Genetic Variants Associated With Hidradenitis Suppurativa

Key Points

Question  What are the genetic variants associated with hidradenitis suppurativa (HS) and the genes that may affect disease pathogenesis?

Findings  A genome-wide association study (GWAS) of 753 patients from the University of North Carolina at Chapel Hill was performed and meta-analyzed with GWAS in 3 large biobanks (UK Biobank, FinnGen, and BioVU). Results identified 1 significant locus and 2 suggestive loci and replicated the HS associations in BioVU for 2 loci near the SOX9 and KLF5 genes.

Meaning  These findings implicated genetic variants in risk of HS and support that nearby genes SOX9 and KLF5, which play important roles in epidermal differentiation and follicular inflammation, may be associated with disease pathogenesis.

Importance  Hidradenitis suppurativa (HS) is a common and severely morbid chronic inflammatory skin disease that is reported to be highly heritable. However, the genetic understanding of HS is insufficient, and limited genome-wide association studies (GWASs) have been performed for HS, which have not identified significant risk loci.

Objective  To identify genetic variants associated with HS and to shed light on the underlying genes and genetic mechanisms.

Design, Setting, and Participants  This genetic association study recruited 753 patients with HS in the HS Program for Research and Care Excellence (HS ProCARE) at the University of North Carolina Department of Dermatology from August 2018 to July 2021. A GWAS was performed for 720 patients (after quality control) with controls from the Add Health study and then meta-analyzed with 2 large biobanks, UK Biobank (247 cases) and FinnGen (673 cases). Variants at 3 loci were tested for replication in the BioVU biobank (290 cases). Data analysis was performed from September 2021 to December 2022.

Main Outcomes and Measures  Main outcome measures are loci identified, with association of P < 1 × 10⁻⁸ considered significant.

Results  A total of 753 patients were recruited, with 720 included in the analysis. Mean (SD) age at symptom onset was 20.3 (10.57) years and at enrollment was 35.3 (13.52) years; 360 (50.0%) patients were Black, and 575 (79.7%) were female. In a meta-analysis of the 4 studies, 2 HS-associated loci were identified and replicated, with lead variants rs10512572 (P = 2.3 × 10⁻¹¹) and rs17090189 (P = 2.1 × 10⁻⁸) near the SOX9 and KLF5 genes, respectively. Variants at these loci are located in enhancer regulatory elements detected in skin tissue.

Conclusions and Relevance  In this genetic association study, common variants associated with HS located near the SOX9 and KLF5 genes were associated with risk of HS. These or other nearby genes may be associated with genetic risk of disease and the development of clinical features, such as cysts, comedones, and inflammatory tunnels, that are unique to HS. New insights into disease pathogenesis related to these genes may help predict disease progression and novel treatment approaches in the future.

Hidradenitis suppurativa (HS) is a common chronic inflammatory disease characterized by painful nodules, abscesses, and tunnels with predilection for intertriginous sites.¹ Strong evidence of heritability has been described in cohort and twin studies,^2-6 and our analysis of 281 patient pedigrees in the US demonstrated that siblings of affected individuals have a nearly 20-fold risk of developing HS.⁷ While γ-secretase complex gene variants have been implicated in less than 5% of patients,⁸ most genetic risk factors are unknown. Genome-wide association studies (GWASs) provide an opportunity to identify common genetic variants that contribute to HS susceptibility.

Understanding of HS pathogenesis has shifted substantially in recent years. Hypotheses focused on apocrine gland inflammation and follicular occlusion have faded, with recent characterizations describing HS as a disease of chronic follicular inflammation and dysbiosis. Inflammatory cytokines, complement C5a, and several cell types have all been implicated in the inflammatory response in HS,^9-23 but the genetic and environmental factors predisposing to this response are unclear.

Pathogenesis of HS is characterized by unique clinical features, such as comedones, cysts, and chronically inflamed and/or epithelialized tunnels in areas of chronic involvement. Tunnels attract a chronic inflammatory response, and the epithelial lining has overlapping characteristics with nearby epidermis.²⁴ The mechanisms leading to these unique structures are unclear and a notable gap in our understanding of disease progression.

