The diversity and commonalities of ... - Springer Link

Langenbecks Arch Surg (2011) 396:273–298 DOI 10.1007/s00423-011-0739-1

REVIEW ARTICLE

The diversity and commonalities of gastroenteropancreatic neuroendocrine tumors Simon Schimmack & Bernhard Svejda & Benjamin Lawrence & Mark Kidd & Irvin M. Modlin

Received: 3 January 2011 / Accepted: 7 January 2011 / Published online: 28 January 2011 # Springer-Verlag 2011

Abstract Background Recent data demonstrate that the incidence of gastroenteropancreatic neuroendocrine tumors (GEP-NETs) has increased exponentially (overall ~500%) over the last three decades, thus refuting the erroneous concept of rarity. GEP-NETs comprise 2% of all malignancies and in terms of prevalence, are the second commonest gastrointestinal malignancy after colorectal cancer. Diagnosis is usually late since there is no biochemical screening test and symptoms are protean and overlooked. As a consequence, 60–80% exhibit metastases with a consequent suboptimal outcome. Discussion The gastrointestinal tract and pancreas exhibit ~17 different neuroendocrine cell types, but neither the cell of origin nor the biological basis of GEP-NETs is understood. This review examines GEP-NETs from the cellular and molecular perspective and addresses the distinct patterns of functional tumor biology pertinent to clinicians. Although grouped as a neoplastic entity (NETs), each lesion is derived from distinct cell precursors, produces specific bioactive products, exhibits distinct chromosomal abnormalities and somatic mutation events and has uniquely dissimilar clinical presentations. GEP-NETs demonstrate very different survival rates reflecting the intrinsic differences in malignant potential and variations in proliferative regulation. Apart from the Supported by NIH: DK080871 S. Schimmack : B. Svejda : B. Lawrence : M. Kidd : I. M. Modlin (*) Gastrointestinal Pathobiology Research Group, Department of Gastroenterological Surgery, Yale University School of Medicine, PO Box 208602, New Haven, CT, USA e-mail: [email protected] S. Schimmack Visceral- and Transplantation-Surgery of Heidelberg, University Hospital of General-, Heidelberg, Germany

identification of the inhibitory role of the somatostatin receptors, there is limited biological knowledge of the key regulators of proliferation and hence a paucity of successful targeted therapeutic agents. IGF-I, TGFβ and a variety of tyrosine kinases have been postulated as key regulatory elements; rigorous data is still required to define predictably effective and rational therapeutic strategy in an individual tumor. A critical issue in the clinical management of GEPNETs is the need to appreciate both the neuroendocrine commonalities of the disease as well as the unique characteristics of each tumor. The further acquisition of a detailed biological and molecular appreciation of GEP-NETs is vital to the development of effective management strategy. Keywords Chromogranin . Enterochromaffin . GEP-NET . Neuroendocrine . Serotonin . Somatostatin

Introduction Although initially considered rare tumors, recent data indicate that the incidence of gastroenteropancreatic neuroendocrine tumors (GEP-NETs) (Fig. 1) has increased exponentially over the last three decades, and they are as common as myeloma, testicular cancer, and Hodgkin's lymphoma [1]. As such, they comprise 2% of all malignancies and in terms of prevalence, GEP-NETs represent the second commonest gastrointestinal malignancy after colorectal cancer [1]. The increase in incidence and prevalence most likely reflects improvement in disease awareness and diagnostic techniques [1]. GEP-NETs present a considerable diagnostic and therapeutic challenge since their clinical presentation is nonspecific. Diagnosis is usually therefore late in the natural history of the disease with metastases evident at presentation in 60–80% [2]. Most GEP-NETs are sporadic lesions,

