Anticancer effect of retinoic acid via AP-1 activity repression is mediated by retinoic acid receptor alpha and beta in gastric cancer cells

To uncover the mechanisms relating to the anticancer effect of retinoic acids in gastric cancer cells, the mediation of activator protein-1 (AP-1) activity repression by retinoic acid receptors (RARs) was investigated. All-trans retinoic acid (ATRA) inhibited AP-1 activity in BGC-823 cells (RARalpha(+), RARbeta(+)), but not in MKN-45 cells (RARalpha(lo), RARbeta(-)). Transient transfection of RARbeta expression vector into MKN-45 cells significantly resulted in direct repression of AP-1 activity in a receptor concentration-dependent manner, and this could be strengthened by ATRA. Stable transfection of RARbeta into MKN-45 cells directly inhibited cell growth and colony formation, and ATRA also enhanced these effects. Transient transfection of RARalpha into MKN-45 cells however, displayed receptor concentration-dependent AP-1 activity inhibition only in the presence of ATRA. Stable transfection of RARalpha into MKN-45 cells resulted in ATRA-dependent inhibition of cell growth and colony formation. For AP-1 binding activity induced by TPA, the repressive effect of ATRA was only observed in BGC-823 and RARalpha and RARbeta stably transfected MKN-45 cells, but not in intact MKN-45 cells. This indicates the necessity for sufficient cellular RARalpha and/or RARbeta in order for AP-1 activity repression to occur. Deletion of DNA binding domain (DBD) of RARbeta, but not ligand binding domain (LBD), eliminated the anti-AP-1 function of RARbeta. It is therefore concluded that both RARalpha and RARbeta are mediators in the anticancer function of ATRA via AP-1 activity inhibition, and that RARbeta, not RARalpha, can inhibit AP-1 activity to a certain extent directly by itself. Thus DBD, not LBD, is critical for anti-AP-1 activity.