Blood T Helper Memory Cells: A Tool for Studying Skin Inflammation in HS?

Hidradenitis suppurativa (HS) is an inflammatory skin disease characterized by painful lesions on intertriginous body areas such as the axillary, inguinal, and perianal sites. Given the limited treatment options for HS, expanding our knowledge of its pathogenetic mechanisms is a prerequisite for novel therapeutic developments. T cells are assumed to play a crucial role in HS pathogenesis. However, it is currently unknown whether blood T cells show specific molecular alterations in HS. To address this, we studied the molecular profile of CD4+ memory T (Thmem) cells purified from the blood of patients with HS and matched healthy participants. About 2.0% and 1.9% of protein-coding transcripts were found to be up- and down-regulated in blood HS Thmem cells, respectively. These differentially expressed transcripts (DETs) are known to be involved in nucleoside triphosphate/nucleotide metabolic processes, mitochondrion organization, and oxidative phosphorylation. The detected down-regulation of transcripts involved in oxidative phosphorylation suggest a metabolic shift of HS Thmem cells towards glycolysis. The inclusion of transcriptome data from skin from HS patients and healthy participants in the analyses revealed that in HS skin lesions, the expression pattern of transcripts identified as DETs in blood HS Thmem cells was very similar to the expression pattern of the totality of protein-coding transcripts. Furthermore, there was no significant association between the extent of the expressional changes in the DETs of blood HS Thmem cells and the extent of the expressional changes in these transcripts in HS skin lesions compared to healthy donor skin. Additionally, a gene ontology enrichment analysis did not demonstrate any association of the DETs of blood HS Thmem cells with skin disorders. Instead, there were associations with different neurological diseases, non-alcoholic fatty liver disease, and thermogenesis. The levels of most DETs linked to neurological diseases showed a positive correlation to each other, suggesting common regulatory mechanisms. In summary, the transcriptomic changes in blood Thmem cells observed in patients with manifest cutaneous HS lesions do not appear to be characteristic of the molecular changes in the skin. Instead, they could be useful for studying comorbidities and identifying corresponding blood biomarkers in these patients.


Introduction
Hidradenitis suppurativa (HS; also referred to as acne inversa) is a chronic inflammatory disease with persistence of characteristic skin alterations [1]. These include painful, deep-seated, inflamed nodules and abscesses that occur in early adulthood in the axillary, inguinal, and perianal skin areas. As the disease progresses, pus-draining sinus tracts develop and the normal skin architecture is destroyed, resulting in scars and contractures [1]. The prevalence of HS in Europe, North America, and Australia is about 0.5-2% [1,2]. In these countries, HS affects both sexes with similar frequencies; however, men more often show lesions in the axillary area and women in the inguinal area [3][4][5]. Patients with HS frequently suffer from anxiety, limitations in many areas of life, including their sexual and professional lives, and stigmatization [6][7][8][9]. Furthermore, patients with HS are often affected by metabolic syndrome, nonalcoholic fatty liver disease, atherosclerosis, spondyloarthritis, inflammatory bowel disease, schizophrenia, and bipolar disorders [10][11][12][13][14][15][16][17][18]. The importance of HS arises not only from its profound negative impact on the health as well as the social and professional lives of patients but also from its strong negative socio-economic impact on society. The latter mainly results from the enormous loss of the gross value added, which was estimated at EUR 13 billion per year for Germany [19].
The treatment options for HS are currently very limited [1]. They include the surgical removal of the altered skin areas, the long-term systemic administration of antibiotics, and administration of the anti-TNF-α antibody adalimumab [1]. Satisfactory, long-lasting results are not currently achieved in most patients treated in this way [6,20]. This makes the search for new treatment concepts imperative [21][22][23][24][25][26]. However, a thorough understanding of HS's pathogenesis is necessary for the development of targeted, highly effective therapeutics.
