BACKGROUND:The human skin offers diverse ecosystems for microbial symbionts. However, the factors shaping skin-microbiome interactions are still insufficiently characterized. This contrasts with the broader knowledge about factors influencing gut microbiota.OBJECTIVES:We aimed to investigate major patterns of association of host traits, lifestyle and environmental factors with skin bacteria in two German populations.METHODS:This is a cross-sectional study with 647 participants from two population-based German cohorts, PopGen (n = 294) and KORA FF4 (n = 353), totalling 1794 skin samples. The V1-V2 regions of the 16S ribosomal RNA (rRNA) gene were sequenced. Associations were tested with two bacterial levels, community (beta diversity) and 16S rRNA gene amplicon sequence variants (ASVs).RESULTS:We validated known associations of the skin microbiota with skin microenvironment, age, body mass index and sex. These factors were associated with beta diversity and abundance of ASVs in PopGen, which was largely replicated in KORA FF4. Most intriguingly, dietary macronutrients and total dietary energy were associated with several ASVs. ASVs were also associated with smoking, alcohol consumption, skin pH, skin type, transepidermal water loss, education and several environmental exposures, including hours spent outdoors. Associated ASVs included members of the genera Propionibacterium, Corynebacterium and Staphylococcus.CONCLUSIONS:We expand the current understanding of factors associated with the skin bacterial community. We show the association of diet with skin bacteria. Finally, we hypothesize that the skin microenvironment and host physiology would shape the skin bacterial community to a greater extent compared with a single skin physiological feature, lifestyle and environmental exposure.

BACKGROUND:Growing evidence suggests that atopic dermatitis (AD) is associated with an increased risk of depressive disorders and anxiety. However, existing studies were observational and may have uncovered correlations but could not easily disentangle noncausal or reverse-causal associations because these associations could be confounded and may not reflect true causal relationships.OBJECTIVES:To examine, in a two-sample Mendelian randomization study, the potential effect of AD on the risk of depressive disorders and anxiety.METHODS:Genetic instruments from the largest available genome-wide association study (GWAS) for AD (10 788 cases and 30 047 controls) were used to investigate the relationship to broad depression (170 756 cases and 329 443 controls), major depressive disorder (MDD; 30 603 cases and 143 916 controls) and anxiety (5580 cases and 11 730 controls). A set of complementary approaches were carried out to assess horizontal pleiotropy and related potential caveats occurring in MR studies.RESULTS:We observed no causal impact of AD on the risk of depressive disorders and anxiety, with close-to-zero effect estimates. The inverse weighted method revealed no associations of AD on broad depression [odds ratio (OR) 1·014; P = 0·431], probable MDD (OR 1·002; P = 0·568), International Classification of Diseases, Ninth/Tenth Revision-based MDD (OR 1·001; P = 0·466) or anxiety (OR 1·097; P = 0·180).CONCLUSIONS:This MR study does not support a causal effect of AD on depression and anxiety.

BACKGROUND:Genomic approaches have revealed characteristic site specificities of skin bacterial community structures. In addition, in children with atopic dermatitis (AD), characteristic shifts were described at creases and, in particular, during flares, which have been postulated to mirror disturbed skin barrier function, cutaneous inflammation, or both.OBJECTIVE:We sought to comprehensively analyze microbial configurations in patients with AD across body sites and to explore the effect of distinct abnormalities of epidermal barrier function.METHODS:The skin microbiome was determined by using bacterial 16S rRNA sequencing at 4 nonlesional body sites, as well as acute and chronic lesions of 10 patients with AD and 10 healthy control subjects matched for age, sex, and filaggrin (FLG) mutation status. Nonlesional sampling sites were characterized for skin physiology parameters, including chromatography-based lipid profiling.RESULTS:Epidermal lipid composition, in particular levels of long-chain unsaturated free fatty acids, strongly correlated with bacterial composition, in particular Propionibacteria and Corynebacteria abundance. AD displayed a distinct community structure, with increased abundance and altered composition of staphylococcal species across body sites, the strongest loss of diversity and increase in Staphylococcus aureus seen on chronic lesions, and a progressive shift from nonlesional skin to acute and chronic lesions. FLG-deficient skin showed a distinct microbiome composition resembling in part the AD-related pattern.CONCLUSION:Epidermal barrier integrity and function affect the skin microbiome composition. AD shows an altered microbial configuration across diverse body sites, which is most pronounced at sites of predilection and AD. Eczematous affection appears to be a more important determinant than body site.