274

Langenbecks Arch Surg (2011) 396:273–298

Fig. 1 Neuroendocrine cell types and tumors. Individual neuroendocrine cells (circle) and their associated tumors (rectangles) represent the most commonly clinically encountered gastroenteropancreatic neuroendocrine tumors (GEP-NETs). Chromogranin A (CgA) positive staining (central immunohistochemical image) in a NET (red = Cy-5

labeled CgA, blue = DAPI (nuclear stain)) is the common denominator histopathology biomarker for NETs. The majority (>90%) of tumors express CgA but poorly differentiated lesions (NEC) may lose their neuroendocrine phenotype and be CgA-negative

although some, especially pancreatic neuroendocrine tumors (pNETs), may occur as part of familial tumor syndromes such as multiple endocrine neoplasia type 1 (MEN1 syndrome), von Hippel-Lindau disease (VHL), neurofibromatosis type 1 (NF-1), and tuberous sclerosis (TSC) [3]. The origin of the cells from which GEP-NETs arise is not well understood. Overall, the gastrointestinal tract including the pancreas has at least 17 different neuroendocrine cell types (Fig. 2). The term “neuroendocrine” is a composite description of a cell type that exhibits mixed morphological and physiological attributes of both the neural and endocrine regulatory systems. The bicameral cell embraces the phenotypic relationship of tumor precursor cells to neural cells in the expression of certain proteins, such as synaptophysin, neuron-specific enolase, and chromogranin A (CgA) and the physiological secretory/regulatory role classically ascribed to endocrine cells [4]. Individual GEP-NETs usually originate from a neuroendocrine cell that is specific to a particular part of the gut or pancreas. In some circumstances, neuroendocrine cells may be part of complex lesions that have adenocarcinomatous elements and the precise lineage of such cells is unclear. This review addresses differences and distinct patterns of each GEP-NET, and focuses on the cell and organ of origin as well as the functionality of the tumor.

ganglia, and the adrenal medulla) and diffusely distributed dispersed cells that constitute a disseminated system (diffuse neuroendocrine system (DNES)) and comprise at least 17 different cells. These cells, either individually or in aggregations, populate the skin, thyroid, lung, thymus, pancreas or gastrointestinal tract (Table 1), biliary tract, and urogenital tract; and are the largest group of hormone-producing cells in the body. In some organs (e.g., stomach, pancreas) the presence of a distinct neuroendocrine system provides a functional duality whereby an endocrine and an exocrine function are biologically assimilated. Thus, the acid pepsin digestive function of the oxyntic part of the stomach is intrinsically dependent upon the endocrine role of the gastrin-secreting antrum. Similarly, the digestive role of the pancreas is enmeshed with the glucose homeostatic role of the pancreatic islets embedded within the exocrine parenchyma of the pancreas. Presumably, such specialized regulatory cell collections represent biology in the process of transformation much as diffuse collections of sympathetic neurons evolved into para-renal endocrine organs (adrenals). There has been considerable and prolonged controversy and debate in respect of the developmental origin of gut neuroendocrine cells [5–7]. In the past, it was considered that the neuroendocrine cell system was based upon migration from the primitive neural crest to specific anatomical sites [8]. Currently the most accepted proposal is the “Unitarian Theory” of intestinal cytogenesis, which opines that gastrointestinal cell lineages are derived from a common stem-cell precursor, located in the base of intestinal crypts or in the neck region of gastric glands (Fig. 3) [9]. Recent studies of gastric and intestinal epithelia have identified that neuroendocrine cells are among the progeny of

Neuroendocrine cell phenotypes: development and embryology Broadly speaking, the neuroendocrine cell system can be divided into two systems, namely aggregations of cells that constitute glands (the pituitary, the parathyroids, the para-


275

Fig. 2 Gut neuroendocrine cell morphology. Electron micrograph of an isolated EC cell demonstrates the typical admixture of electron dense and electron-luscent granules (inset) that characterize neuroendocrine secretory cells (top left). Neuroendocrine cells are largely distributed at the base of the gland (bottom left—yellow arrows: immunofluorescent CgA stain (FITC green), nuclei are blue (DAPI)). A higher magnification of the mucosa (demonstrates EC cells predominantly located at the gland periphery (brown DAB staining of CgA; gland cross section—top right). DAB staining of isolated, fixed EC cells exhibits long, axonal-like structures (bottom right) that “synapse” with either adjacent mucosal cells or neurons within the gut mucosa providing the neural phenotype component

