Staphylococcus aureus from Atopic Dermatitis Patients: Its Genetic Structure and Susceptibility to Phototreatment

ABSTRACT We characterized the population of Staphylococcus aureus from patients with atopic dermatitis (AD) in terms of (i) genetic diversity, (ii) presence and functionality of genes encoding important virulence factors: staphylococcal enterotoxins (sea, seb, sec, sed), toxic shock syndrome 1 toxin (tsst-1), and Panton-Valentine leukocidin (lukS/lukF-PV) by spa typing, PCR, drug resistance profile determination, and Western blot. We then subjected the studied population of S. aureus to photoinactivation based on a light-activated compound called rose bengal (RB) to verify photoinactivation as an approach to effectively kill toxin-producing S. aureus. We have obtained 43 different spa types that can be grouped into 12 clusters, indicating for the first-time clonal complex (CC) 7 as the most widespread. A total of 65% of the tested isolates had at least one gene encoding the tested virulence factor, but their distribution differed between the group of children and adults, and between patients with AD and the control group without atopy. We detected a 3.5% frequency of methicillin-resistant strains (MRSA) and no other multidrug resistance. Despite genetic diversity and production of various toxins, all isolates tested were effectively photoinactivated (bacterial cell viability reduction ≥ 3 log10 units) under safe conditions for the human keratinocyte cell line, which indicates that photoinactivation can be a good option in skin decolonization. IMPORTANCE Staphylococcus aureus massively colonizes the skin of patients with atopic dermatitis (AD). It is worth noting that the frequency of detection of multidrug-resistant S. aureus (MRSA) in AD patients is higher than the healthy population, which makes treatment much more difficult. Information about the specific genetic background of S. aureus accompanying and/or causing exacerbations of AD is of great importance from the point of view of epidemiological investigations and the development of possible treatment options.

The aim of this work was to assess the distribution and genetic variation of S. aureus originating from AD patients and to verify the hypothesis that regardless of the genetic background, skin isolates can be effectively eradicated by photoinactivation. Although some data on the genetic diversity of AD-derived S. aureus exist (17), research on S. aureus, from both its primary ecological niche, anterior nares, and inflamed skin of the same AD patient are scarce. Therefore, we aimed to evaluate the clonal relatedness of S. aureus isolates from nasal carriage and skin lesions of AD patients, both children and adults. We were interested in the distribution of genes coding for selected SAg, staphylococcal enterotoxins SEA-D and TSST-1, and their functionality in in vitro culture conditions. In the studied population, we also analyzed the presence of the genes encoding Panton-Valentine leucocidin (PVL). Furthermore, we investigated in vitro photoinactivation of those AD-derived isolates.

RESULTS
The population of AD isolates. One hundred AD patients from the Department of Dermatology, Venereology and Allergology at the Medical University of Gda nsk, Poland, were sampled in 2014 and 2015 under the local Research Ethics Board (approval no. NKBBN/242-477/2014). Atopic dermatitis was diagnosed following the criteria of Hanifin and Rajka (18), which include pruritus, typical morphology and distribution of eczematous lesions, chronicity of the disease and personal or family history of atopy. From each patient, two samples were taken: a swab from the nose (N) and a swab from the skin (S). These 200 samples were cultured on solid medium to allow S. aureus growth. We obtained bacterial growth on trypticase soy broth (TSB) agar for 139 (69.5% of all) samples that derived from 80 AD patients. In the case of 58 patients (116 clinical isolates, 83% of all grown isolates), we observed growth from both sampled sites, the nose (N) and skin (S). In the case of the remaining isolates, they were cultured from either the nose or the skin samples. All cultured AD isolates were tested on both the phenotypic and the molecular level (e.g., susceptibility testing, spa typing, and virulence gene detection). Fifty-nine isolates (42%) were classified as community-acquired S. aureus (CA-SA). An isolate was considered a CA-SA when it was isolated from an individual who did not have any of the following risk factors: isolation of S. aureus more than 48 h after admission to hospital; a history of hospitalization, surgery, dialysis, or long-term-care facility stay within 1 year of S. aureus culture date; presence of an indwelling catheter or a percutaneous device at the time of culture; or previous isolation of S. aureus. A consort diagram is included for easier tracking of the strains and patients from which they were isolated (Fig. 1).
Apart from the identified MRSA isolates, no multidrug resistance was detected among the methicillin-sensitive S. aureus (MSSA) isolates.
Staphylococcal enterotoxin and tsst-1 gene distribution among the S. aureus AD patient isolates. We studied the presence of genes encoding staphylococcal enterotoxins (SEs) A, B, C, D, and toxic shock syndrome toxin-1 (TSST-1) ( Table 1). According to the data obtained, 65% of all studied isolates contained at least one of the studied genes. In the MRSA group, this percentage accounted for 60%, which was lower than the analogous percentage in the MSSA group (66%). The most prevalent gene in the whole studied population was the sec gene (confirmed in 65 isolates), followed by sea (60), sed (24), and tsst-1 (8). Only 2 seb genes were detected in all the isolates collected.
