Introduction

Endocrine tumors of the gastroenteropancreatic (GEP) axis consist of cells that are capable of amine precursor uptake and decarboxylation and therefore have been named APUDomas. (1) The morphologic similarity of the APUD cells suggested a common embryologic origin, which was believed to be the neural crest but was later revised to include the neuroectoderm or, in the case of endocrine cells, the dorsal placoderm. However, most workers agree that these tumors should be classified according to their secretory products (i.e., carcinoid, gastrinoma, insulinoma, somatostatinomas, glucagonoma, vasoactive intestinal peptide [VIPoma] and pancreatic polypeptide [PPoma]).(2),(3;4) Before presenting their clinical characteristics we will briefly review the currently held views on the embryologic origin of these cells and the factors regulating their growth, differentiation and apoptosis that serve to maintain homeostatic balance. We will also indicate the derangements that result in tumor formation.

Developmental Origin of GEP Cells during Embryogenesis

The pancreas is composed of exocrine tissue that produces enzymes for digestion, and an endocrine system designed to maintain glucose homeostasis within narrow confines. The adult endocrine pancreas contains four different cell types which produce insulin (b cells), glucagons (a cells) somatostatin δ cells) and pancreatic polypeptide (PP cells). These are contained within a highly organized structure with b cells in the interior and the remainder as a surrounding mantle. The vascular supply is elegantly organized to deliver blood to the central core of the islet and to perfuse the outer layers in a centrifugal manner form b to a to δ. (synonym B (b) A (a ) D (δ). The adult pancreas also has a sophisticated ductal drainage system that ostensibly is present as a conduit for enzymes to reach the gastrointestinal tract but appears to retain cells capable of transdifferentiation into exocrine or endocrine components upon appropriate activation. A crucial question has arisen as to whether or not the adult pancreas retains these precursor cells and if they can be identified. In recent years a molecular fingerprint of embryonic islet precursors has begun to emerge (5) and precursor cells within embryos and postnatal mice are found in the ducts (6). These ductal precursor cells can be identified by their expression of Glut 2 (7). Recently cells expressing the neuronal antigen, an intermediary filament protein, were located in pancreatic ducts of adult rats and were found to differentiate into insulin expressing cells in vitro(8), further suggesting that the study of the characteristics of embryonic development might be able to assist in the capture of the elusive precursor in the adult pancreas.

Inductive signals important in the initiation of growth and development of the pancreas to a large extent have their signals enacted or amplified by genes targeted within the embryonic stem cells or cells committed to developing into a variety of pancreatic endocrine, exocrine or ductal cells. Pancreatic islet-specific gene expression is to a large extent controlled at the transcriptional level by the binding of islet enriched transcription factors to sequences in islet genes. (Figure 1).

