Characterization of the skin microbiota in bullous pemphigoid patients and controls reveals novel microbial indicators of disease

Abstract

Introduction: Bullous pemphigoid (BP) is the most common autoimmune blistering disease. It predominately afflicts the elderly and is significantly associated with increased mortality. The observation of age-dependent changes in the skin microbiota as well as its involvement in other inflammatory skin disorders suggests that skin microbiota may play a role in the emergence of BP blistering. We hypothesize that changes in microbial diversity associated with BP might occur before the emergence of disease lesions, and thus could represent an early indicator of blistering risk.

Objectives: The present study aims to investigate potential relationships between skin microbiota and BP and elaborate on important changes in microbial diversity associated with blistering in BP.

Methods: The study consisted of an extensive sampling effort of the skin microbiota in patients with BP and age- and sex-matched controls to analyze whether intra-individual, body site, and/or geographical variation correlate with changes in skin microbial composition in BP and/or blistering status.

Results: We find significant differences in the skin microbiota of patients with BP compared to that of controls, and moreover that disease status rather than skin biogeography (body site) governs skin microbiota composition in patients with BP. Our data reveal a discernible transition between normal skin and the skin surrounding BP lesions, which is characterized by a loss of protective microbiota and an increase in sequences matching Staphylococcus aureus, a known inflammation-promoting species. Notably, Staphylococcus aureus is ubiquitously associated with BP disease status, regardless of the presence of blisters.

Conclusion: The present study suggests Staphylococcus aureus may be a key taxon associated with BP disease status. Importantly, we however find contrasting patterns in the relative abundances of Staphylococcus hominis and Staphylococcus aureus reliably discriminate between patients with BP and matched controls. This may serve as valuable information for assessing blistering risk and treatment outcomes in a clinical setting.

Keywords: 16s rRNA gene sequencing; Autoimmune blistering disease; Bullous pemphigoid; Risk factor; Skin microbiota.

Conflict of interest statement

Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Copyright © 2022. Production and hosting by Elsevier B.V.

Figures

Graphical abstract

Fig. 1

Sampling sites for patients with BP and matched controls; Box plots of Shannon (alpha) diversity. 1a. Grey figure represents age-and sex-matched control; orange figure represents a patient with BP. Sites rarely affected by BP include the forehead (purple), upper back (turquoise), and antecubital fossa (dark blue) are represented on both figures. An example perilesional sampling site (red), unaffected contralateral site (yellow) on the patient, and control-matched corresponding site (green) are shown. 1b. Shannon diversity at the ASV-level for sites rarely affected by BP for controls and patients. 1c. Shannon diversity at the ASV-level for patient perilesional, patient contralateral, and control corresponding sites. For box plots: Boxes represent interquartile range between first and third quartiles; horizontal line defines the median. Whiskers represent the 5th and 95th percentiles and values beyond these bounds are considered outliers, marked with black dots. Kruskal-Wallis test applied to analyze site variation. If an overall significant difference was observed, a pairwise Wilcox test was performed; p-values adjusted using the Benjamini-Hochberg method. Significance represented by: * ≤ 0.05; ** ≤ 0.01; *** ≤ 0.001; **** ≤ 0.0001; ns = not significant. Supplementary Table S6 reports summary statistics.

Fig. 2

Partial constrained principal coordinate analyses of Bray-Curtis 2a to 2c. Body sites rarely affected by BP. (anova.cca, Full model: p = 0.0009; terms: disease status, body site (constrained inertia = 5.04%, conditioned inertia = 4.5%), study center: p < 0.001; axes: CAP1, CAP2: p = 0.09; 1,000 permutations). 2d to 2f. Patient perilesional, patient contralateral sites, and control corresponding sites. (anova.cca, Full model: p < 0.001; terms: disease status, blistering status, study center: p < 0.001; axes: CAP1, CAP2: p = 0.009; 1,000 permutations; see “ecological and statistical analysis” in Methods). “+” represents centroid. SD: standard deviation. Site abbreviations: Budapest, Hungary (BUD); Düsseldorf, (D), Dresden, (DD), Freiburg, (FR), Lübeck, (HL), Homburg, (HOM), Kiel, (KI), München, (M), Würzburg (WUE), all Germany; Oulu, Finland (OU); Reims, (RE), Rouen, (RO), both France, Sofia, Bulgaria (SO), Thessaloniki, Greece, (TH).

Fig. 3

Bar plots of mean relative abundance for the ten most important indicator species. 3a. Bar plot showing relative abundance of important indicator species, at the ASV-level, for controls and patients with BP at sites rarely affected by BP [antecubital fossa (AF), forehead (FH), and upper back (UB)]. 3b. Bar plot showing the relative abundances of important indicator species at the ASV-level for patient perilesional, patient contralateral sites, and control-matched corresponding sites. RDP SeqMatch results for the representative ASV sequences are shown in the legend and provided in full in Supplementary Table S4. Supplementary Tables S7, S8 provide statistical parameters for indicator species analyses and summary statistics of all indicator ASVs, respectively.