Insulin, glucagon, somatostatin and pancreatic polypeptide cells were quantified after immunoperoxidase staining in sections of pancreases obtained from nine control subjects and seven diabetic patients with primary or secondary iron overload. One was normoglycaemic, two had glucose intolerance and four presented insulin-requiring diabetes. The whole pancreas was studied, taking into account the heterogeneous distribution of the endocrine cells. In the diabetic patients, the weight of the pancreas tended to be lower. Iron overload predominated in the exocrine tissue, whereas in islets iron concentration was quite variable from case to case. At the Haemalun-Eosine staining the histological appearance of the islets was normal, their shape and size being unchanged; amyloid deposits were absent, as were atrophic islets. Immunoperoxidase staining revealed a severe reduction in the number of immunoreactive B cells in the four diabetic patients. The mass of immunoreactive B cells was calculated from their volume density and from the weight of each lobe of the pancreas. It averaged 950 mg in control subjects, 1580 mg in the normoglycaemic patient, 1010 mg in patients with glucose intolerance and 180 mg in insulin-requiring diabetic patients. The electron microscopic examination, performed in four cases, revealed that the iron deposits were restricted to B cells and associated with progressive loss of their endocrine granules. The study shows that the pancreatic islet abnormalities in iron overloaded diabetic patients are completely different from those of Type 1 (insulin-dependent) and Type 2 (non-insulin-dependent) diabetic patients. This constitutes a further argument for a specific role of iron in the pathogeny of diabetes in haemochromatotic patients.

Pancreases from insulin-dependent diabetics (IDDM), noninsulin-dependent diabetics (NIDDM), and nondiabetic subjects were analyzed by stereological and morphometrical methods in order to determine the weight of the lobe rich in pancreatic polypeptide (PP) cells in relation to the total weight of the pancreas and the volume density of PP cells in both parts of the gland, those rich and poor in PP cells. In control subjects, neither the relative weight of the lobe rich in PP cells, nor the volume density of PP cells varied significantly with aging. In IDDM and NIDDM, the volume density of PP cells was similar to that observed in control subjects. The total weight of the pancreas was markedly decreased in IDD because of an almost selective atrophy of the lobe poor in PP cells; the relative weight of the lobe rich in PP cells was thus much higher than that in control subjects. A less marked atrophy, restricted to the lobe poor in PP cells, was also observed in NIDDM. It is suggested that PP may exert a local trophic role which protects the lobe of the pancreas rich in PP cells from atrophy in diabetic patients. The results further show that the elevated levels of PP in the plasma of elderly or diabetic subjects cannot be ascribed to a hyperplasia of PP cells.