Adding fuel to the fire? The skin microbiome in atopic dermatitis

Abstract

Atopic dermatitis (AD) is a multifactorial, heterogeneous disease characterized by epidermal barrier dysfunction, immune system dysregulation, and skin microbiome alterations. Skin microbiome studies in AD have demonstrated that disease flares are associated with microbial shifts, particularly Staphylococcus aureus predominance. AD-associated S. aureus strains differ from those in healthy individuals across various genomic loci, including virulence factors, adhesion proteins, and proinflammatory molecules – which may contribute to complex microbiome-barrier-immune system interactions in AD. Different microbially based treatments for AD have been explored, and their future therapeutic successes will depend on a deeper understanding of the potential microbial contributions to disease.

Background on atopic dermatitis (AD)

Atopic dermatitis (AD) is a common chronic inflammatory skin condition marked by recurring, itchy, eczematous lesions that can progress to skin fissuring and thickening (Weidinger et al., 2018). AD is predominantly a disease of early childhood affecting approximately 20% of children, with a later smaller peak in incidence in middle-aged adults. AD substantially impacts quality of life, including negative effects on sleep quality (Ramirez et al., 2019) and mental health (Thyssen et al., 2018). Due to compromised skin barrier, AD patients are at risk of secondary infection by microbes, including herpes simplex virus (Beck et al., 2009) and S. aureus (Alexander et al., 2020).

As a complex multifactorial disease, AD is characterized by skin barrier dysfunction as well as immune system dysregulation, and skin microbiome alterations (Figure 1). Genetic investigations of AD have identified mutations involved in skin barrier function and the immune system (Ellinghaus et al., 2013, Esparza-Gordillo et al., 2013, Hirota et al., 2012, Paternoster et al., 2011, Sun et al., 2011, Weidinger et al., 2013). Although primarily viewed as type 2 inflammation, some AD studies and differential responses to biologics have highlighted potential contributions of non-T helper 2 (TH2) inflammation such as TH1, TH17, and TH22(Brunner et al., 2017). Interestingly, a positive association between AD burden and a country's gross domestic product suggests the potential influence of environmental and lifestyle factors (Laughter et al., 2021). Large cohort studies have identified inverse risk of AD with factors typically associated with microbial exposure, such as having three or more siblings, early daycare attendance, pet ownership, and farm residence (Benn et al., 2004, Gehring et al., 2001, Karadag et al., 2007); however, others have highlighted the heterogeneity of AD and question the role of these exposures (Cramer et al., 2010). A study of electronic medical records identified an increased risk of AD with antibiotic exposure in utero or in early life (Chiesa Fuxench et al., 2023), supporting a possible role of the microbial exposures in AD. This review focuses on the skin microbiome in AD and how the skin microbiome relates to host immunity and the physical barrier function of the skin.

Figure 1. Interplay among skin microbiome, skin barrier, and immune system in AD.

AD is characterized by skin microbiome shifts, e.g. increase of Staphylococcus aureus and Staphylococcus epidermidis during flares. Expansion of S. aureus is associated with higher disease severity. Skin barrier dysfunction observed in AD can be due to genetic factors like mutations in barrier proteins, cell adhesion defects, or skin lipid alteration. Barrier dysfunction can be both exacerbated by and contribute to shifts in the skin microbiome. The immune system engages in a dynamic dialogue with the microbiome, starting with early life microbial exposures, which are key to developing immune tolerance. Skin microbes can modulate the immune response, engaging in interactions that can promote or dampen skin inflammation.

The skin microbiome in health and AD

Skin plays a crucial role as a physical barrier – preventing water loss and protecting from environmental exposures – and is comprised of varied ecological niches (Marples, 1965). These different physiologic microenvironments including sebaceous, moist, and dry skin host distinct microbiomes (Costello et al., 2009, Grice et al., 2009, Oh et al., 2014, Saheb Kashaf et al., 2022). Sebaceous sites (e.g. forehead) support lipophilic microbes such as Cutibacterium acnes and Malassezia species, which have lipase genes necessary to metabolize skin lipids; moist sites rich in eccrine glands support halotolerant Staphylococcus spp. including S. epidermidis whose sphingomyelinase can facilitate host production of protective ceramides (Zheng et al., 2022). Viral colonization with eukaryotic DNA viruses tends to be highly individualized versus anatomically-specific (Oh et al., 2016). Studying the skin microbiome can be challenging due to the skin's low microbial biomass, which increases the risk of contamination. Moreover, variability in sampling, storage, and processing methods as well as different exposures can introduce confounders affecting study findings (Kong et al., 2017).

