The effects of topical tretinoin on collagen synthesis and degradation were studied in 29 volunteers. The subjects applied 0.1% tretinoin cream on their non-sun-exposed abdominal skin once a day for 1 week (n = 10) (experiment 1) or twice a day for 2 weeks (n = 8) (experiment 2) or once a day for 2 months (n = 11) (experiment 3). After the treatments, suction blisters were induced and aminoterminal propeptides of type I and III procollagens (PINP, PIIINP, respectively) (experiments 1 and 3) and carboxy-terminal propeptide of type I procollagen (PICP) (experiment 2) were assayed as an index of de novo collagen synthesis by radioimmunoassays. Matrix metalloproteases 2 (MMP-2) and 9 (MMP-9) were assayed by the zymography method in experiment 2. In experiment 3, histology, immunohistochemistry of type I and III procollagens, tenascin, mRNA levels of type I collagen α1-chain [α1(I)], interstitial collagenase (MMP-1), MMP-2, MMP-9 by slot-blot analysis and the levels of α1(I) collagen mRNA by a quantitative polymerase chain reaction method were studied. The proportional area of elastic fibres visualized in Verhoeff-stained sections was analysed by computerized digital image analysis. The results indicated that treatment with topical tretinoin does not markedly induce de novo synthesis of collagen in vivo or affect matrix metalloproteases. In the immunohistochemical stainings, tenascin was increased in the papillary dermis. As it has been suggested that tretinoin could counteract the atrophogenic effect of corticoids on the dermis, the effect of a combination of betamethasone-17-valerate (once a day) and tretinoin (once a day) on the propeptide levels was also studied. Betamethasone alone caused a 60% decrease in the concentrations of PINP and PIIINP, and a similar decrease was found after the combination treatment, indicating that topical tretinoin administered during short treatment periods does not counteract the inhibitory effect of a potent corticoid on collagen propeptides.