The power and potential of BIOMAP to elucidate host‐microbiome interplay in skin inflammatory diseases

The two most common chronic inflammatory skin diseases are atopic dermatitis (AD) and psoriasis. The underpinnings of the remarkable degree of clinical heterogeneity of AD and psoriasis are poorly understood and, as a consequence, disease onset and progression are unpredictable and the optimal type and time point for intervention are as yet unknown. The BIOMAP project is the first IMI (Innovative Medicines Initiative) project dedicated to investigating the causes and mechanisms of AD and psoriasis and to identify potential biomarkers responsible for the variation in disease outcome. The consortium includes 7 large pharmaceutical companies and 25 non‐industry partners including academia. Since there is mounting evidence supporting an important role for microbial exposures and our microbiota as factors mediating immune polarization and AD and psoriasis pathogenesis, an entire work package is dedicated to the investigation of skin and gut microbiome linked to AD or psoriasis. The large collaborative BIOMAP project will enable the integration of patient cohorts, data and knowledge in unprecedented proportions. The project has a unique opportunity with a potential to bridge and fill the gaps between current problems and solutions. This review highlights the power and potential of the BIOMAP project in the investigation of microbe‐host interplay in AD and psoriasis.


| Atopic dermatitis and psoriasis are the most common chronic inflammatory skin diseases
One of the greatest challenges that health systems will face globally in the twenty-first century is the increasing burden of chronic noncommunicable diseases. 1 The skin is an organ often affected by chronic conditions, in particular inflammatory immune-mediated diseases, either as the primary target or through secondary manifestations. The two most common chronic inflammatory skin diseases are atopic dermatitis (AD) and psoriasis. 2,3 Data from the WHO Global Burden of Diseases initiative indicate that at least 230 and 125 million people worldwide have AD and psoriasis (lifetime prevalence 10-15% and 2-3%, respectively 4 ), with AD being the leading cause of the non-fatal disease burden conferred by skin conditions. 5 At the patient level, both AD and psoriasis have diverse and marked negative impacts on quality of life (QoL) and place a tremendous financial burden on patients and also on healthcare providers. 6,7 AD and psoriasis are associated with a strongly increased risk of comorbidities. Up to one-third of patients with AD suffer from comorbid atopic diseases, such as food allergy, rhinitis and/or asthma, 8 and up to 20% of psoriasis patients are affected by psoriatic arthritis. 9 Inflammatory bowel disease, rheumatoid arthritis, cardiometabolic traits and neuropsychiatric conditions have also been linked with both AD and psoriasis. [10][11][12][13][14] AD can manifest at any point in life but the incidence peaks in early infancy, around age 2 years. 15 After onset, the course may be continuous for long periods, but may also show a relapsing-remitting nature. 16,17 Conventional clinical teaching is that AD clears in more than 50% of affected children, but recent data indicate that the proportion of patients with persistent or adult-onset disease, or with relapses after longer asymptomatic intervals, is much higher than previously thought. 18,19 Psoriasis can also manifest at any age, but onset most commonly occurs between 18 and 39 years and between 50 and 69 years of age. 20 Its natural course is highly variable, but there is little robust epidemiological data on patient trajectories. Both AD and psoriasis are based on a strong inherited predisposition and triggered by environmental factors, ultimately leading to epidermal barrier deficiency and excessive T-cell activation; however, the underlying T-cell polarization is different. Psoriasis is largely driven by Th17 T cells and associated with type 17 responses, and severe disease can effectively be controlled in most patients by blocking the IL-23/Th17 T-cell axis. 21 However, there is significant inter-patient heterogeneity in efficacy and adverse effects to respective biologics, and up to 35% of patients fail their first biologic therapy, 22 possibly reflecting the heterogeneity of the disease. AD has a strong Th2 component but appears to involve multiple immune pathways that might create different disease features. 23,24 As in many other common chronic inflammatory diseases, the underpinnings of the remarkable degree of clinical heterogeneity of AD and psoriasis are poorly understood and, as a consequence, disease onset and progression are unpredictable and the optimal type and time point for intervention are as yet unknown. 3 Thus, the delineation of disease subtypes and their mechanistic basis and molecular signatures, and biomarkers capable of assessing disease-related individual patient trajectories and response to different therapies are key unmet needs. 3 Ideally, current classifications would be replaced with an aetiology-based taxonomy that can be coupled with effective and safe treatment regimens for AD and psoriasis. Major technological advances in recent years include high-resolution 'omics' assay technologies that enable large-scale multidimensional molecular profiling across biological strata. The intelligent integration of 'omics' data with detailed clinical, environmental and lifestyle information has the potential to identify the molecular identity and subclasses of the underlying disease aetiologies. 25, 26 Yet generating,

