flow cytometry showed that c-Fos accelerated the cell cycle kinetics. Following serum stimulation, Cyclin D1 was more abundantly expressed in c-Fos overexpressing cells. Cyclin D1 accumulation did not result from increased transcriptional activation, but from nuclear stabilization. Overexpression of c-Fos correlated with higher nuclear levels of inactive phosphorylated GSK-3β, a kinase involved in Cyclin D1 degradation and higher levels of EGF-R mRNA, and EGF-R protein compared to IHH-C both in serum starved, and in serum stimulated cells. Abrogation of EGF-R signalling in IHHFos by treatment with AG1478, a specific EGF-R tyrosine kinase inhibitor, prevented the phosphorylation of GSK-3β induced by serum stimulation and decreased Cyclin D1 stability in the nucleus. METHODS:IHHs stably transfected with c-Fos (IHHFos) or an empty vector (IHH-C) were grown in medium supplemented with 1% serum or stimulated with 10% serum. Cell proliferation was assessed by cell counts, 3H-thymidine uptake and flow cytometry analyses. The levels of cell cycle regulatory proteins (Cyclin D1, E, A) cyclin dependent kinases (cdk) cdk2, cdk4, cdk6, and their inhibitors p15, p16, p21, p27, total and phosphorylated GSK-3β and epidermal growth factor receptor (EGF-R) were assayed by Western blotting. Analysis of Cyclin D1 mRNA levels was performed by reverse transcription-polymerase chain reaction and real-time polymerase chain reaction (PCR) analysis. Stability of Cyclin D1 was studied by cycloheximide blockade experiments. RESULTS: Stable c-Fos overexpression increased cell proliferation under low serum conditions and resulted in a two-fold increase in [3 H]-thymidine incorporation following serum addition. Cell cycle analysis by

Nuclear factor kappaB (NF-kappaB) and activator protein 1 are transcription factors involved in the regulation of cell proliferation that play important roles in tumorigenesis. We investigated whether these two factors cooperate for transcriptional regulation of cyclin D1 (CCND1), a gene whose deregulation is critical during carcinogenesis. We demonstrate that overexpression of JunD in human hepatocarcinoma cells strongly activates transcription mediated by the kappaB2 site of the CCND1 promoter in reporter assays, in a manner strictly dependent on the presence of NF-kappaB proteins. Serum stimulation increased the expression of p65, p50, c-Fos, c-Jun and JunD and induced the recruitment of p65, p50 and JunD to the kappaB2 site of the promoter in DNA pull-down assays. Chromatin immunoprecipitation (ChIP) analysis confirmed the serum-induced recruitment of JunD to the promoter in vivo and showed that the presence of JunD was dependent on the presence of p65 and p50, indicating a protein-protein-dependent mechanism of JunD recruitment. Serum-induced activation of protein binding to kappaB2 correlated with high levels of phosphoinositide-dependent protein kinase-1 (PDK-1) phosphorylation. Both LY294002, a specific inhibitor of phosphatidylinositol 3-kinase (PI3K), and overexpression of a dominant-negative form of PDK-1 inhibited the JunD-stimulating effect in reporter assays. LY294002 also prevented the serum-induced recruitment of JunD, but not p65 or p50 to the promoter in ChIP assay. JunD-p65 complexes, identified in vivo by co-immunoprecipitation, were decreased by LY294002 and by small interfering RNA inhibition of PDK-1. Taken together, our data demonstrate a PI3K/PDK-1-dependent functional cooperation of NF-kappaB and JunD in the transcriptional regulation of CCND1 by serum.