To better understand genetic factors that may underlie pathogenesis, we performed a GWAS of 720 patients with HS in the Hidradenitis Suppurativa Program for Research and Care Excellence (HS ProCARE) at the University of North Carolina (UNC) Department of Dermatology with controls from the Add Health study.^25,26 To our knowledge this is the first GWAS and meta-analysis for HS and uses HS ProCARE results along with association results from 2 biobanks,^27,28 with replication in BioVU.²⁹ We identified 2 loci associated with HS (P < 5 × 10⁻⁸). The location of HS-associated variants in skin regulatory elements and the functions of nearby genes in HS disease pathways suggest that the variants alter expression of nearby genes to affect disease risk.

Following UNC Chapel Hill Institutional Review Board approval, patients with HS presenting to UNC Dermatology outpatient clinics provided written informed consent and were enrolled in HS ProCARE from August 2018 to July 2021. Clinical and demographic data were collected by skilled interviewers using questionnaires and stored in a RedCAP database. This study followed the Strengthening the Reporting of Genetic Association Studies (STREGA) reporting guideline.

Samples were collected using Oragene_Dx saliva collection kits (DNA Genotek). A subset of patients provided peripheral blood mononuclear cells processed using lymphocyte separation media from blood collection in sodium citrate. Isolation of DNA was performed in the UNC BioSpecimen Processing facility using the PerkinElmer Chemagic Magnetic Separation Module I and magnetic bead separation kits.

Quantity and quality of DNA were determined by NanoDrop OD and Quant-iT PicoGreen dsDNA Assay Kits (Invitrogen). The DNA was analyzed for human content with Femto Human DNA Quantification Kits (Zymo Research) and a QuantStudio 6 Real-time PCR system. Samples were genotyped in the UNC Mammalian Genotyping Core for approximately 654 000 variants using the Infinium Global Screening Array-24, version 3.0 (Illumina). We removed individuals with missing rate greater than 5% or sex mismatch and removed variants with missing rate greater than 5%. We confirmed genotype concordance for 7 pairs of duplicated samples. After quality control, we retained 720 of 753 samples and 644 188 variants. Genotypes were analyzed in build hg38.

The National Longitudinal Study of Adolescent to Adult Health (Add Health) is an ongoing longitudinal study of the health and developmental trajectories for US individuals from early adolescence into adulthood.^25,26 Genotyping data collection was performed in 2008 with participant consent and has undergone extensive quality control and cleaning.²⁵ Covariate data (age, sex, body mass index [BMI, calculated as weight in kilograms divided by height in meters squared], smoking status, and race and ethnicity) were also collected in 2008. Among 5801 Add Health participants with both genotypes and covariates, 3868 had European ancestry and 707 had African ancestry. We performed liftover from hg19 to hg38.

We used variants shared by HS ProCARE and Add Health to perform genotype imputation with the TOPMed freeze 8 reference panel, following our publications,^30-32 and using Eagle, version 2.4³³ for phasing and minimac4³⁴ for imputation. We reestimated imputation quality separately for European-ancestry HS cases, European-ancestry controls, African-ancestry HS cases, and African-ancestry controls and only retained variants with R² of 0.8 or greater in all 4 groups.

We first performed logistic regression to correct for covariates, including age, age squared (age²), sex, BMI, smoking status (current smoker, former smoker, or nonsmoker), race and ethnicity indicators (Asian, Black, Hispanic/Latino, or White), and the first 20 genotype principal components (PCs), and inverse-normal transformed the residuals. Then we performed GWAS on the residuals with EPACTS, version 3.3.0 using the EMMAX test³⁵ to control for sample relatedness,^36,37 which was quantified by a kinship matrix constructed from genotyped variants with missing rate less than 1% and minor allele frequency greater than 1%. To evaluate potential confounding by race and ethnicity, we repeated the GWAS with European-ancestry–only and African-ancestry–only subsets of the data, conditioning on the same covariates except race and ethnicity indicators.

