MAGE1 is expressed by a subset of pancreatic endocrine neoplasms and associated lymph node and liver metastases.

BACKGROUND: MAGE1 was originally isolated from human melanoma cells as a target antigen for autologous cytotoxic T lymphocytes. Expression of MAGE1 has subsequently been identified in a number of neoplastic cell types, including testicular germ cell and breast cancer cells, which has led to the development of antitumor MAGE1 vaccines. AIM OF THE STUDY: To determine if Mage-1 is expressed in pancreatic endocrine neoplasms (PENs) and PEN metastases. METHODS: We utilized immunolabeling analysis for Mage-1 on 49 primary PENs, 11 liver metastases, and 6 lymph node metastases. A semiquantitative labeling index (LI) of 0 (no expression), 1, 2 (moderate expression), and 3 (intense expression, correlating with internal control markers) was used to determine relative amounts of MAGE1 expression in these lesions. RESULTS: We have identified MAGE1 expression in a subset (42 of 49; 86%) of PENs. Normal pancreatic ducts, present in tissue adjacent to PENs, were utilized as a positive control for Mage-1 immunolabeling (index score 3); no other detectable labeling for Mage-1 was evident in normal pancreatic tissue. Primary PENs, with or without metastases (mean LI score 1.2 vs 1.0, respectively), did not demonstrate a significant difference in Mage-1 LI, although intratumoral heterogeneity was apparent in some, but not all, of these lesions. Lymph node metastases (mean score 2.0) demonstrated a significant increase in Mage-1 LI as compared to primary, non-metastatic lesions (p = 0.04984) and primary metastatic lesions (p = 0.02351). In contrast, six patients with a survival of less than one year demonstrated a low Mage-1 LI (mean score, 0.58). CONCLUSIONS: MAGE1 expression is present in a subset of primary PENs and in lymph node metastases, and may therefore serve as a useful marker and potential therapeutic target in PENs. Furthermore, the absence of Mage-1 expression in a subset of primary PENs may indicate a worsened prognosis.

Chromogranin A and pancreastatin-like immunoreactivity in serum of gastrinoma patients.

Gastrin and pancreastatin-like immunoreactivity were determined by radioimmunoassay methods and chromogranin A was determined by enzyme-linked immunoassay in sera from 18 patients with gastrinomas (Zollinger-Ellison syndrome) and in 20 age and sex matched controls. Gastrin serum levels in the gastrinoma patients were in the range 26-80,000 pmol/l, and in the controls 5-31 pmol/l. Chromogranin A serum levels in the gastrinoma group were in the range 6-2,700 ng/ml (mean +/- SEM: 400 +/- 147 ng/ml). The mean value of chromogranin A was significantly higher than in the control group (8 +/- 2 ng/ml, p = 0.008). The serum levels of pancreastatin-like immunoreactivity in the gastrinoma patients were in the range 23-1,994 pg/ml (597 +/- 123 pg/ml). The mean value of pancreastatin-like immunoreactivity in the gastrinoma group was significantly higher than in the control group (104 +/- 25 pg/ml, p = 0.0002). The levels of chromogranin A and pancreastatin-like immunoreactivity were significantly higher in patients with verified metastatic disease (p = 0.04, p = 0.01 respectively). There was a significantly positive correlation between levels of gastrin and pancreastatin-like immunoreactivity (r = 0.7, p = 0.002), while no correlation was found between gastrin and chromogranin A levels or between levels of chromogranin A and pancreastatin-like immunoreactivity. The study demonstrates an elevation of both chromogranin A and pancreastatin-like immunoreactivity in serum of gastrinoma patients. The lack of correlation between gastrin and chromogranin A, however, gives an indication that the gastrinoma cells are not the main source of serum chromogranin A elevation.

Endocrine neoplasms of the pancreas: a clinicopathologic study of 24 cases and immunohistochemical remarks.

