SD: Seborrheic Dermatitis; PCoA: Principal Coordinate Analysis; IGS: Intergenic Spacer

Seborrheic dermatitis (SD) is a common inflammatory skin disorder associated with seborrhea. It is characterized by erythematous patches with yellow-gray scales. These appear most often on the face, especially the nasolabial folds, scalp, and upper trunk, in sebaceous gland-rich areas of the skin. The disorder is present in about 3% of the general population, more prevalent in males than in females, and frequently observed in patients with acquired immunodeficiency syndrome (AIDS) and Parkinson’s disease [1,2]. The disease presents in two age groups: newborn infants up to 5 months of age, and young adults with increased sebum secretion from the skin.

The human skin is populated by various microorganisms, including bacteria and fungi [3,4]. Of these, skin fungi, Malassezia species, play an important role in the development of SD. As Malassezia species require fatty acids for their growth, they colonize the sebaceous glandrich areas of the skin, including the face, scalp, and back, where they feed on fatty acids from human sebum. Malassezia species secrete lipases that hydrolyze sebum into triglycerides, which are further hydrolyzed into fatty acids [5-7]. Fatty acids are utilized as nutrition by skin microorganisms, including Malassezia. However, the unsaturated fatty acid oleic acid causes inflammation of the skin directly by eliciting IL-1α secretion from macrophages or epidermal keratinocytes, and is thought to be the causative agent of SD [7]. In fact, the mRNA expression of a specific Malassezia lipase gene can be detected from lesional sites in patients with SD [8,9]. The clinical condition improves on administering antifungal agents, suggesting that Malassezia species are one of the causative agents of SD [10,11].

Tajima et al. [12] analyzed the Malassezia microbiota in the skin of patients with SD using a DNA-based molecular approach. The genus Malassezia currently includes 14 species; of these, M. globosa and M. restricta are the major species in the skin of patients with SD and these species are more abundant at lesional sites than non-lesional sites.

In the present study, we analyzed comprehensively the skin fungal microbiota of lesional and non-lesional sites of SD patients to obtain basic information to enable understanding of the development of SD and skin fungal microbiota using a pyrosequencing approach.

Patients and sample collection

Twenty-four Japanese outpatients with SD were enrolled in this study (19 males and 5 females; mean age 47.8 ± 18.8 (range 20-77) years). Patients who took antimicrobial agents before involvement in this study were excluded. The study protocol was approved by our Institutional Review Board and informed consent was obtained from each individual.

Skin fungi were collected by applying a 5×7 cm OpSiteTM transparent adhesive dressing (Smith and Nephew Medical, Hull, UK) using the method of Sugita et al. [13]. Briefly, dressings were applied three times each to both lesional and non-lesional sites on the faces of SD patients.

Fungal DNA extraction

The DNA was extracted directly from the dressing using the method of Sugita et al. [13]. Briefly, the dressing was placed in a 1.5 mL Eppendorf tube with 1 mL of lysing solution (100 mM Tris–HCl (pH 8.0), 30 mM EDTA (pH 8.0), and 0.5% SDS) and incubated for 15 minutes at 100°C. The suspension was extracted with phenol– chloroform–isoamyl alcohol (25:24:1, vol/vol/vol), and subsequently, chloroform–isoamyl alcohol (24:1, vol/vol). The DNA was precipitated with 2.5 volumes of ethanol, in the presence of 3 M sodium acetate and Ethachinmate (Nippon Gene, Toyama, Japan), according to the manufacturer’s instructions.

Pyrosequencing

The universal primers NL1 and NL4 containing the A and B sequencing adaptors were used for PCR amplification of the D1/D2 variable regions of the 26S rRNA gene. The primers used were B-NL1 (5’-cctatcccctgtgtgccttggcagtctcaGCATATCAATAAGCG GAGGAAAAG-3’; adaptor B in lowercase) and A-NL4 (5’-ccatctcatccctgcgtgtctccgactcagatcagacacg NNNNNGGTCCGTGTTTCAAGACGG-3’; adaptor A in lowercase letter, N represents a bar code unique to each sample). The cycling conditions were 94°C for 1 min, followed by 30 cycles at 94°C for 30 s, 56°C for 30 s, and 72°C for 30 s, with a final extension at 72°C for 10 min. A negative control containing no template was also used. Following purification of each PCR amplicon, equimolar amounts of each PCR reaction were mixed in a single tube. Then, the purified amplicon mixtures were sequenced in a 454 GS FLX pyrosequencing platform (Roche Diagnostics Japan, Tokyo, Japan), according to the manufacturer’s instructions.

Sequence processing and data analysis

The primer and barcode sequences were removed from the data and possible chimeras were also excluded from the analysis. Sequences ≥400 bp and <1,000 bp in length were subjected to analysis. The D1/D2 LSU sequences were classified to the genus level using the RDP classifier (http://rdp.cme.msu.edu/). The D1/D2 LSU sequence of the type strains of yeast-like fungi was extracted from GenBank. The sequences were classified to the species level using an in-house BLAST search. The R package vegan [http:/CRAN.R-project.org/package=vegan] was used to construct the Shannon diversity index boxplot. Significance was tested using a t-test with a one-tailed distribution and two-sample variance. A three-dimensional principal coordinate analysis (PCoA) plot was normalized using weighed values (http://www.quiime.org).