Atopic dermatitis (AD) affects up to 20% of children and adults worldwide. To gain a better understanding of the pathophysiology of AD we conducted a large scale transcriptomic study of AD and psoriasis using 147 RNA-seq samples obtained from healthy, lesional and non-lesional skin. We found a striking correlation between dysregulated genes in lesional psoriasis and AD, with 81% of AD dysregulated genes also differentially expressed in psoriatic skin. We also demonstrated greater disease heterogeneity of AD compared to psoriasis, and described disease-specific molecular and cellular features. Notably, while lesional and uninvolved psoriatic skin had IL-17 dominance, lesional and uninvolved AD skin, had a striking dominance of IL-13/IL-10 pathways, with near undetectable IL4 expression, suggesting that IL-17/IL-13 define the polar opposites of these two diseases. This study is the first direct comparison between the AD and psoriasis transcriptomes using RNA-seq and demonstrates the highly discordant cytokine signatures of these two inflammatory skin diseases.

markdownabstractIn the version of this article initially published, in Fig. 3, the _y_-axis numbering did not match the log scale indicated in the axis label. The error has been corrected in the HTML and PDF version of the article.

Atopic dermatitis (AD) and psoriasis (PS) are common inflammatory skin diseases driven by an inherited predisposition and characterized by a disturbed epidermal architecture and keratinocyte differentiation as well as recurrent T-cell mediated cutaneous inflammation. However, the clinical phenotypes are mutually exclusive with opposing genetic effects in shared pathways and largely antagonistic immune deviations. Skin global gene expression profiling studies have indicated the presence of distinct gene expression signatures in both diseases, but in part suffered from methodological limitations. In the present study, we assessed the cutaneous nonlesional and lesional transcriptome in 22 AD patients, 25 PS patients and 36 healthy controls matched for age, sex and site of biopsy using next generation sequencing. Differential gene expression (DEG) analysis was done using DESeq2 adjusting for age, gender and experimental conditions. DEGs were defined by absolute log2 fold-change (FC) of above 1 and a Benjamini-Hochberg corrected p-value of below 0.05. The absolute number of detected genes ranged between 23.000 and 29.000. Nonlesional skin from both AD and PS showed little abnormalities and was largely comparable to healthy skin except for slight alterations related to epidermal differentiation especially in AD. In AD, we generally observed a considerably higher variability of the cutaneous expression pattern than for PS, potentially reflecting the more heterogenous phenotype. Apart from a characteristic shared inflammatory signature defined by 2690 overlapping DEGs we could identify disease specific signatures for AD (defined by a set of 784 DEGs) and PS (2119 DEGs). Other than previously reported, the transcriptome of acute and chronic AD lesions did not clearly separate.

Genomic approaches to characterize skin bacteria have revealed characteristic site-specificities. In addition, in children with AD, shifts of microbial community structures were described at creases and in particular during flares, which have been postulated to mirror the disturbed skin barrier function and/or cutaneous inflammation at these sites. But it is yet unclear whether they are restricted to predilection sites or a general feature of the skin of AD patients. Further, the impact of distinct abnormalities of epidermal barrier integrity and function such as an inherited filaggrin deficiency and of the acuteness of local inflammation on cutaneous microbial community structures have not been examined so far. The skin microbiome was determined by bacterial 16S rRNA sequencing at 4 different body sites of 10 AD patients and 10 healthy controls matched for age, sex and FLG mutation status. In addition, in AD patients acute and chronic eczema lesions were analysed. Independly from the presence of AD, filaggrin deficiency was associated with decrease of diversity and Proteobacteria abundance. Irrespective of the body site, patients with AD showed an increase of Staphilococci with a predominance of S. epidermidis, which was particularly marked in predilection sites. The cutaneous microbiome at acute and chronic lesions is highly similar independently from the body site. Compared to nonlesional skin, there were gradual shifts in diversity from acute to chronic lesions. S. aureus showed a higher abundance on chronic than on acute lesions. Distinct disturbances of epidermal barrier function such as filaggrin deficiency are associated with shifts of the microbial community composition characterized by a reduced diversity and Proteobacteria abundance. In AD, there is a generalized increase of Staphilococci also on non-predilection sites. The impact of inflammation, i.e. affection by eczema, appears to overlay locoregional influences on microbiome composition.