Figure 1

These transcription factors are involved in the temporal expression of genes that direct pancreatic development. Cell specific and extrinsic factors present in the endoderm act in a permissive or restrictive manner to direct the formation of the islets and the various cells and structures that comprise the adult islet (9). The PDX-1 encoded homeodomain protein in mammals (STF-1, IDF-1, IDX-1) was isolated as a transcriptional regulator of insulin and somatostatin (10-12). It binds and transactivates the insulin promoter (13). PDX was first detected in embryonic pancreatic and duodenal endoderm. It is detected in all embryonic protodifferentiatd epithelial cells during pancreatic development (14). In the pancreas it becomes progressively restricted to the islets, where it is produced in >90% of beta cells, 15% o delta cells and 3% alpha cells. PDX-1 defines pancreatic gene expression pattern and cell lineage differentiation (15). Mice heterozygous for PDX-1 develop normally but in homozygous PDX – the normal branching outgrowth of the pancreas is arrested at an early stage (13;16); Maturity onset diabetes occurred in patients heterozygous for the gene (17;18). Diabetes develops in aging transgenic mice following suppression of PDX-1{Thomas, 1 A.D. THOMAS01 /id}. The regulation of the PDX-1 gene appears to be central to the development of the pancreatic anlage during embryonic development as well as maintaining islet mass in the adult and contributing to the regulation of insulin secretion from the adult pancreas(19). The PDX-1 gene is initially expressed in exocrine and endocrine pancreatic precursors but later becomes restricted to the beta cells in the islets. Transgenic models leading to loss of PDX-expression, either via double knockouts, dominant negative control or elimination of the target binding protein leads to pancreatic agenesis in the case of the double knockout and with haploinsufficiency, to defects in glucose-stimulated insulin secretion in mice and in humans (13),(18). Abolition of PDX-1 in differentiated bcells in mice results in loss of the bcell phenotype, and impaired expression of Glut 2, glucokinase required for bcell production of insulin in response to glucose(13). PDX-1 is also found where bcell neogenesis is occurring as in duct ligation model of neogenesis (20), partial pancreatectomy(21), overexpression of TGF alpha(22), or interferon gamma (23). In all instances the formation of new bcells is preceded by expression of PDX-1. It has been reported that stimulation of the initiation of transdifferentiation of adult stem cells with INGAP is associated with increased expression of PDX-1 in both ducts and subsequently newly formed islets (24). This endorses the notion that cell differentiation in the adult can recapitulate normal fetal ontogeny. Further support for this notion derives from the observation that transdifferentiated cells stained positive for the neuronal antigen PGP 9.5. PGP 9.5 is an isoform of ubiquitin carboxyterminal hydrolase (UCTH-LI) and is a marker for neurons and neuroendocrine cells in the skin (25-27) as well as the pancreas (28). It was also found in ductal cells during embryonic islet morphogenesis and in our studies on duct ligation of the hamster pancreas (29). Thus the combination of PDX-1 and PGP9.5 suggest evolution from the ductal phenotype to a cell precursor en route to neoislet formation.

Various studies have shown that the hepatic nuclear transcription factors( HNFs) form a hierarchy of transcription factors that exert positive and negative influences on pancreatic islet growth and development (30;31). Of particular relevance to islet development is the interpolation of HNF 3 b, a member of the forkhead/winged helix family of transcription factors, between inductive signals for bcell development and the expression of PDX-1 (32). Similarly HNF1 a binds to a regulatory domain of PDX-1 and knockouts have reduced expression of PDX-1 (32). Cell specific and extrinsic factors are expressed during fetal development that determine the region of the endoderm destined to form the pancreatic bud (33), (34), (35), (36). The initiation of the pancreatic program requires that signals specify the pancreatic region within the developing endoderm. Sonic hedgehog and Indian hedgehog genes dictate an intestinal differentiation, and for pancreatic development to occur these genes must be excluded. Candidate factors for excluding these genes are Activin –b and fibroblast growth factor (37). The dorsal and ventral buds may develop differently and the LIM homeodomain protein Isl I may be an important determinant of pancreatic development (38). Lateral specification of pancreatic development is mediated by Notch signaling by specifying a particular pathway in a field of initially equivalent cells. Notch signaling controls the choice between differentiated endocrine and progenitor cell fates in the developing pancreas and a block in activation of the Notch receptor resulting in high Neurogenin 3 expression and promotes an endocrine fate. These cells upon differentiation migrate into the adjacent mesenchyme where they cluster and upon receiving an inductive signal, for example INGAP; generate distinct endocrine cell types depending on the inductive milieu. This in turn activates PDX-1 which appears to act upstream of fibroblast growth factor (FGF) signaling (13) and induces the FGF1-5 ligands. This is necessary for full maturation of the glucose sensing mechanism of the b cell including expression of the low affinity glucose transporter, Glut 2, and the proinsulin processing machinery, the proinsulin convertase PCI/3 and PC2 responsible for converting proinsulin to insulin. This appears to be conserved between mice and men (33), (34), (36). HNF3 b is a candidate for initiating the positive response to the inductive signal and is expressed in the mouse fetus prior to the expression of PDX-1 at embryonic day 8 (E8.0) in the dorsal endoderm of the fore/midgut before the appearance of the insulin and /or the glucagon expressing cells (19). These primordial cells lack b cell specific markers. Around E13 the number of endocrine cells starts to increase and develop the characteristics of endocrine clusters destined to develop into the organ of Langerhans (39). The primordial cells not yet committed to develop into pancreatic islet cells express a number of neuronal markers including Neurogenin 3, (40), PGP 9.5 , and of Nestin (41). Cells destined to become islet cells appear to express the glut 2 transporter prior to development of hormone secretory capacity and this has been used as a marker for these committed cells in the pancreatic ductal system. (7). The organization into individuala,b, δ and PP islet cells is dependent upon appropriately timed expression of a number of other genes including PAX 4, PAX 6 and PDX-1 (for a detailed review see (42)) and is complete by day 18, but further refinements and development of glucose sensing occurs in the 2 week postnatal period in mice (39). Although HNF 3 b may be necessary for the response to inductive factors in embryonic development, it appears that HNF-1a is necessary to maintain the islet specific expression pattern and is required throughout adult life (43).