The skin microbiome, which is relatively stable in healthy individuals (Oh et al., 2016), exhibits variability in AD patients with staphylococcal species expansion associated with some disease flares. Culture-based studies have linked S. aureus with AD skin lesions, even without clinical signs of infection (Leyden et al., 1974). S. aureus colonization rates in AD vary across study populations with a meta-analysis reporting colonization in 70% (95% CI 66-74) of affected skin and 39% (95% CI 31-47) of unaffected skin (Totté et al., 2016). Next generation sequencing technologies (16S rDNA and shotgun metagenomic sequencing) of AD patient skin samples have demonstrated a relative expansion of staphylococci: S. aureus and coagulase-negative staphylococci (CoNS) (Byrd et al., 2017, Fyhrquist et al., 2019, Kong et al., 2012, Saheb Kashaf et al., 2023, Seite et al., 2014). AD microbiome composition more closely resembled that of healthy subjects after skin-directed treatment (Kong et al., 2012). In addition to S. aureus, other microbes such as Malassezia spp. (Sparber et al., 2019) have been found to promote inflammation, with ongoing debate regarding their contribution to disease (Casagrande et al., 2006, Chng et al., 2016, Darabi et al., 2009, Hiragun et al., 2013).

The skin microbiome also varies by age. Infant skin microbiomes are initially influenced by the mode of delivery (Dominguez-Bello et al., 2010) but demonstrate body site-specificity independent of delivery method by six weeks of age (Chu et al., 2017). In infant AD studies (culture-based and microbiome-based), higher levels of skin commensal staphylococci at 2-3 months of age were associated with lower likelihood of AD within the first two years of life (Kennedy et al., 2017, Meylan et al., 2017); however, whether these bacteria are early signs of, or have any relevance to, eventual health status remains unknown. As children grow, skin microbiomes continue to diversify (Capone et al., 2011, Rapin et al., 2023), with a marked change in composition during puberty (Park et al., 2022). Both children and adult/teenagers with AD have similar puberty-related shifts in their microbiome, specifically the association of skin Streptococcus with younger children and Cutibacterium and Corynebacterium with adult-teenagers (Shi et al., 2016).

Earlier studies suggest that S. aureus may not merely be a bystander but potentially involved in disease pathogenesis. In line with positive correlations between S. aureus by routine culturing and increased AD severity (Simpson et al., 2018), higher S. aureus relative abundances in affected skin was associated with greater AD severity, using Scoring Atopic Dermatitis (SCORAD) (Kong et al., 2012). Of note, the AD skin microbiome can be heterogeneous (Tay et al., 2021) and not all AD patients show an expansion of staphylococci which may reflect different subtypes of the disease (Byrd et al., 2017, Kashaf et al., 2023).

Staphylococcal strains associated with AD

Distinct S. aureus strains have been identified in AD. AD-associated S. aureus in a Spanish cohort were enriched for clonal complexes (CCs) CC5, CC15, CC30, and CC45 versus CC30 in atopic controls with food allergies and asthma (Rojo et al., 2014). In AD children in the United Kingdom, S. aureus CC1 predominated, and S. aureus CC30 and CC45 colonized healthy controls (Harkins et al., 2018). Furthermore, S. aureus CC1 strains have been associated with higher AD severity (Yeung et al., 2011). While many AD patients harbored the same S. aureus types over time, changes in S. aureus CCs in some patients were associated with higher SCORAD (Clausen et al., 2019).

In a global analysis of S. aureus isolates from AD patients and controls, some sequence types (STs), including ST1 (CC1), were significantly enriched in AD (Saheb Kashaf et al., 2023). While ST1 was virtually absent in the United States, the majority of ST1 isolates contained the chromosomally encoded fusC gene – known to confer fusidic acid resistance – suggesting topical fusidic acid (common in Europe and Asia but not approved in the United States) contributes to the relative expansion of ST1, as previously reported (Carter et al., 2018, Edslev et al., 2018). Sharing with household contacts, e.g. caregivers (Chia et al., 2022) and siblings (Saheb Kashaf et al., 2023), also appears to shape AD-associated staphylococcal strains. Tracking staphylococcal mutations over multiple visits showed strains could be shared and stably retained for 2-4 years (Saheb Kashaf et al., 2023).

