The insulin-producing β cell is perhaps the most intensely studied endocrine cell type, largely because of the implications for understanding the pathogenesis and treatment of diabetes.
Many mouse models have clarified the factors necessary for β-cell differentiation, development, and maturation (see Table 4.2). One such factor is the basic helix-loop-helix transcription factor Neurod1. Interestingly, Neurod1 is expressed in all endocrine cell types except the somatostatin-producing δ cell, and targeted disruption of the Neurod1 gene in mice results in severe reduction in α and β cells, and in neonatal diabetes owing to β-cell apoptosis . The Maf family of transcription factors is also involved in the pathway of α- and β-cell differentiation.
Both MafA- and MafB-deficient mouse models have pancreatic phenotypes. Loss of MafB leads to perinatal lethality and, although the total endocrine cell mass is unaffected, the pancreas shows reduced numbers of α and β cells . By contrast, mice with a targeted deletion of gene encoding MafA are born viable and with normal islet cell numbers, but demonstrate β-cell dysfunction with advancing age, leading to glucose intolerance and diabetes . In these mice, β-cell genes, including those encoding insulin, Neurod1, and Glut2, are significantly reduced. Owing to its importance in β-cell function, MafA is considered as a marker for mature, functional β cells.
Interestingly, a number of factors expressed broadly in early pancreas development become restricted to specific endocrine cell types during the secondary transition and acquire additional function in the differentiation or maturation of these cell types. One particular example is Pdx1. Whereas mice with a homozygous deletion of the gene encoding Pdx1 are born without a pancreas, haploinsufficiency of Pdx1 results in glucose intolerance [49,50]. Virtually identical phenotypes are observed in humans with respective homozygous and heterozygous mutations in Pdx1 . After the pancreas is fully developed, Pdx1 is expressed primarily in β cells and is necessary for β-cell function, including transcriptional activation of several β-cell genes . Nkx2.2 is another transcription factor that shows expression in the early pancreatic progenitors but becomes restricted to specific endocrine cell populations later in development . Loss of Nkx2.2 in the mouse results in the complete absence of differentiated insulin-producing β cells, a significant decrease in α and PP cells, and a concomitant increase in ghrelin-expressing ϵ cells .
Mouse models have also identified specific factors necessary for the differentiation of the glucagon-expressing α cell. In particular, deletion of the transcription factor Arx results in hypoglycemia and neonatal lethality. Given that Arx is expressed in all endocrine cells except the β cell, the pancreas of the Arx-null mouse has altered endocrine cell ratios: a complete absence of α cells and an increase of β and δ cells. Conversely, the misexpression of the Arx gene in either Pdx1- or Pax6-expressing cells results in a loss of β and δ cells and an increase in α and PP cells .Moreover, compound mutants have demonstrated the complexity of transcription factor interactions and the importance of Arx function to endocrine cell development. Specifically, deletion of the Arx and Pax4 genes result in the loss of α and β cells but an increase in δ cells , and the Arx/Nkx2.2 compound mutant pancreas showed a restoration of the PP cell population that was lost in the Nkx2.2 null pancreas .
The evidence for the relevance of transcription factors to human pancreas development and disease is highlighted by the discovery that mutations in a number of transcription factors identified to be important in pancreas development in the mouse also display pancreas-related phenotypes in the human.Mutations or deletions of many of these factors have been established as the cause of monogenic forms of diabetes known as maturity-onset diabetes of the young (MODY) , or as the cause of human syndromes that include diabetes [42,43,55–58] (Table 4.2).
Translating pancreas development into cell-based therapies for diabetes
The knowledge gained from decades of pancreas development research has stimulated the translational pursuit of engineering insulin-producing β cells in vitro for therapeutic purposes.
Specific extrinsic factors, including FGF, RA, and inhibitors of BMP or Shh signaling, have been applied to mouse and human embryonic stem cells in culture to successfully drive these malleable cells toward a pancreas fate  (see Figure 4.3). More recently, success in the creation of pluripotent, embryonic-like stem cells from somatic cells has opened the possibility of generating patient-specific β cells as cell replacement therapies for diabetes . To date, however, such techniques have resulted in compound or mixed populations of hormone-producing cells, which have little or no capacity for glucose-stimulated insulin secretion when generated wholly in vitro, suggesting that many intrinsic and/or extrinsic factors involved in the cytodifferentiation of pancreatic progenitors remain to be identified. In the mouse, the genetic manipulation of certain intrinsic factors, that is, Pdx1, Neurog3, MafA , or the induction of severe pancreatic injury , has also identified the capacity of differentiated cells in the pancreas to be “reprogrammed” to β cells. Therefore the continued merging of developmental biology research with in vitro differentiation technology may produce the long awaited therapeutic cure for diabetes.
