Recent advances in molecular biologic techniques have brought a new approach to the field of investigative and diagnostic pathology. It has become possible to gain insight into the cellular and subcellular defects that form the basis for a variety of diseases. It has also become apparent that both heritable and sporadic neoplasms are genetic disorders with discrete cellular defects. In general, a single genetic defect or mutation is not sufficient to initiate tumorigenesis but rather these contribute in a cumulative fashion to the transformation and/or proliferative process. For each cell type, a unique combination of genetic hits results in the pathologic phenotype. In some instances the mutation is oncogenic, i.e it involves a gene whose gain of function is associated with dysregulated synthesis of a normal product; eg. ras and myc. In other situations, it is characterized by loss of tumor suppressor gene (TSG) activity , i.e. it involves a gene whose product normally restrains cell proliferation eg. Rb, p53. Usually it is the combination of gain of oncogenic function coupled with loss of tumor suppressor action which permits autonomous cell proliferation. Other factors including angiogenesis, telomerase activity, gain of survival signals, and metastatic potential all contribute to the ultimate pathologic phenotype that we recognize as carcinoma. While many of operative mechanisms responsible for human neoplasia probably apply to neuroendocrine tumorigenesis, considerably less is known about the molecular events resulting in these lesions.

Morphologic Studies
The morphologic analysis of endocrine pancreatic lesions requires the combination of histology and immunohistochemistry at the minimum^(1-3). It is important to characterize pancreatic tumors as endocrine in nature and to distinguish neoplasms from pseudoneoplastic conditions. Endocrine tumors generally are well delineated but unencapsulated lesions that have a characteristic histopathology. They are composed of small nests, trabecula or sheets of epithelial cells in a highly vascular stroma. They occasionally form gland-like structures. The stroma may, in some instances, form amyloid. The tumor cells usually have poorly-defined cell borders, abundant cytoplasm that may contain eosinophilic, amphophilic or basophilic granules. Characteristically the nuclei of tumor cells are bland; nuclear pleomorphism that defines malignancy in other epithelial tumors is not a reliable indicator of aggressive behavior. Poorly differentiated carcinomas have less cytoplasm, lack granularity and have larger hyperchromatic nuclei. The differential diagnoses include non-endocrine tumors, such as solid and cystic neoplasms, and non-neoplastic proliferative lesions of pancreatic endocrine cells: nesiodioblastosis and pseudoneoplastic proliferation of endocrine cells in chronic pancreatitis.

Determination of the endocrine nature of pancreatic tumors relies on immunolocalization of common markers of neuroendocrine differentiation. They stain for synaptophysin, a 38 Kd molecular weight molecule that is associated with synaptic vesicles of neurons and neuroendocrine cells. Most contain chromogranins, proteins associated with secretory granules. There are two families of chromogranins, A and B; to appropriately classify these lesions one needs to identify both chromogranins. In the pancreas, the most numerous cells are B cells that produce insulin and these are negative for chromogranin A, the only chromogranin recognized by most widely used antibodies applied for immunohistochemistry. Moreover, chromogranin immunoreactivity is directly related to the number of secretory granules that may be scarce in poorly differentiated tumors. Other markers of neuroendocrine differentiation include CD57 (Leu 7), neural cell adhesion molecule (NCAM; CD56), neuron specific enolase (NSE), and Protein Gene Product 9.5 (PGP 9.5) that stain variable subpopulations of endocrine lesions and some, such as NCAM and NSE stain some non-endocrine tumors as well.

The structure-function correlations of these lesions are best defined by immunoreactivity for peptide hormones. The hormones expressed by normal pancreatic endocrine elements show a highly characteristic pattern of expression. Pancreatic A cells express glucagon and have a normal distribution around the periphery of the bands of the normal islet. Insulin is localized to B cells that comprise the majority of cells that fill the bands of the normal islets. Somatostatin is produced by D cells that are scattered in the islets between the A and B cells. Pancreatic polypeptide is localized to PP cells that are randomly found within islets but are more numerous in the head of pancreas and specifically in the posterior portion. This distribution of hormones is lost in abnormal islets and in tumors. Other hormones may be expressed, reflecting either fetal regression (e.g. gastrin) or true ectopia of hormone expression (e.g. expression of GRH, ACTH). The wide spectrum of clinical symptomatology associated with these lesions is attributed to the numerous peptide hormones that can be elaborated by the cells of the diffuse endocrine system.

