Interestingly, work pioneered by Falkmer and others indicate that pancreas phylogeny was founded by the insulin-producing beta cell (see Fig. 1). In the hagfish and lampreys (our most primitive vertebrate species of today), the first sign of ‘a new organ’ is found as collections of endocrine cells around the area of the bile duct connection with the duodenum. These endocrine organs are composed of 99% beta cells and 1% somatostatin-producing delta cells. Compared to the more primitive protochordates (e.g., amphioxus), this represents a stage where all previously scattered insulin-producing cells of the intestinal tissue have now quantitatively migrated to found a new organ involved in sensing blood glucose rather than gut glucose. Only later in evolution, the beta cells are joined by exocrine tissue and alpha cells (exemplified by the rat-, rabbit- and elephant-fishes). Finally, from sharks and onwards in evolution, we have the islet PP-cell entering to complete the pancreas [2,3]. Hagfish and lampreys may have one or more endocrine buds – and later the vertebrate pancreas develop as independent ventral and dorsal buds that eventually fuse to become one organ. In the bony fish, a giant islet – known as the ‘Brockman-body’ is derived of dorsal origin, while the ventral buds give rise to acinar cells, ducts, and smaller islets [4,5].

Interestingly, amphioxus (proto-chordate) Pdx-1 expression is already narrowed to a confined region of the gut [6]. The presence of Pdx-1 may have been instrumental for the subsequent evolutionary accumulation of beta cells in this region – as well as for the elaborate involvement of Pdx-1 as a beta-cell-specific transcriptional regulator of glucose-responsive genes. Lack of Pdx-1 in vertebrates cause pancreas agenesis [7–9]. Pdx-1 [7,8] and Nkx6.1 [10] are both transcription factors with a restricted expression within the mature pancreatic beta cell. In fact, the co-expression of those two markers in adult islets specifically identifies the beta-cell subpopulation [11]. During pancreas ontogeny, Pdx-1 is expressed within the early budding tissue as well as in the duodenum and antral stomach (albeit at lower levels). Again the co-expression of Pdx-1 and Nkx6.1 selectively specifies all of the early pancreatic progenitors found in the buds of ventral and dorsal origin [11,12]. Such progenitors need to activate Ngn3 gene transcription to produce the bHLH-transcription factor required for the initiation of the endocrine-cell maturation program [13] – a process regulated by Notch signalling [14].

Ptf1A/p48 is an exocrine bHLH-type transcription factor involved in pancreatic enzyme gene regulation [15] and is also influenced by Notch signalling [16,17]. P48 expression is highly confined within the endoderm to the pancreatic epithelium. It was therefore not unexpected that acinar cells failed to develop in p48 null-mutant mice [18]. Endocrine cells were reported still to form, and eventually localize within the spleen (the spleen forms during condensation of the dorsal pancreatic mesenchyme) [18]. However, later lineage-tracing studies showed that most endocrine cells actually derive from p48-expressing progenitor cells – but also confirmed that some endocrine cells still form in the absence of p48 [19]. Interestingly, Pdx-1 expression from Ptf1a/p48 cis-regulatory sequences restores pancreas tissue to Pdx-1-null mice [19]. This indicates that p48 and Pdx-1 together are required in the specification of the pancreatic progenitor cell.

In fact, if the p48 expression domain is extended into the duodenal pdx-1 domain, it will result in ectopic pancreas formation containing both endo- and exocrine cells: transgenic mice carrying the Hes1 null mutation (a down-stream effector of Notch signalling) display precocious formation of endocrine (alpha) cells in the pancreatic epithelium due to lack of repression of Ngn3 [14]. However, within the Pdx-1 positive duodenal epithelium the lack of Hes-1 leads to focal activation of p48 – with resulting ectopic pancreas formation [20]. Similar data were recently reported in transgenic Xenopus with pan-endodermal expression of p48, which leads to ectopic pancreas formation in the entire duodenal pdx-1 domain [21].

It may be hypothesized that the early Pdx-1+/Nkx6.1+ cells of the ontogenic pancreas reflect the phylogenetically first appearing beta cells in this region – and that these progenitors subsequently have adopted a wider differentiation potential to cover additional pancreatic cell types, including the enzyme-producing acinar cells (reflecting their subsequent appearance during phylogeny). It is highly plausible that the triple-positive cell (Pdx-1+/Nkx6.1+/p48+) found in the early pancreatic buds indeed represents a multipotent pancreatic stem cell. It may be further speculated that the early ontogenic Pdx-1+/Nkx6.1+/(p48+) pancreatic epithelial cells may constitute a source of progenitor cells carrying a phylogenetically imprinted pre-programming favouring beta-cell formation (default pathway?).

Organ donor islet transplantation has demonstrated the proof of the concept that the restoration of an adequate beta-cell mass can restore euglycaemia in patients with diabetes [22]. Future cell therapy of diabetes will rely on stem-cell-derived therapeutic beta cells (alternatively, in vivo-controlled beta-cell regeneration). Cell therapy of diabetes is envisioned to prevent the development of diabetes' late complications – and may in some ways even be considered a cure [1].

Embryonic stem (ES) cells constitute so far the only reliable source of stem cells with a proven potential (pluripotency) to become insulin-producing cells (reviewed in [1]). Recent progress in generating definitive endoderm [23] – and subsequent further maturation towards insulin-producing cells from human ES-cell cultures is a highly exciting and promising approach [24]. The strategy builds on replicating in vitro each individual step of fate-choice that the cells of the developing embryo are passing through to become glucose-sensing pancreatic beta cells. The ES cells are thus directed through a series of sequential steps designed by translating knowledge from developmental biology. Reaching the triple-positive (Pdx-1+/Nkx6.1+/p48+) cell stage constitutes a milestone along the developmental path. Achieving this stage may not only guarantee for subsequent potential towards pancreatic maturation, but may in addition provide a suitable developmental stage, which can be subjected to massive expansion [25].