Funding information
The research has received funding from the FP7 (MAARS-Grant 261366) and IMI2 (BIOMAP-Grant 821511). and progression are unpredictable and the optimal type and time point for intervention are as yet unknown. The BIOMAP project is the first IMI (Innovative Medicines Initiative) project dedicated to investigating the causes and mechanisms of AD and psoriasis and to identify potential biomarkers responsible for the variation in disease outcome. The consortium includes 7 large pharmaceutical companies and 25 nonindustry partners including academia. Since there is mounting evidence supporting an important role for microbial exposures and our microbiota as factors mediating immune polarization and AD and psoriasis pathogenesis, an entire work package is dedicated to the investigation of skin and gut microbiome linked to AD or psoriasis.
The large collaborative BIOMAP project will enable the integration of patient cohorts, data and knowledge in unprecedented proportions. The project has a unique opportunity with a potential to bridge and fill the gaps between current problems and solutions. This review highlights the power and potential of the BIOMAP project in the investigation of microbe-host interplay in AD and psoriasis.

K E Y W O R D S
atopic dermatitis, biomarkers, microbiome, psoriasis processing, distributing and utilizing sufficiently large, detailed, reliable and robust sample collections and data sets require complementary expertise and collaborative efforts. 27

| The human skin microbiome in AD and psoriasis
The human microbiome, that is the assemblage of microbial genomes on or in our bodies, plays an essential role in maintaining our health via crosstalk with the immune system. 28 Representing the largest organ of our body including various distinct physical and chemical niches, the skin presents a diverse environment for microbial growth.
Furthermore, being a protective barrier against the external environment, skin constantly receives microbial input from the surroundings, consequently hosting the most diverse microbial communities in the body. 29,30 In healthy skin under steady-state conditions, most microbes thrive as commensals and mutualists, hence interacting with dermal cells in a way that maintains homeostasis of cutaneous immunity. 28 An inadequate barrier function can result from endogenous factors such as filaggrin (FLG) loss-of-function mutations, 31,32 or local inflammation, or from exogenous factors such as bathing practices, and may allow for colonization by opportunistic microbes, triggering an undesirable immune activation. Perturbations in this host-microbe network alter both skin microbiome and immune functions. Whether such shifts are apparent even before disease initiation and can drive disease development have been examined only in small studies 33 but during inflammation, such as in AD, the composition of microbiome often shifts substantially. 34 Shifts in the microbiome composition in psoriasis have also been observed. 35