Fungal community composition

This study analyzed 480,186 high-quality sequences ≥400 bp and <1,000 bp in length, ranging from 6,058 to 13,773 reads per sample. The numbers of reads for lesional and non-lesional sites did not differ significantly (lesional sites 9,744 ± 1,864, non-lesional sites 10,263 ± 2,044). The samples were grouped into 30 genera in two phyla: ten yeast-like Ascomycota and 20 filamentous Basidiomycota (Table 1). Malassezia was the most abundant fungus genus at both sites, and predominated at non-lesional sites (83.3 ± 26.2%) compared with lesional sites (57.3 ± 9.6%) (Supplement Figure 1), while the proportion of non-Malassezia yeast-like fungi was higher at lesional sites (36.7 ± 10.7%) than non-lesional sites (7.1 ± 5.6%). Filamentous fungi accounted for 6.0 ± 6.4% and 1.3 ± 1.8% at lesional and non-lesional sites, respectively. Nine taxa accounted for approximately 3% of the sequences at both lesional and non-lesional sites: Candida parapsilosis and C. guilliermondii, Cryptococcus albidus, Rhodotorula mucilaginosa, Malassezia restricta, M. globosa, M. sympodialis, and M. slooffiae, and Cladosporium spp. (Figure 1) and accounted for over approximately 90% of all of the fungal species. The colonization levels of M. furfur, M. sympodialis, and M. slooffiae did not differ significantly between lesional and non-lesional sites, whereas colonization by both M. globosa and M. restricta differed significantly. M. restricta predominated at lesional sites (87.1 ± 11.0%) versus non-lesional sites (58.1 ± 13.9%), while M. globosa predominated at non-lesional sites (33.6 ± 15.6%) versus lesional sites (8.4 ± 6.0%) (Supplement Figure 2).

Figure 1: Distribution of the nine major taxa.
The distributions of 9 of the 30 taxa are shown.

Table 1: The percentage of fungi detected from scale samples from patients with seborrheic dermatitis.

The Shannon diversity index was calculated to determine the diversity of the samples. Lesional samples showed significantly greater diversity than non-lesional samples (p<0.01) (Figure 2). PCA was used to evaluate sample diversity and analyze the relationships among samples, and revealed clear separation between lesional and nonlesional sites (Figure 3).

Figure 2: Shannon diversity index.
Lesional sites were significantly (p<0.05) more diverse than non-lesional sites.

Figure 3: The fungal communities of skin samples obtained from lesional and non-lesional sites.
The fungal communities were generated by principal coordinates analysis (PCoA).
Red: lesional site; blue: non-lesional site.

This study is the first comprehensive analysis of the skin fungal microbiome of SD patients using a pyrosequencing method.

Studies of the relationship between skin microorganisms and SD have focused on Malassezia species, since they secrete lipase to hydrolyze sebum into fatty acids. A molecular-based culture-independent method should be used to analyze the cutaneous Malassezia microbiota in healthy subjects or patients as no culture medium that allows efficient recovery of Malassezia from scale samples is available. Tajima et al. [12] first investigated the skin Malassezia microbiome of SD patients using a culture-independent method. The level of Malassezia colonization at lesional sites was approximately threefold that at non-lesional sites. In addition, M. restricta predominated over M. globosa. Malassezia is part of the normal flora in healthy individuals, but is also associated not only with SD but also with pityriasis versicolor and atopic dermatitis [14,15]. Both M. restricta and M. globosa are major components of various skin diseases, although the ratio of the two is disease-specific [16,17]. In this study, the pyrosequencing assay revealed that M. restricta predominated over M. globosa at lesional sites (Supplement Figure 2). While the ratio of M. restricta to the total Malassezia colonization level at non-lesional sites was significantly higher than that at lesional sites, that of M. restricta was almost the same at both lesional (49.6%) and non-lesional (52.9%) sites in terms of total fungal colonization level. This was due to the fact that the colonization level of non-Malassezia yeast species was higher at lesional sites than at nonlesional sites. Of the 30 taxa identified in this study, 9 accounted for more than 90% of the total colonization. These microorganisms are normal fungal flora in healthy subjects, although Candida parapsilosis, C. guilliermondii, Cryptococcus albidus, and Rhodotorula mucilaginosa rarely cause opportunistic fungal infections in immunocompromised hosts [3,18-20].

The greater quantity of sebum at lesional sites than non-lesional sites is thought to explain why the fungal microbiota was more diverse or greater fungal colonization was present at lesional sites. Since Malassezia species require fatty acids for their growth, they have more lipase genes than other skin fungi, with M. globosa and M. restricta possessing 15 and 10 lipase genes (unpublished data), respectively. Candida species also secrete lipases, so lipases secreted by other fungi might act synergistically in the development of seborrheic dermatitis [9,21].

Bacteria and other fungi might also be involved in the development of seborrheic dermatitis. The fungal microbiomes of the scalps of patients with and without dandruff are similar, whereas the bacterial microbiomes are different [22]. Staphylococcus epidermidis predominated on the dandruff scalps (60%), while Propionibacterium acnes predominated on normal scalps (74%) in French subjects. A discrepancy in the fungal microbiome was observed between that French study and a Korean report. The latter study reported that Filobasidium species were most common in dandruff patients, while Acremonium species were most common in healthy subjects. Surprisingly, no Malassezia species were detected from the scalp samples; this may have been due to differences in the techniques used [23].

The genotype of Malassezia species might also be involved in the development of seborrheic dermatitis. As the intergenic spacer (IGS) region of the Malassezia rRNA gene shows remarkable intra-species diversity, population structure can be elucidated by genotypic analyses. Tajima et al. [12] and Hiruma et al. [24] found patient-specific IGS genotypes of M. globosa and M. restricta in patients with seborrheic dermatitis and dandruff.

In conclusion, this study found that the fungal communities were more diverse at lesional sites than non-lesional sites. Our findings provide useful information for understanding the interactions between skin microorganisms and seborrheic dermatitis.