Pancreatic Organogenesis In Mammalian Models

Modern understanding of pancreatic organogenesis is derived in large part from the seminal reports of Wessels and Cohen and Pictet and Rutter, who reported detailed analyses of pancreatic morphogenesis in the mouse and rat.16,17 The pancreas forms as dorsal and ventral evaginations of early foregut endoderm (Figure 1.2), which subsequently fuse into a single organ during the rotation of the intestinal tract. In contrast to the human pancreas, the ducts of Santorini and Wirsung in the rat do not typically fuse and drain independently into the intestinal lumen.17 These dorsal and ventral buds begin as simple cylindrical evaginations of simple cuboidal epithelium and later elaborate digitations to produce a complex branching tree of folded epithelial sheets.16 These folds and branches are reminiscent of the branching ductal tree seen in adult pancreas, but no true ductal epithelium exists at this point as evidenced by the lack of ultrastructural features of ductal epithelium and the lack of mature duct-specific markers, such as mucins or the cystic fibrosis transmembrane regulator.17-19

In the mouse, the dorsal bud initially appears as a bulge in the intestinal tube at E9.5 (embryonic days post coitus). Classic works describe a single ventral bud that appears later at E10.5 and fuses by day E12.5.17 Recent work, guided by modern markers of pancreatic endoderm (Pdxl, see below), demonstrates two symmetric ventral pancreatic bulges in the E9.5 mouse embryo in intimate association with each of the paired vitelline veins. The left vitelline vein and the left ventral pancreatic bud regress as the mouse embryo develops, leaving the right (portal) vitelline vein and a single ventral pancreatic bud by E10.5.20 Similar observations of three initial pancreatic buds have also been made in Xenopus21 and

Figure 1.2 Dorsal pancreatic anlage in the E10.5 embryonic mouse. Immuno-histochemical staining for PDX-1 identifies developing intestine (arrow) and dorsal pancreatic bud (arrowhead). Image courtesy Dr. Farzad Esni.

chick,15 suggesting that human ventral pancreas may first appear in a similar manner.

Pictet and Rutter noted the appearance of glucagon producing cells as early as 20 to 22 somites (E9.0). These cells appeared to be an integral part of the epithelial sheet, joined at the apical surface by tight junctions to neighboring protodifferentiated cells. Between 20 and 35 somites (E9.5 to E10.25), these glucagon-positive cells were seen to separate and cluster, such that they no longer contacted foregut lumen but continued, like all epithelia, to remain bounded from adjacent mesenchyme by basal lamina.17 Although this work inferred the presence of glucagon based on characteristic granule appearance on electron microscopy, recent work confirms these early conclusions. Reverse transcription polymerase chain reaction (RT-PCR) has demonstrated the appearance of glucagon, as well as other endocrine hormones, at this early stage. Somatostatin messenger RNA (mRNA) is detectable as early as E8.5 (prior to dorsal bud formation), and insulin and glucagon mRNA appear at E9.5.22 These results were also confirmed by antibody-based detection of cells that express either glucagon alone or glucagon and insulin together in early pancreatic buds.23 The initial, reasonable hypothesis was that these cells were direct precursors to the islet forms seen in newborn pups. However, it can be dangerous to infer cell lineage merely from the expression of a few markers. As discussed below, more recent work suggests that these early endocrine cells are not the true precursors to mature islets.24,25

Beginning at about E13.5, there is an exponential increase in measurable digestive enzyme activity and insulin content in the pancreatic rudiments. Pictet and Rutter termed this period the "secondary transition."17 Electron microscopy of pancreatic sections at this time reveals the accumulation of rough endoplasmic reticulum and the first appearance of zymogen granules. A population of insulin and glucagon positive cells emerge and gradually accumulate in clusters adjacent to the epithelial sheet.26 Mature islet-like clusters of endocrine cells do not appear until late in gestation at about E18.5. In contrast to the early appearance of endocrine hormones by RT-PCR analysis, amylase mRNA cannot be detected until E12.5.22 The convoluted folded epithelial sheet of the pancreatic buds acquires a distinct acinar morphology during the secondary transition. By E14.5, these acinar structures are lined by columnar cells bearing apical zymogen granules immu-noreactive for amylase, trypsin, and other zymogens.