| Skin microbiome in AD
The skin microbiome in AD is characterized by increased abundance of Staphylococcus aureus (S. aureus) and reduced diversity of the commensal skin microbiome ( Figure 1A). Most AD subjects are colonized with S. aureus, compared with only 10% of healthy individuals, and the relative abundance of S. aureus correlates with disease flares and severity. 34,[36][37][38][39][40][41][42] S. aureus exacerbates AD through mechanisms that affect the epidermal skin barrier as well as cutaneous innate and adaptive immune responses. 43 Staphylococcal enterotoxins act as superantigens to activate polyclonal T-cell responses and can also act as allergens to stimulate IgE production. [44][45][46] Staphylococcal phenol soluble modulins (PSMs), such as δ-toxin which induces mast cell degranulation, and α-toxin which activates keratinocyte IL-1α and IL-36α production are also likely to drive inflammation in AD. 47,48 Skin microbial diversity is reduced in AD and diversity inversely correlates with disease severity. 34,37,41 Common skin microbiome members, including coagulase-negative staphylococci (CoNS) such as Staphylococcus epidermidis, may aid skin homeostasis and protect against the pathogenic effects of S. aureus. S. epidermidis has been shown to promote TLR2 signalling and antimicrobial peptide (AMP) expression in keratinocytes and also to induce PSM production, which inhibit growth of S. aureus in vitro. 49,50 Topical treatment with CoNS in AD resulted in decreased S. aureus colonization. 51 CoNS have also been shown to reduce S. aureus-driven skin inflammation by producing auto-inducing peptides that inhibit the S. aureus accessory gene regulatory quorum sensing system. This resulted in reduced expression of the S. aureus virulence factor PSMα in vitro and reduced S. aureus-induced skin barrier damage in mice. 52 Furthermore, other skin commensals including Cutibacterium acnes and the gram negative Roseomonas mucosa have also been shown to inhibit growth of S. aureus. [53][54][55] The homeostasis-inducing properties of these commensal species could potentially be harnessed therapeutically to reduce inflammation and treat AD in the future. The very first scaffold of the human skin microbiome is already set at birth and impacts health and disease via early influences on the developing immune system. 65,66 During puberty, the change in hormones and sebum expression in the skin leads to a profound change in the microbial composition of the skin, 67 which then remains largely stable during adulthood, despite lifelong exposure to strongly fluctuating environmental factors. 68 To better understand the fundamental forces that shape the healthy skin microbiota, several studies have investigated the impact of lifestyle and environmental factors on the microbial skin community in the general population. These studies have shown that the strongest influence on the skin microbiome stems from the local skin microenvironment, in particular determined by skin pH, skin hydration, sebum production and epidermal lipid content. 69,70 Additionally, links between the skin microbiome or its members and a variety of intrinsic and extrinsic factors have been observed, including age, 57,71,72 sex, 73,74 BMI, 75 use of cosmetic products, 76,77 exposure to antibiotics, 78 ethnicity and geographical region. [79][80][81] Moreover, variation in the skin microbiome was found to be associated with several environmental factors, including UV exposure, 82 exposure to domestic animals such as dogs, 83 contact with soil and plants, 84 and urbanization of place of residence. 57,85 While many of these results have been replicated in independent studies (eg age), our current understanding is still patchy, with few studies available that aimed to integrate many candidate factors. 71,74 1.3.2 | The effects of the environment on the microbiome and the skin-gut axis in AD and psoriasis There is increasing recognition of the global burden of both AD and psoriasis, with changing epidemiological patterns in high-and low-income countries. 86,87 Both diseases are more prevalent in high-income and in highly westernized countries, but the exact relationship between environmental risk factors and AD and psoriasis remains to be elucidated. Epidemiological studies have implicated hygiene-related factors, urbanization and climate, which are thought to reduce microbial biodiversity, and lifestyle factors, such as diet and obesity, alcohol, smoking and stress, which may impact chronic inflammation. 88 The International Study of Asthma and Allergies in Childhood (ISAAC) studies contributed significantly to our understanding of the global prevalence of AD, and the changing patterns amongst high-and low-income countries. [88][89][90][91] There is conflicting evidence regarding the geographic distribution of psoriasis with some studies reporting higher incidence and prevalence rates of psoriasis with increasing distance from the equator. 92,93 However, this relationship was not confirmed in a recent systematic review and metaanalysis, 87 which attributed increased frequency of psoriasis to higher income levels. Urbanization, air pollution and differences in climate and UV exposure are possible explanations for increased frequency of AD and psoriasis at higher latitudes. Lower UV exposure directly contributes to lower vitamin D levels, which may be relevant to AD and psoriasis given their associations with hypovitaminosis D. 94,95 Vitamin D affects the innate and adaptive immune system, antimicrobial defences and influences skin barrier function. 95,96 UV radiation has been shown to impact the skin microbiome in healthy volunteers, 82 and phototherapy modifies the skin microbiome in patients with psoriasis 97 and with AD. 98 Narrowband UVB and natural sunlight exposure on the skin may even modulate the gut microbiome. 99,100 The skin, gut and household microbiome varies amongst populations living in regions with the same latitude, but varying levels of urbanization. These changes, particularly changes in the mycobiome, are associated with availability of household cleaning products and dwelling type. 85 Further research integrating the impacts of the environment, including temperature, climate, pollution and urbanization, on the skin and gut microbiome and the relationship with AD and psoriasis is required.
In addition to environmental factors associated with urbanization, a Western lifestyle, diet and increasing obesity may play a role in AD and psoriasis. 101,102 Patients with psoriasis are at significantly increased risk of metabolic diseases including hyperlipidaemia, insulin resistance, obesity and the metabolic syndrome. [103][104][105] Mendelian randomization has shown that obesity plays a causative role in psoriasis. 106 The relationship between diet and the gut microbiome is bidirectional 107 : nutrient availability impacts the bacterial community structure and the metabolic effects of the gut microbiome influence the host's energy availability and alter the metabolome. 108 The health consequences of an obesity-associated gut microbiome have been discussed elsewhere, [109][110][111] and associations between the skin microbiome and obesity and diet have more recently been reported. 75,112 Whether the gut and/or skin microbiome have mediating, confounding or bystander roles in the relationship between psoriasis and/or atopic dermatitis and obesity remains to be fully elucidated.

