Unique molecular signatures typify skin inflammation induced by chemical allergens and irritants

Abstract

Background: Skin exposure to chemicals may induce an inflammatory disease known as contact dermatitis (CD). Distinguishing the allergic and irritant forms of CD often proves challenging in the clinic. Methods:To characterize the molecular signatures of chemical-induced skin inflammation, we conducted a comprehensive transcriptomic analysis on the skin lesions of 47 patients with positive patch tests to reference contact allergens and nonallergenic irritants.Results: A clear segregation was observed between allergen-and irritant-induced gene profiles. Distinct modules pertaining to the epidermal compartment, metabolism, and proliferation were induced by both contact allergens and irritants; whereas only contact allergens prompted strong activation of adaptive immunity, notably of cytotoxic T-cell responses. Our results also confirmed that: (a) unique pathways characterize allergen-and irritant-induced dermatitis; (b) the intensity of the clinical reaction correlates with the magnitude of immune activation. Finally, using a machinelearning approach, we identified and validated several minimal combinations of biomarkers to distinguish contact allergy from irritation. Conclusion:These results highlight the value of molecular profiling of chemicalinduced skin inflammation for improving the diagnosis of allergic versus irritant contact dermatitis.

Gene profiling reveals a contact allergy signature in most positive Amerchol L‐101 patch test reactions

“…Our gene profiling study demonstrated that the molecular pathways that are activated in response to Amerchol application largely overlapped with the pathways found in the skin of allergic individuals patch tested with reference allergens, such as nickel, methylisothiazolinone, linalool hydroperoxide, 14 or para‐phenylenediamine and epoxy resin‐bisphenol A 37 . Recently, Dhingra et al suggested that molecule‐dependent profiles could exist between contact allergens, and they reported that a significant T H 2 bias with T H 22 polarization characterized the responses induced by molecules such as fragrances, whereas type‐1/type‐17 phenotypes predominated for molecules like nickel 38 .…”

“…of each donor was subtracted from the corresponding‐treated samples of the same donor. Projecting the resulting data on previous PCA comparing gene expression in chemical allergen (also called hapten) and irritant samples, 14 we found that the gene inductions for positive AL‐101 samples (+/?, + and ++) overlapped for the most part with the inductions recorded for hapten but not irritant samples, typifying an allergen‐like response to Amerchol (Figure 2A, colored symbols).…”

“…We have previously identified and selected 52 DEGs as the basis of a hapten core signature. The latter mainly consisted of several T cell– and cytotoxic‐related transcripts, 14 which are upregulated within the 72 hours following hapten re‐test. We then analyzed the induction of these 52 DEGs in the different AL‐101‐treated skin samples.…”

“…The gene array data set generated in this study was also compared to previous data sets obtained from skin biopsies of positive patch test reactions induced by reference chemical allergens and irritants. 14 Expression values were modeled using different mixed linear regression models, with the group of chemicals for the previous data set (chemical allergens, irritants, controls) and the treatment condition…”

“…We then compared gene profiling of Amerchol-inflamed skin with the transcriptomic signatures of positive patch test reactions to several reference chemical allergens (nickel, methylisothiazolinone, linalool hydroperoxide) and irritants (SLS, nonanoic acid, and cantharidin), evaluated in a previous study from our group. 14 To this end, we executed a hypothesis-free PCA of data from the full set of genes measured on the arrays for each AL-101 sample. To compensate for donor-induced variability, the gene expression level of the control sample (pet.)…”