Pulsatile insulin secretion by human pancreatic islets.

Insulin is secreted in discrete bursts. These pulses are also present when individual or groups of islets are perifused. Interpretation of the measured frequency and magnitude of pulsatile hormone secretion requires an examination of the sensitivity and specificity of the methods for pulse detection and validation of these for the experimental apparatus and hormone assay in which they are applied. In the present study we achieve these aims for a perfusion method for measurement of pulsatile insulin release by human islets. A deconvolution technique previously developed for measurement of pulsatile hormone secretion in vivo was specifically validated for in vitro pulse detection in the present study. Deconvolution analysis reliably (>90%) detected insulin pulses with an amplitude 20% or more above baseline and recovered quantitatively the insulin secretion profile, insulin secretion rate, and insulin pulse mass from single as well as multiple perifused islets. Cluster analysis was less sensitive, but was able to detect most (>80%) pulses with an amplitude of 40% or more above baseline. With this limitation, cluster analysis is potentially useful for groups, but not single perifused human islets. Analysis of single human islets showed that enhanced insulin secretion by increased glucose concentrations in the perfusate is achieved by enhancing insulin pulse mass with no change in pulse frequency. Perfused single or groups of human islets exhibited an interpulse interval ( approximately 6-8 min) comparable to that observed in humans in vivo. Dynamic in vitro perifusion should facilitate studies of the mechanisms driving pulsatile insulin secretion.