| Genetics
The influence of human genetics on the skin microbiome is largely understudied, particularly at the general population level, and available AD studies have mainly focused on mutations in the skin barrier gene filaggrin (FLG). In a seminal work, Si et al. 113 found that the heritability of bacterial clades ranged from 40.9% to 56.4% in a study of 45 individuals, including twins. In addition, they found an association between a single nucleotide polymorphism (SNP) in the FLG gene when searching within a SNPs panel of skin-related genes. FLG encodes a structural protein essential for skin barrier function, 32 and its loss-of-function mutations are the strongest known genetic risk factors for AD 31,114 and the cause of ichthyosis vulgaris. 115 FLG mutations have been associated with distinct skin microbiome profiles of healthy individuals, which instead resembled microbiome profiles observed in AD patients. 116 Furthermore, FLG mutations were associated with Staphylococcus aureus colonization in AD patients 117 and microbial composition in patients' non-lesional skin. 118 Nevertheless, to the best of our knowledge, no systematic survey of possible influences of genes on the skin microbiome has been conducted on the general population nor on patients with AD or psoriasis. This is in strong contrast with the increasing number of genome-wide association studies conducted on gut microbial communities (mGWAS), which now include thousands of participants. 119,120 The interaction between host genetics and gut microbiomes found by mGWAS studies suggest that such interactions may exist for the skin microbiome.
Further findings are, however, more suggestive of a potential impact of host genetics on skin microbiome, such as microbiome-related gene expression profiles of psoriasis patients and AD patients, [121][122][123] the association of skin bacterial communities with ethnicity, 81 and the small proportion of microbial community variation explained by individual, lifestyle and environmental factors combined (around 15%). 76

| Disease initiation and early AD development
The healthy skin microbiota changes considerably throughout life, with Staphylococcus and Streptococcus dominating in infancy, while Cutibacterium and Corynebacterium are more abundant in adulthood ( Figure 2A). 67,70,124,125 Interestingly, the prevalence of AD is highest in the first years of life, with a considerable decline around school age 126,127 (Figure 2B), and while AD can resolve in some cases, for others it becomes a lifelong condition. 86 Interestingly, the skin microbiota in young children is also quite different from that of older children and adults. 67,70,124,125 This could suggest an agespecific skin dysbiosis in infant lesional skin ( Figure 2C), a hypothesis that is supported by a few studies, each with low numbers of participants. 122,128 Skin barrier dysfunction, including that caused by filaggrin mutations, is associated with immunological Th2 skewing, 86 but this relationship is bidirectional as Th2 inflammatory cytokines (such as IL-4, IL-13 and IL-33) can directly disrupt the skin barrier, through alterations in filaggrin breakdown products and stratum corneum lipid mediators. 129 Mode of delivery may exert a small influence on the skin microbiome at birth, but this influence appears to be short-lived. 65

