1 The direct eects of diazoxide on mitochondrial membrane potential, Ca 2+ transport, oxygen consumption and ATP generation were investigated in mouse pancreatic B-cells and rat liver mitochondria. 2 Diazoxide, at concentrations commonly used to open adenosine 5'-triphosphate (ATP)-dependent K + -channels (K ATP channels) in pancreatic B-cells (100 to 1000 mM), decreased mitochondrial membrane potential in mouse intact perifused B-cells, as evidenced by an increase of rhodamine 123¯uorescence. This reversible decrease of membrane potential occurred at non-stimulating (5 mM) and stimulating (20 mM) glucose concentrations. 3 A decrease of mitochondrial membrane potential in perifused B-cells was also caused by pinacidil, but no eect could be seen with levcromakalim (500 mM each). 4 Measurements by a tetraphenylphosphonium-sensitive electrode of the membrane potential of rat isolated liver mitochondria con®rmed that diazoxide decreased mitochondrial membrane potential by a direct action. Pretreatment with glibenclamide (2 mM) did not antagonize the eects of diazoxide. 5 In Fura 2-loaded B-cells perifused with the Ca 2+ channel blocker, D 600, a moderate, reversible increase of intracellular Ca 2+ concentration could be seen in response to 500 mM diazoxide. This intracellular Ca 2+ mobilization may be due to mitochondrial Ca 2+ release, since the reduction of membrane potential of isolated liver mitochondria by diazoxide was accompanied by an accelerated release of Ca 2+ stored in the mitochondria. 6 In the presence of 500 mM diazoxide, ATP content of pancreatic islets incubated in 20 mM glucose for 30 min was signi®cantly decreased by 29%. However, insulin secretion from mouse perifused islets induced by 40 mM K + in the presence of 10 mM glucose was not inhibited by 500 mM diazoxide, suggesting that the energy-dependent processes of insulin secretion distal to Ca 2+ in¯ux were not aected by diazoxide at this concentration. 7 The eects of diazoxide on oxygen consumption and ATP production of liver mitochondria varied depending on the respiratory substrates (5 mM succinate, 10 mM a-ketoisocaproic acid, 2 mM tetramethyl phenylenediamine plus 5 mM ascorbic acid), indicating an inhibition of respiratory chain complex II. Pinacidil, but not levcromakalim, inhibited a-ketoisocaproic acid-fuelled ATP production. 8 In conclusion, diazoxide directly aects mitochondrial energy metabolism, which may be of relevance for stimulus-secretion coupling in pancreatic B-cells.

The effect of glucose stimulation (25 mM for 5 min) on the phospholipid and neutral lipid composition of isolated pancreatic islets was studied to find out whether there is a change in the mass of potential lipid mediators or modulators of insulin secretion. For comparison, the lipid compositions of homogenates and subcellular fractions from RINm5F insulin-secreting tumor cells and of glucose-stimulated streptozotocin/nicotinamide-induced islet cell tumors were analyzed. After separation of the lipid extract into a neutral and an acidic fraction by anion-exchange chromatography, lipids were separated by high-performance thin-layer chromatography and quantitated by in situ densitometry of the cupric sulfate-charred bands. In glucose-stimulated islets, the molar percentages of phosphatidic acid (PA) and of phosphatidylinositol were significantly increased (3.1 vs. 4.7 mol% and 8.6 vs. 11.8 mol%), while those of all other phospholipids and neutral lipids, including 1,2-diacylglycerol, were not significantly changed. In stimulated islet cell tumors, an increase of PA was visible in the microsomal fraction, and there was an increase of lysophosphatidylcholine in the mitochondrial fraction. However, in both tumoral tissues, particularly in RINm5F cells, the lipid distribution pattern showed abnormalities which can be regarded as a loss of differentiation and which limit the usefulness of these tissues for the study of the physiological regulation of lipid metabolism during glucose stimulation. In conclusion, the data are in accordance with a role of PA early in stimulus-secretion coupling. The well-known stimulation of phospholipid synthesis in pancreatic islets during glucose-induced insulin secretion does not result in an increase in the total phospholipid mass.

The role of plasma membrane depolarization as a determinant of the initial phase of insulin secretion was investigated. NMRI mouse islets and beta-cells were used to measure the kinetics of insulin secretion, ATP and ADP content, membrane potential, and cytosolic free Ca(2+) concentration ([Ca(2+)](i)). The depolarization of metabolically intact beta-cells by KCl corresponded closely to the theoretical values. In contrast to physiological (glucose) or pharmacological (tolbutamide) ATP-sensitive K(+) (K(ATP)) channel block, KCl depolarization did not induce action potential spiking. The depolarization by 15 mM K(+) (21 mV) corresponded to the plateau depolarization by 50 or 500 microM tolbutamide; that by 40 mM K(+) (41 mV) corresponded to the action potential peaks. Nifedipine and diazoxide abolished action potentials but not KCl depolarization, suggesting that the depolarizing strength of 15, but not 40 mM K(+) corresponds to that of K(ATP) channel closure. K(+) (40 mM) induced a massive secretory response in the presence of 5 mM glucose, whereas 15 mM K(+), like 50 microM tolbutamide, was only slightly effective, even though a marked increase in [Ca(2+)](i) was produced. Raising glucose from 5 to 10 mM in the continued presence of 15 mM K(+) resulted in a strongly enhanced biphasic response. The depolarization pattern of this combination could be mimicked by combining basal glucose with 15 mM K(+) and 50 microM tolbutamide; however, the secretory response to these nonnutrients was much weaker. In conclusion, the initial secretory response to nutrient secretagogues is largely influenced by signaling mechanisms that do not involve depolarization.

No abstract