Baseline clinical, hormonal and molecular markers associated with clinical response to IL- 23 antagonism in hidradenitis suppurativa: A prospective cohort study

Hidradenitis suppurativa is a complex inflammatory disease in which predicting therapeutic response remains challenging. IL- 23 interacts with sex hormones but the rela-tionships between the two in HS remains uninvestigated. To assess whether baseline clinical, hormonal or molecular markers are associated with clinical response to IL23 antagonism with risankizumab in hidradenitis suppurativa. Twenty six individuals with Hurley stage 2/3 disease were administered risankizumab 150 mg Week 0, 4, 12. Baseline sex hormones and skin biopsies were taken. Clinical response at Week 16 assessed by the HiSCR, and differences between responders and non- responders assessed. Eighteen of 26 participants achieved HiSCR50 at week 16 (69.2%). Clinical response to IL- 23 antagonism was associated with male gender, elevated total serum testosterone and decreased levels of FSH. Stratification by clinical responders/nonresponders identified differentially expressed genes including PLPP4 and MAPK10 . Immunohistochemistry identified elevated numbers of CD11c, IL- 17A and IL- 17F positive cells compared to nonresponders. CD11c + cells significantly correlated with serum levels of total testosterone and inversely correlated with serum FSH. Clinical response to IL- 23 antagonism in HS is associated with serum sex hormones, Th17 polarized inflammation in lesional tissue and CD11c + cells. These potential therapeutic biomarkers require further validation in larger cohorts but may suggest potential targeted HS therapy.

Immunohistochemistry identified elevated numbers of CD11c, IL-17A and IL-17F positive cells compared to nonresponders. CD11c + cells significantly correlated with serum levels of total testosterone and inversely correlated with serum FSH. Clinical response to IL-23 antagonism in HS is associated with serum sex hormones, Th17 polarized inflammation in lesional tissue and CD11c + cells. These potential therapeutic biomarkers require further validation in larger cohorts but may suggest potential targeted HS therapy.

K E Y W O R D S
hidradenitis suppurativa, hormones, inflammation, interleukin-23, monocytes, testosterone in Phase 2 studies, 8 with no significant difference in clinical response when compared to placebo 8 (Table S4). Isolated case reports and case series, however, report benefit in some patients. 6,9 HS also has a significant hormonal component to disease pathogenesis. 10 The link between sex hormones and the Th17 immunological axis is incompletely understood. 11 Perimenstrual disease flares, remission during pregnancy and the effect of comorbidities such as obesity and insulin resistance on hormone levels suggest sex hormones play an important role in modulating inflammation in HS. 12 Hormonal therapies such as spironolactone, oral contraceptives and 5α reductase inhibitors such as finasteride and dutasteride are used, primarily in female patients, with variable levels of efficacy. 12,13 The Th17 immune axis is known to interact with sex hormones through IL-23. 14,15 IL-23 interacts with noncanonical androgen receptor signalling pathways, modulating inflammation in epithelial tissues and monocytes/macrophages. 14,15 Monocyte activation and development into macrophages are altered in the presence of tissue estrogens, (specifically E2 Estradiol) both in in vivo and in vitro investigations. 14,15 Monocytes and macrophages are known to be central players in the pathogenesis of HS, [16][17][18][19] produce IL-23, and are proposed to interact with T cells, B cells and neutrophils to direct and orchestrate chronic inflammation in HS. [16][17][18][19] Recent Phase 2 studies of IL-23 antagonism in HS have suggested disparate clinical responses based upon participant gender, with primarily male cohorts demonstrating higher rates of clinical response to IL-23 antagonism compared with primarily female cohorts. 8 This would be consistent with sex hormones influencing the role of IL-23 directed inflammation in HS; however, clinical and mechanistic evidence to support this hypothesis is currently lacking.
The aim of this prospective cohort study was to assess whether baseline clinical, hormonal or molecular markers are associated with clinical response to IL-23 antagonism with risankizumab in hidradenitis suppurativa.