such multipotential stem cell [10–13]. It therefore seems likely that gastrointestinal neuroendocrine cells are derived from local tissue-specific stem cells, probably through a committed precursor cell. In the pancreas, it is thought that the common precursor cell resides within the ductal epithelium and that this structure provides the basis for the genesis of the pancreatic islets [14]. A second hypothesis suggests that islet neogenesis or development of an islet-precursor cell may occur from an already differentiated pancreatic cell (i.e., transdifferentiation of an acinar cell) [15]. Within the gastrointestinal tract and pancreas, at least 17 individual neuroendocrine cell types have been identified as derived from local multipotent gastrointestinal stem cells. The precise mechanism of differentiation of cells of the DNES is still poorly understood although individual transcription factors including protein atonal homolog 1 (PATCH1), neurogenin-3 (NGN3) and neuroD have been identified as regulatory components responsible for lineage transformation. An example of this cell specification role is the effect of loss-of-function mutations in NGN3 evident in individuals with congenital malabsorptive diarrhea. This is associated with failure to promote neuroD transcription, resulting in a specific loss of intestinal enteroendocrine cell populations [16]. In general, neuroendocrine cells are terminally differentiated and considered non-proliferating as demonstrated by the absence of

proliferation makers for example Ki67 in CgA-expressing cells [17]. An alternative proliferative mechanism, however, probably exists since neuroendocrine cells are able to adapt to pathological and physiological stimuli within their environment [18]. Evidence for this phenomenon is provided in the instance of gastric enterochromaffin-like (ECL) cells [19]. The ECL cells, located in the oxyntic mucosa, interact with antral G-cells, which secrete gastrin and activate ECL-cell histamine production which, in turn, drives acid secretion from parietal cells. Loss of parietal cells (e.g., in atrophic gastritis) or acid suppression results in increased gastric pH, increased gastrin secretion, and culminates in increased ECL-cell hyperplasia and even neoplasia. If ECL cells are terminally differentiated, this suggests that the mechanism of proliferation likely resides within the ECL-cell stem cell or progenitor. Since a similar phenomenon occurs as a component of hypergastrinemiaassociated MEN1 syndrome, it is plausible that an intrinsic gastrin-activated genetic event may also be implicated [20].

Neuroendocrine cell phenotypes: secretory function A characteristic of neuroendocrine cells is the production of a variety of bioactive peptides and amines (Table 1). Secretory products are stored in large dense-core vesicles

276


Table 1 Gastroenteropancreatic neuroendocrine (GEP-NET) cell types: distribution and bioactive products Cell type

STOM Oxy

A B D EC ECL G Gr GIP

n

I L M N P/D1 PP S VIP X

+++ +++ +++

+++

+ +

STOM Ant

DUOD

PANC

+++ +++

+++ +++

+++ +++ +++ +

+++

+++ + +++

+

+ +

+++ + +++ + +++ n +++ +

JEJ

ILEUM

APPX

COLON

+++ +++

+ +++

+ +++

+ +++

+ +++

+ +

+

+

+++ + +++ +++ +++

+ + + +++ +++

RECTUM

+ ++

n

+, n +++ ++

+++ +

+

+

+++

+

+

+

Bioactive product Glugagon Insulin Somatostatin 5-HT Histamine Gastrin Ghrelin GIP/Xenin Cholecystokinin GLI/PYY Motilin Neurotensin Pancreatic Polypeptide Secretin/5-HT VIP Amylin

Although neuroendocrine cells are distributed throughout the gut, there is evidence of some spatial restriction, for example ECL cells—oxyntic gastric mucosa, G cells to the antrum, and duodenum. Neuroendocrine cells of the pancreas are, however, tightly spatially aggregated into islet architecture. The majority of gut neuroendocrine cells are scattered throughout the gastrointestinal tract, for example somatostatin (D) and serotonin (EC) cells. This suggests a more ubiquitous role for the products of these cells. Despite a similarity in distribution, EC cell-derived tumors are ~30× more common than somatostatinomas (yellow rows) STOM = stomach, APPX = appendix, DUOD = duodenum, JEJ = jejunum, PANC = pancreas, Ant = antrum, Oxy = oxyntic mucosa, n = neonatal and fetal period, + = few cells, ++ = cells present, +++, major site

