Identification of novel in vitro test systems for the determination of glucocorticoid receptor ligand-induced skin atrophy.

Topical glucocorticoids (GCs) demonstrate good anti-inflammatory effects but are limited by their side effect potential, with skin atrophy being the most prominent one. Thus, determining the atrophogenic potential of novel compounds is important. The aim of this study was to establish an in vitro skin atrophy model. A screening cascade was applied and GCs with a known atrophogenic potential were used as tool compounds. Five rodent and human cutaneous cell types/cell lines and 2 human skin equivalents were tested. Known and suspected atrophy markers related to collagen metabolism and epidermal thickness were measured. Altogether, a combination of 7 different cellular assays with up to 16 markers each were investigated. A reproducible, more than 2-fold, regulation of the candidate markers by dexamethasone or clobetasol was found for: (a) matrix metalloproteinase (MMP) 1, 2, 3 and 9 expression in human keratinocytes, (b) COL1A1 and COL3A1 expression in 3T3 fibroblasts, and (c) epidermal thickness, collagen and MMP synthesis in the full-thickness skin model (FTSM). These 3 models were further investigated with a panel of 4-5 GCs, demonstrating dose dependency and correlation with the atrophogenic potential of the tool compounds, qualifying them as potentially suitable. Finally, the predictability of these models for the in vivo situation was analyzed, testing a novel selective GC receptor agonist (SEGRA) in comparison to clobetasol. The results from the in vitro models suggested less atrophogenic effects for the SEGRA compound, which indeed was confirmed in the hr/hr rat skin atrophy model. In conclusion, a combination of 3 in vitro models based on 3T3 cells, human keratinocytes and FTSM with several readouts is recommended to determine atrophogenicity of GC receptor ligands. Further experiments are necessary to eventually reduce this panel and to demonstrate the true predictability for the clinic.