The etiopathogenesis of HS is only partially understood [1]. Genetic predisposition, smoking, and obesity are important factors that lead to the onset of HS [27,28]. An attenuated production of antibacterial proteins seems to facilitate the persistence of specific bacteria in HS skin lesions [29,30]. Bacterial products and components of damaged skin cells might be responsible for the permanent stimulation of immune cells, leading to the attraction of different immune cell types from the blood into the skin and their activation there [27]. These processes are reflected at the molecular and clinical levels by the massive production of cytokines and the recurring development of nodules and abscesses with pus discharge, respectively. In fact, analyses of HS lesions demonstrated strongly elevated expression levels of several immune mediators [29,[31][32][33][34][35][36][37][38][39]. Beyond cytokines, activated fibroblasts and neutrophilic granulocytes secrete matrix metalloproteinases, which mediate the tissue destruction frequently observed in HS [35,37]. The results of placebo-controlled clinical trials imply a relevant role of IL-17A in the processes underlying the persistence of skin alterations in HS [40,41]. IL-17 is a cytokine that is mostly produced in humans by a specific CD4 + T memory (Th mem ) cell population [42,43]. However, investigations of Th mem cells in HS are rare [44][45][46]. Therefore, we analyzed blood Th mem cells from patients with HS and control participants in this project.

Results and Discussion
To gain insight into the role of Th mem cells in HS pathogenesis, we aimed to characterize the molecular profile of these cells. Therefore, we used RNA sequencing to analyze the transcriptome of Th mem cells purified from the blood of 15 patients with HS in comparison to healthy participants (n = 13) matched for age, gender, and smoking habit (Table 1).
A principal component analysis (PCA) of the transcriptome data revealed molecular differences between the Th mem cells from patients with HS and healthy participants, as obvious by a sub-clustering of cells from both cohorts (Figure 1a). Importantly, as shown in Figure 1b, a Z-score analysis disclosed that the molecular differences between the HS and healthy donor Th mem cells were accompanied by a substantial similarity in the gene expression pattern of individual samples among each cohort.  A principal component analysis (PCA) of the transcriptome data revealed molecular differences between the Thmem cells from patients with HS and healthy participants, as obvious by a sub-clustering of cells from both cohorts (Figure 1a). Importantly, as shown in Figure 1b, a Z-score analysis disclosed that the molecular differences between the HS and healthy donor Thmem cells were accompanied by a substantial similarity in the gene expression pattern of individual samples among each cohort.  Among 19,890 protein-coding transcripts, 772 transcripts (3.88%) were found to be differentially expressed in HS Th mem cells compared to healthy donor Th mem cells (Figure 1c; Table S1). Among these 772 differentially expressed transcripts (DETs), 393 (1.98%) were up-regulated and 379 (1.90%) were down-regulated (Figure 1c). To the best of our knowledge, there are no data published on blood Th mem cells from patients with other common chronic inflammatory skin diseases such as psoriasis or atopic dermatitis for comparison.
Interestingly, the DETs most profoundly regulated in blood HS Th mem cells (Figure 1d, Tables 2 and 3) comprise products of genes not previously mentioned in the context of HS. Importantly, there was no significant difference in the expression of IL-17A (p = 0.35), IL-17F (p = 0.95), or IL-26 (p = 0.65) between HS and healthy donor Th mem cells. This implies that either blood Th mem cells do not significantly infiltrate established skin lesions of HS patients or that blood Th mem cells obtain their specific pathogenic profile only after having infiltrated the skin. In the next step, we investigated whether the molecular alterations observed in HS Th mem cells were associated with specific cellular processes and performed a gene ontology (GO) term enrichment analysis. The identified DETs were annotated to 86 GO biological process terms. The top 15 enriched cellular processes included purine nucleotide metabolic processes, nucleoside triphosphate metabolic processes, mitochondrion organization, and respiratory chain oxidative phosphorylation (Figure 2a). This might suggest alterations in the oxidative phosphorylation, ATP synthesis, RNA transcription, DNA replication, and signal transduction in these cells. Given the fact that transcripts involved in oxidative phosphorylation were found to be down-regulated in this analysis (Figure 2b), a more pronounced shift towards glycolysis-dependent energy production can be assumed (Figure 2c) for HS Th mem cells compared to healthy donor Th mem cells. Glycolysis has been shown to be enhanced during the activation of naïve CD4 + T cells and the differentiation of these cells into CD4 + effector/memory