Skin Microbiota Perturbations Are Distinct and Disease Severity–Dependent in Hidradenitis Suppurativa

Abstract

Hidradenitis suppurativa (HS) is a prevalent and debilitating inflammatory skin disease characterized by painful and recurrent nodules and abscesses, malodorous purulent drainage, and disfiguring sinus tract and scar formation involving intertriginous body sites. Microorganisms have been implicated in HS pathogenesis, and broad-spectrum antimicrobial therapy is one of the mainstays of HS management. However, bacteria have been identified in only ∼50% of HS lesions using conventional culture-based methods, and no consistent organism has been cultured from HS lesions (Brook and Frazier, 1999Brook I. Frazier E.H. Aerobic and anaerobic microbiology of axillary hidradenitis suppurativa.J Med Microbiol. 1999; 48: 103-105Crossref PubMed Scopus (59) Google Scholar, Gener et al., 2009Gener G. Canoui-Poitrine F. Revuz J.E. Faye O. Poli F. Gabison G. et al.Combination therapy with clindamycin and rifampicin for hidradenitis suppurativa: a series of 116 consecutive patients.Dermatology. 2009; 219: 148-154Crossref PubMed Scopus (197) Google Scholar, Jemec, 2003Jemec G.B. Hidradenitis suppurativa.J Cutan Med Surg. 2003; 7: 47-56Crossref PubMed Scopus (0) Google Scholar, Join-Lambert et al., 2011Join-Lambert O. Coignard H. Jais J.P. Guet-Revillet H. Poirée S. Fraitag S. et al.Efficacy of rifampin-moxifloxacin-metronidazole combination therapy in hidradenitis suppurativa.Dermatology. 2011; 222: 49-58Crossref PubMed Scopus (104) Google Scholar, Leach et al., 1979Leach R.D. Eykyn S.J. Phillips I. Corrin B. Taylor E.A. Anaerobic axillary abscess.Br Med J. 1979; 2: 5-7Crossref PubMed Scopus (54) Google Scholar). Recent studies have used a genomic approach to examine skin microbiota shifts in HS lesional skin and demonstrated relative abundances of Corynebacterium, Firmicutes, and anaerobes at HS lesional sites (Guet-Revillet et al., 2014Guet-Revillet H. Coignard-Biehler H. Jais J.P. Quesne G. Frapy E. Poirée S. et al.Bacterial pathogens associated with hidradenitis suppurativa, France.Emerg Infect Dis. 2014; 20: 1990-1998Crossref PubMed Scopus (93) Google Scholar, Ring et al., 2019Ring H.C. Sigsgaard V. Thorsen J. Fuursted K. Fabricius S. Saunte D.M. et al.The microbiome of tunnels in hidradenitis suppurativa patients.J Eur Academy Dermatol Venereol. 2019; 33: 1775-1780Crossref PubMed Scopus (34) Google Scholar, Ring et al., 2017Ring H.C. Thorsen J. Saunte D.M. Lilje B. Bay L. Riis P.T. et al.The follicular skin microbiome in patients with hidradenitis suppurativa and healthy controls.JAMA Dermatol. 2017; 153: 897-905Crossref PubMed Scopus (157) Google Scholar, Schneider et al., 2019Schneider A.M. Cook L.C. Zhan X. Banerjee K. Cong Z. Imamura-Kawasawa Y. et al.Loss of microbial diversity and body site heterogeneity in individuals with Hidradenitis suppurativa.2019: 612457Google Scholar). How these skin microbiota perturbations relate to overall disease severity, however, is not understood. In this pilot study, we investigated the relationship between overall disease severity and skin bacteria perturbations in HS at sites of HS predilection. Our findings demonstrate increased bacterial diversity in HS subjects as compared with healthy volunteers (HVs) at sites of disease predilection, and reveal broadening divergence of skin microbiota composition from HVs with worsening disease severity primarily driven by increased relative abundances of anaerobes and decreased relative abundances of major skin commensals. We recruited 12 adults with a physician-confirmed diagnosis of HS and age- and sex-matched HVs. All subjects provided written informed consent and underwent a skin