Section 6 -

Pancreatic Endocrine Tumors: Islets of History, Developments and Diagnostics Paul Komminoth Institute of Pathology, Kantonsspital CH-5404 Baden, Switzerland Aurel A. Perren Department of Pathology UniversityHospital CH-8091 Zürich, Switzerland

Islets of History The pancreas was discovered by Herophilus (335-280 BC), a Greek anatomist and a few hundred years later, Ruphos, another Greek anatomist, gave the pancreas its name. For many centuries the function of the pancreas was unknown and only after detailed anatomical studies by Bartholini (1616-1680) and the detection of the ductus Wirsungianus in 1642 it became clear that this organ is involved in the digestion of food (exocrine function). However, it took more than 100 years until the endocrine function of the pancreas emerged. In 1788 Cowley described the relation between diseases of the pancreas and the development of a diabetes mellitus. Peter Langerhans (1849-1888), a Berlin-based pathologist first recognized the islets of the pancreas on the basis of their histologic staining characteristics [1]. In 1927, Wilder et al described the first hormone-producing pancreatic tumor syndrome in a patient with hypoglycemia and a metastatic islet cell tumor [2]. Subsequently four other classic pancreatic endocrine tumor (PET) syndromes have been described: Zollinger-Ellison syndrome (1955) caused by gastrin-producing tumors [3]; Verner-Morrison syndrome (1958), WDHA (watery diarrhea, hypokalemia, and achlorhydria) syndrome or pancreatic cholera, caused by vasoactive intestinal peptide (VIP)-releasing tumors or VIPomas [4]; glucagonoma syndrome, described by Mallinson et al in 1974 [5, 6]; and somatostatinoma syndrome, described by Ganda et al in 1977 [7]. Several other rare clinical syndromes have been proposed as possible functional endocrine syndromes associated with PET. These include calcitoninoma [8], parathyrinoma [9], growth hormone-releasing factor–secreting tumor (GRFoma), adrenocorticotropin hormone–secreting tumor (ACTHoma), and neurotensinoma [10]. Patients with pancreatic neoplasms that have the histologic characteristics of a PET but no associated elevation in plasma hormone levels (excluding pancreatic polypeptide [11]) and those without a recognizable clinical syndrome are considered to have nonfunctional (functionally inactive) PET. Islets of Developments The distinction between benign and malignant PETs is crucial and still an unsolved problem in diagnostic pathology. The only reliable criteria for malignancy are the infiltration of the tumor into adjacent organs (e.g. the spleen) and the presence of lymph node and/or distant metastases. While the currently used morphological parameters outlined in the WHO classification [12] are useful in daily practice, it would be most helpful to have additional molecular or immunohistochemical markers to better classify (especially small) PET and to separate potentially aggressive ("baby lions") from less harmful tumors ("baby cats"). A prerequisite to identify such markers, however, is a better understanding of the initiation, progression and dissemination of PET. While the molecular basis of the familial PETs, which are associated with multiple endocrine neoplasia type 1 (MEN1) and von Hippel-Lindau (VHL) syndrome, has recently been established [13, 14], only little is known about the oncogenesis and the molecular basis of progression of sporadically occurring PETs. A limited number of published studies indicate that in contrast to other human tumors, the activation of oncogenes is not an early event in PETs [15, 16, 17]. Furthermore, it has been demonstrated that the MEN1 gene is mutated in approx. 