Retinoic acid (RA) is widely used for topical therapy ofseveral skin diseases; it also improves the aspect of chronicsolar damage. Topical RA induces irritation of the skin,which precludes its use in some skin diseases that respondto systemic retinoids.Irritation might be explained, in part, by an overload of theRA-dependent pathways with non-physiological amounts ofexogenous RA in the skin. The globally attainable concen-trations of RA in the different layers from therapeuticallyefficient formulations have been determined in human skinin vivo. A steep concentration gradient with high concentra-tions in the epidermis, up to 450 ng/g wet weight (1.5 µM),and relatively low amounts in the dermis (55 ng/g wetweight, 180 nM) is achieved 100 min following applicationof RA 0.1% in isopropanol, corresponding to 15% of theapplied dose of 100 mg [1].Retinoid content analysis in human skin treated with0.1% RA for 4 days under occlusion showed the followingvalues: RA 824 ng/g (2,750 nM), 13-cis-RA 745 ng/g(2,480 nM) and 4-OH-RA 93 ng/g wet weight (310 nM) [2].A recent study on percutaneous absorption of RAshowed that about 2% of a single dose of 100 mg in a 0.05%formulation is absorbed; the same result is obtained after28 days of daily application [3]. Due to the similarity ofchemical structure, a similar pharmacokinetic behaviour isexpected for topical 13-cis-RA [4]. Such high tissue con-centrations of RA are in overexcess of the concentrationsneeded to saturate nuclear receptors [5, 6].Although major advances have been made in the analysisof the molecular events resulting from topical applicationof RA [7], it is still not established if all the therapeuticactivities of topical RA are mediated by nuclear receptors,and if irritation is necessary for obtaining some of theseactivities.One possibility is that significant biological activity maystill be achieved with much lower concentrations of topicalRA; in this case, RA is delivered at a low rate from a largeepidermal reservoir to its intracellular targets. Alternatively,instead of treating the skin with the ligand of nuclear receptorsitself, delivery could be distinctly targeted with ‘proligands’.We have explored the possibility of delivering retinoidactivity to human skin topically with a natural retinoid thatdoes not bind nuclear receptors. Such a precursor should behandled by enzymes of keratinocytes in the epidermis andtransformed into either RA or storage forms such as all-trans-retinol (ROL) and retinyl esters.Epidermis is a differentiating, non-homogenous tissue,made of keratinocytes that are not yet committed to terminaldifferentiation (basal cell layer) and a population of differen-tiating cells (suprabasal cell layers); the need for, and con-centration of, RA may not be identical in all layers and agradient of RA has been considered to be a key event in thematuration of keratinocytes [8]. This is supported by the factthat the conversion rate of ROL into RA by human kera-tinocytes depends on the state of keratinocyte differentia-tion, differentiating keratinocytes being able to oxidise ROLat a higher rate than non-differentiating ones [9]. Thatenzymes transforming the precursors into RA have distinctactivities at different stages of differentiation indicates thepossibility of targeting RA to epidermal cells in a differenti-ation related manner. This would be one approach to reduceside-effects. The use of RA precursors, such as ROL, reti-nyl esters, all-trans-retinaldehyde (RAL) and β-carotene,should therefore be considered in this context.