1.3.1 Transcriptional Machinery in Pancreatic Development

The embryonic events of pancreatic development can be conceptually resolved into three phases.27 First, a restricted portion of multipotential foregut endodermal epithelium is specified to become pancreatic anlagen. Second, the cell fates of these multipotential epithelial cells are determined in a regulated manner. Third, proliferation and organization of these pancreatic precursors ultimately leads to specialized islets of Langerhans and the extensively arborized epithelial tree of the adult pancreas. Advances in molecular biologic and genetic techniques over the past 15 years have yielded insight into the transcriptional machinery that regulates each of these developmental steps. Early Patterning of Pancreatic Anlagen

A number of molecules play important roles in inducing or specifying foregut endoderm to a pancreatic fate, including the Pdxl, Hlxb9, and Ptfla transcription factors and the sonic hedgehog (Shh) signaling path ways. The parahox homeodomain transcription factor Pdxl (pancreas duodenum homeobox-1) was the first gene identified in which inactivation by germline knockout specifically disrupted pancreatic development while sparing most endodermal and all mesenchymal derivatives.28 Orthologs of Pdxl in humans, mice, rats, frogs, and fish are highly conserved in both sequence and function.29,30,31 In the adult mouse and rat, Pdxl expression is largely restricted to the p- and 6-cells of the adult islets of Langerhans, where it contributes to the expression of insulin and somatostatin.29 In the developing mouse, Pdxl is expressed throughout the entire exocrine and endocrine pancreas (including the pancreatic ducts), as well as in the proximal common bile duct and cystic ducts, the pyloric glands of the stomach, and the duodenal epithelium.32

The important role of Pdxl in pancreatic development was demonstrated in the mid-1990s, when Pdxl-' null mice were generated and found to display aborted pancreatic morphogenesis and a lack of differentiated endocrine and exocrine cell types.28 This requirement for Pdxl is cell autonomous to endoderm: recombination of Pdxl-' endoderm with wild type mesenchyme fails to generate pancreas, and recombination of wild type endoderm with Pdxl-' mesenchyme leads to normal pancreatic growth.33 Early work in the ectopic expression of Pdxl involved in ovo electroporation of anterior chick endoderm; in this model, ectopic introduction of Pdxl leads to downregulation of nonpancreatic transcription factors such as Hex and CdxA, but does not lead to bud formation or pancreatic cytodifferentiation.34 In contrast, more recent work involving adenoviral delivery of rat Pdxl to adult mouse liver has found endocrine hormone expression within liver parenchyma and restoration of glucostasis following streptozotocin ablation of pancreatic p-cells.35 36 In addition, transgenic expression in Xenopus of a fusion protein combining a potent VPl6 transactivation domain with Pdxl results in expression of endocrine and exocrine markers within hepatic anlagen.37 Therefore, overexpression of Pdxl in embryonic and mature liver is able to activate pancreas-specific gene expression, although neither experimental model generates characteristic pancreatic structural elements such as islets, acini, or ductal trees.

Although clearly essential to pancreatic organogenesis and eventual cytodifferentiation, initial hopes that Pdxl would be a "master gene" in pancreatic development have not been realized. Importantly, even though the pancreatic buds normally form within the domain of Pdxl positive endoderm, the absence of functional Pdxl does not prevent this initial bud formation.28,32,33 The earliest events in specifying the pancreatic anlagen must, therefore, involve events upstream or independent of Pdxl function. Recent evidence suggests that the homeodomain transcription factor Hlxb9 may play such a role for the dorsal anlagen. HlxbSr/- mice fail to form a dorsal pancreatic bud and do not express Pdxl in dorsal foregut endoderm. Interestingly, ventral bud pancreas continues to develop in Hlxb9r/- mice, albeit with a moderate reduction in insulin positive cells and subtle perturbation of islet architecture.38,39

The patterning of a restricted domain of foregut to a pancreatic fate is also dependent upon Shh signaling. Although Shh is widely expressed in intestinal endoderm of the mouse,40 Shh expression is conspicuously absent in dorsal and ventral pancreatic anlagen.41 Misexpression of Shh in pancreatic anlagen leads to a mixed pancreaticrduodenal architecture in pancreatic buds,41 and abrogation of Shh signaling through cyclopamine treatment of chick embryos leads to heterotopic pancreas throughout the foregut.42 This repression of Shh depends on signals from notochord.43,44 The inductive, propancreatic effect of notochord can be reproduced in vitro by soluble activin-B, a member of the transforming growth factor-family, and fibroblast growth factor 2 (FGF2).43 Of note, even though the notochord spans the entire axial length of the embryo, only anterior endoderm can be induced to express pancreatic markers by notochord.15 These observations may be unified by the observed anteroposterior restriction of activin receptor expression and function to an area of anterior endoderm that ultimately contributes to the dorsal pancreatic bud.45

