Previous studies have revealed that embryonic stem (ES) cells are capable of differentiating spontaneously towards pancreatic P cells. The use of growth factors and specific agents, such as nicotinamide, can further enhance this effect in a process known as directed differentiation. In vivo monitoring of pancreatic development indicates that the pancreas is derived from an endodermal lineage. This thesis investigated the differentiation of an embryonic carcinoma cell line, F9, which is characterised as having minimal spontaneous differentiation, towards endoderm and from there, on to pancreatic lineages. Treatment of monolayer F9 cells with 10'7 M retinoic acid (RA) or a combination of retinoic acid and 10'3M dibutyryl cyclic adenosine monophosphate (db cAMP) for 72 hours generates both primitive and parietal endodermal types. Culture of F9 cells in suspension leads to the formation of embryoid bodies (EBs). Treatment of these aggregates with RA generates visceral endoderm on the outer surface of the EB. These endodermal subtypes were found to express the full catalogue of cytokeratins (CKs) associated with the pancreatic islets i.e. CKs 8, 18, 7 & 19. The F9-derived endoderm was then treated with agents that have been shown to promote endocrine differentiation from progenitor cells i.e. activin A, betacellulin hepatocyte growth factor (HGF), nicotinamide and sodium butyrate. Certain treatments were found to increase the levels of important pancreatic mRNAs i.e. a combination of HGF and activin A increased pancreatic and duodenal homeobox gene 1 (PDX1) and preproinsulin (PPI) expression in RA-treated F9 monolayers. Sodium butyrate was shown to have positive effects on the expression of somatostatin and pancreatic polypeptide (PP) mRNAs. Other p cell genes detected in differentiating F9 cells include Pax6 and indian hedgehog (Ihh). To further study the role of the transcription factor PDX1 in pancreatic development, cell trapping was employed. In this system, a specialised construct was designed with the zeocin resistance gene (Sh ble) under the control of the PDX1 promoter region. This construct was shown to be capable of selectively maintaining the viability of F9 cells with the capability to drive the PDX1 promoter, in the presence of zeocin. The cells purified using this technique were found to exhibit increased levels of PDX1 mRNA and in certain cases PPI also, even though PDX1 protein was not detected. Further evidence for the presence of functional PDX1 in cell trapped lines was provided by the observation that these cell lines co-expressed somatostatin, which is known to be transactivated by PDX1. Upon culturing p cells for prolonged periods in vitro it has been noticed that these cells lose the ability to secrete insulin in response to environmental glucose signals (glucose stimulated insulin secretion, GSIS), in a process sometimes referred to as dedifferentiation. Upon culturing the murine P cell line, Min6, for 40 passages similar effects were noticed. Analysis of the gene expression profiles in continuously cultured Min6 cells via DNA array led to the discovery that a number of genes associated with rapidly proliferating primitive type cells were upregulated in the high passage Min6 cells (e.g. the polycomb group gene EED & p55CDC). This analysis also revealed that a number of genes critical for the regulated secretion of insulin were downregulated, including; nucleobindin, secretogranin II, secretogranin V, chromogranin B and prohormone convertase 2 (PC2). These observations and the associated reduction/loss of terminal differentiation markers such as PDX1, glucagon and somatostatin were all consistent with de-differentiation or with overgrowth of the culture by a more poorly differentiated subpopulation. The human insulin gene has been successfully expressed in a number of cell lines as a possible method for ‘artificial p cell delivery’. BHK-21 cells are characterised as a safe cell line to implant into humans in a suitably encapsulated format. Expression of the human PPI gene in BHK-21 cells resulted in large amounts of proinsulin production. There was no evidence of proinsulin processing to mature insulin and this was associated with an inability to store the protein properly. This also resulted in the continuous secretion of proinsulin at a steady rate of 0.12 pmol proinsulin/hr/105 cells. These cells were found to be unable to regulate proinsulin secretion in response to glucose. Overexpression of glucokinase resulted in a moderate GSIS but overexpression of GLUT2 and glucokinase resulted in the total loss of insulin secretion. This may be due to a deleterious increase in the metabolic burden in the cells. Proinsulin secreting BHK-PPI-C16 cells were also capable of increasing proinsulin secretion upon treatment with cAMP and it is proposed that this increase in secretion is due to a general increase in intracellular proinsulin i.e. the rate of secretion was determined by the availability of the protein.