Preliminary data suggest that embryonic stem cells can be differentiated into insulin secreting cells ex vivo, but these cells do not achieve terminal differentiation and have a low insulin content and poor response to glucose. Their growth is unbridled and despite production of b cells, they fail to cure diabetic mice (44) Israeli scientists found insulin-producing cells in embryoid bodies formed spontaneously from ESCs when they stop growing, but these too do not make sufficient insulin(45). Soria and colleagues used gene-trapping techniques to isolate insulin-producing cells and transfected an antibiotic resistant gene adjoined to the insulin promoter. When these cells formed three-dimensional structures the cells increased insulin to therapeutic levels but this of course remained unregulated (46). Others have created long-lived cell lines from b cells (47), while others have engineered beta cells from non-b cells (48). These have lacked the necessary ingredients of expandability, and physiologic regulation including glucose sensing and an off mechanism in the absence of glucose.

Rather than struggle with the propensity of non-pancreatic stem cells with their reversion to their former state and the difficulties of identifying the necessary control mechanisms for transitioning ESCs into pancreatic stem cells, some researchers have looked to the pancreas as a source of more mature stem cells. Peck and colleagues in a multi-step process, identified islet producing stem cells and transformed these into islet progenitor cells. These grew into islet like structures which increased pancreatic mass 10,000 fold but the cells never fully matured (8). Bonner–Weir and colleagues have applied growth promoting substances to ductal cells in culture and stimulated these to grow and express the IPF-1/PDX-1 protein, the transcription factor necessary for endocrine cell development. At this stage a switch to differentiation factors induces the cells to form cultivated human islet buds which produce a small amount of insulin in response to glucose. The single biggest limiting factor of course was the limited capacity for forming the number of cells required to reverse diabetes (49). Starting with adult human b cells, Levine and colleagues immortalize them by transfection with the SV40 T antigen and the K-ras oncogene to stimulate growth. Cells are induced to transdifferentiate by transfection with PDX-1 and formation of the three dimensional complexes with cell-to-cell contact conducive to insulin production. With the appropriate application of a growth-inducing stimulus these constructs are now capable of secreting insulin in response to glucose. However these structures are not stable. When implanted they metastasize like tumors and lose their insulin secretory capacity (50;51).