The analysis of the SAg gene distribution among adults and children revealed interesting differences ( Table 2 and 3). The detected frequencies of sea, sec, sed, and tsst-1 in both groups differed significantly and were higher in the adult group. The frequency of detection of these genes was approximately twice that of children. In the case of the sed gene, the vast majority, namely, 23 out of 24 genes, were detected in the adult group, thus indicating that it was characteristic of the adult population. We found only two seb genes exclusively in the pediatric group. These differences were not affected by the MRSA distribution, as it was underrepresented in both groups: 3 isolates in children and 2 isolates in adults. We observed another interesting difference between children and adult isolates. In the group of children, most strains only had one out of the studied genes (10 strains, 1 gene; 6 strains, 2 genes; 2 strains, 3 genes), while in the adult group, most strains had . 1 studied gene (48 strains, 2 or 3 genes; 24 strains, 1 gene).
In a group of S. aureus isolated from nonatopic controls, the distribution of SAg genes exhibited a different profile: sea, sec, sed, and tsst-1 were detected less frequently than in the AD group (Table 4, Table S1). In contrast to the AD group, the sea, sec, and tsst-1 genes were less frequent and the seb gene was more frequent in nonatopic controls. It must be emphasized, however, that nonatopic isolates were collected in 2008 and 2009, whereas the current AD isolates were collected in 2014 and 2015.
An interesting observation was the detection of numerous isolates carrying lukS/lukF-PV genes (pvl-genes) encoding the Panton-Valentine leukocidin (PVL) (18/139; 12.9%). Most of the detected genes came from isolates from children (14/18; 77.7%). In the group of 18 pvl-positive isolates, 11 isolates were from the nose and 7 (from children only) from the skin. In 11 out of 18 pvl-positive isolates, the lukS/lukF-PV genes were the only detected ones among the genes tested. In two MRSA isolates (CC59-t437) originating from two different patients, the presence of genes coding for PVL was also detected ( Table 1).
The distribution of SAg genes among nasal isolates versus skin isolates revealed no statistically significant differences. The sea, sec, sed, and tsst-1 genes were detected with a similar frequency in nasal isolates and skin isolates, which was apparent for both groups studied, children and adults. For a more specific comparison, we excluded isolates from the same patient that belonged to the same spa type; however, in this case, no significant differences were observed.   Staphylococcal toxin production. To test whether the strains that possess genes encoding SAgs can produce the proteins, we performed immunodetection of SAgs during the logarithmic phase of growth in randomly chosen isolates. On average, staphylococcal enterotoxin A (SEA) and staphylococcal enterotoxin C (SEC) were produced in the largest amounts. From the experiments performed, we observed no difference between skin and nose isolates ( Table 5). We did not observe any difference in the amount of toxins produced depending on the origin of the isolates (hospital acquired versus community acquired) or the age of the patients from whom the isolates were derived (children versus adults). We observed that 3 MRSA isolates had sea genes. Two of the 3 were from the same patient and belonged to the same spa type (t008). Interestingly, the third of the above-mentioned MRSA isolates, apart from SEA, also produced SEB, which was detected in only two cases in the whole population, and the strain also produced SEC.
Identification of the clonal structure of S. aureus isolates. We obtained 43 different spa types among the studied population, including one new type (t20244). Using Ridom SeqSphere1 software, we assigned the spa types to 12 clusters, including four clusters consisting of two isolates of the same spa type, each, and six singletons. Three spa types (t282, t639, and 18582), shorter than 5 repeats, were excluded from the analysis (Fig. 2).
Brackets indicate the percentage of isolates with a gene detected within a given spa type (%); numbers indicating percentage of isolates with a gene detected within a given clonal complex are in boldface; n denotes the number of genes detected in the whole population studied (139 isolates); the designation CC denotes a clonal complex; and the designation t with a number (e.g., t091) indicates a specific type of spa, according to Ridom nomenclature (http://spaserver.ridom.de/).
Analysis of the relationship between the frequency of genes and the type of MLST-CC and the type of spa revealed that the distribution of the frequency of occurrence of a given gene differs in a statistically significant way depending on the spa/MLST-CC type ( Table 6). For all genes (sea, seb, sec, sed and tsst-1) there is a statistically significant correlation with the type of spa/MLST-CC (statistical significance in this case shows that the frequency distribution of a given gene differs significantly with respect to the spa type and MLST-CC type).
