Our previous studies (Sidhu JS et al. Arch Biochem Biophys 1993: 301, 103-113; Sidhu JS et al. In Vitro Toxicol 1994: 7, 225-242) demonstrated the importance of culturing primary rat hepatocytes with an overlay of extracellular matrix (ECM), together with optimal media formulations (Williams' E or Chee's), to efficiently maintain in vivo-like responsiveness of phenobarbital (PB)-inducible cytochrome P450 genes in vitro. In the present report, we have characterized culture conditions further by examining individual and interactive effects of dexamethasone (Dex) and PB on CYP2B1, CYP2B2, and CYP3A1 expression. Dex alone was not effective in enhancing CYP2B1 or CYP2B2 expression levels. However, together with PB, addition of low concentrations (10(-9)-10(-8) M) of Dex resulted in a marked potentiation of PB-inducible P450 gene expression. In contrast, at levels > 10(-7) M, Dex profoundly inhibited PB induction of the CYP2B1 and CYP2B2 genes. The overall stimulatory response to Dex was more dramatic in cells cultured in Williams' E than in Chee's medium. Similarly, concentrations of PB > 0.5 mM resulted in substantially reduced levels of CYP2B1 and CYP2B2 induction than those attainable at lower PB concentrations. These results suggest that Dex and PB function cooperatively to regulate the CYP2B1 and CYP2B2 genes, and that composite interactions may either negatively or positively regulate expression, in a concentration-dependent manner. CYP3A1 was not regulated in a similar biphasic fashion, as this gene was fully responsive even at high dose levels of PB or Dex. With respect to other marker genes evaluated, high Dex concentrations (> 10(-7) M) were only marginally inhibitory to beta-naphthoflavone-mediated induction of CYP1A1 and CYP1A2 mRNAs, and did not perturb expression of the liver-selective serum albumin gene. Addition of Dex was critical, however, to maintain glutathione S-transferase Pi expression, a marker of hepatocyte dedifferentiation, in the repressed state. Defining optimal culture conditions for maintaining hepatocyte differentiation in vitro are requisite for establishing primary culture models enabling investigation of the molecular mechanisms of PB-mediated gene regulation.