The Class II basic helix-loop-helix (bHLH) transcription factor Ptf1a (pancreas transcription factor 1a, also referred to as Ptf1a-p48 or p48) has also been recently implicated in early dorsal and ventral pancreatic development. Ptf1a-/' mice fail to develop a pancreas, although they do possess a dorsal pancreatic duct remnant similar to that seen in Pdx1-~ knockouts. However, in contrast to Pdx1-~ mice, differentiated endocrine cells are present in Ptf1a/- mice, but are mislocated initially in the pancreatic mesentery and later in the spleen. This phenotype contributed to the initial view of Ptf1a as a key transcription factor in exocrine, but not endocrine, development.46 Recent work has suggested an alternate view, in which Ptf1a plays an early role in the normal development of both endocrine and exocrine lineages. MacDonald, Wright, and colleagues crossed transgenic mice expressing Cre recombinase in the endogenous Ptf1a locus onto a Rosa26R background. In these mice, cells that activate the Ptf1a promoter are permanently labeled by the Cre-mediated genomic excision of loxP flanked sites in the Rosa26R line, an excision that results in a functional lacZ reporter gene; furthermore, all daughter cells are labeled as well, allowing accurate lineage tracing.47 In heterozygous transgenics (Pf1a+/Cre), lacZ activity was observed in both endocrine and exocrine cell types, but not in duodenum or other endodermal derivatives. By breeding these transgenics to homozygosity (Pf1aCre/Cre), the investigators were able to trace the fate of these cells in the absence of functional PTFla protein. Surprisingly, a broad region of duodenal epithelium adjacent to the pancreatic duct remnant expressed lacZ, indicating that pan creatic progenitors, in the absence of PTF1a, revert to an intestinal cell fate.47 These results provide evidence that PTF1a commits foregut endo-derm to a pancreatic fate, and that in its absence, this endoderm defaults to intestinal differentiation. Patterning of Dorsal vs. Ventral Anlagen

Although the dorsal and ventral pancreatic buds eventually fuse to form a single organ, they appear to be histologically (and perhaps evolution-arily) distinct. In developing Xenopus, insulin expression is restricted to the dorsal pancreatic bud and its derivatives21; however, in rats48 and chicks,15 the dorsal bud is enriched in glucagon positive cells. The induction of the dorsal and ventral anlagen also appears to occur via distinct molecular pathways. This fact is not surprising in light of the vastly different mesenchymal environments encountered by the dorsal and ventral buds. The dorsal anlagen develops in intimate contact first with notochord and later with dorsal aorta and pancreatic mesenchyme, and the ventral anlagen develops in contact with hepatic precursors, cardiac mesoderm, and septum transversum.49,50

As mentioned above, Hlxb9r/~ mice completely fail to form a dorsal, but not ventral, pancreatic bud.38,39 Further work focusing on ventral pancreatic endoderm in the mouse suggests a distinct role for hedgehog signaling as well. Naked ventral endoderm harvested from E8 to E8.5 mice grown in collagen gels, unlike dorsal endoderm, thrives in the absence of mesenchymal signals, does not express Shh, and acquires Pdx1 expression. If exposed to FGF2 or cocultured with cardiac mesenchyme, ventral endoderm cultures do not acquire Pdx1 expression, but do express Shh and albumin (a marker of hepatic cell fate). These results propose a model in which ventral endoderm is bipotential and that FGF2 signaling from cardiac mesenchyme patterns the anterior portion toward liver, rather than toward a default of pancreas.50 Thus FGF2 in ventral endoderm appears to block pancreatic differentiation and upregulate Shh, and FGF2 signaling in the dorsal chick endoderm appears to promote pancreatic differentiation and downregulate Shh.43 Endocrine Cytodifferentiation

A number of transcription factors important to endocrine differentiation have been identified, and schematic cascades of transcription factor activation have been proposed to unify existing data and explain the formation of four specialized cell types from foregut endodermal precursors.51 Although these schemata depict linear cascades, the actual interplay is likely to be more complex and modified by as yet unidentified molecules.