The sea gene was most often associated with spa type t091 in nearly 22% of sea-positive isolates, spa types t084, t024, t12375, and t015 in 5% each, and the rest in 2 to 3%. The sec gene was most often associated with spa type t091 in almost 17% of sec-positive isolates; spa type t015 in 14%; t156, t084, and t537 in 8% each; and the rest in 2 or 3% each. The sed gene was most often associated with spa type t091 in nearly 21% of sed-positive isolates; spa types t084 and t024 in 13% each; t005, t1255, t1312, and t156 in 8% each; and the rest in 4% each. In over half (62%) of the tested isolates, the presence of .1 gene was detected in a given isolate. The most common combinations were sea 1 sec (26 isolates), sea 1 sed (9 isolates), sea 1 sec 1 sed (11 isolates), and sea 1 seb 1 sec (2 isolates). The tsst-1 gene was most often associated with the spa types t2223, t012, and t18582 in 25% of tsst-1-positive isolates, and the remaining ones in almost 13% each. The efficiency of aPDI against the studied clinical isolates. To check the studied S. aureus population susceptibility to photoinactivation as a possible treatment option for atopic skin decolonization, we subjected the studied isolates to rose bengal (RB)mediated phototreatment. We were interested in whether the genetic diversity observed in the studied population affects phototreatment efficacy. Specifically, we were interested in checking whether the strains carrying SAg-producing genes are equally sensitive to the applied phototreatment in vitro. Phototreatment of S. aureus isolates from adults and children indicated a very efficient reduction ($3 log 10 CFU/mL; 99.9%) in the viable count of all isolates. The great majority of the studied population (88.4%) was photoinactivated under the following mild conditions: 0.5 mM RB and 40 J/cm 2 . Increasing the concentration of RB to 1 mM resulted in 100% of strains being photoinactivated, even at half of the light dose (20 J/cm 2 ) (Fig. 3). The obtained results indicate a very effective method of photoinactivation, which can be successfully used in research on more complex models.
The conditions of phototreatment that we applied to S. aureus isolates were also studied with respect to a keratinocyte cell line (HaCaT) to check the cyto-and phototoxicity in a eukaryotic model. Under the tested conditions of RB and light dose combinations, the photo-and cytotoxicity against the eukaryotic cell line were negligible. We also checked higher concentrations of RB under dark and light conditions, which resulted in only a slight decrease in keratinocyte survival (Fig. 4). This result means that aPDI is neither cyto-nor phototoxic against this cell line under the studied in vitro conditions.

DISCUSSION
The epidemiology of staphylococcal infections is extremely diverse, with new clones emerging and spreading in different populations at geographically different locations. The isolation rate of S. aureus in the population of AD patients might be as high as 90% in some populations. Given the frequency of isolation of S. aureus from atopic skin as well as the high prevalence of AD (15 to 30% in children and 2 to 10% in adults worldwide) (19), such patients represent a rich reservoir of staphylococcal isolates and potential drug-resistant isolates (20).
Consequently, there is a continuing need to monitor the genetic variability of these isolates as well as their toxigenic properties. Literature data indicate that skin isolates of AD patients produce large amounts of extracellular toxins, including SAgs, which are the subject of this analysis. However, the results of our analyses showed tremendous heterogeneity in the level of protein production among various isolates; moreover, we observed no differences in the number of proteins produced by nose and skin isolates or isolates from nonatopic patients (Table 5). This could be because we used TSB medium for bacterial cultures, whereas the highest protein production was observed in bovine heart media (21,22). In addition, a more thorough selection of the S. aureus study population based on the clinical presentation of AD (mild versus severe) in the patients from whom the isolates were obtained would be more informative with respect to the amount of toxin produced and the search for correlations. Staphylococcal enterotoxins (SE) are responsible for foodborne infections but also act as so-called superantigens that induce dysregulation of the immune response in AD patients. Data in the literature indicate that the type of inflammation and immune responses in AD patients depend on specific S. aureus strains (5). This shows that the diversity of AD strains is a factor that significantly contributes to the exacerbation of disease symptoms and encourages more in-depth research to identify the characteristics of S. aureus (including genetic attributes) that play a role in adaptation to the disease or to living on atopic skin. The fact that S. aureus frequently colonizes nasal epithelial cells without causing disease symptoms causes difficulties in epidemiological studies. On the other hand, S. aureus colonization of the skin is rather rare in the general population but very common in the AD population. The answer to the question of whether the S. aureus residing on the skin is genetically identical to that residing in the nasal epithelium of AD patients is not clearly defined. In previous studies, it was found Relative toxicities are shown according to control cells cultured without the presence of a photosensitizer (rose bengal) and without light treatment. The statistical significance was measured using one-way analysis of variance (ANOVA) and Dunnett's multiple-comparison test. Significance is marked with asterisks (****, P , 0.0001) with respect to untreated samples (cells kept in the dark). that strains isolated from AD patients differ from those isolated from other sources: they induce a stronger immune response than nonatopic strains or common commensal Staphylococcus epidermidis (5, 23, 24), they produce important exogenous proteases (25), and they show distinct cell wall properties (26). By analyzing S. aureus at the genetic level by tracing the relationship based on an analysis of spa types, many authors have demonstrated a distribution of genotypes in an S. aureus population isolated from AD patients (17) that is heterogeneous but insufficiently sensitive to explain the underlying mechanism of the dependence between atopic skin and S. aureus presence. Undoubtedly, the SEs produced by S. aureus play an important role in this situation, but their distribution within clones is inconclusive, although SEs are related to certain clones (27).