subsets such as Th17 cells. However, permanent glycolysis was associated with an exhausted phenotype of CD4 + T cells and the occurrence of autoimmune diseases [47][48][49]. Therefore, our data may indicate that the blood HS Th mem cells have a long history of activation and/or have re-migrated into the blood from the chronically inflamed tissues. The second possibility is particularly interesting but speculative and should be investigated in further studies. blood from the chronically inflamed tissues. The second possibility is particularly interesting but speculative and should be investigated in further studies. ribonucleoside triphosphate metabolic process purine ribonucleoside triphosphate metabolic process purine nucleoside triphosphate metabolic process nucleoside triphosphate metabolic process ATP metabolic process mitochondrion organization purine nucleotide metabolic process purine ribonucleotide metabolic process purine nucleoside monophosphate metabolic process purine-containing compound metabolic process respiratory electron transport chain ribonucleotide metabolic process ribose phosphate metabolic process mitochondrial ATP synthesis coupled electron transport ATP synthesis coupled electron transport  Outer mitochondrial membrane  In order to identify potential master regulators of the DETs of blood HS Th mem cells, we performed an overrepresentation analysis of transcription factor target gene sets using the C3 gene set collection of the Molecular Signature Database. This analysis revealed NFRKB (nuclear factor related to kappa-B binding protein) as the master regulator responsible for the regulation of the largest proportion of the DETs of blood HS Th mem cells (Figure 2d). It was suggested that NFRKB is a telomere-associated protein and might influence cell proliferation, but the biological function of this transcription factor is still not well understood [50]. In addition to NFRKB, SETD1A (SET domain containing 1A) was found to potentially regulate the second largest group of DETs identified for blood Th mem cells ( Figure 2d). As member of the SET1/MLL family methyltransferases, SETD1A was described as influencing chromatin conformation by modulating the methylation of histone H3 at lysine 4 (H3K4me3), resulting in transcriptional activation of the affected genes [51]. Interestingly, it was reported that mutations in SETD1A substantially increase the risk for the development of schizophrenia [51,52].
Next, we asked whether the transcriptomic signature of blood HS Th mem cells can be found in the lesional skin of HS patients. For these analyses, we used our recently published transcriptome data of the skin samples taken from the axillary region of HS patients and from equal sites of healthy participants [35,38]. First, we tested whether individual DETs detected in blood HS Th mem cells were significantly up-regulated, downregulated, or unchanged in the transcriptome of HS lesions compared to healthy donor skin. Surprisingly, most transcripts up-regulated or down-regulated in blood HS Th mem cells were not found to be regulated in the same direction in HS skin lesions ( Figure 3 or Table S2). In fact, the majority of transcripts corresponding to the DETs of blood HS Th mem cells (74.9% of up-regulated DETs; 83.2% of down-regulated DETs) were unregulated in HS skin lesions. Moreover, the proportions of transcripts up-regulated in HS lesions did differ between the following groups: transcripts up-regulated in blood HS Th mem cells (14.6%), transcripts down-regulated in blood HS Th mem cells (13.1%), and the totality of transcripts in HS skin (15.6%) (Figure 3 or Table S2). Even more surprising was that the proportion of transcripts down-regulated in HS lesions among the down-regulated transcripts in blood HS Th mem cells (3.6%) was significantly smaller than among all transcripts (12.5%) ( Figure 3 or Table S2). In order to identify potential master regulators of the DETs of blood HS Thmem cel we performed an overrepresentation analysis of transcription factor target gene sets usin the C3 gene set collection of the Molecular Signature Database. This analysis reveale NFRKB (nuclear factor related to kappa-B binding protein) as the master regulator respo sible for the regulation of the largest proportion of the DETs of blood HS Thmem cells (Fi ure 2d). It was suggested that NFRKB is a telomere-associated protein and might influen cell proliferation, but the biological function of this transcription factor is still not we understood [50]. In addition to NFRKB, SETD1A (SET domain containing 1A) was foun to potentially regulate the second largest group of DETs identified for blood Thmem ce ( Figure 2d). As member of the SET1/MLL family methyltransferases, SETD1A was d scribed as influencing chromatin conformation by modulating the methylation of histon H3 at lysine 4 (H3K4me3), resulting in transcriptional activation of the affected genes [51 Interestingly, it was reported that mutations in SETD1A substantially increase the risk f the development of schizophrenia [51,52].