The UK Biobank (UKB)²⁸ is a prospective biobank with approximately 500 000 individuals aged 40 to 69 years in 2006 to 2010. We identified 247 participants of primarily European ancestry with HS defined with either International Classification of Diseases, Ninth Revision (ICD-9) (705.83, Hidradenitis) or International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) (L73.2, Hidradenitis suppurativa), with ancestry groups defined as in our previous work³⁶ using k-means and self-reported race and ethnicity. All European-ancestry individuals without HS (n = 453 049) were included as controls. GWAS was performed using REGENIE³⁸ with the same analysis strategy as HS ProCARE, accounting for the same covariates and 10 PCs. We included variants with minor allele frequency greater than 0.5% and imputation quality INFO score greater than 0.8 and performed liftover from hg19 to hg38.

FinnGen³⁹ aims to collect and analyze genome and health data from 500 000 Finnish biobank participants. We analyzed version 7 data (June 2022) consisting of 309 154 individuals (173 746 females and 135 408 males), including 673 HS cases (defined by the same ICD-9 and ICD-10 codes as UKB) and 297 544 controls. The GWAS summary statistics from FinnGen were directly downloaded from https://www.finngen.fi/ in build hg38. FinnGen included age, sex, genotyping batch, and 10 PCs in models.

We performed meta-analyses using METAL⁴⁰ with effective sample sizes calculated with the following equation:

We first performed 3-way meta-analysis with UKB, FinnGen, and all HS ProCARE patients analyzed jointly with variants existing in 2 or more studies. After obtaining replication results from BioVU described below, we meta-analyzed the top variants in a 4-way meta-analysis (eTable 7 in Supplement 1). To further validate results and consider population stratification, we performed additional meta-analyses as sensitivity analyses: a 4-way meta-analysis with UKB, FinnGen, European-ancestry HS ProCARE, and European-ancestry BioVU individuals, and a 6-way meta-analysis also including African-ancestry HS ProCARE and African-ancestry BioVU individuals. We defined genome-wide significance as P < 5 × 10⁻⁸.

We tested lead variants from the GWAS meta-analyses for association with HS in the BioVU biobank based at Vanderbilt University Medical Center. BioVU DNA was extracted from blood samples, and genotypes were linked to deidentified medical records. Genotyping was performed on the Multi-Ethnic Global array, and genotype imputation was performed on Michigan Imputation Server using the TOPMed reference panel. Ancestry information was defined by PCs with 1000G as reference. Hidradenitis suppurativa disease status was determined with either ICD-9 or ICD-10 codes. We performed logistic regression in R (R Foundation for Statistical Computing) to test variant associations with HS. Covariates included age, age², sex, smoking status, median BMI, self-reported race (Asian, Black, and White), ethnicity (Hispanic or not), genotype batch, median age, and the first 10 genotype PCs.

To identify potential candidate genes, we performed literature searches on the closest protein-coding genes. We identified candidate variants based on linkage disequilibrium (LD) (r² > 0.8) using TOP-LD⁴¹ with European reference panels and used the University of California, Santa Cruz, genome browser (http://genome.ucsc.edu) to display annotation tracks for ENCODE candidate cis-regulatory elements, conservation across species, and epigenomic marks of accessible chromatin and histone chromatin immunoprecipitation corresponding to enhancers.^42-45 We looked up lead variants in the GTEx eQTL database,⁴⁶ but neither variant was in LD with the lead variant for an eQTL in any tissue (r² < 0.5 in 1000 Genomes European participants).

We performed genetic correlation analysis with LD score regression.^47,48 We considered celiac disease,⁴⁹ inflammatory bowel disease,⁵⁰ polycystic ovary syndrome,⁵¹ psoriasis,⁵² rheumatoid arthritis,⁵³ schizophrenia,⁵⁴ and type 2 diabetes.⁵⁵ GWAS summary statistics for these diseases were downloaded directly from the previous publications (eTable 6 in Supplement 1).

We enrolled 720 patients with HS in an initial phase of an HS ProCARE registry and gathered detailed clinical characteristics (Table 1). Mean (SD) age at symptom onset was 20.3 (10.57) years and at enrollment was 35.3 (13.52) years; 360 (50.0%) patients were Black, and 575 (79.7%) female. More than half were obese (BMI >30), and fewer than 30% were smokers. Most patients had Hurley stage II (48%) or stage III (37%) disease. Almost half reported an affected first-degree relative. Phenotypes as described by van der Zee and Jemec⁵⁶ were categorized at enrollment, with the majority having the regular phenotype.