preparatory regimen consisting of avoidance of detergents, cosmetics, oral and topical antibiotics, and bathing before swabbing. Skin swabs were collected from predetermined sites of HS predilection (axilla, gluteal crease, inguinal crease, and inframammary fold) with collection control swabs. DNA was extracted and amplified from the samples, and the V1–V3 hypervariable region of the 16S ribosomal RNA gene was sequenced to identify important bacterial species common to skin (Conlan et al., 2012Conlan S. Kong H.H. Segre J.A. Species-level analysis of DNA sequence data from the NIH Human Microbiome Project.PLOS ONE. 2012; 7e47075Crossref PubMed Scopus (112) Google Scholar, Meisel et al., 2016Meisel J.S. Hannigan G.D. Tyldsley A.S. SanMiguel A.J. Hodkinson B.P. Zheng Q. et al.Skin microbiome surveys are strongly influenced by experimental design.J Invest Dermatol. 2016; 136: 947-956Abstract Full Text Full Text PDF PubMed Scopus (179) Google Scholar). Negative controls were analyzed to monitor for contamination. A mothur-based pipeline was used for sequence analysis (mothur version 1.39.1; Free Software Foundation, Inc.). Nonparametric Kruskal-Wallis test followed by Wilcoxon test for paired comparisons was performed to examine differences in taxonomic relative abundances and diversity indices between affected HS skin, referred to as HS(A), unaffected HS skin, referred to as HS(U), and HV skin, with Bonferroni correction for multiple comparisons. Spearman correlation was used to assess for relationship between HS severity and taxonomic relative abundances. Statistical analyses were completed using R software (R version 3.5.2; R Foundation for Statistical Computing); P < 0.05 was considered significant. These studies were approved by the University of California, San Francisco Institutional Review Board (Protocol #16-19770). Study subjects represented diverse backgrounds and the range of HS disease severity (Table 1). To determine whether abundance patterns of bacterial communities differed between body sites of HS predilection, we evaluated genus-level taxonomic relative abundances in affected HS skin, unaffected HS skin, and HV skin (Figure 1a). We observed decreased relative abundances of commensal Cutibacterium spp. on the skin of patients with HS as compared with HVs (mean [95% confidence interval]; axilla: HS(A), 9.6 × 10-4%, [7.6 × 10-6%–0.0019%]; HS(U), 0.07% [−0.82% to 0.96%]; HV, 7.03% [−1.10% to 15.16%]; P = 0.017 for HS(A) vs HV, Wilcoxon test for paired comparisons; inguinal crease: HS(A), 0.15% [−0.041% to 0.35%]; HS(U), 0.12% [0.0045%–0.23%]; HV, 18.04% [−24.33% to 60.42%]; P = 0.044 for both HS(A) vs HV and HS(U) vs HV) (Figure 1b). In contrast, Gram negative anaerobes (inguinal crease: HS(A), 19.04% [8.66% to 29.41%]; HS(U), 15.64% [−19.15% to 50.43%]; HV, 0.44% [0.068% to 0.82%]; P = 0.0076 for HS(A) vs HV) (Figure 1c) and Gram positive anaerobes (inguinal crease: HS(A), 20.05% [7.43% to 32.67%]; HS(U), 8.06% [−1.17% to 17.29%]; HV, 1.89% [0.93% to 2.85%]; P = 0.0076 for HS(A) vs HV) (Figure 1d) predominated on the skin of HS subjects as compared with HVs. Bacterial diversity as measured by Shannon diversity, an ecological measure of both community richness and evenness, was significantly increased in the inguinal creases of HS subjects as compared with HVs (Shannon diversity index: HS(A) 2.03, HS(U) 1.22, and HV 0.91, P = 0.015 and 0.036 for HS(A) vs HV and HS(A) vs HS(U), respectively) (Figure 1e). These findings suggest that, even in the absence of clinical lesions, HS skin has a distinct microbiome.Table 1Clinical Characteristics of SubjectsCharacteristicHSHVTotal subjects125Age, mean (SD)33.1 (11.6)37 (9.7)Male:Female5:72:3Race/EthnicityB/NH 1W/H 3W/NH 