Multiple endocrine neoplasia type 1 (MEN-1) is a rare autosomal-dominant disease. It is associated with a broad range of endocrine tumours, most frequently arising in the parathyroid glands, the pituitary and the pancreas. Most neuroendocrine tumours will be diagnosed in the pancreas as non-functioning neuroendocrine tumours or insulinomas. Forty-two percent of the patients will develop a gastrin-secreting neuroendocrine tumour, a gastrinoma. Gastrinomas in MEN-1 tend to be small, multiple and preferentially located in the duodenum. This paper will focus on the specific characteristics of gastrinomas in the setting of MEN-1 compared to sporadic gastrinomas. The developments in understanding the tumorigenesis of these tumours and the consequences for diagnosis and therapy will be discussed.
Multiple endocrine neoplasia type 1 (MEN-1) is a rare autosomal-dominant disease. MEN-1 is associated with a broad range of endocrine tumours, most frequently arising in the parathyroid glands, the pituitary and the pancreas. Penetrance for neuroendocrine tumours is up to 70%–100%. Half of the patients with MEN-1 will present with a pancreatico-duodenal neuroendocrine tumour at the age of 50 years. Most will arise in the pancreas as non-functioning neuroendocrine tumours or insulinomas. Forty-two percent (range 20%–61%) of the patients will develop a gastrin-secreting neuroendocrine tumour, a gastrinoma. In contrast to sporadic gastrinomas that occur predominantly in the pancreas, in MEN-1 most of these tumours reside in the duodenum. Duodenal gastrinomas are small, usually below 1 cm in diameter, multiple and show a preference for the proximal duodenum. Pancreatic gastrinomas associated with MEN-1 are very rare, as are other uncommon extra-pancreatic, extra-duodenal locations that have been associated with gastrinomas.
Independent of the tumour location, gastrinomas are well differentiated endocrine carcinomas. Recently, a specific sequence of events has been suggested to occur during tumorigenesis in gastrinomas due to multiple endocrine neoplasia. In MEN-1 distinctive proliferative hyperplastic gastrin cells in the non-neoplastic duodenum are interpreted as multicentric precursor lesions. Their spectrum of proliferative changes is compared to those seen in entero-chromaffine-like (ECL)-cells in chronic atrophic gastritis. The first step is supposed to be diffuse hyperplasia of gastrin-positive cells, followed by linear hyperplasia, i.e., chains of more than five gastrin-positive cells and more than two chains per mm. With further proliferation micronodular (30–50 µm) hyperplastic lesions, i.e., nodules of more than 5 gastrin-positive cells within glands or crypts of the mucosa followed by enlarged nodules (>90–210 µm) of gastrin-positive cells with solid architecture have been demonstrated. These hyperproliferative and hyperplastic lesions were found only in patients with MEN-1, but were absent in patients with sporadic, not MEN-1 associated duodenal gastrinoma. The occurrence of gastrin cell hyperplasia at various sites in the duodenal mucosa explains the multifocality of gastrinomas in MEN-1. It is however unclear whether additional stimuli, besides the MEN-1 mutation are responsible for the development of hyperplastic lesions. As the next step in the tumorigenesis of MEN-1 associated gastrinomas micro-invasive lesions, small clusters of gastrin-positive cells in the lamina propria between the glands, resulting in micro-tumours (>250 µm) with trabecular growth pattern and fibrosis have been described. In MEN-1 the heterozygous germ line mutations of the MEN-1 gene, a tumour-suppressor gene, are by themselves insufficient for tumour development. According to the two-hit hypothesis of tumour development by Knudson additional somatic inactivation of the wild-type MEN-1 allele on chromosome 11q13 has to occur. This may lead to a loss of a part or all of chromosome 11, termed loss of heterozygosity. By using a combination of fluorescence in situ hybridisation and immuno-fluorescence techniques, Anlauf et al. were able to simultaneously detect allelic deletion and hormone expression in individual duodenal cells. Different tumours from one patient showed different patterns of loss of heterozygosity, suggesting that each gastrinoma arises as a result of an independent second hit. In contrast none of the hyperplastic lesions was found to be positive for loss of heterozygosity. These observations enabled the authors to conclude that the hyperplastic lesions, suggested to be pre-neoplastic, retain both MEN-1 alleles and that allelic loss or other mutation in the wild-type allele inactivating the gene, is responsible for the final development of neoplastic lesions. It still remains to be analysed which factors are responsible for loss of heterozygosity in these individual cells.
To summarize, MEN-1 associated duodenal gastrinomas are preceded by gastrin-positive cell hyperplasia. Allelic deletion in these cells results in multifocal gastrin positive neoplastic lesions.
MEN-1 associated duodenal gastrinomas show a trabecular and pseudo-glandular pattern; immuno-histochemically they express gastrin and occasionally somatostatin. Interestingly only about 50% of these tumours are functionally active i.e., associated with a Zollinger-Ellison syndrome and thus are gastrinomas per definition. They are characterised according to the WHO classification (i) as well differentiated neuroendocrine tumours, i.e., benign, if non-angioinvasive and below 1 cm in size or (ii) benign or low grade malignant, if confined to the mucosa-submucosa with or without angioinvasion and less than 1 cm in size. With invasion of the muscularis propria and beyond or the development of metastases they are characterized as well differentiated neuroendocrine carcinomas. Tumours with a high grade malignancy are termed poorly differentiated neuroendocrine carcinomas. Recently a TNM classification has been suggested for duodenal tumours by the European Neuroendocrine Tumour Society (ENETS).