We performed initial GWAS analyses using HS ProCARE patients and Add Health controls (Methods) in all samples or separately for European- and African-ancestry participants. Eight loci showed suggestive evidence of association in the meta-analysis (P < 1 × 10⁻⁶) (eTables 1 and 2 in Supplement 1; eFigures 1-3 in Supplement 2). Despite use of external controls, we observed minimal inflation, with genomic control lambda ranging from 1.00 to 1.05 (eFigures 4-6 in Supplement 2). To increase power to detect associations, we performed GWAS in the United Kingdom Biobank (UKB; 247 cases, 453 048 controls) and meta-analyzed HS ProCARE GWAS results with UKB and the FinnGen biobank based in Finland (673 cases, 297 544 controls) (Methods). The 3-way meta-analysis revealed 1 genome-wide significant locus (P < 5 × 10⁻⁸), and 2 suggestive loci (Table 2, Figures 1 and 2; eTable 3 and eTable 10 in Supplement 1; eFigure 7 and eFigure 8 in Supplement 2). The most strongly associated variant, rs10512572 (meta-analysis P = 1.8 × 10⁻⁸) on chromosome (chr) 17, showed the strongest association in FinnGen (P = 1.9 × 10⁻⁷), with nominal support in HS ProCARE (P = 3.9 × 10⁻³) and a trend toward significance in UKB (P = .15), all with the same direction of effect. At the second locus, the most strongly associated variant, rs17090189 (meta-analysis P = 1.7 × 10⁻⁷) on chr13, also showed the strongest association in FinnGen (P = 3.7 × 10⁻⁷), with nominal support from HS ProCARE (P = .03) and UKB (P = .08), all with the same direction of effect. At a third locus, the most strongly associated variant, rs5792315 (meta-analysis P = 8.9 × 10⁻⁷) on chr11, showed the strongest association in HS ProCARE (P = 3.3 × 10⁻⁴), with support from FinnGen (P = 1.3 × 10⁻³) and a trend in UKB (P = .13), all in the same direction of effect. Notably, the human leukocyte antigen region on chr6 did not show even suggestive evidence of association with HS in any individual study or meta-analysis.

We tested the 10 most strongly associated variants at the 3 loci for association with HS in the BioVU biobank (Methods). BioVU contains 290 HS cases, including 189 European-ancestry and 101 African-ancestry individuals, with 64 234 and 12 105 controls for the 2 ancestry groups, respectively. The chr17 locus was replicated in BioVU (P = 1.9 × 10⁻⁴) in the same direction of effect (eTable 4 in Supplement 1). One variant at the chr13 locus showed nominal evidence of association (P = .05) in the same direction. In a 4-way meta-analysis of BioVU and the first 3 studies, the chr17 locus became more significant (P = 2.3 × 10⁻¹¹), and the chr13 locus exceeded the genome-wide significance threshold (P = 2.1 × 10⁻⁸) (Table 2; eTable 5 in Supplement 1). The chr11 locus was not replicated in BioVU (P = .27).

We performed 2 additional meta-analyses as sensitivity analyses to validate our findings. A 4-way meta-analysis of all European-ancestry individuals replicated the chr17 (P = 5.5 × 10⁻⁹) and chr13 (P = 1.6 × 10⁻⁸) loci (eTable 8 in Supplement 1). A 6-way meta-analysis adding African-ancestry individuals from HS ProCARE and BioVU showed that the 2 loci were more significant: chr17, P = 4.0 × 10⁻¹¹; and chr13, P = 3.5 × 10⁻⁹ (eTable 9 in Supplement 1). These results suggest that the signals are robust to different analysis strategies.

The HS-associated variants on chr17 and chr13 are located in intergenic regions (Figure 2, Figure 3). At the chr17 locus, the closest protein-coding gene, located more than 600 kb away, is SOX9, which encodes a transcription factor that regulates several developmental processes. SOX9 is expressed in the epidermal basal layer and outer root sheath (ORS) of hair follicles and is highly expressed in psoriasis and nonmelanoma skin cancers.⁵⁷ Variants at this association signal, including rs17226067, rs17825774, and rs17825799, are located in a keratinocyte distal enhancer-like regulatory element (Figure 3) as defined by ENCODE SCREEN,⁴⁵ suggesting that 1 or more variants may alter gene regulation.