6PI/NH 1Mixed race/NH 1A/NH 1B/NH 3W/NH 1Hurley StageStage 1: 2Stage 2: 8Stage 3: 2N/AAbbreviations: A, Asian; B, black; H, Hispanic; HS, hidradenitis suppurativa; HV, healthy volunteer; N/A, not applicable; NH, non-Hispanic; PI, Pacific Islander; W, white. Open table in a new tab Abbreviations: A, Asian; B, black; H, Hispanic; HS, hidradenitis suppurativa; HV, healthy volunteer; N/A, not applicable; NH, non-Hispanic; PI, Pacific Islander; W, white. To investigate similarities between samples, we performed principal coordinates analysis using taxonomic abundance-based Yue-Clayton theta (θ) similarity coefficient (Figure 1f) (Yue and Clayton, 2005Yue J.C. Clayton M.K. A similarity measure based on species proportions.Commun Stat Theor Methods. 2005; 34: 2123-2131Crossref Scopus (364) Google Scholar). Bacterial communities from adults with mild HS (Hurley Stage 1) clustered tightly with HVs, indicating higher similarity between skin bacterial communities in these subjects. Bacterial communities of adults with moderate and severe HS overall (Hurley stage 2 and 3, respectively) clustered separately from mild HS and HVs, with severe HS separating further from mild HS and HVs than moderate HS. Bacterial communities in patients with more severe HS scattered throughout both axes and clustered less tightly than patients with mild HS, suggesting higher within-group variation in more severe disease. Evaluation of the relative abundances of individual taxa by Hurley staging revealed that separation of samples by disease severity moderately positively correlated with relative abundances of mixed anaerobes (Spearman ρ = 0.54, P = 0.07) and moderately negatively correlated with relative abundances of major skin commensals (ρ = −0.54, P = 0.07), including commensal staphylococcal species and Cutibacterium and Corynebacterium spp. (Figure 1g and h, Supplementary Figure S1a–c), with marginal significance. Major skin commensals were defined as genera that were prevalent in more than 95% of individuals with >1% median relative abundance (Oh et al., 2016Oh J. Byrd A.L. Park M. Kong H.H. Segre J.A. NISC Comparative Sequencing ProgramTemporal stability of the human skin microbiome.Cell. 2016; 165: 854-866Abstract Full Text Full Text PDF PubMed Scopus (444) Google Scholar). These findings suggest that skin bacterial communities in HS are distinct, diverse, and divergent from those in HVs, with worsening disease severity associated with increasing skin bacterial perturbations in HS and higher within-group variation in more severe HS. Building upon prior reports of increased relative abundances of anaerobes in HS, we further define a positive association between relative abundances of anaerobes and disease severity in a diverse North American HS population. Our findings support previous reports of Porphyromonadaceae and Prevotellaceae in HS inguinal creases and extend the spectrum of obligate anaerobes described in HS to include Fusobacteria and Clostridiales (Ring et al., 2019Ring H.C. Sigsgaard V. Thorsen J. Fuursted K. Fabricius S. Saunte D.M. et al.The microbiome of tunnels in hidradenitis suppurativa patients.J Eur Academy Dermatol Venereol. 2019; 33: 1775-1780Crossref PubMed Scopus (34) Google Scholar, Ring et al., 2017Ring H.C. Thorsen J. Saunte D.M. Lilje B. Bay L. Riis P.T. et al.The follicular skin microbiome in patients with hidradenitis suppurativa and healthy controls.JAMA Dermatol. 