Next, we asked whether the transcriptomic signature of blood HS Thmem cells can b found in the lesional skin of HS patients. For these analyses, we used our recently pu lished transcriptome data of the skin samples taken from the axillary region of HS patien and from equal sites of healthy participants [35,38]. First, we tested whether individu DETs detected in blood HS Thmem cells were significantly up-regulated, down-regulate or unchanged in the transcriptome of HS lesions compared to healthy donor skin. Surpri ingly, most transcripts up-regulated or down-regulated in blood HS Thmem cells were n found to be regulated in the same direction in HS skin lesions ( Figure 3 or Table S2). fact, the majority of transcripts corresponding to the DETs of blood HS Thmem cells (74.9 of up-regulated DETs; 83.2% of down-regulated DETs) were unregulated in HS skin l sions. Moreover, the proportions of transcripts up-regulated in HS lesions did differ b tween the following groups: transcripts up-regulated in blood HS Thmem cells (14.6% transcripts down-regulated in blood HS Thmem cells (13.1%), and the totality of transcrip in HS skin (15.6%) (Figure 3 or Table S2). Even more surprising was that the proportion transcripts down-regulated in HS lesions among the down-regulated transcripts in bloo HS Thmem cells (3.6%) was significantly smaller than among all transcripts (12.5%) (Figu 3 or Table S2).  In HS skin lesions, the transcripts corresponding to the up-and down-regulated DETs of blood HS Th mem cells are not more frequently up-and down-regulated than the totality of lesional protein-encoding transcripts. DETs of blood HS Th mem cells were determined as described in Figure 1b. The frequencies of transcripts corresponding to up-regulated (a) and down-regulated (b) DETs of blood HS Th mem cells that were up-or down-regulated or not-regulated (cutoff: adjusted p < 0.01; log 2 fold change 0.5/−0.5) in HS skin lesions compared to healthy donor skin were determined based on a previously published data set [35,38]. For comparison, the proportions of up-or down-regulated or not regulated (cutoff: adjusted p < 0.01; log 2 fold change 0.5/−0.5) transcripts among the totality of protein coding-transcripts in HS skin lesions compared to healthy donor skin are shown. Differences in the frequencies of up-/down-/not-regulated transcripts corresponding to DETs of blood HS Th mem cells compared to the totality of protein-coding transcripts in HS lesions were tested using the X 2 test (*** p < 0.001). HS, hidradenitis suppurativa, DETs, differentially expressed transcripts. Second, we correlated the log 2 fold change values of the DETs of blood HS Th mem cells versus the log 2 fold change values of the same transcripts that were calculated for HS skin lesions compared to healthy donor skin. As shown in Figure 4, there was no significant correlation between the changes in blood Th mem cells and in skin for either up-regulated (Figure 4a) or down-regulated (Figure 4b) blood Th mem cell DETs. Overall, these data show a lack of a blood-Th mem -cell-specific transcriptomic signature in lesional HS skin and imply that the molecular alterations identified in HS blood Th mem cells are not involved in HS skin pathology. This suggests three possible scenarios for the massive cutaneous presence of T cells, which definitely play a pathogenetic role in the persistence of HS skin lesions: (i) the T cells immigrate into the skin in large numbers before/early in the development of lesions, (ii) they obtain their pathogenic profile only after having immigrated into the skin, or (iii) they arise from the T cells that were already in the skin (e.g., from resident memory cells). The second and third scenario are supported by the presence of the tertiary lymphoid structures in HS lesions [38]. blood HS Thmem cells compared to the totality of protein-coding transcripts in HS lesions were tested using the X2 test (*** p < 0.001). HS, hidradenitis suppurativa, DETs, differentially expressed transcripts.