Epidemiological studies suggested an association between atopic dermatitis (AD) and cardiovascular disease. Therefore, we investigate associations and potential underlying pathways of AD and cardiovascular disease in large cohort studies: the AOK PLUS cohort (n = 1.2 Mio), the GINIplus/LISAplus birth cohorts (n = 2,286), and the Cooperative Health Research in the Region of Augsburg (KORA) F4 cohort (n = 2,990). In addition, metabolomics in KORA F4 and established cardiovascular risk loci in genome-wide data on 10,788 AD cases and 30,047 controls were analyzed. Longitudinal analysis of patients with AD in AOK PLUS showed slightly increased risk for incident angina pectoris (adjusted risk ratio 1.17 [95% confidence interval 1.12-1.23]), hypertension (1.04 [1.02-1.06]), and peripheral arterial disease (1.15 [1.11-1.19]) but not for myocardial infarction (1.05 [0.99-1.12]) and stroke (1.02 [0.98-1.07]). In KORA F4 and GINIplus/LISAplus, AD was not associated with cardiovascular risk factors and no differences in metabolite levels were detected. There was no robust evidence for shared genetic risk variants of AD and cardiovascular disease. This study indicates only a marginally increased risk for angina pectoris, hypertension, and peripheral arterial disease and no increased risk for myocardial infarction or stroke in patients with AD. Relevant associations of AD with cardiovascular risk factors reported in US populations could not be confirmed. Likewise, patients with AD did not have increased genetic risk factors for cardiovascular disease.

Atopic dermatitis (AD) has been recently reported to increase the risk for cardiovascular diseases in adults, as well as related risk factors in children. These findings were mainly reported for US populations. Therefore, in the present analysis, we aim to investigate the association of AD with cardiometabolic diseases and related risk factors using data of three German cohorts of children and adults: a prospective analysis using data of 1925696 subjects of the cohort of German National Health Insurance beneficiaries, a cross-sectional analysis of 2990 individuals of the KORA F4 cohort and a longitudinal analysis of 3665 participants of the 10 and 15 year follow-ups of the GINIplus and LISAplus birth cohort studies. None of the adjusted regression models yielded significant associations of AD with any of the tested cardiometabolic diseases or related risk factors. In a sensitivity analysis of the German National Health Insurance beneficiaries, after further adjustment for health care utilization behavior (i.e. the total number of physician contacts due to reasons other than AD), AD was associated with decreased risk of stroke, type 2 diabetes, hypertension, and higher risk of angina pectoris in the sub-group of AD patients with systemic medication. These findings were limited to patients more than 40 years old and, except for stroke, stronger in patients with more severe AD. When analysing the GINIplus and LISAplus cross-sectionally, a direct association between HDL and AD was observed at age 15 years. Our findings do not confirm the hypothesis that AD is a risk factor for cardiometabolic diseases and risk factors. To the contrary, sensitivity analyses indicate an inverse association, if any.

Skin global gene expression profiling has indicated the presence of distinct gene expression signatures in patients with inflammatory skin diseases such as atopic dermatitis and psoriasis. However, many studies suffer from small sample sizes, imperfect patient and biopsy matching, and limitations related to the use of microarray technologies. We here assessed the cutaneous transcriptomal architecture in 28 patients and 39 healthy controls matched for age, sex and site of biopsy using next generation sequencing generating ∼26 million (median) paired-end reads per sample. We performed a differential gene expression (DEG) analysis using DESeq2 adjusting for age, gender and experimental conditions. Functional annotation was performed using the R-package goseq and the Gene Ontology database. DEGs were defined by absolute log2 fold-change (FC) of above 1 and a Benjamini-Hochberg corrected p-value of below 0.05. The absolute number of detected genes ranged between 23.000 and 29.000. Digital image analysis was performed including total cell count, CD3, CD8 and neutrophil elastase expression in relation to epidermal and dermal area. First analyses identified 11 and 5554 transcripts differentially expressed in non-lesional psoriasis (PN) versus healthy skin (NN) and lesional psoriasis (PL) versus NN (11/2104 up- and 0/3450 downregulated), respectively. In non-lesional psoriatic skin, upregulated genes were related to epidermal differentiation and neutrophil aggregation. These genes were further strongly upregulated in PL with an approx. 5 fold log2FC. Compared to PN, in PL also an upregulation of multiple genes associated with Th17 inflammation and neutrophil chemotaxis and aggregation was observed. Currently pathway and gene co-expression network analysis as well as correlation analysis with tissue composition based on computer-aided quantitative analysis of digital images is being performed.

Association studies have identified a total of 23 European and Asian genetic susceptibility loci for atopic dermatitis (AD), although these explain only a small fraction of the estimated total heritability.1,2