Neurogenin3 (ngn3) is a Class II bHLH transcription factor that plays an important role in endocrine differentiation. Targeted knockout of ngn3 leads to a complete absence of islets or differentiated endocrine cells.52 Overexpression of ngn3 in transgenic mice leads to hypomorphic, poorly branched pancreatic buds composed principally of endocrine cells, suggesting that ngn3 commits a precursor population to an endocrine fate.53,54 Ngn3 transcripts are first detected in E9.0 embryonic pancreatic buds and peak at E15.5, coincident with the "second wave" of cytodifferentiation. Ngn3 is undetectable in newborn and adult mouse pancreas and cannot be detected in embryonic cells positive for insulin, glucagon, somatostatin, or peptide YY, suggesting that ngn3-expressing cells represent a population of committed but still undifferentiated endocrine precursors.54

Other transcription factors have been identified to be important in endocrine pancreas development, although none have a null phenotype as dramatic and specific as that seen in ngn3 knockouts. NeuroD (also known as BETA2, beta cell E-box transactivator-2) is inducible by ngn3,55 enhances insulin promoter activity,51 and its null phenotype exhibits marked reduction of all endocrine cell types but with preserved ngn3 expression.56 Inactivation of the paired domain homeobox gene Pax4 leads to an absence of mature p- and 6-cells, with a preponderance of a-cells and preservation of exocrine development; this phenotype suggests a role for Pax4 as a switch between a and p/6 cell fates.57 Knockout of the related homeobox gene Pax6 results in a pancreas with few or no a-cells, in addition to dramatic reduction of p-, 6-, and PP-cells.58 Nkx2.259 and Nkx6.160 are members of the NK class of homeodomain proteins, both of which are broadly expressed in the early (E9.5/E10.5) pancreatic bud and then progressively restricted to mature endocrine cells by E15.5. Knockout of either disrupts p-cell development, and analysis of double knockout mice suggests that Nkx6.1 lies downstream of Nkx2.2.60 Exocrine Cytodifferentiation

In contrast to endocrine pancreatic differentiation, few regulatory components responsible for exocrine differentiation have been identified.61 As discussed above, Ptf1a plays an early role in pancreatic development and is required for the appearance of exocrine cell types. Ptf1a is the only known transcription factor specific to the adult exocrine pancreas.62 The trimeric transcriptional activator complex PTF1 and its cognate binding sequence was initially identified through DNA footprint analysis of the promoters of rat amylase, elastase, and trypsin genes.63 Subsequent analysis demonstrated PTF1 to be composed of three subunits, of which only one, PTF1a-p48, is tissue specific. Like several genes important in endocrine pancreatic differentiation (e.g., ngn3, neuroD), Ptf1a is a Class

II bHLH protein.64 Transcripts are first detectable at E9.5 in the pancreatic buds, although transcripts can be transiently found at E8.5 in the developing hindbrain.47,65

Mistl is a recently identified Class II bHLH transcription factor that is only expressed in acinar cells of the pancreas, although it is expressed in a variety of secretory extrapancreatic tissues, such as salivary acini, and serous cells of the stomach, prostate, and seminal vesicles.66,67 In the embryonic pancreas, Mistl is detectable early (E10.5) in the dorsal pancreatic bud; at E14, when acinar structures are histologically distinct, Mistl expression is confined to acinar cells. Knockout mice null for Mistl survive and at birth are reported to be grossly undistinguishable from control littermates. Examination of Mistl~/- mice later in life reveals defects in acinar cell organization and loss of acinar cell polarity; similar defects are also found in salivary and seminal vesicle epithelia. 68 The bHLH domain of Mistl is highly similar to that found in the recently characterized Drosophila transcription factor dimmed, which is required for amplified levels of secretory activity in Drosophila neuroendocrine cells.69 Given the restriction of Mistl to secretory epithelial cell types in mammals, dimmed and Mistl may ultimately prove to be functional orthologs that both promote cellular machinery necessary to maintain a secretory cell type.