| ME THODS
Twenty six individuals with dermatologist diagnosed HS (based upon the modified Dessau criteria 20 ) were included in this cohort study.
All individuals had Hurley stage 2 or 3 disease and a minimum of five inflammatory lesions (abscesses and nodules). Inclusion and exclusion criteria are presented in Table S1. Clinical data including demographics, smoking status, body mass index (BMI), family history of HS and diagnosed insulin resistance were collated (Table S2). Disease severity was assessed using Hurley staging and lesion counts using the International Hidradenitis Suppurativa Severity Score (IHS4) outcome measure. 2 Baseline blood work including serum total and free testosterone, follicular stimulating hormone (FSH), luteinizing hormone (LH), sex hormone binding globulin (SHBG) as well as Creactive protein (CRP) levels. All participants were administered risankizumab at psoriasis dosing of 150 mg week 0, 4 and 12. The primary outcomes of interest was clinical response as measured by the Hidradenitis Suppurativa Clinical Response (HiSCR 2 ) at week 16. Additional outcomes included the IHS4, as well as HiSCR75 (defined as a 75% reduction in abscess and nodule count without an increase in abscesses or draining tunnels), and HiSCR90 (defined as a 90% reduction in abscess and nodule count without an increase in abscesses or draining tunnels) as used in the recent bimekizumab Phase 2 clinical trial in HS. 21 Clinical responders were defined as individuals who achieved HiSCR at Week 16. Nonresponders were defined as those who did not achieve HiSCR 50 at Week 16.

| Skin biopsy collection and processing
Lesional, perilesional and nonlesional skin biopsies were taken prior to the commencement of risankizumab using previously described standardized lesion and site methodologies. 22 Each 6 mm biopsy specimen was bisected, with one section immediately placed in RNA later and frozen at −80 degrees Celsius until processing for RNA extraction. The other section was placed in OCT medium and frozen at −80 degrees Celsius and processed for Immunohistochemistry.
RNA was extracted using the Qiagen RNEasy kit then eluted in 50ug of RNA-ase free water. RNA was processed and analysed using the Nanostring nCounter Fibrosis 2.0 multiplex gene expression assay (gene list in Table S3). nCounter technology uses a pair of gene-specific probes-a capture probe and a reporter probe-with each aligning to their target RNA by complementary base pairing.
The reporter probe contains a target-specific fluorophore, and readout is via an imaging platform that identifies and quantifies probe complexes. 23 Tissue sections in OCT were cut into 5. used with DAB chromophore. Negative controls without the use of primary antibody were also produced for quality control purposes.
IHC quantification was undertaken using semiquantitative measurement by two experienced independent raters (AF, JWF), with any disagreement mediated by a third author.

| Statistical analysis
Power calculation was undertaken to determine the significance of findings. Assuming a clinical response rate of 40% similar to clinical trial data, 8 a sample size of 26 participants will allow detection of significant change in two markers (with log fold change >1. 5) with power greater than 80% if one tailed significance t-tests is performed.
Descriptive statistics were used to collate all demographic and disease severity data in the included patients. Differences between groups (responders/nonresponders) were analysed using chisquared analysis for binary/dichotomous variables and the Wilcoxon rank sum test for continuous variables. p < 0.05 was considered statistically significant and adjustment for multiple comparisons was made using the Benjamini Hochberg procedure. All statistical analysis was completed in Graphpad Prism (9.4.1).

| Analysis of serum sex hormones
Serum sex hormones were measured by the same laboratory across all individuals. All premenopausal female participants underwent baseline blood tests during the first 3 days of their menstrual cycle to account for cyclical variation. As per the exclusion criteria-no women in this cohort study were on active hormonal contraception.
Differences between groups were analysed using Wilcoxon rank sum test for continuous variables. p < 0.05 was considered statistically significant and adjustment for multiple comparisons was made using the Benjamini-Hochberg procedure.

| Immunohistochemistry quantification
IHC staining underwent semiquantitative analysis using previously published methods. 24 Differences between groups were analysed using the Wilcoxon rank sum test with adjustment for multiple comparisons was made using the Benjamini Hochberg procedure.