Others and we have elected to utilize the factors resident in the pancreas to stimulate islet cell growth and proliferation as an alternate to the above approaches (42;52). It has been known for years that factors present in the pancreas mesenchyme may have an important role in islet integrity (39). More recently the close association between islets and their ducts of origin has been established by electron microscopy of pancreases using cytokeratin 20 markers of duct cells and islet hormone markers. (53). The close contact between the islets and duct system has raised interesting possibilities. For example, the open nature of insulin and somatostatin cells allows secretion of hormones into the intestinal lumen, a feature we named “Lumone” many years ago (54) and indeed insulin, serotonin, gastrin, somatostatin and members of the Reg family of peptides have been found in the intestinal lumen and pancreatic juice (55-59) and receptors for insulin have been found on the luminal surface of duct cells (60-62). The reciprocal relationship may have greater consequence for b cell function. Acinar cell proteins such as Reg are found in pancreatic juice (63-67) and their target may be the stimulation of growth and proliferation of duct cell proliferation and differentiation (63-67). Okamoto and colleagues established a model for islet regeneration in 90% depancreatectomized rats by the administration of the poly (ADP-ribose) synthetase inhibitor, nicotinamide and found that the remaining islets underwent considerable hypertrophy. They screened the islet derived cDNA library and found the novel regenerating gene and named it Reg. The rat Reg cDNA encoded a 165 aminoacid protein with a 22 aminoacid signal peptide. Subsequently they isolate the human counterpart which is 165 aminoacids, with 68% homology to that of the rat Reg protein. The recombinant forms of Reg have been shown to expand b cell mass by inducing hypertrophy of existing islets and limited replication (68).They then isolated several Reg and Reg-related genes from human rat and mouse and grouped members of the family into three subclasses. Group I encodes a b cell growth factor and some of the type 111 (a, b,and g ) targets neuronal cells and cells of the epithelial alimentary tract where it is found extensively. In the process of ordering these genes (69) a novel form of Reg, Reg III δ, was found with 6 exons, spanning about 3Kb, encoding a 175 aminoacid protein with 40-52% of homology to other Reg proteins. Unlike Reg I and Reg II which are expressed in hyperplastic islets, Reg III delta was expressed predominantly in the exocrine pancreas. This mouse form of Reg may be the counterpart of hamster and human INGAP gene that is found almost uniquely in the exocrine pancreas, appearing with islet neogenesis and responsible for stimulating proto-differentiated cells in the ductal system to proliferate, differentiate into islets and function physiologically to reverse diabetes (70). However Sasahara and colleagues (71) cloned a novel cDNA from mouse pancreas having a 72% homology to hamster INGAP cDNA and 47-52% homology to other members of the Reg family including the different forms of Reg and pancreatitis associated protein (PAP) and pancreatic thread protein of rat, mouse and man. They refer to this protein as INGAP related protein. In contrast to INGAP which is expressed during neogenesis, INGAPrP was abundantly expressed in the normal mouse pancreas.

The developing pancreas appears as a protrusion from the dorsal surface of the embryonic gut (39). The different islet-cell types appear sequentially during development in vivo. Therefore, it seems reasonable to propose that coordinated growth depends on the specificity of growth factors. Rosenberg and Vinik (72) used a model for new islet formation (i.e., nesidioblastosis) and showed that pancreatic ductal cells are capable of differentiating on stimulation into adult endocrine cells that are capable of secreting insulin in a fully regulated manner. This has led to the notion that endocrine tumors derive from a totipotential stem cell in the gut that is capable of differentiating into any one of a variety of cells that may be responsible for the clinical syndrome (Fig. 2).

Figure 2

A great deal of interest is now being focused on the factors responsible for the initiation of growth, growth proliferation, differentiation into adult endocrine cells, and, in neuronal systems, for growth cessation and cell maintenance. Several models of pancreatic regeneration and tumor formation have been established (63;73-82).

Growth Factors and the development of neoplasms of the gep axis

Multiple growth factors and receptors are frequently expressed in GEP tumors. These growth factors may include insulin-like growth factor-1, platelet-derived growth factor, transforming growth factors (TGF) -a and b basic fibroblast growth factors, and nerve growth factor (83;84). The frequent co-expression of TGF-b and its corresponding receptor, the epidermal growth factor receptor, suggests the presence of autocrine regulatory mechanisms in these tumors(84). TGF-b has been implicated in the desmoplastic reaction associated with carcinoid tumors (84;85). Overall, the precise role of these growth factors and their importance in the growth and progression of GEP tumors is unknown.