The ultrastructure of these lesions is usually highly characteristic. Tumor cells have well developed rough endoplasmic reticulum, prominent Golgi complexes and membrane-bound secretory granules that store hormones for secretion in response to stimulation. The development of these organelles varies with cell differentiation and hormonal activity; the numbers of secretory granules reflect the balance between synthesis, storage and secretion. The morphology of secretory granules is reflective of cell type and hormone content.

The current approach to the classification of these lesions ⁽⁴⁾ is based on three principles: (i) the diagnosis should be based on the microscopic features of the lesion but must incorporate useful immunohistochemical data, (ii) tumors should be distinguished according to the differentiated cell type and site of origin, and (iii) tumors should be subdivided based on biological behaviour, into benign, low grade malignant and high grade malignant lesions; the latter include poorly differentiated endocrine carcinoma and the so-called "small cell" or "oat cell" carcinomas. The determination of malignant potential is based on architectural and cytologic features including invasion. The location and hormone content of individual lesions is of importance, especially in the classification of tumors of histological low grade malignancy, the "well differentiated endocrine carcinoma" where these additional pieces of information aid in predicting the likelihood of metastatic behavior.

Ploidy and Karyotype Studies of Neuroendocrine Tumors
Although DNA image and flow cytometry have shown a relationship between a high DNA index and a poor prognostic outcome, examination of a limited number of endocrine pancreatic tumors has failed to show any karyotypic abnormalities in these tumors ⁽⁵⁾.

Oncogenes and Neuroendocrine Tumorigenesis
Proto-oncogenes are normal cellular genes that play an essential role in the proliferation and differentiation of normal cells. They function at each step of signal transduction pathways as growth factors (eg. c-sis), membrane receptors (eg. c-erb-B, c-neu, c-fms), GTP binding proteins (eg. ras family) and nuclear proteins (eg. c-myc, c-fos). Proto-oncogenes may be activated by point mutations, translocations or by increased expression. Genetic alteration in these genes leads to sustained activation of the gene product in the absence of the normal control mechanisms. Activated oncogenes have been associated with a large number of human tumors, for example N-myc amplification in neuroblastomas and K-ras point mutations in colonic carcinomas.

Ras proteins are involved in transducing signals from the cell surface to a number of ligand-receptor complexes. The commonest mutational sites alter the GTP-binding domain (codons 12/13) or more rarely the GTPase domain (codon 61). Point mutations of all 3 ras genes (H-, K- and N-) are thought to be common in endocrine tumors such as thyroid tumors ⁽⁶⁾ but not neuroendocrine tumors.

G-proteins are heterotrimeric membrane-anchored peptides that play a central role in transducing signals from the cell surface ligand-receptor complexes to the downstream effectors. The a -subunit dissociates from the ß- and gamma-subunits of Gs when GTP displaces its bound GDP, stimulates adenylyl cylase to produce cyclic AMP from ATP. Cyclic AMP (cAMP) in turn activates c-AMP-dependent protein kinases, increases intracellular calcium transport, and may potentiate the effect of activated inositol phospholipid-dependent protein kinases. The weak intrinsic GTPase activity of Gsa and the action of GTPase activating peptides (GAP) dissociates GTP from Gs a and terminates the response. Additionally, the multiple structural and functional isoforms of adenylyl cyclase underscore the complexity of this redundant system of signal transduction coupling and provides some insight into the array of potential sites of somatic mutations which could alter both cell division and hormone production. Indeed one of the earlier and most exciting molecular defects to be described in endocrine tumors involved the single point mutations in two critical domains of the Gs a subunit of GTPase codon 201 where Arg is switched to a Cys or codon 227 where Gln is replaced with Arg ⁽⁷⁾. Substitutions at these codons (the gsp mutation) activate adenylyl cyclase by inhibiting the hydrolysis of GTP and thereby maintaining Gs a in a constitutively activated state. Marked elevation in mRNA levels of the Gs a gene have been documented in insulin-secreting pancreatic endocrine tumors⁽⁷⁾.