| Nanostring nCounter analysis
Extracted RNA was analysed using the nCounter system (Nanostring) using the Human Fibrosis V2.0 gene panel. Raw data were processed using the Nanostring Nsolver (version 4.0.70) analysis software using quality control and normalization procedures derived from the NormqPCR R package as previously described. 25 Differentially expressed genes (DEGs) were defined as >1.5 Log2Fold change with a false discovery rate <0.05 and p value <0.05.

| Clinical response to IL-23 antagonism was significantly associated with male gender, insulin resistance and serum sex hormones
Eighteen of the 26 included participants were classified as responders (69.2%), having achieved HISCR at Week 16. Responders demonstrated dramatic reduction in the number of clinically apparent nodules and abscesses, but also draining tunnels as well as diffuse erythema, pain and swelling ( Figure S1). The demographic and disease characteristics of the included patients (including responders and nonresponders) is presented in Table 1. A statistically significant difference in the proportion of male participants in the responder category when compared with the non-responder category as measured by the chi-squared test. (77.8% vs. 12.5% p < 0.01). There was no significant difference between responders and nonresponders with regards to age, BMI or Hurley stage (Table 1). Significant differences in total testosterone, FSH and LH were also identified between responders and nonresponders. (Table 1, Figure 1).
The univariate association of clinical response as measured by HiSCR (hereafter termed HiSCR50) was maintained when deeper measures of clinical response (HiSCR75 and HiSCR90) were analysed (Appendix S1). Statistically significant differences between serum FSH and serum total testosterone were identified between HiSCR75 responders/nonresponders. Only serum total testosterone was statistically significant between HiSCR90 responders/ nonresponders/( Figures S2-S5).    (Table S5).

| Immunohistochemistry identifies Th17 associated proteins such as IL-17A, IL-17F and IL-23p19 as well as and CD11c + leucocytes as associated with clinical response in HS lesional tissue
Confirmatory IHC indicates significantly increased number of  (Figure 4).

F I G U R E 1
Comparison of baseline serum testosterone and serum FSH levels in responders and nonresponders (at week 16) to risankizumab therapy for hidradenitis suppurativa as well as receiver operating curve for testosterone and HiSCR50.
F I G U R E 2 Baseline multiplex gene expression (Nanostring) data for baseline tissue biopsies in 26 participants, stratified by lesional, perilesional and nonlesional tissue.
Univariate correlation identified a significant association between serum FSH and serum testosterone with semiquantitative IHC cell counts of CD11c and IL23p19 positive cells ( Figure S8).

| DISCUSS ION
Clinical response to IL-23 antagonism with risankizumab in HS is associated with baseline elevation in total testosterone and base- PCOS is known to alter the populations and activity of CD11c + monocytes, 31 and sex hormones can alter monocyte differentiation with testosterone driving a nonclassical monocyte differentiation and estrogens forming a classical differentiation programme. [32][33][34][35] This is manifest by alterations in CD14 and CD16 expression; however, this data has not been replicated in HS.

F I G U R E 3
Log fold change in gene expression of select genes in responders versus nonresponders in hidradenitis suppurativa. Statistically significant elevation in gene expression was seen in T cell, macrophage monocyte and dendritic cell genes compared to nonresponders. This included IL-17A, CXCR1, TRAT1 and LTA. Principal component analysis demonstrating discrete clustering of responders and nonresponders based on gene expression data.
Additionally, estrogens can depress the dendritic cell stimulation of T cells, and testosterone promote inflammatory activation of monocyte-derived dendritic cells. [32][33][34][35] Our presented results would be consistent with previous observations that sex hormones may impact the ratio of classical/intermediate/nonclassical monocytes. 32

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets used in this manuscript are publicly available through Gene Expression Omnibus (GEO) via accession number GSE214820.

S U PP O RTI N G I N FO R M ATI O N
Additional supporting information can be found online in the Supporting Information section at the end of this article.

Figure S1
Representative clinical photographs of response to Risankizumab therapy at Week 16.

Figure S2
Comparison of FSH and total testosterone between responders and non responders stratified by HiSCR75 and HiSCR90.