Apoptosis (i.e., programmed cell death) has been shown to be an important process that may occur under normal physiologic conditions, including embryonic growth and development, the differentiation of B-cell populations, and the involution of cells deprived of necessary growth factors(86). Apoptosis may be induced by a variety of chemotherapeutic drugs and cytokines (87). Several growth factors and substances that are secreted by neuroendocrine tumors, including TGF-b (88) glucocorticoids, and somatostatin (89) have been shown in other model systems to induce apoptosis. The importance of apoptosis in the normal growth and differentiation of neuroendocrine tissues, however, and the importance of apoptosis in the response of GEP tumors to chemotherapy remain unknown.

The multiple endocrine neoplasia (MEN) characterized by the combined occurrence of tumors of the pituitary, pancreas, and parathyroid glands is associated with the loss of a tumor suppressor gene on chromosome 11q13 (90;90;91). This is the same chromosome on which the insulin gene has been located (92).It has been linked to nesidioblastosis in certain families and parathyroid mitogenic activity can be identified in the plasma of patients with MEN-1(93;94). All of this suggests a genetic predisposition to tumor formation based on elaboration of a growth factor. Data from cell lineage analysis of pancreatic islet cells suggest that progenitor cells, which contain catecholamines, are present in pancreatic ducts and give rise to the glucagon and insulin cells of adult islets;(95) and they can be stimulated to grow by plasma from patients with MEN-1. The findings that patients with MEN-1 also might elaborate into their plasma mitogenic factors for pancreatic islet cells led McLeod and colleagues to postulate a genetically determined, circulating growth factor in the initiation of GEP tumor growth(96). It has been suggested, but not proven, that allelic loses in the MEN-1 tumor suppressor gene located in the 11q13 region also might be responsible for sporadic parathyroid, pituitary, neuroendocrine tumors of the stomach, pancreas, and intestine(97). The few cases of carcinoid tumors studied have not shown losses in this region.

In addition, MEN-2a, (98); (99) MEN-2b, (100); (101) and familial medullary thyroid carcinoma are associated with mutations of the RET proto-oncogene, which is a conventional dominant oncogene located on 10q11.2. Although mutations in this region have been associated with sporadic medullary thyroid carcinoma, the role, if any, of this gene in sporadic GEP tumors is not known.

Characteristics Of Neuroendocrine Cells

A number of peptides originally isolated from gut endocrine tissues have been shown to occur in nerves. These include gastrin, cholecystokinin, vasoactive intestinal polypeptide (VIP), and substance P (SP). As a corollary, peptides that have been found primarily in nervous tissues have now been identified in gut endocrine cells and include somatostatin, enkephalins, SP, neurotensin, and thyrotropin-releasing hormone (TRH) (102;103);(104). Because many of these peptides occur both in endocrine cells and nerves, “endocrine” tumors of the gut may, in fact, be endocrine or neurocrine. Unique to the GEP axis is the ability of the endocrine cell to secrete a variety of peptides and amines. Hormonal peptides not only have been found within the same cell (e.g., motilin and serotonin in the enterochromaffin [EC] cell), but they have also been localized to the same secretory granule. Whether these act within the secretory granule in a paracrine manner or are co-regulated in some way is not clear. At any one point in time several hormones and amines are co-secreted,and the symptom complex derives from one or more of the peptides and amines produced and cannot simply be ascribed to a single factor. Thus, a tumor may secrete one peptide, recur, and secrete yet another, and its metastases may secrete still other peptides. In the British National Supra-Regional Survey of National Health Service Hospitals, 58% of 353 patients with neuroendocrine tumors had increased serum levels of two or more hormones at diagnosis. Nine percent of patients had clinical symptoms related to different hormones, and four patients developed new symptoms from secretion of a second hormone after diagnosis (105).

Rick Lloyd has recently reviewed the use of various markers to identify neuroendocrine characteristics (106). Although there are many broad-spectrum neuroendocrine markers, chromogranin and synaptophysin are the principal ones used in diagnostic pathology. He has added to the armamentarium different keratins in the differential diagnosis and particularly the low molecular weight keratins such as CAM 5.2 to avoid false negative results in the workup of some neuroendocrine tumors. For low levels of expression he emphasizes the need for in situ hybridization to establish the presence of the message if not the protein.