Studies comparing S. aureus strains in AD and controls have proposed various virulence factors in disease pathogenesis. Adhesion experiments showed higher ClfB ligand binding activity in AD-associated skin strains, particularly CC1, than health-associated nasal carriage strains (Fleury et al., 2017), which corresponds with microbial genome-wide association analysis identifying skin adhesion genes, e.g. sdrC/D/E and clfA/B (Saheb Kashaf et al., 2023). In epicutaneous mouse models studying S. aureus toxins, phenol-soluble modulin α (PSMα) promoted IL-17-associated skin inflammation (Liu et al., 2017, Nakagawa et al., 2017, Williams et al., 2019), and δ-toxin elicited allergic inflammation via mast cell degranulation (Nakamura et al., 2013). S. aureus can also produce proteases, including V8, that can compromise the skin barrier integrity. The serine protease V8 was shown to directly activate pruriceptor sensory neurons and, in mouse models, induced itch and subsequent scratch-induced skin damage (Deng et al., 2023). Regulators of the Agr pathway, such as sarU and sarT, help regulate α-hemolysin, a toxin associated with AD (Brauweiler et al., 2014, Niebuhr et al., 2011, Wichmann et al., 2009). Additional studies have explored the role of the S. aureus capsule (Key et al., 2023) and biofilm (Di Domenico et al., 2018, Gonzalez et al., 2017). Collectively, these studies highlight the multifaceted mechanisms through which S. aureus strains may contribute to AD pathogenesis, from disrupting the skin barrier to promoting inflammation.

While commensal staphylococci are typically considered beneficial, some AD flares have relative increases in CoNS including the pathobiont S. epidermidis, which may suggest a potential contribution to skin disease or a milieu conducive to general staphylococcal growth. For example, S. epidermidis that produces protease EcpA is associated with an inflammatory response in in vitro and mouse models (Cau et al., 2021). Knocking out PSMs and protease EcpA in S. epidermidis resulted in less mouse skin inflammation (Williams et al., 2023). These studies underscore CoNS strain diversity.

Skin microbiome-immune system interactions in AD

Many mouse studies have been critical in elucidating microbiota-immune cross talk in skin, including microbial tuning of skin immunity, host immunity shaping of skin microbiota, and microbial modulating of skin inflammation (Kobayashi et al., 2019, Linehan et al., 2018, Ridaura et al., 2018, Wang et al., 2017). For example, mouse skin commensals can influence local effector T cell function through activation of the IL-1 signaling pathway as evidenced by studies in germ-free mice, which showed reduced cutaneous cytokine production and weakened immune response after parasitic infection that was restored following topical introduction of human skin commensal S. epidermidis (Naik et al. 2012). In addition, lipoteichoic acid (LTA) from S. epidermidis can suppress TLR3-associated local skin inflammation in a mouse injury model through a mechanism dependent on TLR2 (Lai et al. 2009) and, when injected in mouse skin, can induce TLR2 activation in mast cells, leading to enhanced production of the antimicrobial peptide (AMP) cathelicidin and protection from vaccinia virus (Wang et al., 2012).

Since AD patients are frequently colonized with S. aureus, which is associated with higher disease severity, various models have been used to study the role of skin microbes in promoting the dysregulated immune response in AD (Kim et al., 2019). In conventional mice, S. aureus strains derived from AD patients induced increased epidermal thickness and TH2 and TH17-associated cells in contrast to S. aureus from healthy individuals or USA300 (a community-acquired methicillin-resistant S. aureus), suggesting clinical S. aureus strains could elicit differential skin responses (Byrd et al., 2017). In a Adam17^fl/flSox9-^Cre mouse model which is characterized by antibiotic-responsive dermatitis, eczematous skin disease was associated with Corynebacterium mastitidis, S. aureus, and Corynebacterium bovis (Kobayashi et al., 2015). When ADAM17-deficient mice were re-exposed to S. aureus, eczema scores were more severe; C. bovis re-exposure was associated with a TH2 immune response, suggesting different bacteria in an AD mouse model led to distinct host responses. In topical MC903 and OVA-sensitized mouse models, TH2 inflammation was associated with reduction in antibiotic-producing CoNS and increases in non-antibiotic-producing CoNS and S. aureus, similar to microbiome changes observed in AD patients; however, TH17 inflammatory conditions featured higher AMP expression which limited growth of S. aureus (Nakatsuji et al., 2023). Additionally, IL-17RA deficient mice exhibited skin inflammation with eosinophils and TH2 cytokines, a defective skin barrier, and shifts in the skin microbiome (Floudas et al., 2017). JunB-deficient mice also developed AD-like features, including TH2/TH17 immune responses, barrier dysfunction, and an increase in S. aureus (Uluçkan et al., 2019). Prophylactic antibiotics prior to the development of microbial shifts in JunB-deficient mice reduced IL-17A expression, further supporting a role of the microbiome in shaping the immune response.

