Stratum corneum levels of inflammatory mediators and natural moisturizing factor in patch test reactions to thiurams and fragrances and their possible role in discrimination between irritant and allergic reactions to hapten mixtures

Patch test (PT) reactions to thiuram mix (TM) and fragrance mix (FM) I or II without concomitant reactions to their single constituents are potentially caused by the irritant properties of the mixes.

Patch testing is the gold standard for diagnosis of contact allergy. Due to an inherent irritant potential of many haptens, differentiation between true allergic and irritant patch test reactions could be challenging, in particular when hapten mixtures are tested.
Fragrances and rubber accelerators such as thiurams are among the most common causes of contact allergy. [4][5][6][7] They are patch tested as screening mixtures in the European and most national baseline series. When a patch test reaction to thiuram mix (TM) or fragrance mix (FM) I or FM II occurs, testing their single ingredients is recommended. A patch test reaction to a mix without concomitant reaction to at least one of its single ingredients is fairly common when testing the TM (7.3%-53.6%), [8][9][10] FM I (42.7%-67.3%), 4,11,12 or FM II (17.0%-43.0%). 4,11,12 This phenomenon occurs more frequently when the reaction to the mix is weak, and it may indicate an irritant reaction to the hapten mixture. 13 However, also false-negative reactions to the single components have been discussed.
Haptens have different physicochemical properties eliciting different cutaneous inflammatory responses. 2 Inflammatory profiles may help to distinguish between haptens, and possibly even between allergic and irritant responses. 14 RNA microarrays have been used to investigate changes in gene expression in patch test reactions to various haptens showing a selective upregulation of several chemokines in nickel-induced ACD, 15 and different types of immune polarization with respect to Th1/Th17, Th22, and Th2 components for patch testing with haptens including fragrances and thiurams. 16 Previously it was demonstrated that the levels of various inflammatory mediators in the stratum corneum (SC), the uppermost layer of the epidermis, differed after patch testing with nickel, chromium, methylchloroisothiazolinone/methylisothiazolinone (MCI/MI), and the skin irritant sodium lauryl sulfate (SLS). Levels of interleukin (IL)-16 were increased in patch test reactions to all haptens but not to SLS rendering it a potential biomarker to differentiate between ACD and irritant contact dermatitis (ICD). 17 Moreover, SC levels of the natural moisturizing factor (NMF), which mainly consists of degradation products of the epidermal protein filaggrin, were decreased after patch testing with SLS, and similarly with MCI/MI, which was explained by its irritant properties. 18 In the present study we investigated the inflammatory SC profiles of patch test reactions to TM, FM I, FM II, and their respective constituents. In addition, we assessed the suitability of theses profiles to potentially discriminate between irritant and allergic reactions to the mixes.

| Selection of patients and patch testing
The experimental protocol followed the Declaration of Helsinki principles and was approved by the ethics committee of the University of Osnabrück. As part of the routine diagnostic procedure, patients with predominantly work-related hand dermatitis were patch tested at the Institute for interdisciplinary Derma-  21 They were tested in the mix at a concentration of 1% pet. each, resulting in a combined concentration of 8%. For breakdown testing they were tested also at 1% pet. The test preparations of FM I and several of its single ingredients contained the emulsifier sorbitan sesquioleate (SSO) at concentrations of 5% and 1%, respectively. 22 FM II consisted of citral (1%), citronellol (0.5%), coumarin (2.5%), farnesol (2.5%), α-hexyl cinnamal (5.0%), and hydroxyisohexyl 3-cyclohexene carboxyldehyde (HICC) (2.5%). They were tested in the mix at a concentration of 14% pet. 23 Its single components were tested in pet. at double the concentration that they have in the mix. 24 Patients with patch test reactions (?+, + or ++) to FM I, FM II, or TM at D3 were invited to participate in the study, and written informed consent was obtained from each participant.