A comparison of our results regarding the genetic structure in the studied population with data published in other countries on AD populations shows some differences. For example, CC1 has been shown in some studies to be the most frequently isolated genotype from AD patients (28,29). In contrast, in our case, CC1 was detected sporadically, and CC7 was the most frequently isolated clone. We observed that the S. aureus population colonizing AD patients, studied at a genetic level based on commonly used molecular methods such as spa typing and MLST-CC, is consistent with the population observed in a given geographical area and is not specific to the type of disease. The most frequently detected genotype in our studied population was CC7 (ST7-t091), like the overall European trend (27,30). It is worth noting, however, that the clone CC7 (ST7-t091) was present in our population with a 4-fold higher frequency (22.5% versus 5.3% or 5.8%) compared to that for bloodstream-isolated S. aureus or isolates circulating in the healthy carrier population (27,30). Data on the relationship of some CCs to clinical symptoms are currently rather scarce, and research on this subject is quite limited. For example, CC30 is known to be associated with endocarditis (31), while no clear-cut correlations have been observed in AD. The only known example in the literature is the correlation between more frequent CC1 detection in AD patients with FLG mutations (32).
The question remains whether CC7 is somehow better adapted to colonizing atopic skin than other clones much less frequently detected in this study. Answers to these questions are being sought by studying virulence factors whose genetic determinants are located on mobile genetic elements. An example of such factors is SAgs. Based on our genetic analyses of the 4 classical enterotoxins and TSST-1, we showed that the gene distribution was heterogeneous among the different spa or MLST-CC types but also clonal. We were able to observe some trends that seem interesting. More than 60% of the sec genes, which were the most frequently detected SA-encoding genes in our population, were present mainly in CC45, CC7, and CC12. The high frequency of the sec gene was associated with many isolates belonging to CC7, as it was the most abundant clone in our population. In turn, the clones characterized by a high frequency of sec gene detection were dominated by CC12 (100% of isolates were positive for sec), CC72 (100%), and CC45 (75%). In the case of CC7, the frequency of gene detection was 37% and was close to CC15 (38.5%) and CC97 (33%). The same phenomenon was observed for spa types, i.e., 54% of detected sec genes were attributed to only 5 spa types (Table 6).
In the population we studied, we identified 5 MRSA isolates. They belong to CC8 (2), CC59 (2), and CC5 (1), the most popular genetic backgrounds of MRSA clones circulating in Polish hospitals in the first decade of this century (33). Such a low number of MRSA isolates seems to be in accordance with the numbers published from other European countries, 3% and 7.9% (4,34), and in contrast to some observations from outside Europe, 30.8% (35).
A meta-analysis of 95 studies indicated that the rate of toxin-producing S. aureus on the lesional skin fluctuated between 31.5 and 80% and was higher than that on healthy control skin (36). Nevertheless, the percentage distribution of particular toxin-encoding genes varied depending on the geographical population studied. In the case of the selected toxin-encoding genes examined in this study, we did not observe a random distribution but rather an association with specific CCs. The SEs we examined were assembled on mobile genetic elements (i.e., phages or pathogenicity islands) and can therefore be transferred to other bacterial cells horizontally and vertically. However, data from the literature show that the SAg distribution is rather vertical (37). Like previously published data, we observed that tsst-1 was associated with CC30 (37). The sec gene, which predominated in our studied population, was detected mainly in CC45 and CC7 isolates. All these results suggested that horizontal transfer of these genetic elements is not common. In contrast, sea, which is located on phage U3, was distributed much more broadly, and detected mostly in CC30 (5 of 7 isolates), CC97 (8 of 15 isolates), CC7 (16 of 35 isolates), and CC45 (8 of 25 isolates), but occasionally also in CC15, CC12, and CC1. The SAg-encoding gene sed (plasmid origin) was mainly detected in CC22 and in the same clones as sea (CC7, CC15, CC45, CC12, CC8, and CC30).