Broadly speaking, the exocrine pancreas encompasses both the specialized acinar cells, which secrete digestive zymogens, the ductal epithelium, and the centroacinar cells at the junction of acinar and ductal elements. Even though many studies consider acinar and ductal differentiation under the umbrella of exocrine pancreas, it may be erroneous to link ductal and acinar differentiation, as demonstrated by a recent study also employing Cre labeling techniques. Melton and colleagues created a transgenic mouse using the Pdxl promoter to drive expression of a tamoxifen-inducible Cre recombinase. By administering a tamoxifen pulse to pregnant mothers at varying stages of gestation and later examining the embryos for Cre reporter activity, the investigators noted early divergence between ductal labeling and acinar or islet labeling. If given tamoxifen between E9.5 and E11.5, all three cell types were labeled. However, if tamoxifen was administered earlier (E8.5) or later (E12.5) only acinar and islet cell types were labeled.70 Furthermore, by creating a transgenic in which Pdxl expression could be inhibited by administration of tetracycline to pregnant mothers, MacDonald and colleagues demonstrated that inhibition of Pdxl after E12.5 prevents acinar and islet differentiation, but does not prevent main pancreatic duct formation.71 These results suggest that ductal precursors may diverge from a common acinar/islet precursor during early development of the pancreatic buds. Regulation of Endocrine and Exocrine Differentiation

Several lines of evidence point to a common progenitor cell for endocrine and acinar cells. All epithelium in the pancreatic buds express Pdx1 at early stages, but with cytodifferentiation acinar cell types lose high-level Pdx1 expression. Conversely, Ptf1a is only found in acinar cells of the adult pancreas, but Cre-loxP lineage tracing demonstrates that a large proportion of islet cells express Ptf1a at some point in their ontogeny. The events that govern this determination toward exocrine or endocrine cell fates is especially relevant to the goal of directing an as-yet-unidentified pancreatic stem cell to adopt p-cell fates for cell replacement therapy in diabetes.

A recent report regarding the "transcriptome" of dorsal bud pancreatic cells provides some insight into the cascade of events that diverge to become endocrine or exocrine cytodifferentiation.72 These investigators trypsinized dorsal pancreatic buds from E10.5 mice to yield single epithelial cells. The mRNA isolated from an individual epithelial cell was then amplified and hybridized on a custom microarray containing cDNAs of transcription factors previously implicated in pancreatic development. By analyzing dozens of individual cell "transcriptomes," several consistent patterns emerged. Although ordered relationships of precursors and progeny cannot be established using this method, conceptually ordering the transcriptomes based on stepwise accumulation or loss of a transcription factor provides a logical and reasonable estimate of ordered relationships. In this manner, Chiang and Melton proposed that a population of Pdx1+ cells (which ubiquitously coexpress Nkx2.2 and Nkx6.1) first acquire Ptf1a expression and then diverge into two paths:

1. An exocrine path, in which amylase and trypsin expression is initiated, followed by later loss of Nkx2.2 and Nkx6.1

2. An endocrine path, in which ngn3 is activated and Ptf1a is lost24,72

Notch signaling appears to play an important role in regulating the commitment of this common progenitor to endocrine and exocrine cell fates. In the developing nervous system, Notch signaling mediates the phenomenon of lateral inhibition, in which a differentiated cell instructs its neighbors to maintain an undifferentiated state. Interruption of Notch signaling through targeted knockout of its intracellular mediator RBP-J or the Notch ligand Dll1 (Delta-like ligand 1) led to apparent expansion of ngn3 positive precursors in the early foregut endoderm.53 Mice null for the Notch downstream effector Hes1 have a hypoplastic pancreas, with a predominance of endocrine tissue.73 Finally, although HES-1 is uniformly detected throughout the early pancreatic buds, by E9.5 a HES-1 negative population emerges; these HES-1 negative cells, in which Notch signaling is presumably inactive, acquire markers of early endocrine differentiation like ngn3, neuroD, and Pax6. These early observations were unified in a model in which active Notch signaling is required to prevent a population of pancreatic precursors from undergoing early endocrine differentiation, and thus be available for expansion and later exocrine differentiation. In this model, ablation of Notch signaling results in precocious and early endocrine differentiation, depleting the pool of pancreatic precursors and resulting in a hypoplastic, endocrine-dominant pancreas.73 This model is informed by more recent work involving transgenic misexpression of the constitutively active Notch1 intracellular domain (NotchICD) in Pdx1-positive cells.74 The pancreata of these mice are severely deficient in both differentiated endocrine and exocrine cell types, consistent with a role for active Notch signaling in maintaining a precursor-like state in pancreatic anlagen.