2017; 153: 897-905Crossref PubMed Scopus (157) Google Scholar, Schneider et al., 2019Schneider A.M. Cook L.C. Zhan X. Banerjee K. Cong Z. Imamura-Kawasawa Y. et al.Loss of microbial diversity and body site heterogeneity in individuals with Hidradenitis suppurativa.2019: 612457Google Scholar). Colonization of the inguinal crease with these gut and vaginal commensal anaerobes and common wound colonizers is plausible, given the proximity to these anatomic sites and the chronic nature of HS lesions. Our findings also corroborate recent findings describing increased relative abundance of Porphyromonadaceae and Prevotellaceae in HS dermal tunnels, a characteristic of moderate and severe disease (Ring et al., 2019Ring H.C. Sigsgaard V. Thorsen J. Fuursted K. Fabricius S. Saunte D.M. et al.The microbiome of tunnels in hidradenitis suppurativa patients.J Eur Academy Dermatol Venereol. 2019; 33: 1775-1780Crossref PubMed Scopus (34) Google Scholar). Finally, the role of anaerobes in HS is supported by previous studies of effective antibiotic regimens which include anaerobic coverage (Gener et al., 2009Gener G. Canoui-Poitrine F. Revuz J.E. Faye O. Poli F. Gabison G. et al.Combination therapy with clindamycin and rifampicin for hidradenitis suppurativa: a series of 116 consecutive patients.Dermatology. 2009; 219: 148-154Crossref PubMed Scopus (197) Google Scholar, Join-Lambert et al., 2011Join-Lambert O. Coignard H. Jais J.P. Guet-Revillet H. Poirée S. Fraitag S. et al.Efficacy of rifampin-moxifloxacin-metronidazole combination therapy in hidradenitis suppurativa.Dermatology. 2011; 222: 49-58Crossref PubMed Scopus (104) Google Scholar). This study is limited by its cross-sectional design and limited sample size; a larger cohort will be required to validate these findings. We measured relative abundances of taxa to understand community structure and patterns in bacterial abundance; however, absolute abundances are needed to understand bacterial bioburden in future studies. Further studies are needed to investigate the role of skin microbiota in the pathogenesis and exacerbation of this debilitating condition. All sequencing data were deposited and are available at the NCBI Sequence Read Archive (SRA) under BioProject number PRJNA543491. Haley B. Naik: https://orcid.org/0000-0003-2760-3904 Jay-Hyun Jo: https://orcid.org/0000-0002-2039-3425 Maia Paul: https://orcid.org/0000-0002-6018-1558 Heidi H. Kong: https://orcid.org/0000-0003-4424-064X HBN has received grant support from Abbvie, consulting fees from 23andme, and is a board member of the Hidradenitis Suppurativa Foundation. This publication was supported by the National Center for Advancing Translational Sciences , National Institutes of Health , UCSF-CTSI Grant Number KL2 TR001870 (HBN), as well as the American Acne and Rosacea Society Research Grant (HBN), the American Skin Association Research Grant in Inflammatory Diseases (HBN), the Dermatology Foundation Medical Dermatology Career Development Award (HBN), and the Intramural Research Program of the National Institute of Arthritis and Musculoskeletal and Skin Diseases (J-HJ, HHK). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH. The authors thank Julia A. Segre and other members of the Segre and Kong labs for their underlying efforts, and the healthy volunteers and patients with hidradenitis suppurativa for their contributions. Conceptualization: HBN, HHK; Data Curation: HBN, J-HJ, MP; Formal Analysis: J-HJ, HHK; Funding Acquisition: HBN, HHK; Investigation: HBN, J-HJ, MP; Methodology: HBN, HHK; Project Administration: HBN; Resources: HBN, HHK; Software: J-HJ; Supervision: HBN, HHK; Validation: HHK; Visualization: J-HJ, HHK; Writing - Original Draft Preparation: HBN; Writing - Review and Editing: HBN, J-HJ, MP, HHK. Twelve adults with dermatologist-confirmed diagnosis of hidradenitis suppurativa (HS) based on diagnostic criteria adopted by the Second International Hidradenitis Suppurativa Research Symposium were recruited for the study (Zouboulis et al., 2015Zouboulis C.C. Del Marmol V. Mrowietz U. Prens E.P. Tzellos T. Jemec G.B. Hidradenitis suppurativa/acne inversa: criteria for diagnosis, severity assessment, classification and disease evaluation.Dermatology. 