F I G U R E 2
The development of the skin bacterial community structure and the nature of perturbations during AD lesions in different developmental periods. A, Schematic view of the healthy developing skin microbiota, shown through the relative abundance of the four most prevalent bacterial genera present on the skin during infancy, childhood and adulthood. B, Schematic view of the prevalence of AD during infancy, childhood and adulthood, highlighting that the major disease burden of AD occurs in early life. C, The red box marks the age-specific dysbiosis associated with lesional AD skin, resulting in higher (up-arrow) or lower (down-arrow) relative abundances of certain bacterial genera. The nature of this dysbiosis in infancy is largely unknown and is therefore marked with a question mark with the initiation and early development of AD and other atopic diseases remains to be defined. It is plausible that there is crosstalk between the infant gut and skin microbiota and the developing immune system.

| Gene-microbiome networks underlying cutaneous inflammation
Abnormal host-microbe interactions are associated with cutaneous disorders like AD and psoriasis, 34 [174][175][176] Streptococcal peptidoglycan (PG) was also proposed to participate in p by binding to innate immune receptors. 173 In addition, PGcontaining cells were detected to be increased in chronic plaque skin lesions and associated with PG-specific CD4+ T cells. Since there is mounting evidence supporting an important role for microbial exposures and our microbiota as factors mediating immune polarization and AD and psoriasis pathogenesis, flares and chronicity, an entire BIOMAP work package is dedicated to the investigation of skin and gut microbiome linked to AD or psoriasis. Microbiome research in this area has been more focused on AD but remains fairly limited for both diseases. In addition to microbial patterns and signatures associated with AD and psoriasis, BIOMAP investigates potential disease subtypes based on microbial heterogeneity. Furthermore, variability across time scales due to disease-and disease activityrelated changes, and host molecular constituents related to normal and pathological shifts will be dissected.

| Aims and perspectives of the BIOMAP project related to host-microbe interplay
Our efforts in investigating the causes and mechanisms of AD and psoriasis, in identifying biomarkers which may be responsible for the variable disease outcome, and in understanding the role of the human microbiota in disease pathogenesis are summarized in Figure 3. In the subsections below we list main aims of the BIOMAP project.

| Expand current knowledge regarding AD and psoriasis-associated microbiomes and their role in pathogenesis
To date, most microbiome studies have been relatively small, with cohorts that include different disease subtypes. The inherent heterogeneity of skin inflammation, disease diagnostic criteria and the skin microbiome makes it difficult to draw firm conclusions from these small studies on differences in the microbiome between health and disease.
Moreover, the lack of sufficiently sized longitudinal studies hampers the possibilities to gain insight into causality. In addition, significant variability in the methods used to study the skin microbiome has made comparing findings between studies difficult and limits the potential that can be learned from these studies. 180,181 The BIOMAP consortium has the opportunity to integrate information from several large paediatric and adult cohorts, while accounting for these sources of variation, to more precisely determine the microbiome in psoriasis and AD. 16S rRNA gene amplicon data from several BIOMAP cohorts will be harmonised, allowing standardization across studies for downstream analysis steps.
Moreover, the choice of 16S rRNA primer pair(s) will be carefully considered. 181 In parallel with 16S rRNA gene sequencing, the large MAARS cohort within BIOMAP uses whole metagenomic shotgun (WMS) sequencing to study the microbiome in AD and psoriasis. 122 The BIOMAP WMS data will provide species-and strain-level taxonomic information for eukaryotes, prokaryotes as well as viruses and enable the profiling of their functional potential. Constructing microbial genomes from WMS and contrasting their functional potential associated with diseased and healthy skin will provide us further mechanistic insights into how the skin microbiome may function in AD and psoriasis.