15-20% and that the VHL-gene is rarely involved in sporadic PETs by mutation [18, 19, 20]. In the following paragraphs we summarize our results of molecular studies on PET obtained using a candidate gene approach and genome wide approaches such as comparative genomic hybridization (CGH) and cDNA expression arrays. Comparative Genomic Hybridization By comparative genomic hybridization (CGH) we have screened small (early) PETs as well as large (advanced) benign and malignant lesions including metastasis in order to identify common chromosomal gains and losses which might be relevant in the initiation and progression of PETs. We have studied 44 PETs and identified chromosomal aberrations in 36 tumors. Chromosomal losses were slightly more frequent (mean: 5.3) than gains (mean: 4.6). The most frequent losses involved the Y chromosome (45% of male PETs), 6q (39%), 11q (36%), 3p, 3q, 11p (30%), 6p (27%) and 10q and Xq (25%), whereas most common gains included 7q (43%), 17q (41%), 5q and 14q (32%), 7p, 9q, 17p, 20q (27%) and 12q and Xp (25%). A correlation was found between the total number of genetic changes per tumor and both tumor size and disease stage. In particular, loss of 3p and 6 and gain of 14q and Xq were found to be associated with metastatic disease. Furthermore, characteristic patterns of genetic changes were found in the various PET subtypes, e.g. 6q loss in malignant insulinomas, indicating that these tumor subtypes evolve along genetically different pathways [21]. In order to search for genetic alterations that play a role in early tumor development, we have further studied 38 small PETs with a diameter < 2 cm, including 24 insulinomas and 10 non-functioning PETs. Such small neoplasms are classified as tumors with a clinically benign behavior, if no clear morphological signs of malignancy are observed [22]. We identified chromosomal aberrations in 24 PETs (mean: 4). Interestingly, the number of gains differed significantly between non-functioning and functioning PETs (3.4 vs. 1.3, respectively; P=0.033), as did the numbers of aberrations in the benign (n=30) and malignant (n=8) tumors (2.8 vs. 8.4, respectively; P=0.0016). In insulinomas, particularly 9q gain was common (33%), whereas most frequent losses involved 1p (21%), 1q, 4q, 11q and Xq (17%). In contrast, particularly loss of 6q was observed in the non-functioning PETs (50%), as were gains of 4p (40%). Loss of 3pq, 6q and gain of 17pq, in addition, proved to be strongly associated with malignant behavior (p<0.005). Our results indicate that functioning (insulinomas) and non-functioning PETs evolve along different genetic pathways, and that tumor progression is associated with specific chromosomal gains and losses [22]. In a third CGH study we examined 17 paired specimens of primary PETs and their metastases as well as 28 nonmetastatic PETs. The mean number of genomic changes was 17.3 in metastases, 12.5 in their primary tumors, and 4.5 in nonmetastatic PETs. The genomic changes which were enriched in metastases included gains of chromosomes 4 and 7 and loss of 21q. These results point towards chromosomal loci harboring genes contributing to the metastatic progression of PETs [23]. In summary data of CGH indicate that tumors of larger size, tumors with malignant potential and especially metastases harbor more genetic alterations than small and clinically benign neoplasm. These findings point toward a tumor suppressor pathway together with genomic instability as important mechanisms associated with tumor progression [12]. Candidate Gene Approach In addition and in parallel to the genome wide CGH studies, we used a candidate gene approach to examine specific genomic regions and genes. Combined results of LOH studies revealed similar chromosomal regions with genetic losses, however, in general, the rate of LOH is roughly two times higher than allelic losses detected by CGH and at regions 3p23, 6q22, 9p, 11q13, 18q21 and 22q12.1 the differences are even more pronounced, indicating that small deletions not detectable by CGH are also involved [24, 25]. In order to investigate the role