2015; 231: 184-190Crossref PubMed Scopus (187) Google Scholar). All subjects were systematically phenotyped for demographic, clinical, and disease characteristics including disease severity through the University of California, San Francisco Hidradenitis Suppurativa Registry (Supplementary Table S1). Disease severity scoring was determined based on overall disease assessment. Sex- and age-matched healthy adults within ± 5 years of HS subjects were also recruited. The study was approved by the University of California, San Francisco Institutional Review Board, and all subjects provided written informed consent. All subjects confirmed adherence to a skin preparatory regimen in the days leading up to sampling. The skin preparatory regimen consisted of avoidance of the following agents and activities: oral antibiotics for 8 weeks; topical antibiotics for 4 weeks; detergents, deodorants, and topical cosmetics for 7 days; swimming or hot tubs for 7 days; topical emollients or solutions for 48 hours; and showering or bathing for 24 hours. Swabs (Epicentre Catch-all swabs #QEC89100) were premoistened in lysis buffer (Lucigen MasterPure Yeast DNA Purification Kit #MPY80200) and then vigorously rubbed on a 4-cm2 area of predetermined moist body sites of HS predilection (axillary vault, inguinal crease, inframammary fold, and gluteal crease) for 20 seconds. Affected skin was defined as sampling sites with inflammatory papules, nodules, sinus tracts, scarring, tunnels, and/or drainage; unaffected skin was defined as skin without clinical lesion involvement. Skin swab samples were collected into lysis buffer from a 4-cm2 area of predetermined moist body sites of HS predilection (axillary vault, inguinal crease, inframammary fold, and gluteal crease) and stored at −80 °C to sequence in batches. Two additional collection control swabs were obtained for each session to assess for sampling and reagent contamination. Samples and collection controls were incubated in enzymatic lysis buffer and lysozyme for 30 minutes at 37°C, followed by standard protocol for lysing Gram positive bacterial cells (Invitrogen PureLink Genomic DNA kit, Carlsbad, CA #K1820-02). Purified genomic DNA was resuspended in PureLink Genomic Elution Buffer and stored at −20 °C for batched processing (Grice et al., 2009Grice E.A. Kong H.H. Conlan S. Deming C.B. Davis J. Young A.C. et al.Topographical and temporal diversity of the human skin microbiome.Science. 2009; 324: 1190-1192Crossref PubMed Scopus (1729) Google Scholar). The V1–V3 hypervariable regions of the 16S ribosomal RNA gene were amplified from purified genomic DNA using primers 8F (5'-AGAGTTTGATCCTGGCTCAG-3') and 534R (5'- ATTACCGCGGCTGCTGG) (Fadrosh et al., 2014Fadrosh D.W. Ma B. Gajer P. Sengamalay N. Ott S. Brotman R.M. et al.An improved dual-indexing approach for multiplexed 16S rRNA gene sequencing on the Illumina MiSeq platform.Microbiome. 2014; 2: 6Crossref PubMed Scopus (956) Google Scholar). PCR reactions were performed in duplicate for 30 cycles, purified using Agencourt AmpureXP (Beckman Coulter, Brea, CA), and quantified using the Quant-IT dsDNA Kit (Invitrogen). Pooled amplicons were purified with MinElute PCR purification kit (Qiagen, Hilden, Germany), and next generation sequencing using the Illumina MiSeq platform with Reagent kit v3 (600 cycle) was performed (Illumina, San Diego, CA). A mothur-based pipeline was used for V1–V3 sequence analysis (Schloss et al., 2009Schloss P.D. Westcott S.L. Ryabin T. Hall J.R. Hartmann M. Hollister E.B. et al.Introducing Mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities.Appl Environ Microbiol. 