Increased Levels of Inflammatory Proteins, Including <scp>TARC</scp>/<scp>CCL17</scp>, in Skin of <scp>AD</scp> Patients During <scp>JAK</scp> Inhibitor Treatment

Atopic dermatitis (AD) is a complex and heterogeneous inflammatory skin disease that not only involves T helper (Th)2 responses, but also Th1, Th17 and Th22 cytokine pathways. Thymus and activation-regulated chemokine (TARC)/CCL17 is a type 2 chemokine that is highly expressed in AD skin and blood. Previous research showed that serum TARC levels significantly correlate with disease severity [1]. To date, TARC has been identified as the most reliable clinical biomarker to measure AD severity and to evaluate treatment response [2]. However, we recently reported that serum TARC might not be an adequate biomarker in AD patients treated with Janus kinase (JAK)-inhibitors (JAKi), as we found persistently high serum TARC levels despite a good treatment response [3]. To further investigate this, we aimed to assess local skin levels of TARC and other inflammatory proteins of AD patients treated with JAKi. We measured 11 inflammatory proteins, including TARC, in tape strips from adult AD patients with a good clinical response to JAKi (upadacitinib or abrocitinib, Eczema Area Severity Index [EASI] ≤ 7), who had either persistently elevated (JAKh, n = 5) or normalised (JAKn, n = 5) serum TARC levels. Patients were retrospectively selected by serum TARC levels measured during routine diagnostics. Tape strips were collected of lesional and non-lesional skin (time points [Tx] vary per patient, ≥ 1 year of treatment) and were compared to tape strips of lesional skin from patients with active AD without systemic treatment (ADcontrol, n = 5). Additionally, tape strips were collected from nonatopic healthy controls (HCs, n = 3). Besides TARC, disease-related proteins (pulmonary and activation-regulated chemokine [PARC], cutaneous T-cell-attracting chemokine [CTACK], interleukin [IL]-13 and periostin), pro-inflammatory proteins (IL-18, IL-8, IP-10 and IL-1α), and tissue remodelling proteins (matrix metalloproteinase-1 [MMP-1], IL-15) were measured. Tape stripping of stratum corneum was performed using D-Squame tape strips (3.8 cm2, Standard D-Squame, Clinical & Derm TX USA) as previously described [4]. Tape strips 5–7 were used for analysis and eluted overnight at 4°C in a PBS elution buffer containing 0.5% Tween 20 and complete protease inhibitor cocktail (Roche Diagnostics). Protein levels were measured by Luminex multiplex immunoassay [5]. In addition, clinical effectiveness was measured by the EASI. For statistical analysis, the Mann–Whitney U test was used to identify differences in skin protein levels, EASI scores and serum TARC levels between patient groups and the Wilcoxon Rank test to compare lesional and non-lesional skin within subgroups. All patients provided written informed consent and participated in the Dutch BioDay registry. All patients were treated at the University Medical Center Utrecht (November 2021–February 2024). Patient characteristics included the following: JAKh patients: all males, median age 30.0 years [interquartile range (IQR) 24.5–31.0], median baseline EASI 14.7 [IQR 12.1–24.6]; treatment: 15 mg [n = 2] and 30 mg QD upadacitinib [n = 1] and 200 mg QD abrocitinib [n = 2]; mean treatment duration: 20.0 months [SD 1.9]; JAKn patients: 60% male, median age 37.8 years [IQR 24.5–52.0], median baseline EASI 12.8 [IQR 11.2–30.4]; treatment: 15 mg QD upadacitinib [n = 3] and 200 mg QD abrocitinib [n = 2]; mean treatment duration: 18.4 months [SD 6.8]; AD control patients: 60% male, median age 42.0 years [IQR 22.5–58.0], median baseline EASI 14.7 [IQR 12.1–24.6]. EASI scores and serum TARC levels at baseline (T0) and Tx are depicted in Figure 1A, demonstrating persistent elevated TARC levels in the JAKh group compared to the JAKn group, despite a comparable decrease in EASI score. In lesional skin of both JAKh and JAKn patient groups, levels of TARC and other disease-related proteins (PARC, CTACK and periostin) and pro-inflammatory proteins (IL-8, IL-18, IP-10 and MMP1) were elevated compared to non-lesional