP genes included S100A7 (psoriasin), S100A8, S100A9, DEFB4A (βdefensin 2), DEFB103A (β-defensin 3), RNASE7 (ribonuclease 7, RNase7), IL18, and TGFB1 (transforming growth factor β 1). Expression of AMP genes was found only in uninfected skin and MS skin (inoculated skin without OA). No expression of AMP genes was found in samples receiving OA. We can confirm that reference genes were amplified and expressed in samples receiving OA verifying that the lack of expression of AMP genes was not due to a technical problem. In MS skin, expression of DEFB103A (p≤0.03) and RNASE7 (p≤0.03) was increased 193 (IQR 139 to 843) and 7 (IQR 2 to 343) times, respectively, compared to non-inoculated, negative control skin (Figure 4). Meanwhile, S100A9 was significantly (p≤0.03) decreased -10.2 (IQR -15 to -4) fold.
The four different experimental conditions (described above) were also analysed by

AMP levels were analysed separately and compared to non-inoculated control by
Kruskal-Wallis test and Dunn's test, as well as using the CRAPome online tool.

Secreted cytokine responses of human skin tissue exposed to M. sympodialis is characterised by high levels of TGFB1 and IL-18
After two days of incubation supernatants were recovered from the four experimental conditions and analysed for seven different cytokines (IL1-β, IL-6, IL-8, TNFR1, TNFα, TGF1β, IL-18) and compared to non-inoculated skin as the negative control.

DISCUSSION
Oleic acid conditions were added to this model to evaluate the host response to M. sympodialis in oily skin conditions. Previous studies found fatty acids (oleic acid, palmitic acid and lauric acid) to be protective, resulting in the upregulation of β defensin 2 expression by sebocytes in cell culture (Nakatsuji et al, 2010). This is contrary to what was observed with this model in non-inoculated and MSOA skin and may be due to the high concentration of OA used for the experiments in the current report. Skin damage by OA has been documented previously with concentrations as low as 5% and is macroscopically evident as dermatitis in healthy volunteers and histological damage in reconstructed epidermis (Boelsma et al, 1996). Therefore the lack of AMP expression in OA or MSOA skin may be due to epidermal damage as a result of OA supplementation and would require further validation. The direct effect of OA on skin was analysed here by proteomics and the majority of proteins (including keratins involved in cornification, keratinocyte differentiation and wound healing responses) were detected at lower levels in OA skin compared to untreated skin.
Only a handful of proteins had increased levels in OA skin but the differences in fold change were not statistically significant. It is known that OA can 1) damage and produce desquamation of the stratum corneum (DeAngelis et al, 2005), which was also observed in histological sections in our model and 2) histological damage facilitates penetration of M. sympodialis so that it contacts and damages keratinocytes in deeper skin layers. These two consequences could result in the absence of AMPs produced by the epidermis in MSOA skin.
The lack of AMP protein expression in MSOA skin in this model is similar to what has been reported in skin lesions of AE individuals, where previous studies have documented decreased or no expression of β -defensin 2 and LL37 in acute and chronic skin lesions of AE individuals (Ong et al, 2002;Clausen et al, 2018). Another report also showed both AMP gene and protein expression in the epidermis was lower in atopic dermatitis compared to psoriasis patients (de Jongh et al, 2005).
High levels of certain S100 proteins were found in MS skin (S100A2, S100A4 and S100B). These have not been reported to have a role in Malassezia infections or allergic reactions. However, these three S100 proteins have been reported to have a role in macrophage migration, cell proliferation and migration, and as apoptosis inhibitors and regulators of p53 protein (Donato et al, 2013). The increase of these S100 proteins should be confirmed with different techniques, such as gene expression, immunoassays or immunofluorescence as LC-MS/MS could misidentify these peptides with other similar proteins sharing S100 domains such as filaggrin (Bunick et al, 2015). This differentiation will be important as filaggrin is essential for epidermal barrier formation (Hänel et al, 2013) and the loss of its function is already recognised as a causative factor for AE (Nutten, 2015). Further study of the role of filaggrin in the response to M. sympodialis is required. In addition, the lower expression of S100A9 gene found in skin inoculated with M. sympodialis has not been reported before. The expression of S100A9 gene was expected to be increased or unchanged as seen with the rest of S100 proteins. In order to investigate whether decreased expression of S100A9 is a feature of M. sympodialis skin response, a follow up of the dynamics of S100A9 gene and protein expression in earlier and later stages of M. sympodialis infection would be required.
The effect of Malassezia species on cytokine production can vary depending on the clinical and experimental context (Watanabe et al, 2001;Pedrosa et al, 2019;Faergemann et al, 2001). High levels of IL-1β, IL-6, IL-8 and TNFα have been found in supernatants of human keratinocyte cell cultures after 3 to 6 h of co-incubation (Watanabe et al, 2001). However, the cytokine response to Malassezia sp. varies depending on the species; M. pachydermitis induced the highest levels of all these cytokines and M. furfur induced almost no response (Watanabe et al, 2001). In our study, none of these cytokines had high levels when MS skin was compared to noninoculated controls and TNFα levels were significantly decreased in MSOA skin. This finding can be explained in three ways. Firstly, the time point for cytokine analysis differs. In a recent study using reconstructed epidermis, gene expression of IL1B, IGFB1 and TNFA in tissue was increased after 6 h incubation with M. sympodialis, but these same genes were downregulated after 48 h of co-incubation (Pedrosa et al, 2019). Secondly, models to study host response to fungal infection can differ from the response in real human infections. Faergemann et al. (2001) reported no difference in skin TNFα levels in individuals with Malassezia folliculitis and seborrheic dermatitis when compared to healthy volunteers, similar to the current ex-vivo skin model and contrary to what has been reported from monolayer keratinocyte culture.
In addition, lower TNFα serum levels were found in individuals with chronic AE, along with low serum levels of IL-10, β -defensin 3 and high levels of β -defensin 2 (Kanda & Watanabe, 2012). Finally, the lack of AMP expression seen in the MSOA skin can explain the low levels of some cytokines as these AMPs (especially, β -defensins, S100 proteins and cathelicidin) are key for inducing cytokine responses (Niyonsaba et al, 2017).
IL18 and TGFB1 were highly expressed in MSOA skin, whilst only IL-18 levels were significantly higher in the analysed supernatants. IL-18 belongs to the IL-1 cytokine family, along with IL-1α and IL-1β, and is cleaved by caspase-1 after being activated by 3NLRP inflammasome (Fenini et al, 2017). This inflammasome pathway is crucial for inducing Th1/Th2 responses. IL-18 induces the formation of high serum levels of IgM, IgG 1 , IgG 2a and very high levels of IgE antibodies (Enoksson et al, 2011). The production of these antibodies depends on CD4 + T cell-derived IL-4 and the autoreactivity of these antibodies is regulated and depends on NK T cells (Enoksson et al, 2011). The production of high levels of IL-18, along with high levels of IL-4 (Th2 biased response), has been associated with worse prognosis in other infections such as leishmaniasis (Gurung et al, 2015). High serum levels of IL-18, IL12/p40 and IgE antibodies have been found in individuals with atopic eczema and their serum levels correlate proportionally with clinical severity of AE skin lesions (Zedan et al, 2015).
As mentioned previously, M. sympodialis allergens play a crucial role in the pathogenesis of atopic dermatitis (Gioti et al, 2013). In this study, higher numbers of allergens were identified in the MSOA skin compared to MS skin. This finding could be explained by the higher number of yeasts present on the surface of MSOA skin.
Due to the skin damage caused by OA, it is possible that these allergens were in contact with inner epidermal cells and contributed to the host response seen in MSOA skin. The role of allergens in M. sympodialis pathogenicity is a future avenue of investigation with this ex-vivo human skin model.
In conclusion, the local host response to M. sympodialis can be characterised using