Over-expression of genes which act as inhibitors of the apoptotic process may also act as oncogenes. In some studies, the oncogenes c-myc, bcl-2, c-erb B-2, and c-jun have been shown to be frequently expressed in human gastroenteropancreatic tumors. The expression of these oncogenes may represent pathogenic events in the generation, malignant transformation, and progression of gastroenteropancreatic endocrine tumors.

Tumor Suppressor Genes in Endocrine Tumorigenesis
Products of tumor suppressor genes (TSG) act as sequestering agents for transcription factors, thereby, modulating physiologic growth by arresting cell division in the G₁ phase. This delay may allow for repair of genomic damage or may trigger apoptotic cell death. Deletion or reduced expression of TSGs appear to be a commonly shared mechanism in human tumorigenesis. Multiple Endocrine Neoplasia (MEN) type 1 is an autosomal dominant disorder characterized by multiple tumors of endocrine tissues, including pancreas. In patients with MEN-1, loss of heterozygosity in tumors was mapped to 11q13. The MEN-1 tumor suppressor gene was cloned at this site; patients in kindreds with MEN-1 have been shown to have inherited inactivating mutations of one allele with loss of heterozygosity (LOH) of the normal allele in tumors. MEN-1 gene mutations have also been found in a significant number of sporadic tumors of the dispersed endocrine system, primarily gastroenteropancreatic endocrine tumors ^(8-10) including 44% of sporadic gastrinomas and 19% of insulinomas ⁽¹¹⁾. This region of 11q13 contains several other genes that are also known to be associated with tumorigenesis and may be implicated in the development of tumors that show a high frequency of LOH in this region.

Von Hippel-Lindau disease is an autosomal dominant disorder characterized by the development of tumors in multiple organs. The pancreas is affected in up to 75% of patients, but the majority have benign cysts or microcystic adenomas. The endocrine pancreas is affected in 12-17% of patients who develop multiple endocrine tumors ⁽¹²⁾. The VHL gene on chromosome 3p encodes a 213 amino acid protein that is a TSG. Patients with VHL inherit one mutant gene and their pancreatic tumors show loss of the normal allele, proving the role of the VHL gene in their development ⁽¹²⁾. Although sporadic pancreatic endocrine tumors show significant LOH of chromosome 3p ⁽¹³⁾, somatic mutations of the VHL gene have not been identified.

Patients with von Recklinghausen's disease (neurofibromatosis type 1, NF-1) develop somatostatin-producing endocrine tumors of the duodenum but rarely have pancreatic endocrine lesions. The reason for this distribution is not clear ⁽¹⁴⁾. There is little known about the biology of NF1 in sporadic pancreatic tumorigenesis.

The p16^ink4a/CDK2A gene has been implicated in nonendocrine tumorigenesis where this TSG is frequently altered by deletion, mutation or methylation. Mutation or promoter methylation of the p16 gene has been reported in a significant proportion of sporadic pancreatic endocrine tumors, predominantly gastrinomas ^(15,16).

The p53 protein plays a role in cell cycle regulation; mutations, deletions, or rearrangements of the p53 gene are among the commonest genetic mutations in human neoplasms. Although p53 mutations may play a role in the pathogenesis of some endocrine tumors of the appendix, they have not been reported in pancreatic endocrine neoplasia.

Case reports implicate deletions and possible rearrangements of the Rb gene in insulin-producing carcinomas ⁽¹⁷⁾, deletions of the Wilms' tumor and transformation suppressor gene krev-1 in an insulinoma ⁽¹⁸⁾. The APC gene on chromosome 5q21 has been implicated by association with sporadic pancreatic endocrine tumors ⁽¹⁹⁾ but there is no evidence for a casual relationship. A putative tumor suppressor gene on chromosome 1 may predict prognosis in pancreatic endocrine carcinomas ^(20,21). Other significant genomic mutations in human pancreatic endocrine tumorigenesis include chromosome 10 or 17 monosomy⁽²²⁾, and loss of the adhesion molecule DCC (deleted in colonic carcinoma) ⁽²³⁾.