A total of 24 patients with endocrine neoplasms of the pancreas were clinicopathologically and immunohistochemically studied. They consisted of 18 patients with adenoma and 6 with carcinoma. Of the 24 patients, 13 developed attacks of hypoglycemia due to hyperinsulinemia, and 1 developed an uncontrollable duodenal ulcer caused by the hypersecretion of gastrin, however, the remaining 10 were asymptomatic. No prediction could be made as to the site of origin of the tumors. A clear difference was seen between adenoma and carcinoma in the size of the mass, the mean greatest diameter of the 18 adenoma cases being 1.7 cm, while that of the 6 carcinoma cases was 7.3 cm. One of the 13 insulinomas and a gastrinoma was malignant, while all 24 tumors were positive for neuron-specific enolase. The 13 insulinomas were diffusely positive for insulin and 5 were also shown to be focally immunoreactive for gastrin, with 3 also being immunoreactive for somatostatin and 2 for pancreatic polypeptide. The gastrinoma showed immunoreactivity for somatostatin, insulin, pancreatic polypeptide, and glucagon in addition to a positivity to gastrin. The above findings thus indicate the multiple hormone synthesis of endocrine neoplasms of the pancreas.

Natural killer cell activity in patients with neuroendocrine tumours of the gastrointestinal tract; relation with circulating gastrointestinal hormones.

We have evaluated the peripheral blood natural killer (NK) cell activity and the in vitro effect of recombinant gamma-interferon (r gamma-IFN) on NK cell activity in 23 patients with a neuroendocrine tumour of the pancreas, small intestine or liver, and 23 healthy controls. Patients with a gastrinoma showed a NK cell activity which was not different from that of the control group, whereas patients with another type of neuroendocrine tumour had a decreased NK cell activity compared to the controls (p less than 0.05) and the gastrinoma patients (p less than 0.02). The impaired NK cell activity in these patients was as such not related to the presence of liver metastasis or performance status of the patients. r gamma-IFN significantly stimulated the NK cell activity in patients and controls. However, the cytotoxic response of the patients with a hormone production other than gastrin remained lower than in the two other groups. Follow-up studies in 8 patients showed NK cell activities not to vary with stable disease, to decrease with progressive disease, and to increase with regression of disease. In conclusion, NK cell activity is suppressed in patients with neuroendocrine tumours that produce hormones other than gastrin. This impairment is not related to the presence of metastasis but seems to be related to the course of the disease.

Immunohistochemical characterization of gastrinomas with antibodies specific to different fragments of progastrin.

The characterization of the tumors and their metastasis in patients with the Zollinger-Ellison syndrome is currently based on the immunohistochemical identification of gastrin cells. However, sometimes tumoral cells fail to react with common C-terminal gastrin antibodies. In order to clarify this failure, we carried out morphologic, morphometric and immunocytochemical analyses performed on light and electron microscope levels of 6 pancreatic and 1 metastatic gastrinomas, using antibodies raised against various sequences of human progastrin. On the basis, in light microscopy, of qualitative analysis of immunostaining within cells and of immunostained cell numbers, gastrin 34 residue seemed to be the prominent form in 2 of the tumor tissues, G-17 in 1 tumor which was not responsive with C terminus progastrin and N terminus G-34 antisera, and progastrin in the metastatic tissue that did not contain typical gastrin (G-like) cells. Two tumors failed to react with all antisera used. At the electron microscope level, immunogold staining revealed that progastrin was present only in the progranules and gastrin 34 in both progranules and intermediate granules. Quantitative studies performed on 3 tumors showed that, within a given tumoral cell, about 25 percent of progranules contained progastrin while 75 percent contained gastrin 34. We concluded that different forms of gastrin can be immunodetected in a gastrinoma tissue, depending on the regions, and that the distribution of progastrin fragments is variable from tumor to tumor. So, specific antibodies to different fragments of progastrin may help to the characterization of gastrinomas.