ACKNOWLEDGMENTS
Thanks to the Technology hubs at the University of Aberdeen (Microscopy and Histology, qPCR, Proteomics) for their support, sample processing and training.
Special thanks to Professor Annika Scheynius from the Karolinska Institute, Stockholm, Sweden for sharing her expertise and constructive discussions and for giving us the inspiration to work on Malassezia. Probe ** ACATGTCTCGATCCCAC ++ Primers and matching probes were selected using Universal Probe Library Assay Design centre. All Roche probes were hydrolysis probes labelled at the 5 end with fluorescein (FAM) and at the 3 end with a dark quencher dye. * Primers were designed to amplify the β 2-microglobulin gene, which was used as the reference gene ** Modified probe at the 3 end with QXL 670, and at the 5 end with Cy5. For skin samples, dual hydrolysis probe assays were analysed in the same well, with a FAM probe used for target genes primer pairs and a Cy5 probe for the reference gene. M. sympodialis (1 x 10 6 yeasts) was inoculated in b & d, with a & c uninoculated. All samples were incubated at 37 C in 5% CO 2 for 6 days. Scale is indicated by the ruler, with the space between each bar = 1 mm.     . Supernatant cytokine levels from M. sympodialis-inoculated skin at two days incubation. Cytokine levels were measured by immunoassay. Fold change was calculated by comparing levels to negative control, non-inoculated skin. Data is shown as median and IQR, and analysed by Kruskal-Wallis test. If p value <0.05 after Kruskal-Wallis test then post-hoc analysis by Dunn's test was performed and indicated in the graphic as *p≤0.05. n=4 biological replicates, each analysed in triplicate. Figure 8. Comparison of TGFB1 and IL-18 expression in tissue and supernatants. For gene expression relative quantification was performed by qRT-PCR using RNA extracted from samples at 6 days post inoculation. Results obtained for each target gene were normalised against β 2-microglobulin gene expression levels. Protein levels in supernatant at 2 days post inoculation were measured by immunoassay. In both cases, fold change was estimated relative to expression in non-inoculated skin. Data is shown as median and IQR and statistical analysis was done with Kruskal-Wallis test. If p value <0.05 after Kruskal-Wallis test then post-hoc analysis by Dunn's test was performed and indicated in the graphic as *p≤0.05, ** p≤0.001. n=4 biological replicates, each analysed in triplicate.