RCJ C 5.18 (C 5.18) is a chondrogenic clonal cell line which, under standard culture conditions, develops chondroblastic features including the production of a cartilagenous matrix. Retinoic acid (RA) is known to inhibit the chondrogenic differentiation of C 5.18 cells and this may parallel the teratogenic effects of retinoids in vivo; however, the question as to which of the 3 retinoic acid receptors (RAR alpha, beta, gamma) or the 3 retinoid X receptors (RXR alpha, beta, gamma) mediate this RA-induced inhibition remains unanswered. We tested several retinoids with different receptor binding characteristics. Cartilage formation in C 5.18 cultures was evaluated by counting the number of cartilage nodules formed, and by quantitating the glycosaminoglycan content of the cultures using alcian blue staining. All of the retinoids prevented cartilage formation in a dose-dependent manner. Treatment with the retinoids did not affect cell number, thereby ruling out any toxic effects. RA, which binds to all 3 RARs with similar affinity, produced a 50% inhibition (IC50) of cartilage formation at 4 x 10(-10) M. We also tested Ch55, which also binds to all 3 RARs, but with higher affinity than RA. This compound was approximately 10 times more potent than RA (IC50 2 x 10(-11)M). 9-cis RA, which binds to the 3 RARs with affinities similar to RA and also binds to the 3 RXRs, was less active (IC50 8 x 10(-9) M), suggesting that RXR binding interferes with the inhibitory effect of ligand-activated RARs. 9-cis retinal, for which the binding characteristics are unknown, had the same effect as 9-cis RA. The synthetic retinoids Am80 (IC50 6 X10(-11) M) and Am580 (IC50 4 X 10(-11) M) were more potent than RA in inhibiting cartilage formation. Since these compounds bind only to RAR alpha and beta, and with higher affinity than RA, the observed inhibition of cartilage formation suggests that interaction of the ligand with RAR alpha and/or beta is sufficient to induce the effects on cartilage development. If reports suggesting that Am580 binds exclusively to RAR alpha are correct, these findings could imply that RAR alpha alone is sufficient to mediate the inhibitory effect of the retinoids. Surprisingly, Re80, which binds to RAR alpha and beta with a binding affinity similar to RA (and does not bind significantly to RAR gamma or the RXRs), was the most potent retinoid in our system (7 x 10(-13) M). This finding supports mediation of the retinoid effect through alpha and beta receptors. Our results suggest that the retinoids act via RAR alpha and/or beta to inhibit chondroblast differentiation. (C) 1994 Wiley-Liss, Inc.