The pancreas plays a central role in metabolism. The exocrine pancreas, composed of ductal and acinar cells, secretes and delivers digestive enzymes into the duodenum where they contribute to food digestion. The endocrine pancreas, constituted by the islets of Langerhans, secretes hormones in the blood to control glucose levels. In particular, insulin is the hormone produced by β-cells that instructs the peripheral tissues to uptake circulating glucose. Diabetes mellitus is a heterogeneous disease characterized by deregulated glucose homeostasis due to an impaired insulin function. The advancements in clinical practice have made of diabetes a chronic, instead of a lethal, disease; yet no definitive cure is available. Beta-cell transplantation offers promising results, especially for type 1 diabetes where β-cells are selectively eliminated by an autoimmune attack. Unfortunately its application suffers from the scarcity of transplantable pancreas/islets from cadaveric donors. Human embryonic stem cells (ESCs) or induced pluripotent stem cells (iPSCs) would constitute unlimited sources of transplantable β-cells, but the available differentiation protocols are not yet optimal, especially with regards to the numbers and the functionality of the generated β-cells. It is anticipated that a stepwise protocol would succeed in generating mature β-cells in vitro if it is capable of fully mimicking the embryonic development of these cells. Many aspects of pancreatic organogenesis have been elucidated and could serve as a guide for driving the commitment to β-cells. Unfortunately, pancreatic progenitors in vitro do not behave as in the intricate context of an embryo. This severely limits both the understanding of pancreatic development at the single-cell level and our ability in manipulating the process. The aim of our work was to develop in vitro methodologies to sustain pancreatic progenitor culture and expansion and to establish conditions to manipulate their progeny. To this aim, we combined an informed approach based on developmental biology and an empirical one. We first observed that pancreatic epithelial explants from embryonic day 10.5 mice could be cultured in presence of Matrigel™ and exogenous FGFs even in the absence of the mesenchyme that normally surrounds the epithelium. Notably, the removal of mesenchyme did not affect the endocrine commitment pattern of pancreatic progenitors. We then defined culture conditions allowing for the expansion of dissociated embryonic pancreatic progenitors in a three-dimensional Matrigel™-based environment. Under these conditions, progenitors proliferated and self-organized to generate pancreatic organoids containing both progenitors and differentiated cells, mostly exocrine, after 7 days of culture. When grafted into recipient pancreatic explants, the cultured progenitors contributed to the three epithelial pancreatic lineages (ductal, endocrine, acinar), integrating seamlessly into the endogenous cellular network. Thus, cultured progenitors retained their potential to differentiate into endocrine cells, but expressed it only in the appropriate niche. Subsequent removal of single components from the culture system led to the identification of some essential requirements for the maintenance/expansion of dissociated pancreatic progenitors, such as a strong FGF signaling and the Rho-associated kinase (ROCK) inhibitor Y-27632. In addition to this, we observed that only small clusters of pancreatic progenitors, but not single cells, were able to generate pancreatic organoids, suggesting a pivotal role for intercellular signaling in progenitor maintenance and expansion. By manipulating the components of the medium, we could affect fate commitment. In particular, the removal of FGF1 increased the yield of endocrine cells generated in vitro. In our search for a full control over the in vitro niche, we explored chemically defined matrices to replace Matrigel™. We showed that pancreatic progenitors could be cultured on laminin-functionalized hydrogels, although less efficiently than in Matrigel™. Moreover, differentiation into exocrine and endocrine cells occurred spontaneously under these conditions. By comparing Matrigel™, hydrogel microwells and two-dimensional culture systems, we uncovered the importance of a tridimensional architecture for pancreatic progenitor maintenance. We showed that pancreatic progenitors lost their identity as they flattened onto 2D surfaces, whereas 3D culture systems maintained the epithelial character of pancreatic progenitors and allowed the acquisition of apical polarization. To conclude, our work provides the first detailed characterization of in vitro culture systems for expansion and manipulation of dissociated embryonic pancreatic progenitors. Further implementation will hopefully allow the establishment of a fully-defined in vitro niche for developing pancreas with potential fundamental implications for the expansion of ES-derived pancreatic progenitors and their differentiation into functional β-cells.