Knowledge of the eukaryotic microbial community in AD is limited [8, 9]. however, cutaneous fungi and parasites might be of importance in AD pathogenesis, as seen for the bacterial community [18]. In the present study, analysis of similarity implied that the overall eukaryotic microbial community composition on skin and in nares was significantly different between AD patients and non-AD healthy individuals. The community composition was similar on AD LS and NLS, despite clinically differences between the two sample sites. No significant association was found with neither AD disease severity (SCORAD), nor carriage of loss-of-function FLG mutations, as previously reported for the bacterial community on AD skin [5, 19]. The present study confirms, that the eukaryotic microbial richness is significantly greater on AD skin compared to healthy control skin [8, 9]. This might be due to regular use of TCS among patients, as recent TCS treatment was associated with an increased richness on AD LS. Another possible explanation might be that antibacterial therapy alters the microbiome by reduction of bacteria thereby contributing to proliferation and changed virulence characteristics of the remaining microbes such as fungi, similar to what have previously been described for Candida vaginitis after antibiotic treatment [20].
To some extent, the observed variation between AD and healthy control skin might be explained by a significantly higher frequency of the ectoparasite D. folliculorum and the fungus G. candidum on AD skin. D. folliculorum is a common skin mite that predominantly lives in sebaceous areas of the face, where it consumes sebum [21]. Though D. folliculorum is a commensal inhabitant of the skin, increased mite densities and penetration into the dermis have previously been associated with inflammation and skin barrier disruption [22, 23]. Also, D. folliculorum has been associated with the inflammatory skin disease rosacea [21, 24]. D. folliculorum was identified at a significantly higher prevalence on AD skin (LS as well as NLS) in this study, despite the fact that our samples primarily were collected from moist and dry skin areas. A possible explanation for this could be that AD patients regularly use moisturizing lotions as part of their treatment, which might create an advantageous environment for D. folliculorum. This hypothesis can be supported by our observation that D. folliculorum was more prevalent on skin among patients reporting use of TCS, which is in alignment with findings from a study of patients with perioral dermatitis [25]. A second hypothesis could be that the skin conditions in AD itself favour D. folliculorum colonization, as slightly higher colonization frequencies have been observed among people with higher skin pH and lower skin hydration [26], as present in AD.
G. candidum, the other species found to be overrepresented on AD skin, is considered a common skin colonizing yeast [27, 28], However, other DNA sequencing based studies on skin fungal communities, including communities on AD skin, have not reported the presence of this organism [8–11], which might be due to previously small study populations. Whether the distorted skin ecology in AD, including increased skin pH, reduced skin hydration and altered composition of free fatty acids, favours the growth of G. candidum, need to be further investigated.
M. globosa and M. restricta were some of the most frequently observed fungal spp. at all studied sample sites, which is in accordance with previous findings [8, 9], suggesting that these species can be considered as commensals of these anatomical areas. Malassezia is a lipophilic yeast, formerly thought only to reside on seborrheic skin areas (scalp, face and thorax), but after the introduction of molecular based detection methods it has been isolated from most body sites [10]. There was no significant difference in the prevalence of Malassezia spp. on skin between AD patients and healthy controls, which is somewhat surprising as AD patients have a dry skin and the genus Malassezia is lipid dependant. A significantly higher detection rate of M. restricta was found in the nares of healthy controls compared to the nares of AD patients. The explanation for this observed reduction of M. restricta in AD nares is not clear, and need further examination.
Candida is a commensal of mucosal membranes and colonizes moist skin areas, and it is thus surprising that this genus tended to be more frequently detected in AD patients, as patients with AD have drier skin. High prevalence of Candida with the potential of causing has previously been associated with long-term or repeated use of antibiotics [29], however, antibiotic treatment was not associated with significant changes in Candida presence on AD skin in the present study.
The major strength of this study is the large cohort, by far the largest published to date. Furthermore, AD-diagnosis was verified by a specialized dermatologist, and all patients had active disease. This ensures high quality and clinical relevance of the presented data. A well-known limitation in the field of microbiome research is the difficulties in discriminating between colonizing organisms and environmentally derived organisms. Thus, some of the identified species could be environmentally derived contaminants of the human epidermis rather than true colonizers of skin and nares, e.g. Cladosporium and Aspergillus are common spore-forming molds in both indoor and outdoor environments [17]. However, increased sensitivity in AD patients to these molds [17], together with their high prevalence found on AD and healthy control skin [8, 10], support a possible role for mold sensitivity and AD inflammation.