At the chr13 locus, the closest protein-coding genes are the transcriptional regulators KLF12 and KLF5, located approximately 260 kb and approximately 340 kb away, respectively. Deletion of KLF5 alters the skin barrier, and its expression is lower in human diseases with disrupted epidermal sphingolipid secretion.⁵⁸ A variant at this locus, rs61957178, is located in a keratinocyte regulatory element (Figure 3) that may have a long-range effect on KLF5 and/or KLF12. This signal has also been associated with asthma and respiratory diseases.^59,60

To investigate evidence of shared genetic components between HS and autoimmune and inflammatory diseases, we performed genetic correlation analysis using LD score regression.^47,48 Genetic correlation provides an estimate of the overlap in terms of evidence of associations for a pair of complex traits with existing genome-wide summary statistics, which could help understand the genetic components of HS comorbidities. We examined genetic correlation with celiac disease,⁴⁹ inflammatory bowel disease,⁵⁰ polycystic ovary syndrome,⁵¹ psoriasis,⁵² and rheumatoid arthritis,⁵³ as well as schizophrenia⁵⁴ and type 2 diabetes.⁵⁵ We found nominally significant (P < .05) genetic correlation between HS and inflammatory bowel disease (rg = 0.39, P = .04), psoriasis (rg = 0.49, P = .03), and type 2 diabetes (rg = 0.18, P = .04) (eTable 6 in Supplement 1), although none of these results were significant at a stringent Bonferroni-corrected threshold (P < .05/7). These results show that inflammatory bowel disease, psoriasis, and type 2 diabetes may have correlated genetic segments with HS.

Our GWAS meta-analysis of a cohort of patients with HS ascertained by the UNC Department of Dermatology, along with HS genetic data in large biobanks, identified variants associated with risk of HS. Variants at these loci are located in keratinocyte regulatory elements near the genes SOX9 and KLF5, suggesting that common variants may alter regulation of these or other nearby genes to influence HS risk. These genes have roles in skin and follicular inflammation, but have not previously been associated with HS pathogenesis.

The HS candidate gene SOX9, which encodes SRY-box transcription factor 9, is highly expressed in human epidermis and the ORS of hair follicles and is necessary for follicular differentiation and maintenance. The SOX9 gene upregulates MMP1, MMP2, and IL-8, which are linked to tumorigenesis and inflammation.⁶¹ Mouse models lacking SOX9 demonstrate follicular degeneration, epidermal thickening, and eventual dermal scarring. These mice also lost keratin 1 (K1) and keratin 10 (K10) expression in the basal layer of the follicular sheath,⁶² similar to loss of epidermal K1/K10 function in epidermolytic ichthyosis that leads to epidermal fragility, raising the possibility that follicular fragility and dissection play a role in HS pathogenesis. Blocking SOX9 expression in adult mice resulted in development of keratin pearls and loss of K1 expression at the upper ORS, leading to epidermal differentiation and loss of hair follicle stem cells.⁶³ This differentiation toward epidermis may explain the development of characteristic cysts, comedones, and epithelialized tunnels.

Similarly, the candidate gene KLF5, which encodes Kruppel-like factor 5, may affect HS pathogenesis. Overexpression of mouse KLF5, a critical regulator of sphingolipid-metabolizing and secreting enzymes in the skin, led to hyperkeratosis, follicular occlusion, and skin erosions, which are characteristic of HS.⁶⁴ Deletions of KLF5 disrupt barrier function and have been associated with atopic dermatitis, nonhealing wounds,⁵⁸ and IL-17–driven colitis in mice.⁶⁵ Colitis induced by KLF5 deletion was associated with altered gut microbiota and stimulation of the IL-22/STAT3/IL-17 pathway. This colitis was relieved by antibiotics and IL-17–neutralizing antibodies, similar to HS treatment responses.^66-68 In human colitis, reduced intestinal KLF5 expression is linked to increased disease severity.⁶⁵