The microbiome may also produce metabolites that can dampen the immune response. Targeted metabolomics of AD skin showed lower levels of microbial Trp metabolites including Indole-3-aldehyde, which reduced mouse skin inflammation via aryl hydrocarbon receptor (AHR) activation (Yu et al., 2019). A different mouse dermatitis model (Tmem79^−/− mice) showed S. epidermidis and certain strains of S. cohnii reduced skin inflammation potentially via immunoregulatory and glucocorticoid-related pathways (Ito et al., 2021). Several clinical studies using targeted biological therapeutics (dupilumab and trakolinumab) have demonstrated combined reductions in disease severity and S. aureus (Beck et al., 2023, Simpson et al., 2023). However, it is unclear whether observed microbial changes may be a bystander effect from the resolving skin disease itself or may be directly related to the immune-based therapies.

Early life exposures continue to be explored in the context of AD development. The skin microbiome is thought to play an important role in fostering immune tolerance during early life to distinguish between beneficial microbes and harmful pathogens. For example, neonatal mouse exposure to the human commensal S. epidermidis promoted the development of S. epidermidis-specific Tregs which mitigated inflammation in abraded skin during later exposures, whereas initial S. epidermidis exposure in adulthood led to an inflammatory response in mice driven by effector CD4+ T cells upon re-exposure (Scharschmidt et al., 2015). The higher concentrations of Tregs in neonatal mouse skin, especially in hair follicles, were abrogated in mice without fully developed follicles or commensal microbes, highlighting the importance of hair follicles (Scharschmidt et al., 2017). Thus, the importance of timing of microbial exposures in these mouse studies lend support for epidemiological studies linking microbe-related exposures in early life with the inverse risk of developing AD.

Interestingly, eczematous dermatitis is a feature of several inborn errors of immunity (IEI), which could provide further insights into microbial-immune contributions to the pathogenesis of eczematous skin disease. For example, IPEX syndrome patients (FOXP3) have absent or dysfunctional Tregs and can present with eczema in early life. Both Wiskott-Aldrich syndrome (WAS) and dedicator of cytokinesis 8 (DOCK8) deficiency have abnormalities in cytoskeletal functioning, and autosomal dominant hyper-IgE syndrome (STAT3-HIES) stems from altered cellular signaling. The skin microbiomes in STAT3-HIES patients showed relative expansions of Staphylococcus and Corynebacterium spp. as well as presence of bacteria and fungi not typically found on healthy controls (Oh et al., 2013). Additionally, eczema in STAT-HIES improves clinically with therapies targeting Staphylococcus aureus. Microbial alterations in some IEI cohorts include decreased stability of the microbiome over time, reduced skin site-specific patterning, increased fungal richness and abundance of potentially opportunistic fungi, and increased atypical bacteria (e.g. Serratia marcescens in STAT-HIES). DNA viruses, including human polyomaviruses, human papillomaviruses, and molluscum contagiosum, were found in higher relative abundances on the skin of DOCK8-deficient patients as compared to healthy controls, underscoring the important role of host immunity in shaping the skin microbiome (Tirosh et al., 2018).