| Sequential tape stripping of stratum corneum
Tape stripping was performed at D3 with round adhesive discs (3.8 cm 2 , D-Squame, CuDerm, Dallas, Texas) from patch test reactions to the mixes, and if present, to single constituents of the respective mixes as well as skin sites patch tested with pet. 25 Eight consecutive discs from each skin site were collected. Each disc was pressed on for 10 seconds with standardized force using a disc pressure applicator (CuDerm).

| Natural moisturizing factor (NMF) analysis
NMF was defined as the sum of the concentrations of histidine, 2-pyrrolidone-5-carboxylic acid, and trans-and cis-isomers of urocanic acid. NMF was extracted from tape strips number six with 600 μL of millipore water and subsequently analyzed by high-performance liquid chromatography (HPLC-UV). NMF levels were corrected for the amount of protein on the tape, which was determined with the D-Squame Scan 850A instrument (Heiland electronic, Wetzlar, Germany). 25,26

| Multiplex analysis
The samples were extracted from the tape strips number eight by 0.5 mL of phosphate-buffered saline containing 0.05% Tween 20 and sonicated for 15 minutes, as described previously. 17 After vortexing, the extract aliquots were distributed in vials and stored at −80 C until analysis. The analysis of inflammatory mediators from the extracts was performed using the V-Plex multiplex assays and a MESO QuickPlex SQ 120 reader (both MSD, Rockville, Maryland). Based on a previous study, 17

| Patients and patch test reactions
We included 36 patients (25 women, average age: 47 years) with 75 patch test reactions in the study (Table S1). Two patients were excluded (#4 and #17) because they withdrew from the study before sample collection was completed. In 14 patients, a patch test reaction to TM was tape stripped (group T-mix). All of these patients had a concomitant patch test reaction to ≥1 single constituent of the TM (Table S2) Tables 1 and S1. Most patch test reactions were + (73.3%).
The highest share of doubtful reactions was in group F-mix w/o (50.0%).

| Inflammatory mediators
Only the results of + and ++ patch test reactions were included in the first analysis (Figure 1 and S1). Compared with the corresponding pet. Abbreviations: F-mix w/o , reaction to fragrance mix I or II without positive breakdown testing; F-mix with , reaction to fragrance mix I or II with positive breakdown testing; F-single, reaction to single fragrance of fragrance mix I or II; T-mix, reaction to thiuram mix; T-single, reaction to single thiuram of thiuram mix.
inflammatory mediators compared with the corresponding pet. controls were observed in patch test reactions to FM I/II without subsequent patch test reaction to a single constituent (F-mix w/o , n = 6).
To investigate whether the observed differences between patch test reactions (+, ++) to FM I/II with and to FM I/II without subsequent reaction to a single fragrance were related to the different strengths of reactions in both groups, a second analysis was done that included + reactions only of F-mix w/o (n = 6) and F-mix with (n = 10) compared with the corresponding pet. controls.
In addition, the doubtful (?+) patch test reactions of F-mix w/o (n = 6) were compared with their corresponding pet. controls. This was done for all inflammatory mediators that had shown significant differences for F-mix with in the first analysis (IL-16, IL-8/CXCL8, IP-10/CXCL10, TARC/CCL17, MDC/CCL22, and IL-1β). In F-mix with with subsequent reaction to a single fragrance, the levels of all mediators were still significantly higher than in the pet. controls ( Figure 2). No significant differences compared to corresponding pet. controls were found for doubtful (?+) or + reactions of F-mix w/o . In some + patch test reactions of F-mix w/o , the levels of inflammatory mediators increased. However, these changes were overall not as consistent and strong as in F-mix with . The median levels detected in + patch test reactions of F-mix with were significantly higher than the ones in doubtful (?+) reactions of F-mix w/o for IP-10/CXCL10 (P < .05) and CCL17 (P < .05). Comparing the median levels in + patch test reactions between F-mix with and F-mix w/o , only IP-10/CXCL10 was significantly higher in F-mix with (P < .05).

| Natural moisturizing factor (NMF)
In  Figure 3B). When excluding this outlier, still no significant difference compared to pet. controls was found for the remaining levels of NMF (P = .07).