(LDCV) and in small synaptic-like vesicles (SSV), and some proteins associated with these vesicles (e.g., CgA or synaptophysin) have been utilized as specific markers of NECs [21]. Peptide hormones for regulated secretion are packaged into secretory granules (LDCV) that bud from the trans-Golgi network where prohormones and proneuropeptides are stored and processed (Fig. 4). The size, shape, and electron density of the secretory granules have been used with a varying degree of efficacy to characterize individual NEC types. Different granules store individual peptide hormones; however, in some neuroendocrine cells, several different peptides or amines may co-localize in the same granule [22]. A key protein in the genesis of vesicles is CgA which, among other functions, regulates the biogenesis of dense-core secretory granules. Other granins (e.g., chromogranin B (CgB)) adjust proteolytic processing of peptide precursors and promote aggregation-mediated sorting into mature secretory granules, enabling granules to mature into regulatable exocytotic carriers. In certain neuroendocrine cell types, for example ECL cell—histamine, or enterochromomaffin (EC) cell— serotonin (5-HT), amines are co-packaged with chromogranins in secretory vesicles. This process is energy dependent,

driven by proton gradients and involves vesicular monoamine transporters (VMAT) [23]; ECL cells are identified by VMAT2 and EC cells by VMAT1 [24, 25]. Neuroendocrine cell secretion is regulated by a complex variety of G-protein-coupled receptors, ion-gated receptors, and receptors with tyrosine-kinase activity [1]. Secretagogue-evoked stimulation (via cAMP/PKA signaling, through MAPK or via ion channel-mediated depolarization [26]) induces actin re-organization through sequential ordering of carrier proteins at the interface between granules and the plasma membrane. This calcium-dependent step is a prerequisite for regulated exocytosis, and it allows granule membrane trafficking and release of neuroendocrine contents. Regulators include neural, for example α- or β-adrenergic, muscarinic (both stimulatory and inhibitory), and VPAC/PAC1 receptors; hormonal, for example gastrin/CCK2, histamine H1–4, or 5HT1–7 receptors and somatostatin (usually types 2 and 5) which are invariably inhibitory (Fig. 5) [26]. Exocytosis, the mechanistic process by which bioactive products are delivered from the cell into the adjacent milieu, comprises a series of sequential intracellular events. In general the exocytotic process involves three steps: (1)


277

Fig. 3 Neuroendocrine cell differentiation in the gastrointestinal tract. Basal crypt stem (totipotential and pluripotential) cells give rise to a variety of mucosal cell types: Math1 expression directs cells to the secretory lineage and NGN3 to the neuroendocrine lineage. Specific hormone transcription is regulated by several transcription factors such as Pax4, Pax6, and BETA2 (from [9], with permission)

the transport of dense core vesicles to the plasma membrane (recruitment step), (2) their initial interaction with the plasma membrane (docking step), and (3) their subsequent fusion with the plasma membrane (fusion step). Molecular mechanisms leading to regulated exocytosis have been

Fig. 4 Calcium-dependent exocytosis. CgA is a key protein in the genesis of vesicles and regulates the biogenesis of dense-core secretory granules. Secretory products are stored in large dense-core vesicles (LDCV) and in small synaptic-like vesicles (SSV). Proteins associated with these vesicles (e.g., CgA or synaptophysin) have been

summarized in the SNARE (synaptosomal-associated protein receptor) hypothesis [27]. These include the following. 1. Docking of the vesicle to the plasma membrane. SNAREs are present on both the vesicle membrane

utilized as biomarkers of neuroendocrine cells [21]. Prohormones and proneuropeptides are stored and processed in the trans-Golgi network prior to packaging into secretory granules (LDCV) as bioactive peptides for regulated secretion (from [2], with permission)