Second, we correlated the log2 fold change values of the DETs of blood HS Thmem cells versus the log2 fold change values of the same transcripts that were calculated for HS skin lesions compared to healthy donor skin. As shown in Figure 4, there was no significant correlation between the changes in blood Thmem cells and in skin for either up-regulated (Figure 4a) or down-regulated (Figure 4b) blood Thmem cell DETs. Overall, these data show a lack of a blood-Thmem-cell-specific transcriptomic signature in lesional HS skin and imply that the molecular alterations identified in HS blood Thmem cells are not involved in HS skin pathology. This suggests three possible scenarios for the massive cutaneous presence of T cells, which definitely play a pathogenetic role in the persistence of HS skin lesions: (i) the T cells immigrate into the skin in large numbers before/early in the development of lesions, (ii) they obtain their pathogenic profile only after having immigrated into the skin, or (iii) they arise from the T cells that were already in the skin (e.g., from resident memory cells). The second and third scenario are supported by the presence of the tertiary lymphoid structures in HS lesions [38]. Since HS is a complex, systemic disease with multiple comorbidities [1], we wondered whether the DETs detected in blood HS Thmem cells were associated with signatures known from other diseases. As presented in Figure 5, a gene ontology enrichment analysis for human disease association demonstrated a link of the DETs detected in blood HS Thmem cells with different neurological disorders, as well as adipose-tissue-involving thermogenesis, oxidative-stress-mediated carcinogenesis, non-alcoholic fatty liver disease, and virus infection. Interestingly, depression, schizophrenia, and bipolar disorder as well as metabolic alterations and non-alcoholic fatty liver disease were demonstrated to be associated with HS [10,11,17,53,54].
It should be noted that in these analyses, we did not observe any enrichment of transcripts associated with skin disease (Figure 5), supporting our observation described above that in patients with manifest HS skin alterations, changes in the blood Thmem cells are no indicators of the changes in the skin. Since HS is a complex, systemic disease with multiple comorbidities [1], we wondered whether the DETs detected in blood HS Th mem cells were associated with signatures known from other diseases. As presented in Figure 5, a gene ontology enrichment analysis for human disease association demonstrated a link of the DETs detected in blood HS Th mem cells with different neurological disorders, as well as adipose-tissue-involving thermogenesis, oxidative-stress-mediated carcinogenesis, non-alcoholic fatty liver disease, and virus infection. Interestingly, depression, schizophrenia, and bipolar disorder as well as metabolic alterations and non-alcoholic fatty liver disease were demonstrated to be associated with HS [10,11,17,53,54].
It should be noted that in these analyses, we did not observe any enrichment of transcripts associated with skin disease (Figure 5), supporting our observation described above that in patients with manifest HS skin alterations, changes in the blood Th mem cells are no indicators of the changes in the skin.
The expression levels of most of these DETs showed a strong positive correlation to each other, suggesting similar regulatory mechanisms of their expression in the blood HS Thmem cells (Figure 6a). Importantly, most of the 21 transcripts common to all the neurological diseases were down-regulated in blood HS Thmem cells compared to the Thmem cells obtained from healthy participants (Figure 6b). Interestingly, a down-regulation of the expression of several ATP-synthase-encoding genes and numerous proteasome subunit genes was recently demonstrated in the brain of patients with schizophrenia [55,56]. More importantly, Song et al. found that the transcripts associated with mitochondrial oxidative phosphorylation have similar expression patterns in blood immune cells and brain tissues in patients with schizophrenia [57]. It should be mentioned that despite distinct clinical symptoms, the accumulation of misfolded or aggregated proteins in the brain is characteristic of many neurodegenerative disorders, such as amyotrophic lateral sclerosis, Huntington's disease, Parkinson's disease, or Alzheimer's disease. Growing evidence suggests that dysregulation in intracellular protein degradation mechanisms (the ubiquitin-proteasome system or autophagy) leads to proteostasis, contributing to the onset and persistence of these diseases [58][59][60][61]. Thus, the enhancement of proteasome-mediated protein degradation is a promising new therapy option for neurodegenerative disorders that is already being investigated [62][63][64]. The identification of the DETs of blood HS Thmem cells associated with neurological diseases raises the hope that some of these molecules might be suitable as blood biomarkers that are predictive for neurological comorbidities in HS. The expression levels of most of these DETs showed a strong positive correlation to each other, suggesting similar regulatory mechanisms of their expression in the blood HS Th mem cells (Figure 6a). Importantly, most of the 21 transcripts