Adalimumab therapy is associated with increased faecal short chain fatty acids in hidradenitis suppurativa

Abstract Altered gut microbiota composition has been observed in individuals with hidradenitis suppurutiva (HS) and many other inflammatory diseases, including obesity, type 1 and type 2 diabetes. Here, we addressed whether adalimumab, a systemic anti‐inflammatory therapy, may impact the microbiota biochemical profile, particularly on beneficial metabolites such as short‐chain fatty acids (SCFAs). We conducted an observational single‐arm pilot trial to assess gut microbiota composition by 16S rRNA gene sequence analysis and to detect metabolite signatures by gas chromatography in stool samples from participants with HS prior to and 12 weeks after commencing adalimumab therapy. HS individuals that better responded to adalimumab treatment showed a shift in the composition and function of the gut microbiota with significantly increased SCFA acetate and propionate compared to age, gender and BMI‐matched healthy controls. A positive correlation was observed between propionate with Prevotella sp and Faecalibacterium prausnitsii. Increased SCFAs, changes in gut microbiota composition, function and metabolic profile following 12 weeks of adalimumab suggest that targeting SCFAs may be considered a potential biomarker to be evaluated as a complementary protective factor or as a diagnostically relevant signal in HS.

include microbial metabolites such as short-chain fatty acids (SCFAs) acetate, propionate and butyrate. They are produced from fibre fermentation by the gut microbiota and play essential roles in regulating inflammation, metabolism and gut homeostasis. 11,12 Promoting the stimulation of gut SCFA-producing bacteria or increasing the production of SCFA acetate and butyrate play a critical role in tuning the human immune system with the potential to improve glycaemic control, as shown in T1D and T2D. 13,14 Likewise, in mouse models, targeted-SCFA supplementation has shown to improve infection, colitis, type 1 and type 2 diabetes and facilitate anti-cancer therapy. [13][14][15][16][17][18] Although anti-inflammatory therapies are known to modulate the Psoriasis-associated skin microbiota, 19,20 their effects on the microbiota biochemical profile, in particular on beneficial microbial metabolites such as SCFAs, remain unexplored.
Here, we assessed whether adalimumab therapy impacts faecal microbiota and SCFAs in HS.

| Participant characteristics
The demographic and disease-specific variables included 10 participants with dermatologist-diagnosed moderate to severe HS and six matched healthy controls presented in Figure 1. All individuals were over 18 years of age, not currently or previously treated with a biologic agent, and not treated with antibiotics for at least 8 weeks prior to enrolment. Also, none of the participants were treated with antibiotics during the study. The protocol for antibiotics was based on previous microbiota-targeted studies in HS and T1D. 6,13 In addition, six matched healthy control participants without inflammatory skin disease were recruited from the same clinic. Individuals with HS received 12 weeks of subcutaneous adalimumab at standard FDAapproved dosage (160 mg Week 0, 80 mg Week 2 and 40 mg every 2 weeks after that). Stool samples were collected at baseline (W0) and Week 12 (W12) after adalimumab in HS participants and healthy controls.

| Adalimumab delivery is linked with changes in the gut microbiota composition and metabolic profile
A significant difference in the faecal microbial composition was detected between HS and controls after accounting for the timepoints and interaction (Factorial PERMANOVA, R 2 = 0.08, p = 0.002).
Principal component analysis (PCA) illustrates a shift towards control samples. (Figure 2A). HS participants showed reduced alphadiversity (Pieolou's evenness, Shannon's richness index and Simpson's diversity index, p = 0.008-0.011) compared to healthy controls at W0, but the differences between healthy controls and HS at Week 12 were largely absent ( Figure 2B). Differential abundance analysis using linear discriminant analysis Effect Size (LefSE) identified significantly higher in Rumminococcus bromii and Coprococcus eutactus in healthy controls versus HS ( Figure 2C). Taxonomic composition of the microbiota at baseline (W0) between healthy controls and HS F I G U R E 1 Demographics of the recruited participants. Ten hidradenitis suppurutiva (HS) subjects and 6 healthy controls were recruited for this study.
participants showed reduced members of the phylum Firmicutes and Euryarchaeota and increased proportion of Actinobacteria and Bacteroidetes ( Figure 2D).
Treatment with 12 weeks of adalimumab at the FDA-approved dosing regimen (160 mg at Week 0, 80 mg at Week 2, then 40 mg weekly from Week 4 to Week 12) showed a significant increase in alpha diversity similar to healthy controls (Shannon's richness index, p = 0.049; Figure 2B). HS faecal microbiota after 12 weeks of adalimumab therapy was more prevalent of the genus Bifidobacterium, Bacteroides, Agathobacter, Prevotella_9, Faecalibacterium and Blautia ( Figure 2D).  PICRUSt2 functional bacterial profiles showed significant changes in carbohydrate energy production pathways between HS and control ( Figure 3A). For instance, gluconeogenesis I and nucleotide biosynthesis were significantly elevated in HS, while UDP-N-acetyl-D-glucosamine biosynthesis I significantly changed in controls compared to HS participants. It is noteworthy that care must be taken when integrating this result as the pathways were imputed not based on the HS-specific microbial genomes database.
Subsequent shotgun metagenomics or targeted quantitative PCR is warranted to confirm this observation. Twelve weeks of adalimumab were followed by changes in metabolic pathways, including glycolysis I ( Figure 3B).