of 11q, the chromosomal localization of the MEN1 gene, we analyzed 52 PETs for LOH at markers distributed over the long arm of chromosome 11. Loss of 11q was detected in 23/52 (44%) PETs, 6 of which showed isolated MEN1 gene loss and 17 loss of MEN1 plus distal markers. In 16 of these latter PETs, the whole chromosome 11 was lost as detected by CGH and FISH (all monosomies). Our results suggest the existence of a second tumor suppressor gene telomeric of the MEN1 gene playing a role in endocrine tumorigenesis [26] located in a second smallest region of allelic deletion (SRAD) distal of the MEN1 locus comprising 11q23. Succinate dehydrogennase subunit D (SDHD) is a candidate gene on this chromosomal region: It is responsible for familial paragangliomas type 1 (PGL1). We detected allelic loss in 29% of 14 PET, but could not identify any somatic mutations. However, as SDHD is an imprinted gene, a minor role in PET can not fully be excluded [27]. Losses of 3p were identified in 54.9% of the 99 examined PETs. A common region of LOH was 3p23-p25.3. In addition, a strong correlation was found between the loss of alleles on chromosome 3p and clinically metastatic disease. PETs from these patients showed a tendency towards losing large parts or the entire short arm of chromosome 3 including the VHL gene associated with tumor progression [24]. However, VHL mutations appear to be rare in sporadic PET. Loss of chromosomal material on the long arm of chromosome 6 (6q13-6q25-27) was also found to correlate with tumors of malignant clinical behavior and large tumors (>2cm). We identified two SRADs at 6q22.1 and 6q23-24, which point towards tumor suppressor gene loci. Possible candidates are AIM1, CCNC, PTPRK and ZAC [25]. However, we could not identify ZAC mutations in 37 PET by PCR-SSCP and direct sequencing [12]. Peroxiredoxin 1(PRDX1), located on chromosome 1q is an intracellular anti-oxidant protein inactivating H2O2, and reactive oxygen species are known to alter functions of proteins regulating DNA repair, cell cycle and apoptosis. PRDX1 knock out mice develop tumors including PET, and 1p is frequently deleted in human PET (32%). Therefore we examined 17 PET with 1p LOH for PRDX1 mutations by PCR-DGGE. As mutations were absent, we could exclude this gene as candidate tumor suppressor gene in PET [28]. The gene DPC4 is deleted in 40% to 60% of ductal adenocarcinomas of the pancreas and one group described deletions in 50% of non-functioning PET and concluded a common genetic pathway of exocrine and endocrine pancreatic tumors. We examined 31 PET by immunohistochemistry, mutation analysis as well as LOH analysis. DPC4 mutations were absent in all examined tumors. The LOH rate was low (11%) and homozygous deletions were excluded by FISH. On the protein level, DPC4 expression was lost in only one malignant tumor. These results argue against an important role of this tumor suppressor gene in PET and support the concept of different genetic pathways in exocrine and endocrine pancreatic tumors [29]. Germline PTEN mutations are responsible for the Cowden and Bannayan-Riley-Ruvalcaba syndromes and somatic mutations of this tumor suppressor gene are found in a wide range of sporadic cancers. We therefore examined 33 PET for mutations, allelic deletions and protein expression. We found a novel PTEN mutation in one malignant non-functioning PET which was accompanied by loss of PTEN immunoreactivity. PTEN inactivation by mutation and loss of expression seems to be a rare and presumably late event in PET. Instead, either an impaired transport system of PTEN to the nucleus or some other means of differential compartmentalization could account for impaired PTEN function, as 19 of 23 PET exhibited a predominantly cytoplasmatic PTEN immunoreactivity compared to a nuclear staining of normal islets [30]. We examined 130 endocrine tumors including 25 PET for