common to all the neurological diseases were down-regulated in blood HS Th mem cells compared to the Th mem cells obtained from healthy participants (Figure 6b). Interestingly, a down-regulation of the expression of several ATP-synthase-encoding genes and numerous proteasome subunit genes was recently demonstrated in the brain of patients with schizophrenia [55,56]. More importantly, Song et al. found that the transcripts associated with mitochondrial oxidative phosphorylation have similar expression patterns in blood immune cells and brain tissues in patients with schizophrenia [57]. It should be mentioned that despite distinct clinical symptoms, the accumulation of misfolded or aggregated proteins in the brain is characteristic of many neurodegenerative disorders, such as amyotrophic lateral sclerosis, Huntington's disease, Parkinson's disease, or Alzheimer's disease. Growing evidence suggests that dysregulation in intracellular protein degradation mechanisms (the ubiquitin-proteasome system or autophagy) leads to proteostasis, contributing to the onset and persistence of these diseases [58][59][60][61]. Thus, the enhancement of proteasome-mediated protein degradation is a promising new therapy option for neurodegenerative disorders that is already being investigated [62][63][64]. The identification of the DETs of blood HS Th mem cells associated with neurological diseases raises the hope that some of these molecules might be suitable as blood biomarkers that are predictive for neurological comorbidities in HS. Unfortunately, when recruiting the HS patients in this project, the presence of neurological diseases was not inquired about, so we cannot correlate the expression levels of the identified DETs with the presence or absence of the corresponding neurological disorders. In addition, there was no follow-up in this project; thus, we cannot make any statement as to whether the corresponding HS patients developed neurological diseases. Finally, as the results of the study were not foreseeable when it was planned, patients with only neurological diseases where not recruited as a control group. These are certainly limitations of our research project. as to whether the corresponding HS patients developed neurological diseases. Finally, as the results of the study were not foreseeable when it was planned, patients with only neurological diseases where not recruited as a control group. These are certainly limitations of our research project. In summary, we identified numerous transcriptomic changes in Thmem cells isolated from the blood of patients with HS. To our knowledge, this is the first publication demon- Figure 6. Expression levels of DETs of blood HS Th mem cells associated with different neurological diseases show positive associations with each other. DETs of blood HS Th mem cells were determined as described in Figure 1b and subjected to KEGG pathway enrichment analysis as described in Figure 5. (a) Expression levels of DETs of blood HS Th mem cells that were associated with all neurological diseases shown in Figure 5 (amyotrophic lateral sclerosis, Huntington disease, Parkinson disease, Prion disease, Alzheimer disease, Pathways of neurodegeneration-multiple diseases) were correlated to each other. Data are presented as correlation matrix. Spearman's correlation coefficient R and p-values are indicated by the color scale (see legend) and size (maximum size: p < 0.001, 75%: p < 0.01, 50%: p < 0.05, 10%: p ≥ 0.05) of circles, respectively. (b) Log 2 of fold change expression (Th mem cells isolated from the blood of HS patients versus those from healthy donor participants) values of DETs shown in (a) are given using a visual color scale. HS, hidradenitis suppurativa, DETs, differentially expressed transcripts.
In summary, we identified numerous transcriptomic changes in Th mem cells isolated from the blood of patients with HS. To our knowledge, this is the first publication demonstrating the results of such an analysis in HS patients. The transcriptomic alterations concern proteins involved in nucleotide metabolism and ATP synthesis through oxidation in the mitochondria. The changes identified in HS blood Th mem cells were not found in the HS skin lesions and are unlikely to contribute to the persistence of skin alterations.
We consider it a very important finding of this project that the analysis of the blood cells obtained from patients with apparent skin alterations did not provide much insight into the cutaneous HS pathology. Rather, one has the impression that the Th mem cells found in HS patients with developed lesions could be responsible for potential comorbidities. Future studies are required to prove this hypothesis.

Patients and Healthy Donors
Blood samples were obtained from 15 healthy adult volunteers and 15 adult HS patients who did not receive any systemic therapy for at least 3 months before participating in the study. Both participant groups were matched to each other in terms of gender distribution (healthy/HS: 60% female, 40% male), age (healthy: 36.9 ± 2.4; HS: 37.5 ± 3.1) and smoking habit (healthy: 60% smokers, 40% non-smokers; HS: 60% smokers, 40% nonsmokers). Skin samples were obtained from healthy participants (n = 4) and patients with HS (n = 3; lesional skin from axillary region).