The penetrance of duodenal gastrinomas in MEN-1 lies between non-functioning pancreatic tumours and insulinomas. Twenty to 40% of patients, between 50 and 70 years of age, are diagnosed with a gastrinoma. Data from the NIH series indicate a slightly earlier diagnosis with a maximum occurring at 40 years, preceded by symptoms of a Zollinger-Ellison syndrome by 5–7 years. In comparison to sporadic gastrinomas, gastrinomas in the setting of MEN-1 are diagnosed 10 years earlier and the primary is more often located in the duodenum. No other lesion than the primary was found at the time of diagnosis in 9% of the patients with MEN-1 as compared to 27% of those with sporadic gastrinomas. While recent data localize most MEN-1 related gastrinomas within the duodenum, earlier studies reported primary lymph node gastrinomas and significantly more pancreatic gastrinomas than today. This may be due to the high frequency (60 to 80%) of large regional lymph nodes metastases at the time of diagnosis. These were either misinterpreted as primary lymph node gastrinomas due to the very small duodenal lesions that have escaped detection in patients with MEN-1, or were possibly considered as pancreatic tumours, especially if they were located at the upper margin of the head of the pancreas. Duodenal gastrinomas may metastasize to the liver in about 10% of the cases, but this occurs rather late in the course of the disease. Fast-growing poorly differentiated and metastasizing duodenal gastrinomas are extremely rare.
The percentage of patients with gastrinomas as part of the MEN-1 syndrome is about 23% (10%–48%) in ten large series compiled by Berna and co-workers. Most patients present with long-standing upper abdominal pain (66%), diarrhoea (76%), heartburn (52%), nausea (38%) and vomiting (24%) or weight loss (12%) similar to patients with sporadic gastrinoma. Due to the notion that Zollinger-Ellison syndrome complications occur slightly less in MEN-1 compared to sporadic gastrinomas, the number of patients with bleeding as the presenting symptom of Zollinger-Ellison syndrome was significantly lower (12% vs. 27%) in MEN-1 gastrinomas patients. The widespread use of proton-pump inhibitors may mask the classical symptoms of Zollinger-Ellison syndrome and delay the diagnosis of a gastrinoma. Associated nephrolithiasis occurs more often in MEN-1 patients and cutaneous lesions (angiofibromas and collagenomas) have a sensitivity of 75% and specificity of 95% to predict the presence of MEN-1 in patients with Zollinger-Ellison syndrome. Interestingly, the prevalence and probability of Zollinger-Ellison syndrome in MEN-1 is sex-related, with a probability of developing Zollinger-Ellison syndrome at the age of 60 years of 55% in men and 33% in women.
The diagnosis of Zollinger-Ellison syndrome is made by demonstrating an increased gastrin concentration in fasting patients in the presence of proven hyperchlorhydria. In a large prospective investigation in 309 patients with Zollinger-Ellison syndrome (including 30% MEN-1 patients), Berna et al. compared their findings with 2,209 cases from the literature. Both, normal or very high concentrations (>100-fold normal) of gastrin concentrations in fasting patients were uncommon (0.3%–3% and 4.9%–9%, respectively). Forty to 60% of gastrinoma patients in the fasting state had a gastrin concentration below 10 times normal and this overlaps with gastrin concentrations seen in Helicobacter pylori infection. No difference could be detected between sporadic and MEN-1 associated gastrinoma patients. This increase of the gastrin concentration in fasting patients overlaps with hypergastrinaemia seen in fasting patients with idiopathic peptic disease or gastro-oesophageal reflux disease. Furthermore an increased gastrin concentration in fasting patients is observed in patients with pernicious anaemia or on proton pump inhibitor medication. However in both conditions gastric acid secretion is reduced in contrast to gastrinoma patients. In 20%–30% of these patients the gastrin concentration can be up to five times normal, comparable to the gastrin concentration seen in 60% of Zollinger-Ellison syndrome patients.
As expected increased basal acid output, tumour location (pancreas > duodenum), tumour size (large > small) and extent (liver metastases > local disease) all were positively correlated to the basal gastrin concentration.
Proton pump inhibitors have to be stopped at least 1 week prior to the determination of the gastrin concentration and this may pose a serious problem in terms of pain and risk of bleeding in symptomatic patients. While histamine H2 receptor antagonists are needed in higher doses to achieve symptom control in Zollinger-Ellison syndrome compared to idiopathic ulcer disease, this is not the case with proton pump inhibitors. Thus, the use of proton pump inhibitors can delay the diagnosis of Zollinger-Ellison syndrome because of symptom control with conventional dosage.
Hypercalcaemia (due to hyperparathyroidism in MEN-1 patients, which predeces the manifestation of gastrinoma in most patients) has been shown to increase the serum gastrin concentration in fasting patients with Zollinger-Ellison syndrome. However, in the NIH series no correlation between hypercalcaemia and the gastrin concentration in fasting patients was observed, probably due to the low tumour load in MEN-1 patients.
Hypochlorhydria will increase gastrin secretion. Patients with atrophic gastritis, pernicious anaemia, post-vagotomy patients or patients on proton pump inhibitors can be excluded due to the high intragastric pH or low basal acid-output. On the other hand patients with hyperchlorhydria and hypergastrinaemia due to Helicobacter pylori infection, antral G-cell hyperfunction, short bowel syndrome or those with renal failure have to be differentiated from those with Zollinger-Ellison syndrome.
Thus, while the gastrin concentration in fasting patients