The chromogranin/secreotogranin (Cg/Sg) family is composed of several acidic proteins present in secretory granules of neuroendocrine cells. The three major Cg/Sg proteins are currently designated as chromogranin (Cg) A and B and secretogranin 11 (Sg11). Others include Sg111, Sg1V and Sg V. the distribution of CgA has now been studies extensively. It is present in most neuroendocrine cells and neoplasms. A few neoplasms with only a few endocrine secretory granules such as the small cell carcinoma of the lung and Merkel cell carcinomas do not react strongly with CgA. The widespread distribution and high degree of specificity of Cg/Sg make these excellent markers for endocrine cells and their neoplasms. Cg A is endocrine specific but has limited sensitivity; for example for hindgut carcinoids it is only positive for 25-50%, and adding CgB to it or a cocktail will increase the sensitivity for these tumors.

Synaptophysin, a 38kDa protein molecule is a component of the membrane presynaptic vesicles. It is widely distributed in neurons, neuroendocrine cells and their neoplasms and is a good broad spectrum neuroendocrine marker. Synaptophysin can be examined in formalin fixed tissues, which allows tumors to be revisited if initially not thought to be neuroendocrine. Although present in synaptic vesicles in tumors, it is found diffusely in the cytoplasm of the cell. It has however been found in adrenal cortical adenomas and carcinomas so although sensitive is not very specific. It therefore should always be used in conjunction with CgA. Synaptophysin belongs to a family of synaptic proteins that include synaptogamin (p65), SNAP-25, SNAP receptor (SNARE), Syntaxin and Rab3A. However the utility of these proteins in routine diagnostic pathology has not been established to date.

The proconvertases (PC) are enzymes that process propeptides into active peptides within cells. Some of these including PC1/PC3 and PC2 are highly specific for neuroendocrine cells and tumors and can be used as specific markers. Other such as PC4 is present in the testis whereas PC5/6 is more prevalent in the gastrointestinal tract.

Neuron Specific Enolase (NSE) is a very sensitive, but not very specific marker for neuroendocrine cells and tumors. It is commonly found in nerves, and neuroendocrine cells, but some non- neuroendocrine cells react with antisera to NSE. Therefore NSE should only be used as a broad spectrum marker in the diagnosis of these tumors.

Gastrin releasing peptide or Bombesin, which is a tetrapeptide originally isolated from amphibian skin, is present in many endocrine cells as well as neurons. GRP the mammalian analog of Bombesin is found in many lung tumors and gastrointestinal endocrine tumors and can also be used as a broad-spectrum marker.

PGP-9.5 is a soluble protein that was originally isolated from brain. It has now been shown to be a general marker for neuronal and neuroendocrine tissues. Interestingly about 50% of melanomas stain for PGP9.5 whereas these are negative for Cg/Sg.

Peptidylglycine a Amidating Monooxygene (PAM). Amidation is an important step in the maturation of neuropeptides. PAM catalyzes the post translational modification of many neuropeptides. The PAM proteins are usually released along with other peptides during exocytosis whereas membrane bound PAM remains in association with the cell. PAM expression is found in all neuroendocrine cell types (107), (108), (109;110) (111). Scopsi found a close correlation between PAM expression and at least one of the three principal grunion proteins (CgA, CgB or Sg11). It is not clear that this protein provides additional information (111).

Pancreatic endocrine tumors are usually positive for cytokeratins in more than 90% of cases (112). The low molecular weight keratins such as CAM5.2 are more sensitive for neuroendocrine tumors than the keratin cocktails such as AE1/AE3 and thus should be added to any regime that routinely examines pancreatic neoplasms to preclude misdiagnosis of an adenocarcinoma a for a neuroendocrine tumor. The latter have a very much more benign prognosis.