References

1. Asa SL. The endocrine pancreas and its tumors. Endocr Pathol 1993;4:120-30.
2. Apel RL, Asa SL. Endocrine tumors of the pancreas. Pathology Annual 1995;1:305-49.
3. Lubensky IA. Endocrine Pancreas. In: LiVolsi VA, Asa SL, editors. Endocrine Pathology. Philadelphia, PA: Churchill Livingstone; 2002. p.205-35.
4. Solcia E, Kloppel G, Sobin LH, et al. Histological typing of endocrine tumours. 2 ed. Berlin: Springer; 2000.
5. Bardi G, Hohansson B, Pandis N, et al. Karyotypic abnormalities in tumours of the pancreas. Br J Cancer 1993;67:1106-12.
6. Suarez HG, du Villard JA, Severino M, et al. Presence of mutations in all three ras genes in human thyroid tumors. Oncogene 1990;5:565-70.
7. Lyons J, Landis CA, Harsh G, et al. Two G protein oncogenes in human endocrine tumors. Science 1990;249:655-9.
8. Shan L, Nakamura M, Yokoi T, et al. Somatic mutations of multiple endocrine neoplasia type 1 gene in the sporadic endocrine tumors. Lab Invest 1998;78:471-5.
9. Toliat M, Berger W, Ropers HH, Neuhaus P, Wiedenmann B. Mutations in the MEN1 gene in sporadic neuroendocrine tumours of gastroenteropancreatic system. Lancet 1997;350:1223.
10. Wang EH, Ebrahimi SA, Wu AY, et al. Mutation of the MENIN gene in sporadic pancreatic endocrine tumors. Cancer Res 1998;58:4417-20.
11. Debelenko LV, Zhuang Z, Emmert-Buck MR, et al. Allelic deletions on chromosome 11q13 in multiple endocrine neoplasia type 1- associated and sporadic gastrinomas and pancreatic endocrine tumors. Cancer Res 1997;57:2238-43.
12. Lubensky IA, Pack S, Ault D, et al. Multiple neuroendocrine tumors of the pancreas in von Hippel-Lindau disease patients: histopathological and molecular genetic analysis. Am J Pathol 1998;153:223-31.
13. Chung DC, Smith AP, Louis DN, et al. A novel pancreatic endocrine tumor suppressor gene locus on chromosome 3p with clinical prognostic implications. J Clin Invest 1997;100:404-10.
14. Chetty R, Essa A. Heterotopic pancreas, periampullary somatostatinoma and type I neurofibromatosis: a pathogenetic proposal. Pathol 1999;31:95-7.
15. Muscarella P, Melvin WS, Fisher WE, et al. Genetic alterations in gastrinomas and nonfunctioning pancreatic neuroendocrine tumors: an analysis of p16/MTS1 tumor suppressor gene inactivation. Cancer Res 1998;58:237-40.
16. Serrano J, Goebel SU, Peghini PL, et al. Alterations in the p16INK4a/CDKN2A tumor suppressor gene in gastrinomas. J Clin Endocrinol Metab 2000;85:4146-56.
17. Nakamura T, Iwamura Y, Kaneko M, et al. Deletions and rearrangements of the retinoblastoma gene in hepatocellular carcinoma, insulinoma and some neurogenic tumors as found in a study of 121 tumors. Jpn J Clin Oncol 1991;21:325-9.
18. Iwamura Y, Futagawa T, Kaneko M, et al. Co-deletions of the retinoblastoma gene and Wilms' tumor gene and rearrangement of the Krev-1 gene in a human insulinoma. Jpn J Clin Oncol 1992;22:6-9.
19. Stewart CJ, Imrie CW, Foulis AK. Pancreatic islet cell tumour in a patient with familial adenomatous polyposis. J Clin Pathol 1994;47:860-1.
20. Ebrahimi SA, Wang EH, Wu A, et al. Deletion of chromosome 1 predicts prognosis in pancreatic endocrine tumors. Cancer Res 1999;59:311-5.
21. Gortz B, Roth J, Kr Ahenmann A, et al. Deletion of chromosome 1 predicts prognosis in pancreatic endocrine tumors. Am J Pathol 1999;154:429-36.
22. Beghelli S, Pelosi G, Zamboni G, et al. Pancreatic endocrine tumours: evidence for a tumour suppressor pathogenesis and for a tumour suppressor gene on chromosone 17p. J Pathol 1998;186:41-50.
23. Bartsch D, Hahn SA, Danichevski KD, et al. Mutations of the dPC4/Smad4 gene in neuroendocrine pancreatic tumors. Oncogene 1999;18:2367-71.