We have previously shown that retinaldehyde (RAL), a natural metabolite of beta-carotene and retinol (ROL), can be used topically in human skin and exerts biological activity; it may be a convenient way to deliver multipotential vitamin A activity in epidermis. RAL can be converted enzymatically into 2 pathways: one leads to ROL (and then retinyl esters), the other to retinoic acid (RA). The aim of the present study was 2-fold: (i) to see if RAL is metabolised in vivo when topically applied on mouse skin, and (ii) if so, to analyse the occurrence and relative importance of the 2 metabolic pathways as compared to ROL. We studied by HPLC the metabolites detectable in mouse tail skin upon topical application of RAL and ROL. As compared to vehicle-treated controls, RAL-treated mouse skin contained low amounts of all-trans RA and 13-cis-RA, whereas ROL content increased 10-fold and retinyl esters 30-fold after RAL application. As compared to RAL, ROL-treated mouse skin showed no detectable RA, slightly less retinyl esters but a significant amount of 14-hydroxy-4, 14-retro-ROL (14-HRR), a metabolite not previously reported in the skin. 14-HRR was the predominant polar metabolite of ROL. These data indicate that keratinocytes metabolise topical RAL, thus confirming the concept of using RAL as a precursor. Both pathways are used but in significantly different proportions. Thus, only a low proportion of RAL is metabolised into all-trans-RA, which may explain the low irritancy profile of topical RAL and supports the concept of a controlled delivery of ligands. That keratinocytes predominantly channel RAL into storage forms indicates that RAL should also be considered as a convenient way to load the epidermis with vitamin A. The detection of 14-HRR, a metabolite not previously reported in skin, that promotes growth of B Iymphocytes and activation of T Iymphocytes, suggests distinct potentials of topical ROL and RAL.