Some authors suggest that PVL-producing S. aureus strains play a key role in relapse of dermatitis in AD patients (38). According to recent reports, the lukS/lukF-PV genes are more common in isolates from pediatric patients and are detected in both MRSA and MSSA strains (39)(40)(41). Additionally, PVL-producing S. aureus corresponds to the severity of AD symptoms, as it was more often detected in children with moderate and severe SCORing Atopic Dermatitis (SCORAD) values (39). In our study, lukS/lukF-PV genes were found in 12.9% of isolates (18 of 139) and were more common in children (14 of 18 isolates, 77.7%). Our findings correspond with other observations regarding low lukS/lukF-PV detection rates (4,40). According to our results, in the Polish S. aureus population from AD patients, lukS/lukF-PV genes were most common in CC45, CC22, CC7, and CC59. It is worth noting that the S. aureus 079N strain (ST1027) with the newly identified spa type also carried this gene. On the other hand, genes coding for PVL dominated in CC30 in the Brazilian AD population, and in CC5, CC8 and CC30 strains in the Canadian AD population (40,41). The above data indicate no correlation between the clonal complex and the presence of the lukS/F-PV genes among S. aureus from AD patients.
Our results show that a large percentage (40%) of the tested isolates among skin/nose pairs are nonrelated or distinguishable strains, i.e., belonging to different spa types. This result may be partly because a high percentage of nonrelated pairs of isolates originated from children (48%). Unlike the second group of skin/nose paired strains, which were genetically indistinguishable and belonged to the same spa types, the percentage of isolates from children was only 28.5%. Such a high frequency of occurrence of various isolates in the same AD patient has not been previously observed and reported in the literature. In studies described in 2021 by Masiuk et al. (42), only two of 16 S. aureus from AD patients had a different genotype when nose and skin isolates were compared. Similarly, most AD patients have been found to be persistent S. aureus carriers with identical spa types in the nose and skin (43). The authors of previously published studies have also noted a high percentage of related skin and nasal isolates when they originated from one patient (4,44,45).
Due to the frequent use of topical and systemic antibiotics, AD patients are exposed to more antibiotics, which may select for more antibiotic resistant organisms (46). Our studies showed a relatively low percentage of resistance to local staphylococcal mupirocin (10%) and fusidic acid (6.5%). Such a low percentage may be because mupirocin is more often used for the decolonization of MRSA (20), which accounted for a small part of the studied population of isolates. For fusidic acid, a low rate of resistance may be due to short-term treatment with antibiotics. In the treatment of skin infections, a short period of using fusidic acid is recommended to avoid the development of resistance (47). On the other hand, our studies showed a high rate of erythromycin resistance (39%). The use of erythromycin may be limited due to the increasing resistance to macrolides (48). According to literature reports, macrolides are a first-line alternative to penicillin antimicrobials for penicillin allergic AD patients (28). Growing resistance to these groups of antibiotics may limit therapeutic options.
Antimicrobial photoinactivation is a viable option for treating S. aureus. In our studies, we showed that S. aureus is vulnerable to aPDI with the use of RB in combination with visible light (l max = 515 nm). Practically all the AD strains studied were effectively photoinactivated. More importantly, the efficacy of the photoinactivation process was independent of the presence and production of SAgs. This result means that RB-mediated aPDI applied against toxigenic strains is equally effective as it is against nontoxigenic strains. This finding agrees with studies published previously regarding SE-producing strains but with the use of different photoactive compounds, namely, porphyrin derivatives (49). RB is a commonly available compound that has been shown to be safe in animal studies on rabbit corneas (50). RB-based aPDI demonstrated efficacy in decreasing the parasitic load (Acanthamoeba castellanii) and clinical severity of actinic keratitis (51). Moreover, its photodynamic efficacy was positively verified in a mouse skin abrasion model of Pseudomonas aeruginosa infection (52,53). Additionally, RB derivatives have been tested in a mouse wound healing model and shown to promote antistaphylococcal effects in a complex tissue model (54). Research on various RB delivery systems is also being conducted to modulate the action of RB (55). PH-10 is a topical RB hydrogel formulation that selectively delivers RB to the epithelial tissue and is currently being tested in clinical trials (NCT00941278, NCT1247818, NCT12322086, and NCT00690807) for the treatment of psoriasis and atopic dermatitis with encouraging results, although the mechanisms of action of RB in these diseases are not yet fully understood. RB can be used in the clinic as a topical drug, both as a light-free and lightactivated compound. In the latter case, appropriate clinical studies are needed to verify the safety of RB as a potentially phototoxic compound. To date, there are no publicly available data from the only clinical trial on the use of a photodynamic method based on PH-10 in combination with green light (544 nm) in patients with psoriasis. Previous research together with our in vitro results ( Fig. 3 and 4) indicate that RB can be a good candidate for antistaphylococcal treatment. It must be emphasized that RB is active not only against S. aureus but also against other microbial species in vitro, e.g., Bacillus subtilis, Staphylococcus epidermidis, Enterococcus faecalis, Streptococcus salivarius, Streptococcus pneumoniae, Mycobacterium smegmatis, Mycobacterium avium, Pseudomonas aeruginosa, Salmonella Thyphimurium (56). This means that the potential application of RB is much broader than in the presented manuscript, where only S. aureus was used as a model organism.
