Peroxisome proliferator-activated receptor gamma (PPARgamma) is involved in the control of cell proliferation, apoptosis and differentiation in various tumor cells. Among PPARgamma ligands, 15-deoxy-Delta12,14-prostaglandin J2 (PGJ2), the ultimate metabolite of PGD2, plays a role in the biology of brain tumors. It is still unclear to which extent the anti-proliferative and differentiation-promoting activity of PGJ2 is mediated through PPARgamma. We compared the effects of PGJ2 with those of rosiglitazone - the synthetic agonist with the highest affinity for PPARgamma - in 4 human glioblastoma cell lines (A172, U87-MG, M059K, M059J). All cell lines expressed high levels of PPARgamma, consistent with the high levels of PPARgamma protein in 5 tumor samples. Both PGJ2 and rosiglitazone inhibited proliferation of all cell lines with a G2/M arrest and apoptosis, but only PGJ2 up-regulated p21Cip/WAF1. The growth inhibitory effect was partially reversed by the PPARgamma antagonist GW9662. We studied the time sequence of selected molecular events, that lead glioblastoma cells to apoptosis and/or differentiation, after treatment with both agonists. M059K cells committed to undergo apoptosis by PGJ2, initially up-regulated PPARgamma, and then down-regulated PPARgamma as they began apoptosis. Apoptotic cells also increased their expression of retinoic acid receptor beta (RARbeta) and retinoid X receptor alpha (RXRalpha). PGJ2 increased expression of glial fibrillary acidic protein (GFAP) and decreased levels of vimentin, structural proteins modulated during astrocytic differentiation. Unexpectedly, PGJ2 up-regulated the expression of cyclooxygenase-2 (COX-2). Rosiglitazone caused the same pattern of PPARgamma, RARbeta and RXRalpha expression as PGJ2, but no significant modulation of p21Cip/WAF1, cytoskeletal proteins or COX-2 occurred. Our data indicate that PGJ2, and rosiglitazone suppress cell proliferation and cause apoptosis in glioblastoma cell lines, most likely through a PPARgamma-dependent pathway. By contrast, the modulation of differentiation-associated proteins by PGJ2, but not rosiglitazone, suggests that PGJ2 promotes differentiation of glioblastoma cells independently of PPARgamma activation.