2009; 75: 7537-7541Crossref PubMed Scopus (13794) Google Scholar). Sequences were preprocessed to remove primers and barcodes, identify duplicates, reduce noise by preclustering, and subsample to 3,000 sequences per sample. The negative control swabs yielded significantly lower numbers of sequences than skin swabs (1,563 ± 1,161 sequences per control sample versus 54,285 ± 3,548 per skin swab), demonstrating sample collection, processing, and sequencing without major contamination and, therefore, were removed from further analyses. Chimeras were identified and removed from PCR artifacts using UCHIME in mothur (Edgar et al., 2011Edgar R.C. Haas B.J. Clemente J.C. Quince C. Knight R. UCHIME improves sensitivity and speed of chimera detection.Bioinformatics. 2011; 27: 2194-2200Crossref PubMed Scopus (9578) Google Scholar). Taxonomic classifications were assigned to all remaining sequences down to the genus level with reference database from the Ribosomal Database Project, and these classifications were used for diversity analyses and community comparisons. Community analysis was done with genus-level taxonomic classifications. R software was used to generate plots, including relative abundance of bacterial genera and species. To estimate sampling saturation and subsampling size, rarefaction curves were generated for each sample using the mothur command rarefaction.single. Shannon index was used to quantify alpha diversity (Kong et al., 2012Kong H.H. Oh J. Deming C. Conlan S. Grice E.A. Beatson M.A. et al.Temporal shifts in the skin microbiome associated with disease flares and treatment in children with atopic dermatitis.Genome Res. 2012; 22: 850-859Crossref PubMed Scopus (1025) Google Scholar, Von Felten et al., 2012Von Felten S. Niklaus P.A. Scherer-Lorenzen M. Hector A. Buchmann N. Do grassland plant communities profit from N partitioning by soil depth?.Ecology. 2012; 93: 2386-2396Crossref PubMed Scopus (33) Google Scholar). All statistical analyses in figures were represented as mean and standard error of the mean unless otherwise indicated. Values for relative abundance were represented as mean and 95% confidence interval. Nonparametric Kruskal-Wallis test (P < 0.05) was first used to determine statistical difference among groups, followed by Wilcoxon test for paired comparisons, to examine differences in taxonomic relative abundances and diversity indices between HS(Affected), HS(Unaffected) and HV samples, with Bonferroni correction for multiple comparisons (Excoffier et al., 1992Excoffier L. Smouse P.E. Quattro J.M. Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data.Genetics. 1992; 131: 479-491Crossref PubMed Google Scholar). Spearman correlation was used to evaluate the relationship between abundance and disease severity. All tests used P < 0.05 as the criterion for significance.Supplementary Table S1Per Subject Clinical CharacteristicsSubjectAge/SexRace/EthnicityHurley StageHidradenitis Suppurativa HS122/FW/H3 HS229/MW/NH1 HS328/MW/NH2 HS444/MPI/NH2 HS559/FW/NH2 HS624/FW/NH2 HS722/FW/NH3 HS837/FW/H2 HS935/FB/W/NH2 HS1023/MW/NH2 HS1146/MB/NH1 HS1228/FW/H2Healthy Volunteer HV148/MB/NHN/A HV234/FA/NHN/A HV336/MB/NHN/A HV443/FB/NHN/A HV524/FW/NHN/AAbbreviations: A, Asian; B, black; F, female; H, Hispanic; HS, hidradenitis suppurativa; HV, healthy volunteer; M, male; N/A, not applicable; NH, non-Hispanic; PI, Pacific Islander; W, white. Open table in a new tab Abbreviations: A, Asian; B, black; F, female; H, Hispanic; HS, hidradenitis suppurativa; HV, healthy volunteer; M, male; N/A, not applicable; NH, non-Hispanic; PI, Pacific Islander; W, white.