GENERAL
AND
COMPARATIVE
2-Deoxyglucose
ENDOCRINOLOGY
81,276-283 (1991)
Stimulates the Release of Insulin and Somatostatin from the Perfused Catfish Pancreas PETER RONNER
Department of Biochemistry and Biophysics, and Diabetes Research Center, School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104 Accepted March 5, 1990 Glucokinase was proposed to function as a glucose sensor in pancreatic B-cells, acting possibly as a pacemaker of the rate of glycolysis. Glucose, mannose, and 2-deoxyglucose are good substrates of glucokinase which are easily taken up into B-cells. Glucose and mannose are well-known stimuli of insulin release in mammals and fish. I report here that 2-deoxyglucose is also a strong stimulus of insulin and somatostatin release from the in vitro perfused pancreas (i.e., splenic Brockmann body) of channel catfish (Ictaturus punctatus). This is surprising because the product of the glucokinase-catalyzed phosphorylation of 2-deoxyglucose, 2-deoxyglucose-6-phosphate, cannot be metabolized further at an appreciable rate. 3-0-Methylglucose, which does not bind appreciably to mammalian glucokinase, stimulated neither insulin nor somatostatin release. Glucosamine, which binds tightly to glucokinase hut is phosphorylated only at a very low rate, did not stimulate insulin release either, but did cause a small amount of somatostatin to be released. The results suggest that glucose-activated glucokinase itself may serve as a signal molecule in glucose recognition by B- and D-cells. 0 FBI Academic Press, hc.
Glucose stimulates the release of insulin and somatostatin in mammals (reviewed by Matschinsky et al., 1979; Hedeskov, 1980; Gerich, 1981) and in fish (Ince, 1979; Ince and So, 1984; Ronner and Scarpa, 1982, 1984, 1987). Glucose enters glycolysis after phosphorylation by either hexokinase or glucokinase (Meglasson and Matschinsky, 1984). Glucokinase has been proposed to act as a glucose sensor in pancreatic endocrine cells (Meglasson and Matschinsky, 1984, 1986). This is plausible in light of the following facts. (a) Mammalian glucokinase is half-maximally active at about 9 mM glucose, while insulin release is half-maximally stimulated by about 10 n&f glucose (Meglasson and Matschinsky, 1984). (b) Glucokinase resembles the ancestral hexokinase B of yeast which undergoes a conformational change upon binding glucose and is also involved in “glucose repression” (i.e., the inhibition of the expression of other sugar metabolizing enzymes as
long as glucose is present) (Andreone et al., 1989; Frohlich et al., 1985; Stachelek et al., 1986; Steitz et al., 1981). (c) In mammals, the expression of glucokinase appears to be confined to the liver and the pancreatic islets (Iynedjian et al., 1986). (d) Islet glucokinase differs from liver glucokinase in its N-terminus and may lack a central 17amino acid sequence (Magnuson and Shelton, 1989). (e) Starvation leads to a reduction in the levels of both glucokinase and glucose-induced insulin release (Burch et al., 1981). (f) Glucokinase is a mnemonic enzyme with respect to its kinetic behavior: it shows cooperativity while being monomeric (Storer and Cornish-Bowden, 1976; Pollard-Knight and Comish-Bowden, 1982; Cornish-Bowden and Pollard-Knight, 1985; Cardenas et al., 1978; Ricard et al., 1974). (g) Glucose increases the intrinsic fluorescence of glucokinase which strongly suggests a conformational change (Cardenas et al., 1985). 276
0016~6480/91 $1.50 Copyright All rights
0 1991 by Academic Press, Inc. of reproduction in any form reserved.
2-DEOXYGLUCOSE-INDUCED
Current models generally ignore the potential signaling properties of glucokinase molecules; rather, they focus on the ability of glucokinase to set the rate of the glycolytic flux (Meglasson and Matschinsky, 1986). The glycolytic flux is thought to affect ATP-sensitive potassium channels via the cytosolic concentration of ATP (Cook et al., 1988). The ATP-sensitive potassium channels in turn participate in the regulation of the membrane potential which ultimately affects the rate of exocytosis of secretory granules (Ashcroft, 1988). We have shown previously that the perfused cattish pancreas releases insulin and somatostatin in response to glucose and mannose, but not in response to either fiuctose or galactose (Ronner and Scarpa, 1987). This pattern fits the substrate specificity of mammalian glucokinase (Salas et al., 1965; Parry and Walker, 1966; Ureta et al., 1979). 2-Deoxyglucose is the third known and effective substrate of mammalian glucokinase that can enter the pancreatic endocrine cells via the glucose transporter (Salas et al., 1965; Parry and Walker, 1966; Ureta ef al., 1979; Bamett et al., 1973). If indeed glucokinase functions as a glucose sensor in fish endocrine cells, I surmised that 2-deoxyglucose might stimulate the release of insulin and somatostatin. This was indeed the case while other analogs of glucose (glucosamine, 3-O-methylglucose), which are not efficient substrates of glucokinase, did not induce any hormone release. These results may be an indication of a direct role of glucokinase in glucose recognition in fish pancreatic B- and Dcells. Future work will have to show whether cattish pancreatic endocrine cells contain glucokinase-like enzyme activity that is capable of phosphorylating 2-deoxyglucose. It will also be of interest to determine why 2-deoxyglucose is an effective secretagogue in cattish but not in rats (Grodsky et al., 1963; F. M. Matschinsky, personal communication).
INSULIN
277
RELEASE
MATERIALS
AND METHODS
Animals. Gray channel catfish (Zcralurus punctatus) were purchased from the Kurtz Fish Hatchery (Elverson, PA) and maintained indoors in aquaria at 21-23” (Ronner and Scarpa, 1987). Newly acquired fish were preventively treated in a solution of 50 mM NaCl, 0.25 mM formaldehyde, and 7.5 &liter of malachite green for 30 min. The fish were fed a pelleted “pool fish food” (Carolina Biological Supply Co., Burlington, NC). Blood samples taken from anesthetized fish at the time of surgery showed blood glucose levels of 3-6 mM (mean 2 SD: 3.6 +- 1.6 mM; n = 9), as measured with the glucose-oxidase technique in a Beckman ghtcase analyzer. The blood had a hematocrit of 26-38% (31 * 4; n = 10). The fish used for experiments were 23-33 cm long (28 f 3 cm; n = 11) and weighed 96-280 g (186 f 66; n = 11). Perfusion of the pancreas. Fish were anesthetized with MS-222 (3aminobenzoic acid ethylester methanesulfonate) and the splenic Brockmann body was microsurgically prepared for in vitro perfusion as described before (Ronner and Scarpa, 1982, 1987). The basic perfusion medium had the following composition (in nuV): NaCI, 130; KCl, 3; MgSO,, 0.5; CaCI,, 1; glucose, 2; N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid (Hepes)/NaOH, 10; BSA (containing