1) Treatment with the differentiating agent, 5-Bromo-2-deoxyuridine (BrdU) was capable of inducing keratin (K8/ K18) expression in the poorly differentiated stem cell-like lung cancer cell line, DLKP and the more differentiated adenocarcinoma line, A549. The absence of keratin induction by BrdU in the Leukaemic cell line, HL60 suggested that this effect may be epithelial-specific. Immunocytochemical analysis of DLKP, treated with BrdU for one week and then continuously cultured in the absence of BrdU for 3 months, revealed that keratin filament formation was maintained in these cells. 2) RT-PCR and Northern blotting analysis revealed equal levels of K8/18 mRNAs before and after BrdU treatment suggesting that induction occurred at a posttranscriptional level. Since treatment of DLKP with a cocktail of protease inhibitors failed to induce K8 and K18, it would appear that proteolytic mechanisms are not in operation and that control is likely to be at the level of translation. 3) BrdU-treated lung cancer cell lines exhibited significant increases in the levels and phosphorylation of the important translation initiation factor, eIF-4E. On the other hand, levels and phosphorylation of this factor were decreased in BrdU-treated HL60s. eIF-4E may represent a key effector molecule in BrdU-mediated induction of K8/18 expression in epithelial lines, but since little increase in eIF-4E mRNA is observed it may not be the primary target gene for BrdU activation. BrdU almost certainly acts at the transcriptional level, ultimately. 4) Protein levels of the general translation initiation factor, eIF-2a also increased following BrdU treatment of both epithelial lung cancer cell lines. 5) Levels of c-Mycl protein, the growth inhibitory c-Myc isoform, increase following treatment of epithelial lines with BrdU. (Expression of this factor is known to be translationally regulated and dependent upon eIF-4E availability. In addition, c-myc has been reported to increase gene transcription of both eIF-4E and eIF-2a). 6) Since the transcription factor, YY1, has been shown to be induced by BrdU treatment, and it is known to be a regulator of c-myc gene transcription and c-Myc protein activity, we investigated its expression in this system. Levels of YY1 protein increase significantly in BrdU-treated epithelial cell lines, while they decrease in BrdU-treated HL60 cells. 7) Investigations using in-vitro translation suggested that DLKP may contain an inhibitor of K8/18 translation, whose activity is abrogated by BrdU treatment. 8) Transfection of DLKP cells with eEF-4E cDNA results in increased expression, not only of eIF-4E protein, but also of K8/18, YY1 and c-Mycl proteins. 9) Transfection of DLKP with YY1 cDNA resulted in increased expression, not only ofYYl protein, but also ofK8/18, eIF-4E and c-Mycl proteins. 10) From these results we have outlined a possible differentiation-inducing cascade that may be a target for BrdU activation. Emanating from the transcription factor, YY1 it culminates in the activation of a proposed differentiation-inducing feedback loop revolving around e!F-4E and c-Mycl expression, and the induction of keratin expression. 11) Exposure of DLKP to the physiological differentiating agent, Retinoic Acid (RA) resulted in growth inhibition but did not influence the expression of keratins 8/18. The growth inhibitory affects of RA were attributed to the expression of Retinoic Acid Receptor-a in DLKP. The lack or expression of a truncated form of the Retinoic Acid Receptor-P is offered as an explanation for the inability of RA to affect keratin expression in DLKP. RA was shown to be capable of altering K8/18 expression in the RAR-(3-positive cell line, A549. These findings may offer an insight into the poorly differentiated and aggressive nature of DLKP.