skin, with considerable interpatient variation (Figure 1B). Corresponding to the differential TARC levels in serum, TARC levels in lesional skin were significantly higher in the JAKh group than in the JAKn group, and also higher than in the lesional skin of untreated AD patients. As expected, IL-1α levels were relatively low in lesional skin [6]. Furthermore, IL-13 and IL-15 could not be detected. Together, these findings indicate that during JAKi treatment not only TARC but also other severity-associated and pro-inflammatory proteins can remain elevated or become induced in lesional skin of AD patients, despite showing a good clinical response. Additionally, the elevated skin TARC levels correspond with the persistently high serum TARC levels observed in the JAKh group and were even higher than those observed in the skin of untreated AD patients [3]. This is in contrast to previous research showing that skin proteins correlated with AD severity (i.e., TARC, PARC and CTACK) in lesional tissue at respectively RNA-seq and mRNA level, decreased during either dupilumab or abrocitinib treatment [7-9]. The relatively lower levels of TARC in skin compared to PARC/CTACK likely reflect a chemokine profile dominated by local epidermal production of PARC and CTACK, whereas high serum TARC levels reflect its deeper dermal and systemic origin [1]. In our patients treated with JAKi, other pro-inflammatory cytokines such as IL-8 and IL-18 were also elevated, suggesting potential (persisting) immune dysregulation during treatment. Our study is limited by the small sample size and the retrospective selection of JAKi patients based on serum TARC levels. Moreover, the biomarkers were eluted from tape strips, which might underrepresent protein levels in the epidermis. Nevertheless, our results indicate that JAKi incompletely suppress the immune dysregulation in the epidermis of (a subgroup of) AD patients, despite a favourable clinical response. These patients may therefore be prone to disease relapse upon cessation of therapy. Larger and prospective studies are needed to further explore potential changes in biomarker levels in both serum and (epi)dermis in terms of safety and possible ongoing immune dysregulation during JAKi treatment. Conceptualisation: C.M.B., L.F.G., D.S.B., M.G., E.F.K., M.S.B.-W., F.W. Formal analysis: C.M.B. Funding acquisition: M.G., M.S.B.-W. Investigation: C.M.B., L.F.G., D.S.B., M.G., M.S.B.-W., E.F.K. Methodology: C.M.B., M.S.B.-W., F.W., E.F.K. Resources: C.M.B. Supervision: M.G., E.F.K., M.S.B.-W., F.W. Visualisation: C.M.B., E.F.K., M.S.B.-W., F.W., E.F.K. Writing – original draft preparation: C.M.B., L.F.G. Writing – review and editing: L.F.G., D.S.B., M.G., E.F.K., M.S.B.-W., F.W. We would like to thank our patients included in this manuscript for participating in the BioDay registry. This study has no funding, but the BioDay registry is supported by Eli Lilly and Company, Sanofi, Leo Pharma and Pfizer. The study was part of the BioDay registry which was approved by the local Medical Research Ethics Committee as a non-interventional study (METC 18–239) and was performed according to the declaration of Helsinki. C.M. Boesjes is a speaker for AbbVie and Eli Lilly. L.F. van der Gang is a speaker for AbbVie and Sanofi. Dr. D.S. Bakker is a speaker for Sanofi, Novartis, Eli Lilly and LEO Pharma. Dr. M. de Graaf is a consultant, advisory board member and/or speaker for AbbVie, Almirall, Eli Lilly, Janssen, Leo Pharma, Novartis, Pfizer, Regeneron Pharmaceuticals and Sanofi. Prof. Dr. M. S. de Bruin-Weller is a consultant, advisory board member and/or speaker for AbbVie, Almirall, Aslan, Arena, Eli Lilly, Galderma, Janssen, Leo Pharma, Pfizer, Regeneron Pharmaceuticals and Sanofi. Prof. F. van Wijk is a speaker and/or consultant for Janssen, Johnson&Johnson, and Takeda and has received grants from Regeneron Pharmaceuticals, Leo Pharma, Sanofi, BMS, Galapagos and Takeda. Dr. E.F. Knol is a speaker and/consultant for Sanofi, Thermo Fisher Scientifc and GSK. The data that support the findings of this study are available from the corresponding author upon reasonable request.