Genome-Wide Expression Analysis of Human In Vivo Irritated Epidermis: Differential Profiles Induced by Sodium Lauryl Sulfate and Nonanoic Acid

Abstract

The pathogenesis of irritant contact dermatitis (ICD) is poorly understood, and genes participating in the epidermal response to chemical irritants are only partly known. It is commonly accepted that different irritants have different mechanisms of action in the development of ICD. To define the differential molecular events induced in the epidermis by different irritants, we collected sequential biopsies ((1/2), 4, and 24 hours after a single exposure and at day 11 after repeated exposure) from human volunteers exposed to either sodium lauryl sulfate (SLS) or nonanoic acid (NON). Gene expression analysis using high-density oligonucleotide microarrays (representing 47,000 transcripts) revealed essentially different pathway responses (1/2)hours after exposure: NON transiently induced the IL-6 pathway as well as a number of mitogen-activated signaling cascades including extracellular signal-regulated kinase and growth factor receptor signaling, whereas SLS transiently downregulated cellular energy metabolism pathways. Differential expression of the cyclooxygenase-2 and matrix metalloproteinase 3 transcripts was confirmed immunohistochemically. After cumulative exposure, 883 genes were differentially expressed, whereas we identified 23 suggested common biomarkers for ICD. In conclusion, we bring new insights into two hitherto less well-elucidated phases of skin irritancy: the very initial as well as the late phase after single and cumulative mild exposures, respectively.

Anchoring molecular mechanisms to the adverse outcome pathway for skin sensitization: Analysis of existing data

“…To this end, PubMed and GEO DataSets, a repository for gene profi ling datasets, were explored for studies that include human in vivo toxicogenomic data after exposure to any skin sensitizer. Unfortunately, only two suitable and accessible gene expression profi ling studies were retrieved (Pedersen et al 2007, Clemmensen et al 2010, Hansen et al 2005. Two of these studies use nickel or chromium, which induce ACD via a unique mechanism and might not be representative for other low -molecular-weight chemical sensitizers.…”

“…To investigate whether the responses included in the ACD molecular map are specifi c for sensitization a third human in vivo study was evaluated (Clemmensen et al 2010). Here the gene changes after exposure to the irritants sodium dodecyl sulfate and nonanoic acid, which are known false positives in the LLNA, were evaluated (Montelius et al 1998, Gerberick et al 2005).…”

Effects of Cosmetic Formulations Containing Hydroxyacids on Sun-Exposed Skin: Current Applications and Future Developments

“…Further, acute doses with full-spectrum solar-simulated radiation and solar-simulated UVA with or without SPF 15 sunscreen treatment were also assessed [33]. However, there are virtually no published microarray studies evaluating HA formulations with relation to the effects of UV on skin, but there have been some interesting studies that have evaluated skin responses with or without UV when pretreated with skin lightening or irritating compounds [34, 35]. This is an area that has great potential to become the method of choice for safety screening of cosmetic preparations.…”

A Growth Curve Model with Fractional Polynomials for Analysing Incomplete Time-Course Data in Microarray Gene Expression Studies

“…The epidermal response to chemical irritants was investigated by Clemmensen et al [19] using genome-wide expression analysis for 47,000 transcripts or genes in a time-course design applied to human in vivo irritated epidermis. We apply our method to a subset of their data as an example.…”

“…Although the duration of time for an array-based laboratory experiment is usually much shorter (can be in hours) than that in a longitudinal survey (can be in years); however, this should analytically make no difference in terms of growth curve modelling. With this consideration, we exemplify application of the growth curve model with FPs in the analysis of microarray time-course expression data with missing observations from an experiment of human in vivo irritated epidermis [19]. We show that our procedure can automatically identify the significant time trajectories in gene regulatory response through model comparison and statistical testing while efficiently handling missing values.…”