1.3.2 Epithelial-Mesenchymal Signaling

The seminal experiments of Golosow and Grobstein and Wessels and Cohen indicate an important role for endodermal-mesenchymal interactions in pancreatic development.16,75 Golosow and Grobstein dissected intact dorsal pancreatic buds from E11 mouse embryos and cultured them in specialized Eagle's media. Pancreatic buds with intact, attached mes-enchyme grow and eventually differentiate into exocrine and endocrine cell fates with appropriate acinar architecture. Removal of the mesen-chyme, however, results in pancreatic buds that regress in size and show no histologic evidence of differentiation. Surprisingly, recombining these naked buds even with heterotopic (nonpancreatic) mesenchyme restores growth and differentiation. Finally, Golosow and Grobstein demonstrated that separating a pancreatic bud and salivary mesenchyme by a Millipore filter did not eliminate this trophic effect, suggesting that soluble factors are involved as opposed to direct ligand-receptor interactions.75 This work was expanded upon by Wessels and Cohen, who tested the ability of pre-E11 endoderm to thrive in culture.16 Through exhaustive tissue recombination experiments they noted that foregut endoderm from early embryos (3 to 13 somites, E8.0) develops acini in tissue culture if recombined with pancreatic mesenchyme, but fail to do so if recombined with nonpancreatic mesenchyme. Endoderm taken from later stage embryos (14 somites, ~ E9.0), in contrast, is able to form pancreas when provided with nonpancreatic mesenchyme.

These experiments demonstrated three principal concepts:

1. That mesenchyme is required for growth and differentiation of the pancreatic bud at all stages.

2. Mesenchymal trophic factors are likely permissive (and not instructive) for endoderm beyond 14 somites of age, as nonpancreatic mesenchyme serves equivalently well.

3. As a corollary to the previous conclusion, the instructive signals specifying these endodermal buds to become pancreas have already taken place by the time of bud formation at E9.5.

A recent demonstration of the permissive role of mesenchyme comes from mice null for the LIM homeodomain gene Isll; these knockouts failed to form dorsal pancreatic mesenchyme, but retain a normal ventral mes-enchymal environment (septum transversum). In turn, Isll_/- mice exhibited severe failure of dorsal, but not ventral, pancreatic development. In vitro tissue recombination experiments further demonstrated that the defect is intrinsic to the mesenchyme, as recombination of dorsal foregut endoderm from Isll-~ mice with wildtype mesenchyme led to normal pancreatic development.76

Efforts to purify the responsible "mesenchymal trophic factor" in the 1960s and 1970s were unsuccessful, partly due to the purification technology available at the time and partly due to the likely fact that this "factor" must in fact comprise multiple molecules.5,77 The recent recognition of notochord repression of Shh in early foregut endoderm, via soluble factors such as activin-pB or FGF2, suggests a likely molecular correlate for the early instructive signals identified in the early experiments of Wessels and Cohen. Elegant experiments in the mouse have also demonstrated a required role for endothelial tissue in pancreatic specification.20 Isolated foregut endoderm from E8.5 embryos grown in vitro fail to differentiate into pancreas and develop an intestinal-like epithelium. Coculturing this endoderm with notochord, however, led to expression of Pdxl; coculturing with isolated dorsal aortae or other sources of endothelial tissue led to the expression of Pdxl and insulin.

Other experiments have also revealed a role for mesenchymal factors in biasing pancreatic epithelia toward exocrine or endocrine pancreas. Whereas pancreatic buds with intact mesenchyme survive and generate acinar, ductal, and islet cells when transplanted underneath renal capsule, buds transplanted without their mesenchyme fail to form acinar cells, but do form dense clusters of insulin and glucagon positive cells.78 In vitro experiments in which E12.5 pancreatic buds are grown in collagen gels with mesenchyme demonstrate predominantly acinar epithelium by Day 7 of culture, with a 5 to 1 ratio of amylase-positive to insulin-positive cells. In contrast, naked buds grown without mesenchyme, while smaller, have approximately equal ratios of amylase- and insulin-positive cells. Addition of follistatin to cultures of naked buds restored the acinar predominance.79 Follistatin is a soluble protein known to bind to and inactivate activin and other members of the transforming growth factor-p (TGF-p) family (see also Chapter 13), suggesting that active TGF signaling may favor endocrine over exocrine cell types. In fact, treatment of in vitro pancreatic bud cultures with TGF-p1 leads to increased endocrine and decreased acinar cell mass, with little effect on ductal cell mass.80 Transgenic expression of a dominant negative TGF-p receptor leads to increased acinar cell proliferation and apoptosis.81 Together, these results suggest that pancreatic mesenchyme, at least in vitro, may negatively regulate endocrine pancreas development through inhibition of soluble TGF-p family members.