278


Fig. 5 Regulation of serotonin release from enterochromomaffin (EC) cells. Tryptophan is transported into the cell from the apical luminal compartment and converted to serotonin (5-HT) by tryptophan hydroxylase. Serotonin accumulates in secretory vesicles which undergo exocytosis in response to cell activation. Positive regulators (green) include noradrenaline, dopamine, serotonin itself and pituitary adenylate cyclase-activating peptide (PACAP), which excite secretion

through receptor activation and either cAMP or calcium ([Ca2+]) signaling pathways. Inhibitors (red) of secretion include serotonin, dopamine, acetylcholine, glutamic acid, and somatostatin. Therapeutic activation of somatostatin receptors by somatostatin analogs have proved effective in the inhibition of excess serotonin secretion in “carcinoid” syndrome

(v-SNAREs) and on target membranes (t-SNAREs). The docking site is formed by the tight binding of one vSNARE, synaptobrevin or vesicle-associated membrane protein (VAMP), with two t-SNAREs, syntaxin, and synaptosome-associated protein (SNAP-25), resulting in a stable trimeric core complex (Fig. 6). 2. Priming of the exocytotic machinery. After docking, vesicles are not immediately competent to fuse with their target membrane. The trimeric core complex

serves as a binding site for N-ethylmaleimide-sensitive fusion protein (NSF) and thereafter NSF crosslinks multiple core complexes leading to hemifusion of vesicle and target membrane. 3. Triggering of exocytosis by calcium. Synaptotagmins are v-SNAREs that comprise a component of a clamping apparatus that prevents spontaneous fusion of vesicles with their target membrane. Synaptotagmins most likely function as calcium sensors and after a sub-plasmalemmal

Fig. 6 Mechanism of vesicle docking and exocytosis at the plasma membrane. SNAREs have been identified both on the vesicle membrane (v-SNAREs) and on target membranes (t-SNAREs). The docking site is formed by the tight binding of one v-SNARE, synaptobrevin or vesicle-associated membrane protein (VAMP), with two t-SNAREs, syntaxin and synaptosome-associated protein (SNAP-25), resulting in a stable trimeric core complex. At the completion of docking, vesicle content is released into the paracellular space


279

rise in intracellular calcium, interact with syntaxins (t-SNAREs) as well as membrane phospholipids and other yet-unidentified target proteins to allow. 4. Fusion of the vesicular membrane with the plasmalemma. This final step results in exocytosis [28].

Neuroendocrine tumors: etiology and pathogenesis NETs are generally thought to represent malignant transformations of either terminally differentiated neuroendocrine cells or a precursor/stem cell. The mechanism of these events is largely unknown. It is postulated that damage to early, neuroendocrine precursor cell types leads to the development of high grade or poorly differentiated neuroendocrine carcinomas (NECs). G1-NETs (previously called NETs) and G2-NETs (previously called WDNEC (well differentiated neuroendocrine carcinoma)) develop from later stage or partially differentiated cells (Fig. 7). The mechanisms for “damage” are not known but the sequelae are largely considered to be either epigenetic modifications, for example differences in histone acetylation or chromosomal methylation, or spontaneous mutations in critical genes, for example MEN1 [29]. Although an attractive pathological concept, there is little evidence to support the proposal of progression from a GEP-NET-G1 to NET-G2 and finally to a high-grade NEC (the so-called “NET–NEC sequence”) [30]. While the majority (>95%) of GEP-NETs are sporadic [1], a small percentage are either familial or associated with five independent autosomal dominant inherited syndromes. Evidence for the familial basis of gastrointestinal NETs has been assessed in large cancer databases, for example Swedish Family Cancer database [31] which indicated that the risk of developing NETs was significantly higher among individuals with a parental history of NETs (relative risk (RR): 4.33) and in individuals with a sibling history of NETs (RR 2.88). Parental NETs were strongly associated with the development of small intestinal (RR 11.80) and colon NETs (RR 2.78) in the offspring. Although this type of analysis cannot identify candidate genes, it indicates that there exists an individual predisposition to GEP-NET development. A separate defined group of NET genetic disorders includes MEN types 1 and 2, which are the most common forms, VHL disease, von Recklinghausen disease or neurofibromatosis (NF1), tuberous sclerosis (TSC), and Carney complex (CNC). MEN1 is an inherited disease classically constituted by parathyroid hyperplasia/adenoma, pancreatic endocrine tumors, and pituitary tumors. Variations of MENI include in addition adrenocortical secreting or nonfunctional