| Correlation analysis
Significant correlations were found between levels of several inflammatory mediators and strength of patch test reactions (Table S3)

| DISCUSSION
Immune mechanisms of ACD are incompletely understood and involve a complex interplay between dendritic cells, keratinocytes, T cell activation, and regulatory T cell-mediated suppression. 2 bind to the skin-homing receptor CCR4 on T cells, which subsequently promotes their migration into inflamed skin. 30,31 CCR4 is expressed to a greater degree on Th2 cells relative to Th1 cells. 32 An amplification loop has been suggested as TARC/CCL17 and MDC/CCL22 attract IL-4 releasing Th2 cells, which in turn enhance Th2-related chemokine production. 33 In addition to mediating T cell migration, TARC/CCL17 plays an important role in cutaneous dendritic cell migration into draining lymph nodes. 34 In mouse models, TARC/CCL17 was shown to promote contact hypersensitivity. 35,36 In humans, a pronounced increase of TARC/CCL17 expression was detected in patch test reactions to nickel, and it was suggested that it plays a major role during the late elicitation phase of ACD. 37 Similarly, Meller et al found an upregulation of TARC/CCL17 and IP-10/CXCL-10 in skin biopsies from patch test reactions to nickel, but not from ICD. 33 In a study by ACD as compared to atopic dermatitis. 15 CXCL8 is a marker of innate immunity and plays a role in the initial hapten-induced chemokine response. 27 Activation by haptens upregulates the production of F I G U R E 3 (A) Levels of natural moisturizing factor (NMF) in patch test reactions (+, ++) to thiuram mix (T-mix), single thiurams (T-single), fragrance mix I/II without subsequent reaction to a single constituent (F-mix w/o ), fragrance mix I//II with subsequent reaction to a single constituent (F-mix with ), and single fragrances (F-single) compared with their corresponding pet. controls. Data are given as median with interquartile ranges. A non-parametric Wilcoxon matched-pairs signed-rank test was used. (B) Levels of NMF in patch test reactions (?+ or +) to fragrance mix I/II without subsequent reaction to a single constituent (F-mix w/o ) and patch test reactions (+) to fragrance mix I//II with subsequent reaction to a single constituent (F-mix with ) compared with their corresponding pet. controls. The labels indicate the individual number of each patient. Patient #38 reacted both to fragrance mix I (38A) and fragrance mix II (38B). A non-parametric Wilcoxon matched-pairs signedrank test was used for comparison between patch test reaction and corresponding pet. controls. Kruskal-Wallis test followed by Dunn's multiple comparison test was used for comparison between reactions (?+ or +) to F-mix w/o and reactions (+) to F-mix with . *P < .05 CXCL-8/IL-8 in monocyte-derived dendritic cells, whereas irritant exposures leads to decreased CXCL-8/IL-8 production. 40 Similarly, in the study by Dinghra et al, IL-8/CXCL-8 was upregulated significantly in patch test reactions to most haptens tested. 16 Another marker for innate immunity is IL-1β, which is important for mobilization and migration of epidermal Langerhans cells. 41 Its expression was increased in punch biopsies from patch test reactions to nickel and FM I, but not from ICD. 38 In the present study, levels for L-1β were only significantly increased in patch test reactions to FM I/II with pos- This certainly hampers the potential of NMF for differentiation between irritant and allergic patch test reactions to these haptens. In conclusion, the present study shows that the inflammatory profiles of patch test reactions to thiurams and fragrances are similar and indicate a Th2-skewed inflammation. The potential of IL-16 as biomarker for ACD was confirmed, whereas a decrease of NMF may indicate skin irritation. Our results support the hypothesis that doubtful or weak positive patch test reactions to FM I or FM II without concomitant reaction to a single ingredient are rather caused by an irritant than an allergic reaction.

ACKNOWLEDGEMENTS
We acknowledge support from the COST Action CA16113 Clini-MARK. We would like to thank Ulrike Vollmer and Ira Hülshoff for their support in patch testing and taking the samples. Open Access supervision; writing-original draft; writing-review and editing.

CONFLICT OF INTEREST
There are no conflicts of interest.

DATA AVAILABILITY STATEMENT
Data available on request from the authors.