Information on the response of skin cells to light and rose bengal is of interest to scientists involved in the application of photodynamic methods, but data on keratinocytes are quite sparse in this regard. Extensive studies have been performed on the HeLa cell line exposed to light-activated RB (530 nm), showing that apoptosis and autophagy are the crucial events responsible for cell death, while necrosis is negligible (57,58). Other cancer cell lines have also been studied (e.g., HL60, MCF-7, A431), while studies on the mechanisms of cell death in healthy skin cells have not yet been published. Green light, per se, is generally not considered toxic to keratinocytes. However, as mentioned above, when combined with certain photosensitizing agents, such as rose bengal, green light can induce cell death in keratinocytes due to oxidative damage that may ultimately lead to cell death. It should be emphasized, however, that our proposed RB-based photodynamic treatment had no mutagenic effect either in combination with green light or when green light was used alone (Fig. S1).
Whether aPDI approach would improve the functioning of atopic skin, prevent periodic worsening of the disease, or reduce its frequency requires further studies in murine models of atopy and in humans over the long term. Nevertheless, in the case of chronic diseases, such as AD, periodic, repeatable treatment is undoubtedly required, which raises the risk of selecting for resistance when using classic antibiotics. Such risk is minimized in the case of aPDI due to its completely different mechanism of action (59). In this respect, aPDI may constitute a viable alternative to antibiotics.

MATERIALS AND METHODS
Bacterial strains and growth conditions. Three groups of S. aureus strains were used in the study. The first group comprised 139 clinical strains from patients with AD, isolated from the skin and nose from each patient in years 2014 and 2015 (Department of Dermatology, Venerology, and Allergology, Medical University of Gda nsk). The second group was 39 nonatopic isolates from the collection of S. aureus strains in the Laboratory of Photobiology and Molecular Diagnostics (IFB). The last group included 14 S. aureus reference strains, which have been characterized according to toxin production and genetic background (National Medicines Institute, Warsaw, Poland).
Antimicrobial susceptibility testing. The MIC was determined using the broth microdilution method according to the CLSI guidelines (ref. CLSI 2018). Assays for daptomycin were performed in medium supplemented with Ca 21 (50 mg/L). Polypropylene 96-well plates with bacteria at an initial inoculum of 5 Â 10 5 CFU/mL and exposed to the tested antibiotics were incubated for 18 h at 37°C. The following antibiotics Methicillin resistance was assessed by culturing S. aureus isolates on Mueller-Hinton (4% NaCl) agar plates with oxacillin (Graso Biotech, Poland). Isolates were also screened by PCR for the presence of the mecA and mecC genes, as described earlier (62,63).
Isolation of genomic DNA. Bacterial DNA was isolated from 1.5 mL of the overnight culture. Cells were pelleted by centrifugation at 13,400 rcf for 5 min (centrifuge 5415 D, Eppendorf). The pellet was suspended in 100 mL of lysis buffer (20 mM Tris-HCl pH 8.0, 2 mM EDTA pH 8.0, 1.2% Triton X-100) and 10 mL of lysostaphin (0.4 U/mL, A&A Biotechnology, Poland). For a more efficient isolation procedure, 120 mg of glass beads was added (Glasperlen, 0.10 to 0.11 mm, Sartorius StedimBiotech GmbH, Goettingen, Germany). The mixture, after mixing well by vortexing for 1 min, was incubated at 37°C for 40 min. During the incubation process, every 10 to 15 min, the mixture was briefly vortexed (10 to 15 s). The remaining DNA isolation steps were carried out using a genomic minikit (A&A Biotechnology, Poland) according to the manufacturer's instructions with the addition of an RNA digestion step (RNase, 10 mg/mL, A&A Biotechnology, Poland). The DNA concentrations were measured using a NanoDrop ND-1000 (Thermo Scientific, USA).
Detection of staphylococcal toxin proteins: Western blotting. For preparation of bacterial protein lysates, the overnight bacterial cultures were reinoculated at a ratio of 1:100 in 5 mL of sterile TSB medium and incubated until the late log phase of growth was reached (6 to 7 h; OD 600 = 1.5 to 1.6). A total of 200 mL of each sample were suspended in a ratio of 1:1 in 2Â Laemmli sample buffer (with the addition of b-mercaptoethanol; BioRad, USA) and incubated for 5 min in a thermoblock heated to 95°C. After incubation, the samples were briefly centrifuged (1 min, 16,100 rcf). The prepared bacterial protein lysates were stored at 220°C until further analysis.
The total protein concentration in protein lysates was measured. Due to the presence of reducing agent (b-mercaptoethanol), detergent (SDS) and dye (bromophenol blue) in Laemmli buffer, the total protein concentration in protein lysates was measured using an RD DC protein assay kit (BioRad, USA). A g -globulin standard (BioRad, USA) was used as the protein standard at three different concentrations: 0.37 mg/mL, 0.74 mg/mL, and 1.48 mg/mL. The amount of total protein was measured spectrophotometrically using a UV-VIS SPECORD 2000 PLUS spectrophotometer (Analytik Jena, Germany) at a wavelength of 750 nm. The absorbance values of the tested samples were compared with the pattern of the curve of the g -globulin protein standard, yielding the values of total protein concentration in the prepared bacterial lysates.
For sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), 5 mL of a protein marker (PageRuler prestained protein ladder, 10 to 180 kDa; Thermo Fisher Scientific, USA), 5 mL of the protein standard of the tested SAgs at a concentration of 25 mg/mL (Toxin Technology Inc., USA), and 10 mg of protein lysates of the tested samples were applied to the wells of a polyacrylamide gel. SDS-PAGE was Electrotransfer and immunodetection. Wet transfer was performed using a polyvinylidene fluoride (PVDF) membrane (BioRad, USA). A cassette with sandwiches was placed in a Mini-PROTEAN Tetra Cell apparatus (BioRad, USA) along with a cooling cartridge. Membrane transfer was performed on ice in 1Â concentrated transfer buffer at 100 V for 60 min. The PVDF membrane was blocked for 30 min in 30 mL of Tris-buffered saline 1 Tween (TBST) solution with 1% skim milk and shaken on a rocking shaker at room temperature. The membrane was washed twice with 30 mL of TBS-Tween, incubated with a 1:10,000 dilution of the primary antibody (antistaphylococcal enterotoxin IgG; Toxin Technology Inc., USA) in 30 mL of TBST solution with 1% skim milk overnight at 4°C with gentle shaking. The next day, three 5-min washes were performed with 30 mL of TBST with gentle shaking on a rocking shaker. Horseradish peroxidase (HRP)-labeled secondary antibodies at a ratio of 1:10,000 (peroxidase-conjugated AffiniPure Alpaca antirabbit IgG; Jackson ImmunoResearch Laboratories Inc., USA) in 30 mL of TBST solution with 1% skim milk were added and incubated for 30 min on a rocking shaker at room temperature. After this time, three 5-min washes were performed with 30 mL of TBST. Finally, the membrane was washed twice with 30 mL of Tris-buffered saline (TBS). The membrane was covered with Clarity Max ECL substrate (BioRad, USA), and the chemiluminescence signal was read on a ChemiDoc XRS 1 documentation system (BioRad, USA). The obtained image was analyzed in Image Lab (version 6.0, BioRad, USA). The obtained band intensities for individual samples were analyzed against the standard curve of the enterotoxin protein standard. The amount of protein of a given isolate was tested in three independent experiments (from three independent cell cultures).
Antimicrobial photodynamic inactivation (aPDI). (i) Chemicals. Rose bengal (RB) (4,5,6,7-tetrachloro-29,49,59,79-tetraiodofluorescein disodium salt; Sigma-Aldrich, Germany) was used in this study as a light-activated molecule (photosensitizer). A 1 mM RB stock solution was prepared in sterile double-distilled water and stored at 4°C for up to a month. Before use, the stock solution was diluted in sterile double-distilled water to the appropriate concentration.
(ii) Light source. The illumination process was carried out using an LED lamp (SecureMedia, Poland) emitting green light (l max = 515 nm) (66). The output power was 35 mW/cm 2 . The bacterial cultures were irradiated with a total dose of light of 10, 20, or 40 J/cm 2 (duration of the illumination process of 333, 666, or 1332 s, respectively).
(iii) Phototreatment experiment. Bacterial cultures were grown overnight (18 6 2 h) at 37°C in nutrient TSB (bioMérieux, France) and then diluted with fresh medium to a density of 0.5 McFarland units (approximately 10 7 CFU/mL). Diluted S. aureus cultures (200 mL) were incubated in the dark at 37°C with shaking (150 rpm) for 10 min, either with or without the photosensitizer. The final concentrations of the photosensitizer were as follows: 0.5 mM and 1 mM. After the incubation process, 100 mL of each culture was transferred into a 96-well plate and irradiated with different light doses of 10, 20, and 40 J/cm 2 . The remaining 100 mL of S. aureus cultures incubated with and without the photosensitizer in the dark during the illumination process served as controls. After the irradiation time, 10 mL of each sample was taken to perform 10-fold serial dilutions (from 10 21 to 10 24 ) in sterile PBS. Then, 10 mL of each dilution were streaked horizontally onto TSA plates (bioMérieux, France) and incubated for 24 h at 37°C. After overnight incubation, the bacterial colonies were counted, and the results were statistically analyzed. Survival fractions were expressed as ratios of CFU of treated bacteria (with light and photosensitizer) to CFU of untreated bacteria. Each experiment was performed in triplicate (from three independent cultures of each isolate).