Remarkably, genes at the 2 loci appear to have directly opposing and/or complementary roles in epithelial differentiation. The KLF5 gene is present primarily in epidermal keratinocytes that do not express SOX9, which is primarily expressed in the hair bulge.⁵⁸ The KLF5 gene functions as a regulator of epidermal stem cells, while SOX9 is a primary regulator of hair follicle stem cells. More specifically, KLF5 and SOX9 expression demarcate epidermal and hair follicle differentiation, respectively. Coexpression is induced transiently during wound healing as stem cell populations are activated for repair, and induced expression of KLF5 downregulates SOX9 expression. Disruption of homeostasis in these expression patterns may explain the formation of chronic wounds such as inflammatory tunnels and ulcers in HS, as well as cysts and epithelialized tunnels that lack typical differentiation toward hair follicles or epidermis.

Genetic correlations of HS with common comorbidities such as psoriasis, inflammatory bowel disease, and type 2 diabetes suggest that shared genetic mechanisms may predispose to risk. Further investigations are needed with better-powered HS GWASs to evaluate other comorbidities and to uncover the mechanisms of shared risk.

An advantage of our work is the multiple large cohorts to validate association results, including ancestrally diverse patients with well-characterized disease and background. Careful selection of controls is vital for case-only participants. We used considerable effort to match the HS ProCARE cases with appropriate external controls, and Add Health controls resulted in negligible P value inflation. However, some Add Health participants may have HS, which would lead to misclassification bias toward the null. While HS ProCARE includes patients with ancestry from Europe and Africa, inadequate sample sizes limit our ability to detect significant loci within ancestry groups or assess generalizability in other ancestries.

In this genetic association study, as with all GWASs, identifying variant associations near candidate genes with plausible roles in disease biology does not prove a causal effect of these variants or genes on disease risk. Functional studies must assess whether the HS-associated variants and surrounding regulatory elements can alter SOX9 and KLF5 function or expression and, subsequently, affect HS pathology. These genes are strong candidates because they affect hair follicle and epidermal differentiation and inflammation. Other nearby genes such as KLF12 could also play a role. Larger sample sizes and sequencing are needed to detect rare gene variants, such as γ-secretase complex variants, that are implicated in small numbers of patients. Ultimately, understanding how genes lead to tunnel formation and an overactive inflammatory response may lead to therapeutic breakthroughs that target HS pathogenesis.

Accepted for Publication: April 25, 2023.

Published Online: July 26, 2023. doi:10.1001/jamadermatol.2023.2217

Corresponding Author: Christopher J. Sayed, MD, Department of Dermatology, University of North Carolina, 410 Market St, Ste 400, Chapel Hill, NC 27516 (christopher_sayed@med.unc.edu).

Author Contributions: Drs Sayed and Mohlke had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Quan Sun and K. Alaine Broadaway contributed equally as first authors. Yun Li, Karen L. Mohlke, and Christopher J. Sayed contributed equally as senior authors.

Concept and design: Fajgenbaum, Levitt, Liu, Thomas, Li, Mohlke, Sayed.

Acquisition, analysis, or interpretation of data: Sun, Broadaway, Edmiston, Fajgenbaum, Miller-Fleming, Westerkam, Melendez-Gonzalez, Bui, Blum, Lin, Hao, Harris, Liu, Cox, Li, Mohlke, Sayed.

Drafting of the manuscript: Sun, Melendez-Gonzalez, Blum, Lin, Hao, Li, Mohlke, Sayed.

Critical revision of the manuscript for important intellectual content: Sun, Broadaway, Edmiston, Fajgenbaum, Miller-Fleming, Westerkam, Bui, Levitt, Harris, Liu, Thomas, Cox, Li, Mohlke, Sayed.

Statistical analysis: Sun, Broadaway, Miller-Fleming, Levitt, Cox, Li.

Obtained funding: Fajgenbaum, Harris, Li, Mohlke, Sayed.

Administrative, technical, or material support: Edmiston, Fajgenbaum, Westerkam, Melendez-Gonzalez, Bui, Blum, Levitt, Lin, Hao, Liu, Thomas, Sayed.

Supervision: Edmiston, Thomas, Cox, Li, Mohlke, Sayed.