activating mutations of the BRAF oncogene. This gene has recently been reported to be mutated frequently in melanomas, a tumor of neural crest derived cells. We found a high rate of V559E mutations in papillary thyroid carcinomas (47%), one V599E mutation in a well differentiated gastric endocrine carcinoma (malignant carcinoid), but no activating BRAF mutations in all other examined endocrine tumors including 25 PET. These results point towards different pathways in tumorigenesis of endocrine tumors of various localizations and only rare involvement of the MAPK pathway (downstream of BRAF) in a subset of malignant neuroendocrine tumors [31]. In summary, published data indicate that so far only few possible candidate genes located at some of the above mentioned chromosomal loci have been investigated and many genes remain to be identified. Point mutations in genes such as VHL, p16, PTEN, k-RAS, p53 appear to be extremely rare (1-3%) and no mutations were identified in DPC4/SMAD4, RET, ZAC, BRAF and SMAD3. Table 1: Genomic aberrations detected by LOH, CGH and mutation analysis Locus LOH Gene Mutation CGH 1p36- 10/29 (34%) 21/102 (21%) 1q32- 8/29 (28%) 16/102 (16%) 3p23- 23/31 (74%) 19/102 (19%) 3p25-26- 31/73 (42%) VHL 1/75 (1%) 19/102 (19%) 6q22- 43/69 (62%) 29/102 (28%) 9p- 12/37 (32%) p16 1/44 (2%) 0/102 (0%) 9q+ 29/102 (28%) 10q23- 8/16 (50%) PTEN 1/31 (3%) 14/102 (14%) 11p14- 28/102 (27%) 11q13 75/111 (68%) MEN1 33/155 (21%) 31/102 (30%) 11q22-23 20/37 (54%) SDHD 0/20 (0%) 31/102 (30%) 12p12+ K-Ras 1/39 (3%) 23/102 (23%) 15q- SMAD3 0/18 (0%) 6/102 (6%) 17p13- 15/40 (38%) p53 1/40 (3%) 2/102 (2%) 17p+ 32/102 (31%) 18q21- 23/68 (34%) DPC4 0/41 (0%) 6/102 (6%) 22q12.1 9/12 (75%) 4/102 (4%) Xq- 11/23 (48%) 14/46 (30%) Y- 5/14 (36%) 14/56 (25%) cDNA Expression Arrays Another helpful genome wide screening method to identify genes possibly involved in the initiation and progression of a tumor type, are cDNA expression arrays. These nylon membrane-based arrays are designed to provide expression data using spotted cDNA fragments (200-600 bp) of several thousand known human genes. Recently we have performed a cDNA expression array study on 20 insulinomas, where fresh frozen tissue was available. This tumor type represents a homogenous group according to our CGH data. The MEN1 mutation status is negative in 19 of these tumors and one additional tumor carries a somatic MEN1 mutation. Unsupervised cluster analysis revealed two groups of insulinomas: Three of the five malignant (metastasizing) insulinomas are localized in a cluster of 9 tumors. One tumor of uncertain behavior (size >2cm) is also localized in this cluster. The second cluster included the sample of normal islets and consisted of 6 benign insulinomas, one insulinoma of unknown behavior and only one malignant tumor. The latter malignant tumor was classified as such according to the WHO criteria due to a micro-metastasis in a peripancreatic lymph node; the patient is well and without evidence of disease 18 months after tumor resection. The cluster containing the majority of malignant tumors is characterized by 896 overexpressed and 249 underexpressed genes. This result seems to be surprising at a first glance, as tumor suppressor pathways with underexpression of transcripts are suspected in PET. It reveals that malignant transformation with accumulation of genomic defects seems to be reflected by overexpression of many genes. Interestingly, transcripts of both BRCA1 and ATM, genes being part of a large multi-subunit protein complex of tumor suppressors acting as DNA damage sensors are overexpressed in malignant tumors known to be genomically instable from CGH analysis. Of interest are genes overexpressed in all tumors, as they might represent common alterations in all insulinomas and thus may be related to events of tumor initiation. 