Postnatal beta-cell growth and maintenance
Following the secondary transition, the total mass of the pancreas increases substantially, but β cells comprise only about 1–2% of this cellular mass in the mature adult pancreas. Despite this relatively small percentage, the states of β-cell mass and function represent perhaps the greatest determinants in overall glucose and lipid homeostasis in virtually all types of metabolic disorders . Two fundamental concepts regarding postnatal β cells have emerged over the past two decades: (1) although β cells were considered to be a postmitotic cell type in an adult mammal, it is now understood that they indeed exhibit a slow rate of turnover that decreases with age , and (2) the mass and function of β cells (and therefore the balance between new cell formation and death) can be dynamically altered to an extent to compensate for the physiologic or pathologic state of the organism . A major focus and area of controversy has been the mechanisms that underlie postnatal β-cell growth and maintenance. From the discussion in the preceding sections on prenatal development, it is evident that formation of most β cells during embryogenesis occurs through a process known as neogenesis, in which new cells arise from the differentiation of stem or progenitor cells. Although some studies suggest the existence of multipotent stem cells within the postnatal rodent pancreatic epithelium, such cells normally do not give rise to significant numbers of new β cells in adult animals.
Neonatal beta-cell turnover
The neonatal period between birth and weaning in rodents is characterized by a high rate of β-cell turnover and net increase in β-cell mass. Turnover is defined as the dynamic formation and loss of cellular mass . Because there is no longitudinal, noninvasive way to measure β-cell turnover in a given animal, the techniques that estimate β-cell turnover are based on cross-sectional studies from cohorts of animals that analyze steady-state β-cell mass, new β-cell formation (primarily by replication), and β-cell death. Nonetheless, studies using thymidine analog (BrdU) incorporation estimate that in the neonatal rat the rate of β-cell replication is as high as ∼20% new cells per day at 2 days of age, and falling to about ∼10% new cells per day by the time of weaning . By contrast, replication rates in adult rats and mice are much lower, in the range of 0–2% new cells per day [63,64]. Although replication is thought to be the primary source of new β-cell formation during this period, studies of thymidine analog incorporation cannot detect specific contributions from neogenesis, which is thought to play a role during the neonatal period .
Balancing this rate of replication is, in part, the rate of apoptosis that appears to be elevated during the neonatal period, with the frequency of apoptotic cells rising as high as ∼4% (compared to less than 0.4% in adult rats) . However, it should be noted that the true rate of β-cell death is very difficult to measure because dead/dying cells may be cleared more rapidly than can be measured by tissue morphometry, and other forms of death (necrosis) are not typically measured.
The foregoing studies in rodents appear to also reflect the dynamics of β-cell turnover in humans. Based on autopsy studies, the replication rate of β cells appears highest in children, especially infants (coincident with increases in β-cell mass during early life), then declines in adulthood . Taken together, these studies suggest a dynamic remodeling of β cells and their mass in the neonatal/early postnatal period, and mechanisms underlying the increase in β-cell replication rate have been the focus of intense investigation . The growth factors insulin and insulin-like growth factors (IGFs) are obvious candidates, given the autonomous production of both by β cells in the early postnatal period. Elimination of IRS-2 (a key protein in growth factor signaling) results in the failure to maintain β-cell mass in the face of increasing insulin resistance as mice age, suggesting a potential β-cell growth-promoting effect of this signaling cascade . However, elimination of either of the two genes encoding mouse insulin, the insulin receptor in β cells, or the IGF-1 receptor in β cells in mice does not affect neonatal replication or accrual of β-cell mass . Curiously, key cell cycle activators (Cyclins D1 and D2 and Cdk4) also appear dispensable for neonatal β-cell replication, but not for maintenance of β-cell mass in adulthood . These results collectively suggest that the early signals driving replication of β cells during embryogenesis differs from signals that drive accrual of β-cell mass in early life (see later). The physiologic significance of the high neonatal β-cell turnover is a matter largely of speculation. Considering that neonatal islets show diminished or absent responsiveness to glucose-stimulated insulin secretion suggests that the neonatal turnover may be important for the eventual refinement and maturation of β cells . Intriguingly, it has been suggested that this early turnover of β cells may result in exposure of β-cell autoantigens and trigger the pathogenesis of type 1 diabetes in susceptible individuals , although more recently this hypothesis has been challenged in mouse models .