| Provide novel insights into disease initiation and early AD development
The dynamic changes of the skin microbiome during infancy, childhood and around puberty, followed by the relative stability 68 during F I G U R E 3 Potential outputs of the BIOMAP project. The IMI Consortium dedicates an entire work package to host-microbiome interplay in atopic dermatitis and psoriasis, with the potential to provide novel insight into disease mechanisms and classification, novel treatments and strategies for disease prevention. In a large-scale and wellcoordinated manner, BIOMAP partners bring together biological samples, clinical information and 'omics' data from existing patient collections to be integrated with skin and gut microbiomes. for AD. A collaborative approach, using standardized laboratory and analytical pipelines, will facilitate comparisons and replication of findings between these cohorts, aiming to identify possible microbial alterations that precede the development of AD. 138 We will also examine for any infant-specific dysbiosis of eczematous skin lesions ( Figure 2B,C), and finally, we will investigate whether there are specific bacterial biomarkers that predict disease persistence and/or severity in later life.
The dynamic changes of the skin microbiome during infancy, childhood and around puberty, followed by the relative stability 68 during adulthood, raise the possibility that perturbations of the early-life skin microbiome 68 could have long-lasting effects. A better understanding of the factors influencing the early-life skin microbiome may provide insights into the relationship between hygienerelated environmental exposures and the increasing global incidence of AD and allergic diseases, as well as guiding novel preventative and therapeutic strategies for AD.

| Explore pathomechanisms of host-microbe interplay in AD and psoriasis, using cutting-edge bioinformatics and omics technologies
The involvement of streptococci in psoriasis suggests that there could be the interplay between microbes and host genes in psoriatic skin. Moreover, WMS data have shown that S. epidermis strains specific to psoriasis lesions produced virulence factors in lesions but not in unaffected skin implying that there could be microbial participation in psoriasis skin beyond Streptococcus. 185 However, the knowledge on gene-microbe interactions in skin is still scarce. To our knowledge, only the MAARS cohort belonging to BIOMAP 122 has utilized the integration of psoriatic skin microbiome and transcriptome and found weak associations, and no study to date has integrated transcriptome and microbiome for non-lesional and lesional sites separately. Furthermore, there are no studies investigating interactions between host genes and microbial functional genes, which requires WMS. WMS is also required for detecting strain heterogeneity between lesional and non-lesional sites, recently proposed 185 and gene-microbe strain level associations.
The easy accessibility of skin makes it an excellent target for simultaneous sampling of the microbiome and host tissue samples from exactly the same anatomical location to explore the relationship between the skin microbiome, gene regulation and disease activity within the BIOMAP project. During the life course of BIOMAP, we will expand our focus from 16S rRNA gene amplicon sequencing to metagenomics, which is already available in some of the BIOMAP cohorts (eg MAARS) and integrate it with genetic and skin transcriptomics data. Taking advantage of the multiple data layers (eg microbiome, transcriptome and methylome), we have the possibility to address the interplay between specific microbes or their functional properties and host tissue responses in AD and psoriasis.  186 Although bacterial or fungal co-culture approaches seem rather straightforward, 187 the modelling of longterm interactions and intervention studies are limited by technical challenges, while donor-dependent differences limit the power to detect meaningful interactions. Therefore, standardized experimental models with defined genomic background amenable for genome editing, co-culture, omics sampling and longitudinal biophysical measurements are in high demand. The immortalized N/TERT keratinocytes could provide such a resource given their high similarity to primary keratinocytes 188 and accurate disease modelling using CRISPR-Cas technology. 189 Co-cultures of organotypic skin models with a selection of key microbial species and strains identified in multi-omics analysis within the BIOMAP framework can provide detailed insights into molecular interactions, and possibly guide further research into the 'homeostatic' skin microbiota.

| Conclusions
Despite the significant advance in our current understanding of the human skin microbiome and skin inflammatory diseases, many questions remain. What are the factors that ultimately shape the composition of the human microbiota? Is the composition of the human microbiota a cause or just a consequence of disease? Nevertheless, enormous progress has been made, and recent technical advances in the field of omics technologies combined with intelligent integration of the various layers of data will pave the way for further ground-breaking discoveries. The advent of a large collaborative project like BIOMAP will enable the integration of patient cohorts, data and knowledge in unprecedented proportions. Several challenges remain, however, including how to properly handle biological variability between individuals and over time, disease heterogeneity and various technical issues. The BIOMAP consortium constitutes a unique opportunity with a potential to bridge the gap between current problems and solutions, filling important gaps of knowledge.

CO N FLI C T O F I NTE R E S T
Authors have no conflict of interest.