USE OF DERMAL EQUIVALENT AND SKIN EQUIVALENT MODELS FOR IN-VITRO CUTANEOUS IRRITATION TESTING OF COSMETIC PRODUCTS - COMPARISON WITH IN-VIVO HUMAN DATA
C. Augustin et al., USE OF DERMAL EQUIVALENT AND SKIN EQUIVALENT MODELS FOR IN-VITRO CUTANEOUS IRRITATION TESTING OF COSMETIC PRODUCTS - COMPARISON WITH IN-VIVO HUMAN DATA, Journal of toxicology. Cutaneous and ocular toxicology, 17(1), 1998, pp. 5-17
The development of new cosmetic formulations requires precise assessme
nt of their safety and efficacy. Today, legislation imposes increasing
measures of safety as well as the limitation of animal use for such t
esting (European Community directive 93/35/CEE). Subsequently, safety
assessment protocols now focus on in vivo human volunteer tests and in
vitro methods. In this study, in vivo testing consisted of 48 h patch
tests on human volunteers followed by a clinical evaluation of irrita
tion based on a visual scoring system including evaluation of erythema
, edema, dryness, and vesicles. For in vitro testing to substantiate t
he safety of cosmetic products, we propose two three-dimensional model
s, a dermal equivalent (DE) and a skin equivalent (SE). The DE is comp
osed of a porous collagen-glycosaminoglycans-chitosan dermal substrate
populated by normal human fibroblasts. The SE is completed by a fully
differentiated epidermis made by seeding normal human keratinocytes o
nto the DE. To evaluate the usefulness of such in vitro models for cyt
otoxicity trials, 14 cosmetic products reflecting a range of galenic f
orms (oil, mascara, cream, lotion) were tested both in vivo and in vit
ro. In vitro testing consisted of a topical application (10 mu l) of p
ure or appropriate dilution of cosmetic products, reflecting future us
e, onto DEs and SEs (n = 6). After a 24 h contact, measurement of resi
dual cellular viability using the MTT test as well as LDH and IL-1 alp
ha release measurement allowed an evaluation of the skin irritation po
tential. The in vitro results were compared to the in vivo data using
a binary correlation based on the irritation potential prediction (non
irritating/irritating). A concordance of 71% and 79%, respectively, fo
r the DE model and the SE model were obtained versus in vivo data.