| Adalimumab treatment is followed by increased SCFAs in stools from HS participants
There were no significant statistical changes in the faecal concentrations of SCFA acetate, propionate and butyrate at baseline between HS compared to age, gender and BMI-matched healthy controls ( Figure 4A). Stratification of HS participants by gender, smoking status, Hurley stage or presence of epithelialized tunnels did not identify any significant differences between groups regarding faecal acetate or butyrate at baseline using Wilcoxon rank sum test ( Figure S1). Remarkably, increased faecal acetate and propionate were observed after 12 weeks of subcutaneous adalimumab therapy in HS, similar to healthy controls at W12 without treatment

| DISCUSS ION
Adalimumab altered faecal microbiota composition, production of SCFAs, and metabolic function of participants with HS compared to age, gender and BMI-matched healthy controls. While 16S ribosomal gene sequencing of the faecal microbiota in HS has recently been reported, 3,6 we demonstrate that the adalimumab therapy in HS shifts the microbiota and SCFAs production towards healthy controls.
However, it was not associated with clinical response.
Inflammation in HS is heterogeneous, with a degree of systemic inflammation associated with the presence of epithelialized tunnels as measured by serum proteomics. 21  pathway is also active during yeast metabolism and is highly associated with severe HS. 27 Twelve weeks of adalimumab were associated with changes in metabolic pathways, including glycolysis I.
Reductions in purine and L-ornithine biosynthesis were seen at W0 in HS compared to W12. It is worthy to note that we use PICRUSt2 as a bioinformatic tool for predicting functional content from 16S rRNA gene sequence data, which uses phylogeny to predict genomic content. Although it has been used in other microbiota-targeted clinical trials, 28,29 it is not without limitations and used on the assumption that it can adequately reconstruct metagenomic data. Therefore, the results provided only a direction for subsequent gene expression studies.
Restoration of the gut microbiota suppresses the clinical signs of atopic dermatitis (AD) associated with increased faecal SCFAs, induced Tregs, reduced IgE levels and the numbers of mast cells, F I G U R E 3 Adalimumab impact the faecal microbiota function and metabolic pathways. (A) An overall higher abundance of pathways related to carbohydrate biosynthesis and nucleoside and nucleotide biosynthesis were detected in hidradenitis suppurutiva (HS) than in healthy controls. When the comparison was made between time points (B), significant increases in the glycolysis I pathways at Week 12 compared with baseline were found in HS after adalimumab treatment. In contrast, no significant difference was seen in healthy controls. The features presented are significant by Benjamini-Hochberg (BH) adjusted p < 0.01. eosinophils and basophils. 30 It is known that adalimumab alters the faecal microbiota in rheumatoid arthritis and Crohn's disease. 31,32 Overall, our data are the first to demonstrate that adalimumab modulates the biochemical profile of the gut microbiota by increasing the production of faecal SCFAs in HS. Indeed, the pathways driv-

| Limitations of the study
The study included a small sample size and limited follow-up of the study. Without a placebo control group, this study could not de-

| Adalimumab therapy
Subcutaneous adalimumab therapy was provided at the FDAapproved dose (160 mg Week 0, 80 mg Week 2, and 40 mg Week 4, 6, 8, 10 and 12). All participants underwent appropriate screening for HIV, hepatitis B, hepatitis C and quantiferon gold testing for latent tuberculosis prior to enrolment in the study. All participants continued on adalimumab therapy as a standard of care after completing involvement in the trial. Healthy controls were not exposed to adalimumab therapy.

| Short-chain fatty acid analysis
Acetate, propionate and butyrate were measured in duplicates by

| 16S rRNA gene sequence analysis
Total DNA was extracted using the QIAamp DNA stool mini kit (Qiagen). Briefly, the stool was homogenized in Buffer ASL using a

| Statistical analysis
The statistical significance of the alpha diversity indices between different timepoints was derived using Wilcoxon signed rank test. The same test was also conducted to compare the expression of SCFAs across timepoints. For beta diversity, the data were first transformed into Aitchison distance, and the differences between the bacterial composition across timepoints and subject groups (i.e. control and HS) were compared using Factorial, Repeated Measure and Pairwise PERMANOVA separately (vegan R and pairwiseAdonis R). 38,39 In addition, differentially abundant taxa were identified using LEfSE 40 while the differentially abundant PICRUSt2 pathways were inferred using LIMMA R Package 41 adjusting for multiple corrections using the Benjamini-Hochberg approach. The associations between the bacterial taxa with SCFAs and pathways with SCFAs were established using the sparse partial least square analysis (sPLS) implemented by the mixOmics R package. 42 Statistical Analysis for Clinical Associations between clinical characteristics and clinical outcomes (HiSCR) was completed using the Wilcoxon rank-sum test. 43 Adjustment was made using the Bonferroni method for multiple comparisons. p < 0.05 was considered statistically significant. All statistical analysis for clinical associations was undertaken using Prism (8.4.2).

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.