The microbiome and skin barrier function in AD

Skin barrier dysfunction with increased transepidermal water loss is characteristic of AD and has been associated with genetic factors, immune dysregulation, and shifts in the skin microbiome. Mutations in the skin epidermal protein filaggrin (FLG) have been associated with AD (Irvine et al., 2011, Palmer et al., 2006, Sandilands et al., 2007), though only a fraction of individuals with FLG-null mutations develop AD (Morar et al., 2007). Mouse models with complete loss (Flg^−/−) of filaggrin exhibited increased allergen skin penetration and a heightened inflammatory response to environmental exposures but do not have spontaneous dermatitis (Kawasaki et al., 2012). Flg^−/− mouse pups exposed to the human commensal bacterium S. epidermidis demonstrated a persistent effector T cell (TH17) response instead of the Treg response in wild-type mice (Gonzalez et al., 2023). Flaky tail Flg^ft/ft mice (with reduced filaggrin and the matted phenotype) also exhibited increased allergen skin penetration and a heightened inflammatory response to environmental exposures (Fallon et al., 2009, Scharschmidt et al., 2009). Consequently, pre-existing disrupted skin barrier dysfunction involving filaggrin – or other skin barrier proteins (e.g. envoplakin, periplakin, and involucrin) (Natsuga et al., 2016) – might increase exposure to bacteria and their products, driving a dysregulated inflammatory response associated with AD.

The skin lipid layer is another important component of the skin barrier that can be dysfunctional in AD. AD patients have been shown to have skin lipid alterations versus controls (Emmert et al., 2021). Increases in certain ceramides, sphingomyelins, and lysophosphatidylcholines in lesional AD skin as compared to nonlesional and control subject skin were associated with increased transepidermal water loss and an increase in abundances of staphylococci(Kim et al., 2023). Tmem79^−/− mice lack the gene responsible for the matted phenotype(Saunders et al., 2013); the initial dermatitis in Tmem79^−/− mice exhibited altered lipid metabolism in their sebaceous glands and a TH17 immune response, whereas the second-phase dermatitis was microbiota-dependent (Morimoto et al., 2022).

Skin barrier function, microbiota, and immunity are deeply interwoven with complex interactions. In AD skin biopsies and in vitro studies, TH2 cytokines, such as interleukin-4, interleukin-13 (Howell et al., 2009), and interleukin-25 (Hvid et al., 2011), were associated with decreased filaggrin and suppressed AMP expression (Howell et al., 2006). A Flg^ft/ft skin injury model demonstrated that altering microbiome composition significantly decreased IL-1α expression in keratinocytes and subsequent chronic skin inflammation (Archer et al., 2019). Additionally, Netherton syndrome patients who have protease inhibitor LEKTI deficiency had a predominance of S. aureus and S. epidermidis on skin; in mouse models, Netherton syndrome-associated staphylococcal strains elicited skin inflammation, and bacterial products led to increased skin protease activity and skin barrier disruption (Williams et al., 2020).

Microbes may also contribute to strengthening the skin barrier. Ceramides are important skin barrier lipids which can be produced by sphingomyelinase cleavage of membrane lipids. S. epidermidis-produced sphingomyelinase was both non-cytotoxic in vitro in contrast to pathogen-produced sphingomyelinase and capable of increasing skin ceramides in a mouse model (Zheng et al., 2022).

Skin cell adhesion is also critical for barrier function. For example, DSG1 mutations have been associated with severe dermatitis, along with multiple allergies, and metabolic syndrome (SAM syndrome)(Samuelov et al., 2013), and loss of corneodesmosin can lead to eczematous dermatitis, allergies, and predisposition to S. aureus infection (Oji et al., 2010). In vitro human keratinocyte studies showed exposure to S. aureus peptidoglycan strengthened tight junctions via TLR2 signaling, suggesting the skin barrier may adjust defenses in response to pathogens (Yuki et al., 2011). Furthermore, in contrast to standard laboratory mice, germ-free mice and K14^CreAhr^f/f mice had impaired skin barrier function restored with an AHR agonist and with a human skin commensal consortium, suggesting that microbiota-AHR interaction is important for skin barrier function (Uberoi et al., 2021). The skin microbiome, thus, significantly influences the epidermal barrier's formation, maintenance, and repair by orchestrating cellular changes.

Commensal microbiota may also act as a barrier, providing “colonization resistance” by inhibiting growth of pathogens. High-throughput screening of skin bacteria revealed that CoNS with antimicrobial activity were more common in healthy individuals than AD and S. aureus-colonized subjects (Nakatsuji et al., 2017). The autoinducing peptides from these different CoNS suppressed S. aureus accessory gene regulator (agr) activity and decreased PSMα expression; however, ratios of S. aureus and autoinducing peptides shifted during AD flares, likely diminishing effectiveness of CoNS peptides. Thus, skin microbiota may bolster the physical barrier protection via pathogen resistance.