Fig. 7 Transformative events (putative) in the development of GEPNETs. NETs develop in inherited/familial tumors of the stomach (gastric type II) and pancreas (pNETs) as a consequence of either a second hit or LOH. Somatic mutations, the most common event, perhaps due to environmental damage at a committed neuroendocrine precursor stage, lead to well-differentiated NETs (NET-G1). If damage occurs early in stem cell progress (e.g., stem cell 1), poorly differentiated neuroendocrine carcinomas develop. If damage occurs at a later stage, for example to a pluripotent cell (stem cell 2), then a well-differentiated NEC (G2-NET) is the consequence. There is little evidence for evolution from a NET to a NEC in this schema [29]

tumors, thymic NETs, and bronchial NETs [3, 32]. The diversity of MEN1-related lesions and the divergent embryonic origins of affected tissues implicate the MEN1 gene as exhibiting a critical role in early embryogenesis. The MEN1 gene is located on the long arm of chromosome 11, band q13 [33], and comparative genomic analysis of tumoral and constitutional genotypes has identified evidence of somatic loss of heterozygosity (LOH). This is consistent with the likelihood that development of MEN1associated tumors is a two-step process, a germline mutation affecting the first MEN1 allele, and a second somatic inactivation of the unaffected allele (LOH). Tumorigenesis in MEN1 likely involves loss of function of the growth-suppressor gene MEN1 [34]. Menin is a 610amino acid nuclear protein encoded by the MEN1 gene and interacts with Jun D and the AP1 transcription factors to modify growth-regulatory signaling. It also interacts with a putative tumor metastasis suppressor nm23H1/nucleoside

280

diphosphate kinase (nm23), and exerts GTPase activity [35]. Truncation or instability of MEN1 gene products has been proposed to culminate in loss of transcriptional regulation and/or GTP hydrolysis, thereby providing a possible mechanism of MEN1-driven tumor formation [34]. Germline mutations of the RET proto-oncogene encoding a transmembrane tyrosine-kinase-receptor, confer predisposition to clinical variants of MEN2, which can be subdivided into 2A (Sipple’s syndrome), 2B (Gorlin’s syndrome), and Familial Medullary Thyroid Cancer [36, 37]. In MEN2A, medullary thyroid carcinoma is associated with pheochromocytoma (30–50%) and primary hyperparathyroidism (10– 20%). In MEN2B, the major clinical features are medullary thyroid carcinoma, pheochromocytoma, mucosal neuromas, and cranio-skeletal abnormalities sometimes associated with a marfanoid habitus. Additionally, angioneuromatosis of the gastrointestinal tract may also occur [38]. It is noteworthy that the majority of lesions associated with the MEN2/RET abnormality are NETs outside of the gastrointestinal system. VHL disease is an autosomal dominant syndrome whose cardinal features include a predisposition to renal cancers, retinal and/or cerebellar hemangioblastoma, pheochromocytoma, and cystic and/or pancreatic endocrine tumors [39]. Ten percent to fifteen per cent of patients with VHL develop pancreatic islet or ductal endocrine cell tumors [40] and more than 50% exhibit multiple tumors. The VHL gene is located on chromosome 3p35-26 [41], and its product interacts with the elongin family of proteins to regulate transcriptional elongation [42]. Other functions involving the VHL protein are hypoxia-induced cell regulation and extracellular matrix fibronectin expression and localization [43]. NF1- and TSC -related GEP-NETs include multiple tumors in the pancreas and/or duodenum with psammomatous glandular histological features and immunohistochemical expression of somatostatin and/or insulin [3]. The NF gene, located on chromosome 17q11.2, acts as a tumor suppressor. Mutated (non-functional) neurofibromin, the NF-1 gene product, results in a loss of normal function, downregulation of the P21ras signaling pathway; this loss leads to a constitutively activated GTP which results in abnormal cell proliferation [44]. TSC-determining loci have been mapped to chromosomes 9q34 (TSC1) and 16p13 (TSC2). The protein products of the tuberous sclerosis complex genes, hamartin (TSC1), and tuberin (TSC2), have important cellular regulatory functions. These include a role in cell signaling in growth and translation regulation via the PI3K/AKT pathway, in cell adhesion via the glycogen synthase kinase 3 pathway, and in proliferation via the mitogen-activated protein kinase (MAPK) pathway [45]. The CNC, described in 1985 [46], is an autosomal dominant disease comprising skin pigmentation, myxomas,