The collection of all blood and skin samples was approved by the clinical institutional review board of the Charité University Hospital, Berlin, and written informed consent was obtained from all participants. The study was conducted according to the principles of the Declaration of Helsinki.

Cell Isolation
Peripheral blood mononuclear cells (PBMCs) were isolated from the venous blood of patients and healthy donors using density gradient centrifugation as described earlier [65]. CD4 + memory T cells were purified from the obtained PBMCs via magnetic cell sorting using the MACSTM system and the memory CD4 + T cell isolation kit (negative selection) from Miltenyi Biotec. The purity of the isolated T mem cells, as analyzed by flow cytometry (see below), was 97.63 ± 0.18 (healthy donors) and 96.72 ± 0.52 (HS patients) ( Figure S1 and Table S3).

Flow Cytometry
An analysis of the cell purity of the isolated T mem cells was performed via flow cytometry, using the following fluorescent labeled antibodies directed against CD45RA (clone HI10), CD45RO (clone UCHL1), CD3 (clone SK7) (all purchased from BD Biosciences, Heidelberg, Germany), and CD4 (clone 13B8.2) (purchased from Beckmann Coulter, Krefeld, Germany), using a FACS Calibur device (BD Biosciences) as described earlier [65]. Data processing and analysis were performed using associated CellQuest Pro software (version 4.0.2).

Transcriptome Analysis
RNA was isolated from the purified Th mem cells using a Trizol lysing solution (Thermo Fisher, Waltham, MA, USA) according to the manufacturer's instructions. A library preparation from the total RNA was performed using the TruSeq stranded total RNA library prep kit with ribo-zero gold (Illumina, San Diego, CA, USA) according to the manufacturer's recommendation, followed by the sequencing of libraries (50 bp single reads; 30 million reads per sample) using a HiSeq 2000 platform (Illumina). The transformation of the fluorescence images into BCL files was accomplished using the Illumina BaseCaller software followed by the demultiplexing of samples to FASTQ files using CASAVA (version 1.8.2). The sequencing quality was checked using FastQC software (www.bioinformatics.babraham. ac.uk/projects/fastqc, accessed on 3 May 2023). Reads were aligned to the hg38 using Salmon, which is available from https://anaconda.org/bioconda/salmon (accessed on were determined after the data were fitted to models of negative binomial distributions. Raw p-values were fdr-corrected for multiple testing. Shrinkage of the transcriptome data was performed using the lfcShrink function (type = "normal") using DESeq2 in R. A heatmap was generated using the inbuilt heatmap function of R with a predefined order of the samples.
Skin samples from healthy donors and HS patients were processed for transcriptome analysis as previously described [35,38]. The obtained Th mem cell and skin transcriptome data were filtered for protein coding genes, and cutoffs for the differentially expressed transcripts (DETs) were set to an adjusted p < 0.01; log 2 fold change 0.5/−0.5.

Gene Set Enrichment Analysis
An overrepresentation analysis of the differentially expressed genes in terms of gene ontology and in KEGG pathways after DAVID annotation was performed with the clus-terProfiler package. The visualization of the DETs of blood HS Th mem cells involved in oxidative phosphorylation and glycolysis was performed with Qiagen IPA software (v01-22-01). For this analysis, the DETs involved in the intracellular processes shown in Figure 2a were used.

Overrepresentation of Regulatory Target Gene Set Analysis
To identify possible master regulators of target genes found to be differentially regulated in the HS Th mem cells, the msigdbr R-package was used (msigdbr: MSigDB Gene Sets for Multiple Organisms in a Tidy Data Format (https://cran.r-project.org/ web/packages/msigdbr/msigdbr.pdf, accessed on 3 May 2023); R package version 7.5.1, https://CRAN.R-project.org/package=msigdbr, accessed on 3 May 2023). The enrichment of genes within these sets was determined by the enricher function of the clusterProfiler R-package [66]. Raw p-values were fdr-corrected for multiple testing, and an adjusted p-value below 0.05 was considered significant.

Further Statistical Analysis
A correlation analysis of the log 2 fold-change values of the blood Th mem cells versus the HS skin (Spearman's rank-order correlation test; Figure 4) and a comparison of DET frequencies (X 2 test; Figure 6a) were performed using SPSS software (version 29.0.0.0(241)).