Other growth factors have been investigated in in vitro pancreatic bud cultures. E11.5 rat dorsal pancreatic buds cultured for 7 days in collagen gels fail to grow or develop exocrine differentiation in the absence of mesenchyme. Treatment with FGF1, FGF7, or FGF10, however, leads to marked growth and expansion of the exocrine, but not endocrine cell mass82; misexpression of FGF10 in pancreatic epithelia (under control of the Pdx1 promoter) leads to a hyperplastic pancreas deficient in differentiated endocrine and exocrine cell types.83,84 Treatment of naked E13.5 rat dorsal pancreatic buds cultured in collagen gels with epidermal growth factor (EGF) leads to a similar increase in epithelial cell mass, but with an apparent downregulation of amylase, insulin, and glucagon expression. Withdrawal of EGF in these cultures is followed by the appearance of insulin expression throughout the enlarged bud.85 Knockout mice lacking functional EGF receptor have smaller pancreas with reduced endocrine cell mass; at birth, their pancreas lack islets, but have organized islet-like streaks of endocrine tissue along pancreatic ducts.86 All in vitro experiments must be interpreted with due caution, however, given the necessarily artificial conditions. For example, culturing pancreatic buds in collagen gels tends to give a predominantly acinar epithelium, and culturing buds in a basement membrane-like matrix (Matrigel ™) results in a predominantly endocrine epithelium.80 In vivo experiments, such as tissue-specific overexpression of cytokines or their receptors, will help to dissect the role of known cytokine families in pancreatic development. The effects of the TGF and EGF family on the adult pancreas are described in Chapter 13.

1.3.3 The Lineage of Endocrine Cell Types

Given the medical importance of the pancreatic islet and the p-cell, interest has been intense in the endocrine stem cell and in the factors controlling endocrine cytodifferentiation (see also Chapter 28). Surprisingly, the cell of origin for the mature islet is not clear. As described above, previous work in the early pancreatic bud (circa E9.5) had demonstrated the presence of cells expressing glucagon alone, cells coexpressing glucagon and either insulin or pancreatic polypeptide, or multihormonal cells expressing all three endocrine hormones.24 Logically, these cells were thought to give rise to the wave of differentiated endocrine cells arising after E13.5, but several lines of evidence suggest that these early cells give rise only to the a-cell mantle of mature islets and not to mature por PP-cells.

These early glucagon-positive endocrine cells do not coexpress Pdxl or Nkx6.1, two markers usually found in developing endocrine cells, and during the secondary transition they come to surround the developing wave of insulin-positive cells.24 Disruption of pancreas formation by knockout of Pdxl,33 knockout of Ptfla,46 or by overexpression of HlxbS^1 does not prevent the formation of these early glucagon-positive endocrine cells. Finally, Herrera has reported two suggestive lineage tracing studies. In the first, transgenic expression of diphtheria toxin under a glucagon promoter eliminated a-cells, but not p-cells, and expression under an insulin promoter reduced a-cell populations, but eliminated p-cells from mature islets. Questions regarding the cell autonomy of the toxic effect led to a second experiment, in which lineage tracing was performed using Cre-loxP mediated genomic recombination events. These results suggested that the ontogeny of mature p-cells does not include activation of the glucagon promoter.25 Taken together, these data suggest that these early multihormonal cells contribute to the a-cells of mature islets, and a distinct group of progenitors contribute to mature p-cells.

Supplements For Diabetics

Supplements For Diabetics

All you need is a proper diet of fresh fruits and vegetables and get plenty of exercise and you'll be fine. Ever heard those words from your doctor? If that's all heshe recommends then you're missing out an important ingredient for health that he's not telling you. Fact is that you can adhere to the strictest diet, watch everything you eat and get the exercise of amarathon runner and still come down with diabetic complications. Diet, exercise and standard drug treatments simply aren't enough to help keep your diabetes under control.

Get My Free Ebook

Post a comment