28 genes have been found to be overexpressed in all insulinomas by at least a factor 2 compared to normal islets. Islets of Diagnostics In a clinic-pathological study we have recently re-evaluated over 200 PET and classified them using the criteria proposed by the current WHO classification. Using a tissue micro array (TMA) containing 110 of these PETs with a follow-up time of up to 30 years we additionally examined immunohistochemical markers which have been reported helpful in daily diagnostics to better classify PET and to separate clinically aggressive from less harmful tumors. The WHO criteria combining size, mitotic index, Ki67 proliferation index and angioinvasion turned out to perfectly separate PET into four different risk groups. The usefulness of immunohistochemical markers such as p27. cox2, CD99 could not be confirmed using our sample set. Only CK 19 expression was associated with an adverse outcome. Acknowledgments We are grateful to P. Saremaslani, S. Schmid, T. Locher for excellent technical assistance, to the team of D. Zimmermann for performing cycle sequencing analyses and to J. Roth for helpful discussions. References Langerhans P: Beitrage zur mikroskopischen Anatomie der Bauchspeicheldrüse. Berlin, Gustav Lange, 1869 Wilder RM, Allan FN, Power MH, Robertson HE: Carcinoma of the islands of the pancreas; hyperinsulinism and hypoglycemia. JAMA 1927, 89:348-355 Zollinger RM, Ellison EH: Primary peptic ulcerations of the jejunum associated with islet cell tumors of the pancreas. Ann Surg 1955, 142:709-723 Verner JV, Morrison AB: Islet cell tumor and a syndrome of refractory watery diarrhea and hypokalemia. Am J Med 1958, 25:374-380 Mallinson CN, Bloom SR, Warin AP, Salmon PR, Cox B: A glucagonoma syndrome. Lancet 1974, 2:1-5 Mallinson CN, Salmon PR, Borrowman J, Bloom S: Proceedings: The association of a specific skin lesion with islet-cell tumours of the pancreas. Gut 1973, 14:827 Ganda OP, Weir GC, Soeldner JS, Legg MA, Chick WL, Patel YC, Ebeid AM, Gabbay KH, Reichlin S: "Somatostatinoma": a somatostatin-containing tumor of the endocrine pancreas. N Engl J Med 1977, 296:963-967 Howard JM, Gohara AF, Cardwell RJ: Malignant islet cell tumor of the pancreas associated with high plasma calcitonin and somatostatin levels. Surgery 1989, 105:227-229 Mao C, Carter P, Schaefer P, Zhu L, Dominguez JM, Hanson DJ, Appert HE, Kim K, Howard JM: Malignant islet cell tumor associated with hypercalcemia. Surgery 1995, 117:37-40 Meko JB, Norton JA: Endocrine tumors of the pancreas. Curr Opin Gen Surg 1994:186-194 Langstein HN, Norton JA, Chiang V, O'Dorisio TM, Maton PN, Marx SJ, Jensen RT: The utility of circulating levels of human pancreatic polypeptide as a marker for islet cell tumors. Surgery 1990, 108:1109-1115; discussion 1115-1106 Heitz PU, Komminoth P, Perren AA, Klöppel G: Tumours of the endocrine pancreas (chapter 4). WHO classification of tumours: pathology and genetics of tumours of endocrine organs. Edited by DeLellis RA, Heitz PU, Lloyd RV, Eng C. Lyon, IARC Press, 2004, pp 177-208 Chandrasekharappa SC, Guru SC, Manickam P, Olufemi S, Collins FS, Emmert-Buck MR, Debelenko LV, Zhuang Z, Lubensky IA, Liotta LA, Crabtree JS, Wang Y, Roe BA, Weisemann J, Boguski MS, Agarwal SK, Kester MB, Kim YS, Heppner C, Dong Q, Spiegel AM, Burns AL, Marx SJ: Positional cloning of the gene for multiple endocrine neoplasia-type 1. Science 1997, 276:404-406 Latif F, Tory K, Gnarra J, Yao M, Duh FM, Orcutt ML, Stackhouse T, Kuzmin I, Modi W, Geil L, et al: Identification of the von Hippel-Lindau disease tumor suppressor gene (see comments). Science 1993, 260:1317-1320 Höfler H, Ruhri C, Pütz B, Winsberger G, Hauser H: Oncogene expression in endocrine pancreatic tumors. Virchows Arch B 1988, 55:355-361 Iwamura Y, Futagawa T, Kaneko M, Nakagawa K, Kawai K, Yamashita K, Nakamura T, Hayashi H: 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 Lohmann DR, Funk A, Niedermeyer