Pulmonary tumors include typical and atypical carcinoid and large and small-cell carcinomas. About 85% are reactive with cytokeratins, and the chromogranin are usually positive in carcinoids, atypical carcinoids and large-cell neuroendocrine carcinomas, while only about 50% of small cell carcinomas are positive. However other broad spectrum markers of neuroendocrine are positive in these tumors (113). While this may be helpful in delineating endocrine function of these tumors it is not clear that this alters treatment strategies or dictates a different prognostic value.

Anatomic Distribution

More than 50% of neuroendocrine tumors in clinical practice are of the so-called carcinoid variety and are found incidentally at operation, after metastasis has occurred, in the small intestine (especially the appendix). The remaining fraction comprises approximately 50% gastrinomas, 30% insulinomas, 13% VIPomas, 5 to 10% glucagonomas, and, rarely, less than 5% neurotensinomas, somatostatinomas, and ectopic hormone-secreting tumors. Nonsecretory tumors were thought to make up the bulk of pancreatic tumors. However with better immunohistochemical stains for endocrine cells, especially for neuron-specific enclose (NSE), chromogranin, synaptophysin, and receptors for somatostatin (114), there is increasing recognition that tumors masquerading as carcinomas of liver, small cell carcinoma of the lung, and the like are endocrine tumors. Most of these nonsecretory tumors actually store and secrete pancreatic polypeptide (PP), but because it has so little, if any, in the way of biologic activity, the tumor often remains silent until it is quite large.

Approximately 60% of pancreatic gastrinomas are concentrated in Pasarro’s Triangle, an area subtended by the head of pancreas, gastric antrum, and first portion of the duodenum. Other neuroendocrine tumors may be distributed more evenly across the pancreas or in ectopic sites such as the adrenal medulla, whereas carcinoid tumors most frequently occur in the appendix and small intestine.

The tumors are proliferative in nature and may take the form of hyperplasia or neoplasia (adenoma, adenomatous hyperplasia, microadenomatosis, nesidioblastosis, or carcinoma). Hyperplasia is relatively uncommon in benign sporadic tumors, but it is the rule in MEN-1 syndrome and often is present in the area of the pancreas surrounding a benign tumor.

The tumors may be further subdivided into (a) orthoendocrine, when they secrete the normal product of the cell type (e.g., a-cell glucagon), and (b) paraendocrine, when they secrete a peptide or amine that is foreign to the organ or cell of origin. Paraendocrine tumors are found in the adrenal medulla, kidney, lymph nodes, or liver and as a part of MEN-1 when a variety of peptides or amines are secreted. When tumors metastasize, they do so to local lymph nodes, liver, peritoneum, and, rarely, to bone, but this seems to be increasing in frequency as the natural history of these tumors changes with aggressive treatment. Metastases are notoriously highly vascular, which is a telltale sign of a GEP tumor. The occurrence of MEN-1 syndrome may be as frequent as one-third of the cases of GEP tumors, depending on the endemic area. In high-risk areas, measurements of ionized calcium, prolactin, and PP are important. Nonetheless these tumors are rare and slow growing. As Moertel once said, the study of neuroendocrine tumors of the gut is like an Odyssey in the land of slow growing tumors. Their characteristics are shown in Table 1.

Table 2 lists the common clinical syndrome, the tumor types and the hormones or peptides/amines that are produced. The sections that follow focus on the specific syndromes that are ascribed to GEP hyperfunction.

Figure 3

The great majority of these tumors are carcinoid tumors accounting for more than half those presenting each year ( Figure 3). Insulinomas, gastrinomas and PPomas account for 17, 15 and 9, % while the remainder are around the 1% mark. These tumors are what is known in common parlance as “Zebras” because of their rarity, but physicians are fascinated by their complexity and the unusual nature of their presentations. For the most part the endocrinologist makes his living not by diagnosing one of these and treating it, but by excluding conditions that masquerade as a neuroendocrine tumor. For this reason it is probably more appropriate to consider the clinical presentations rather than the tumor types.

By far the most frequent clinical manifestations found in practiceare flushing and diarrhea (Table 3 ), which are the cardinal presentations of the most common tumor syndrome, carcinoid, and this will therefore be discussed first.

Figure 4