Conflict of Interest Disclosures: Dr Fajgenbaum reported grants from Howard H. Holderness Distinguished Medical Scholars Program (scholarship program and curriculum for UNC medical students pursuing research) during the conduct of the study. Dr Bui reported grants from National Institutes of Health (NIH) National Institute of Arthritis and Musculoskeletal and Skin Diseases (1R21AR075996-01A1) during the conduct of the study. Dr Cox reported grants from NIH National Human Genome Research Institute (NHGRI) during the conduct of the study. Dr Mohlke reported grants from NIH during the conduct of the study. Dr Sayed reported grants (education grant, investigator fees to institution) and personal fees (speaking, consulting) from AbbVie, grants (fees to institution as an investigator) and personal fees (speaking, consulting) from Novartis, grants and personal fees (investigator fees paid to institution) from UCB, grants (investigational fees paid to institution) and personal fees (consulting) from Incyte, grants (investigator fees paid to institution) and personal fees (consulting) from InflaRx, personal fees (consulting) from Alumis, and grants (investigator fees paid to institution) from ChemoCentryx outside the submitted work; and serving as a volunteer member of the board of the Hidradenitis Suppurativa Foundation and member of the European Hidradenitis Suppurativa Foundation. No other disclosures were reported.

Funding/Support: The project is supported by NIH grants 1R21AR075996, R01DK072193, T32HL129982, 1T32GM135128, UL1TR002489, P01CA206980, R01CA233524, and P30CA016086.

Role of the Funder/Sponsor: The National Institutes of Health had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: Its contents are solely the responsibility of the authors and do not necessarily represent official views of the National Center for Advancing Translational Sciences or the National Institutes of Health.

Data Sharing Statement: See Supplement 3.

Additional Contributions: The authors thank the HS ProCARE patients for their contributions. The authors also acknowledge the University of North Carolina BioSpecimen Processing Facility Lab, particularly Patricia Basta, PhD (https://bsp.web.unc.edu), for blood/sample processing, nucleic acid extraction, storage, and disbursements, and the University of North Carolina Mammalian Genotyping Core, particularly Amanda Gerringer, MSFS (https://mgc.unc.edu/), for generating genotype data. The Biospecimen Processing Facility Lab is supported by Cancer Center Support Grant P30 CA016086 and the Center for Environmental Health and Susceptibility (CEHS) Grant P30ES010126 that support the core (BioSpecimen Processing Facility) and core staff. The Mammalian Genotyping Core is supported by the Lineberger Comprehensive Cancer Center, the CEHS, and facility users. Additional sources of support include, but are not limited to, the UNC School of Medicine, School of Public Health, University Cancer Research Fund, and the North Carolina Biotechnology Center. We acknowledge the participants and investigators of the FinnGen study. The UKB data analysis was conducted using the UK Biobank Resource under Application No. 25953. We acknowledge the ENCODE Consortium, the ENCODE production laboratory(s) for generating data and the laboratory(s) of Zhiping Weng, PhD, and Manolis Kellis, PhD, for generating annotation file sets on the keratinocyte female donor ENCDO268AAA. The Genotype-Tissue Expression (GTEx) Project was supported by the Common Fund of the Office of the Director of the National Institutes of Health, and by National Cancer Institute, NHGRI, National Heart, Lung, and Blood Institute, National Institute on Drug Abuse, National Institute of Mental Health, and National Institute of Neurological Disorders and Stroke. The data used for the analyses described in this article were obtained from the GTEx Portal in September 2022. We acknowledge the participants of the BioVU resource. The BioVU project at Vanderbilt University Medical Center is supported by numerous sources: including the NIH-funded Shared Instrumentation Grant S10OD017985 and S10RR025141 and CTSA grants UL1TR002243, UL1TR000445, and UL1RR024975 from the National Center for Advancing Translational Sciences. Genomic data are also supported by investigator-led projects that include U01HG004798, R01NS032830, RC2GM092618, P50GM115305, U01HG006378, U19HL065962, and R01HD074711 and additional funding sources listed at https://victr.vumc.org/biovu-funding/. We acknowledge the participants and investigators of the Add Health GWAS study funded by Eunice Kennedy Shriver National Institute of Child Health and Human Development Grants R01 HD073342 and R01 HD060726.