(iv) Spa typing. Amplification of the variable-number tandem repeat (VNTR) region of the spa gene in S. aureus strains from AD patients was conducted by using PCR with specific primers based on Airesde-Sousa et al. (67) (Table 8). The amplification method was performed as follows: a total volume of 50 mL of the reaction mixture contained 1Â DreamTaq Buffer (including 2 mM MgCl 2 ; Thermo Scientific, USA), 100 mM dNTPs (Thermo Scientific, USA), 1 pmol of each primer (Eurofins Genomics, Ebersberg, Germany), 0.05 U of DreamTaq polymerase (Thermo Scientific, USA), and 1 mL of template DNA.
Amplification was carried out in a Gene Amp PCR System 9700 (Applied Biosystems) using the following steps: initial denaturation (94°C, 5 min) followed by 30 cycles of denaturation at 94°C for 30 s, annealing of primers at 63°C for 30 s and extension at 72°C for 45 s. The reaction was terminated with extension at 72°C for 3 min. PCR products were analyzed by electrophoresis through a 1% agarose gel in 0.5Â TBE buffer (diluted from 10Â TBE buffer, Thermo Scientific, USA), stained with a safe replacement for ethidium bromide (Simply Safe; EURx, Poland) and visualized under UV illumination using a ChemiDoc imaging systems instrument (BioRad, USA). The PCR products were sequenced by Eurofins Genomics AT GmbH (Austria) using the primers used in the PCR. Sequencing of both complementary strands of the PCR product was performed.
A minimum spanning tree (MST) was created based on spa type profiles using Ridom SeqSphere1 software version 8.3.1 (Ridom GmbH; Münster, Germany) with default parameters enabled during comparison: MST Clusters cost distance -5 and exclude spa types shorter than 5 repeats, options turned on in the SeqSphere1 software during comparison (68). Multilocus sequence typing (MLST) was performed based on Enright et al. (69) for 13 MSSA isolates characterized by a new spa type or spa type for which sequence type (ST) could not be predicted. Allele numbers and STs were assigned through the S. aureus MLST database (http://pubmlst.org/organisms/ staphylococcus-aureus). For remaining isolates, ST and MLST-clonal complex (MLST-CC) were inferred from the spa type in the Ridom SpaServer (http://spaserver.ridom.de/) and from the literature (70).
Cells (1 Â 10 4 ) were seeded in a 96-well plate and allowed to adhere overnight. RB at the final concentration of 0 to10 mM was examined. The cells were incubated with RB for 15 min at 37°C in the dark, washed once with Dulbecco's phosphate-buffered saline (DPBS; Sigma-Aldrich, Germany), and dissolved in 100 mL of DMEM (Gibco; Thermo Fisher Scientific, USA). Next, irradiation was performed using green light (l max = 515 nm, 40 J/cm 2 , 35 mW/cm 2 , 1332 s). Cell survival was determined after 24 h of incubation by an MTT assay using MTT reagent (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide, Sigma-Aldrich, Germany). Next, 10 mL of MTT reagent (12 mM stock solution) were added to each well, and then the cells were incubated for 4 h at 37°C. Then, cells were lysed in 100 mL of dimethyl sulfoxide (DMSO; Sigma-Aldrich, Germany), and the absorbance of the formazan solution was measured at 550 nm using a plate reader (Victor 1420 Multilabel Plate Reader, Perkin-Elmer). All experiments were carried out in three biological replicates for each condition (cells treated with light and untreated cells stored in the dark). Cell viability (expressed as a percentage) is presented as the mean of the three independent biological replicates with reference to the untreated cells.
Statistical analysis. All statistical calculations were performed using the StatSoft Inc. statistical package (2014) STATISTICA (data analysis software system, version 12.0, www.statsoft.com) and an Excel spreadsheet. Quantitative variables were characterized by the arithmetic mean, standard deviation, median, minimum, and maximum value (range) and 95% confidence interval (CI). On the other hand, the qualitative variables are presented by means of the count and percentage values (percentage). The significance of difference between two groups was tested by means of Student's t test or the Mann-Whitney U test. Chi-square tests of independence were used for qualitative variables (using Yates correction for cell counts below 10, checking Cochran conditions, Fisher's exact test, respectively). In all calculations, P = 0.05 was adopted as the level of significance.
Conclusions. The S. aureus population colonizing the skin of AD patients is heterogeneous and largely reflects the distribution of S. aureus isolates circulating in the general population. Noteworthy, however, is the 4-fold increased number of isolates belonging to CC7 that we identified in the Polish population. The question of whether CC7 is particularly predisposed to colonization of atopic skin remains open.
The most frequently detected superantigen genes among our subjects were sec and sea; we also observed a difference in the distribution of genes encoding the tested superantigens between the group of adults and children. Regardless of the presented pattern of genes encoding enterotoxins, S. aureus is highly sensitive to phototreatment under conditions that are safe for human keratinocytes. The heterogeneous S. aureus population is effectively inactivated in vitro by the action of aPDI, which provides a reasonable basis for validating aPDI as an antimicrobial method in more complex in vivo models.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 1 MB.