melanotic Schwannomas and endocrine tumors of the adrenal glands, Sertoli cells, somatotrophs, thyroid, and ovary [47]. The CNC gene, located on chromosome 17q22q23, encodes PRKARIA, the protein kinase A (PKA) regulatory subunit 1α (R1α), and is a tumor suppressor gene. The role of this gene in GEP-NETs is unclear.

Neuroendocrine tumors: “functional” versus “non-functional” Some NETs are associated with specific symptomatology consequent upon the release of bioactive peptides and amines, for example insulin and 5-HT into the systemic circulation. Such tumors have in the past been designated as “functioning” tumors and recognized as the cause of a variety of syndromes, for example hypoglycemia related to insulinoma (insulin), peptic ulceration and gastrinoma (gastrin), and the diarrhea, abdominal pain, sweating, flushing, bronchospasm, tachycardia, and fibrotic heart disease of "carcinoid syndrome" (5-HT). Other functioning tumor syndromes reflect pancreatic primary lesions associated with either excessive secretion of glucagon (glucagonomas) or vasoactive intestinal polypeptide (VIPomas). In contrast, many NETs (~50%) are not associated with a clinically defined “hypersecretory” symptom complex (syndrome) and were previously termed “non-functioning”. The wide variation previously reported (~10–85%) reflects the limitations in current reporting systems and the difference between older series and more recent ones where more sophisticated biological tools are available. This clinical distinction has artificially led to the conclusion that there are two separate types of NETs. In general, however, NETs are indistinguishable at pathological, immunohistochemical, and transcriptomic level, while therapeutically they appear to respond in an almost indistinguishable fashion. There currently appears to be little scientific information to support the concept that functional NETs are in any biological fashion different to non-functional NETs.

Gastric nets—predominantly ECL-cell tumors The stomach contains at least five types of endocrine cells which collectively comprise ~2% of the cells in the gastric mucosa [48]. Each endocrine cell secretes a “dominant” chemical messenger: ECL cells secrete histamine, G cells secrete gastrin, while EC, D and P cells secrete 5-HT, somatostatin, and ghrelin, respectively [49]. The histaminesecreting ECL cells are the most common gastric neuroendocrine cell type, constitute up to 80% of oxyntic mucosal neuroendocrine cells, and constitute the predominant neuroendocrine tumor type in the stomach. Gastric (ECL


281

cell) NETs are classified into three subgroups based upon their responsiveness to gastrin: Type I and type II tumors occur in the setting of hypergastrinemia (Table 2). Type I lesions represent a physiological response to low acid states such as chronic atrophic gastritis (CAG) and pernicious anemia. Type II tumors are driven by autonomous gastrin secretion from a gastrinoma, almost always in the setting of MEN1. These ECL tumors are similar to type I in respect of their gastrin-responsiveness but the MEN1 genetic abnormality renders them more susceptible to malignant transformation. Type III tumors occur in the absence of hypergastrinemia, are not always of ECL origin, and include a more malignant subtype described as “atypical”. These three types of tumors are usually considered distinct from a poorly differentiated subtype (type IV in some classifications), which had been previously been regarded as anaplastic, high-grade NET, or small cell carcinoma of the stomach [50]. Etiology/pathogenesis Gastrin is the most important growth factor in type I and II gastric NETs. In type I tumors, a loss of parietal cells is associated with a diminution of acid production and an elevation in luminal pH with consequent loss of negative Table 2 Gastroenteropancreatic neuroendocrine (GEP-NET) tumor type, distribution and 5-year survival Cell Type

Tumor

Incidence (%)

Five-year survival (%)

EC

30

68

L β ECL α G δ or D I PP

Intestinal NET (Carcinoid) NET Insulinoma Gastric NET Glugagonoma Gastrinoma Somatostatinoma CCKoma Ppoma

16 7 6 3 2.5 1