HP, Haupel S, Hofler H: Identification of p53 gene mutations in gastrointestinal and pancreatic carcinoids by nonradioisotopic SSCA. Virchows Arch B Cell Pathol Incl Mol Pathol 1993, 64:293-296 Chung DC, Smith AP, Louis DN, Graeme-Cook F, Warshaw AL, Arnold A: A novel pancreatic endocrine tumor suppressor gene locus on chromosome 3p with clinical prognostic implications. J Clin Invest 1997, 100:404-410. Gortz B, Roth J, Krahenmann A, de Krijger RR, Muletta-Feurer S, Rutimann K, Saremaslani P, Speel EJ, Heitz PU, Komminoth P: Mutations and allelic deletions of the MEN1 gene are associated with a subset of sporadic endocrine pancreatic and neuroendocrine tumors and not restricted to foregut neoplasms. American Journal of Pathology 1999, 154:429-436 Zhuang Z, Vortmeyer AO, Pack S, Huang S, Pham TA, Wang C, Park WS, Agarwal SK, Debelenko LV, Kester M, Guru SC, Manickam P, Olufemi SE, Yu F, Heppner C, Crabtree JS, Skarulis MC, Venzon DJ, Emmert-Buck MR, Spiegel AM, Chandrasekharappa SC, Collins FS, Burns AL, Marx SJ, Lubensky IA: Somatic mutations of the MEN1 tumor suppressor gene in sporadic gastrinomas and insulinomas. Cancer Research 1997, 57:4682-4686 Speel EJ, Richter J, Moch H, Egenter C, Saremaslani P, Rutimann K, Zhao J, Barghorn A, Roth J, Heitz PU, Komminoth P: Genetic differences in endocrine pancreatic tumor subtypes detected by comparative genomic hybridization. Am J Pathol 1999, 155:1787-1794. Speel EJ, Scheidweiler AF, Zhao J, Matter C, Saremaslani P, Roth J, Heitz PU, Komminoth P: Genetic evidence for early divergence of small functioning and nonfunctioning endocrine pancreatic tumors: gain of 9Q34 is an early event in insulinomas. Cancer Res 2001, 61:5186-5192. Zhao J, Moch H, Scheidweiler AF, Baer A, Schaffer AA, Speel EJ, Roth J, Heitz PU, Komminoth P: Genomic imbalances in the progression of endocrine pancreatic tumors. Genes, Chromosomes & Cancer 2001, 32:364-372 Barghorn A, Komminoth P, Bachmann D, Rutimann K, Saremaslani P, Muletta-Feurer S, Perren A, Roth J, Heitz PU, Speel EJ: Deletion at 3p25.3-p23 is frequently encountered in endocrine pancreatic tumours and is associated with metastatic progression. J Pathol 2001, 194:451-458. Barghorn A, Speel EJ, Farspour B, Saremaslani P, Schmid S, Perren A, Roth J, Heitz PU, Komminoth P: Putative tumor suppressor loci at 6q22 and 6q23-q24 are involved in the malignant progression of sporadic endocrine pancreatic tumors. Am J Pathol 2001, 158:1903-1911. Speel EJ, Zhao J, Meier D, Matter C, Muletta-Feurer S, Saremaslani P, van Asseldonk M, Roth J, Heitz PU, Komminoth P: Analysis of chromosome 11q in sporadic endocrine pancreatic and neuroendocrine tumors: presence of another tumor suppressor gene telomeric of MEN1? JMD 1999, 1:39 (Abstract S22) Perren A, Barghorn A, Schmid S, Saremaslani P, Roth J, Heitz PU, Komminoth P: Absence of somatic SDHD mutations in sporadic neuroendocrine tumors and detection of two germline variants in paraganglioma patients. Oncogene 2002, 21:7605-7608 Perren A, Schmid S, T. L, Saremaslani P, Heitz PU, Komminoth P: Analysis of peroxiredoxin (PRDX1) on 1p34 in human pancreatic endocrine tumors. Pathology International 2004:A122 Perren A, Hürlimann S, Saremaslani P, Schmid S, Bonvin C, Roth J, Heitz PU, Komminoth P: DPC4/Smad4 expression is lost in a subset of ductal adenocarcinomas of the pancreas but not in endocrine pancreatic tumors and chronic pancreatitis. Mod Pathol 2002, 15:119A (Abstract 492) Perren A, Komminoth P, Saremaslani P, Matter C, Feurer S, Lees JA, Heitz PU, Eng C: Mutation and expression analyses reveal differential subcellular compartmentalization of PTEN in endocrine pancreatic tumors compared to normal islet cells. Am J Pathol 2000, 157:1097-1103. Perren A, Saremaslani P, Bonvin C, Schmid S, Locher T, Heitz PU, Komminoth P: Absence of BRAF Exon 11 and 15 mutations in neuroendocrine tumors. Mod Pathol 2003, 16:108A (Abstract 487)