Microbial-based therapeutic approaches in AD

Given AD-associated staphylococcal shifts, microbially-focused therapeutics have been investigated. Most AD treatments target skin barrier dysfunction (e.g. emollients) and inflammation or the immune system (e.g. topical steroids, biologics) with potential indirect effects on the skin microbiome. This section focuses on therapies that aim to directly target the skin microbiome. Therapies focused on S. aureus elimination include topical antimicrobials (Bath-Hextall et al., 2010), systemic antibiotics, and dilute bleach (sodium hypochlorite) baths (DBB) (Chopra et al., 2017, Hon et al., 2016, Khadka et al., 2021). The inconsistent reductions of S. aureus with DBB have raised questions about DBB’s antimicrobial effects; interestingly, topical hypochlorite decreased skin NF-kB signaling in in vitro and mouse studies, raising the concept of non-microbial effects of DBB on skin. Twice-daily topical niclosamide (ATx201) 2% ointment for seven days significantly reduced S. aureus colonization in AD patients but did not significantly improve clinical symptoms (Weiss et al., 2022). A placebo-controlled phase 2 trial of 28-day topical antimicrobial peptide omiganan decreased Staphylococcus species but failed to significantly alleviate clinical symptoms or inflammation markers (Niemeyer-van der Kolk et al., 2022). These two studies’ reduction of staphylococcal levels without significant clinical improvements suggests that targeting Staphylococcus predominance alone in AD may be insufficient. If staphylococci indeed contribute to AD flares, microbially-focused therapies may help reduce frequency of exacerbations but may be inadequate for controlling existing or worsening inflammation.

Bacteriotherapy approaches have also been tested in AD. In studying the concept that reduction of commensal staphylococci contributes to AD, a human skin commensal Staphylococcus hominis A9 (ShA9) reduced S. aureus and inflammatory toxin expression in mice (Nakatsuji Teruaki et al., 2021). The double-blind, vehicle-controlled, phase 1 clinical trial with topical ShA9 was safe and reduced S. aureus colony-forming units but did not reduce AD severity or clinical symptoms in the treated patients. Results of the phase 2 open-label trial are pending. An open-label phase 1/2 study of twice-weekly topical Roseomonas mucosa for six weeks reported reduction in localized disease in AD patients (Myles et al., 2018); however, the phase 2 study did not meet its endpoint. In addition, multiple investigations have delved into the effectiveness of oral probiotics for AD, but the results have been mixed. Since skin microbiomes are highly individualized, future studies may need to consider personalized approaches (Nakatsuji T. et al., 2021). While microbially based therapies have been tried in various diseases, the most notable positive outcomes have been in infections: Clostridium difficile colitis and neonatal sepsis (Panigrahi et al., 2017). Since AD is inherently considered an inflammatory condition rather than an infection, it remains unclear whether microbiome-based approaches will be sufficiently effective in AD to control clinical skin inflammation and symptoms.

Conclusions

AD is a complex, heterogeneous condition marked by impaired epidermal barrier function and aberrant immunity. In health, the skin hosts different microbial communities that engage in a continuous crosstalk with both the skin barrier and immune system. In AD, shifts in the skin's microbial composition are observed in disease flares, most commonly featuring predominance of S. aureus. Notably, AD patients harbor distinct S. aureus strains that vary in genes linked to skin barrier disruption and immune dysregulation, such as virulence factors, adhesion proteins, and pro-inflammatory molecules. Correspondingly, the relationships between the skin microbiome, host immunity, and the skin barrier are quite complex in AD, and continue to be revealed. With great interest in the association between microbiome shifts and AD, microbiome-targeting therapeutic strategies have been investigated in AD. However, more research is needed to better understand the roles of skin microbiota in AD and how these may lead to potential treatments.

Abbreviations

AD

atopic dermatitis

AHR

aryl hydrocarbon receptor

AMP

antimicrobial peptide

CoNS

coagulase-negative staphylococci

CC

clonal complex

FLG

filaggrin

LTA

lipoteichoic acid

PSM

phenol-soluble modulin

SCORAD

Scoring Atopic Dermatitis

ST

sequence type

TH

T helper

Biographies