Insulin inhibits its own secretion from isolated, perifused human pancreatic islets

It is still a controversial question whether insulin suppresses its own secretion. We prepared pure human islets from three pancreases by collagenase digestion and density gradient purification. Aliquots of 200 islet equivalents (IE, 150-μm sized-islets) were sequentially perifused at 37°C with 3.3 mmol/l glucose (3.3G, 40 min), 16.7 mmol/l glucose (16.7G, 30 min) and again 3.3G (30 min) after 24 h, 37°C culture in CMRL 1066 medium with or without the addition of either 200 or 400 μU/ml human insulin in the incubation medium (6 replicates each). Insulin secretion was assessed by C-peptide (Cp) measurement in the perufusate. Without added insulin (C) and with 200 (Ins200) or 400 (Ins400) μU/ml added insulin, basal Cp release was 0.12±0.03, 0.14±0.02 and 0.14±0.04 ng/ml, respectively. At 16.7G, the first-phase secretion peak (expressed as Cp value) was significantly lower with Ins200 (0.47±0.13 ng/ml,P<0.02) and Ins400 (0.68±0.15 ng/ml,P<0.05) than C (0.83±0.15 ng/ml). The second-phase secretion peak was also significantly (P<0.05) reduced with added insulin (Ins200: 0.47±0.08 ng/ml; Ins400: 0.45±0.07 ng/ml) than in its absence (C: 0.65±0.09 ng/ml). Accordingly, total Cp secretion was lower with Ins200 (10.6±2.3 ng/ml,P=0.03) and Ins400 (11.8±2.3 ng/ml) than with C (16.0±2.2 ng/ml). Thus, the addition for 24 h of either 200 or 400 μU/ml insulin in the culture medium caused a significant decrease of insulin (as assessed by Cp measurement) secretion from perifused human islets, suggesting that feedback suppression of insulin release is at least in part due to a direct action of insulin on the islets.