The present study was aimed at localizing gamma-aminobutyric acid (GABA) and its enzyme of synthesis, glutamic acid decarboxylase (GAD), in the mouse pancreas by immunocytochemical methods. The influence of GABA on hormone release was also studied with normal mouse and rat islets and the isolated perfused rat pancreas. Particular attention was paid to glucagon release to test a recent hypothesis suggesting that GABA mediates the still unexplained glucose-induced inhibition of glucagon release. GABA and GAD were identified only in islet cells and never in the exocrine tissue. Exogenous GABA, baclofen (agonist of GABAB receptors), muscimol (agonist of GABAA receptors), or bicuculline (antagonist of GABAA receptors) did not affect insulin and somatostatin release by isolated mouse or rat islets. GABA was also without effect on glucose-induced electrical activity in mouse B-cells. Glucagon secretion by mouse islets was only slightly inhibited (approximately 20%) by GABA. Since muscimol had a similar effect, and baclofen was ineffective, the inhibition by GABA probably involves GABAA receptor activation. Bicuculline, however, did not antagonize the inhibitory effects of GABA and muscimol, probably because the antagonist alone also decreased glucagon secretion. In contrast to GABA, low (3 mM) and high (20 mM) concentrations of glucose strongly inhibited (approximately 50-65%) glucagon release; this inhibition was not prevented by bicuculline. Similar results were obtained with the perfused rat pancreas; muscimol slightly inhibited glucagon release under various conditions, and bicuculline did not reverse the strong inhibition produced by 16.7 mM glucose. In conclusion, GABA does not affect insulin and somatostatin secretion, but inhibits A-cells, probably by acting on GABAA receptors. It is unlikely, however, that this small inhibitory effect can account for the inhibition of glucagon release produced by glucose.

The transformation of fenugreek subfractions, rich in steroid saponins, was studied upon their passage through the digestive tract to determine the contribution of saponins and/or diosgenin and other steroid sapogenins to the hypocholesterolemic effect of fenugreek seeds. Feces of alloxan diabetic dogs fed fenugreek subfractions were analyzed, and diosgenin, smilagenin and gitogenin were identified and measured using capillary gas chromatography/mass spectrometry. Our results show that saponins are, in part (about 57%), hydrolyzed into sapogenins in the digestive tract. It appears that saponins may be implicated, alone or together with diosgenin, in the observed hypocholesterolemic effect of fenugreek seeds in diabetic dogs.

The effect of l‐glutamate has been studied on insulin secretion by the isolated perfused pancreas of the rat. The glutamate receptor subtype involved has been characterized. In the presence of a slightly stimulating glucose concentration (8.3 mm), l‐glutamate (5 × 10−5−4 × 10−3 m) induced an immediate, transient and concentration‐dependent insulin response. On the other hand, in the presence of a non stimulating glucose concentration (2.8 mm), l‐glutamate (10−3 m) did not modify the basal insulin secretion. The three non‐NMDA receptor agonists, kainate (10−4−10−3 m), α‐amino‐3‐hydroxy‐5‐methyl‐4‐isoxazolepropionic acid (AMPA, 5 × 10−5−10−4 m) and quisqualate (5 × 10−6−5 × 10−5 m) all provoked a transient and concentration‐dependent insulin response from pancreas perfused with 8.3 mm glucose. Compared with glutamate, kainate exhibited a similar efficacy, whereas AMPA and quisqualate elicited only a 3 fold lower maximal insulin response. In contrast, NMDA (10−4−10−3 m) was ineffective. An antagonist of non‐NMDA receptors, 6‐cyano‐7‐nitroquinoxaline‐2,3‐dione (CNQX; 5 × 10−5 m) totally prevented the stimulatory effect of l‐glutamate (4 × 10−4 m) and kainate (2 × 10−4 m). In contrast, the NMDA receptor antagonist, (+)‐5‐methyl‐10,11‐dihydro‐5H‐dibenzo[a,d]cyclohepten‐5,10‐imine ((+) MK801) was without effect. The insulin secretory effect of glutamate (4 × 10−4 m) was not affected by atropine (3 × 10−7 m) or tetrodotoxin (3 × 10−6 m). Quisqualate at a high maximally effective concentration (4 × 10−4 m) inhibited glutamate (10−3 m) or kainate (4 × 10−4 m)‐induced insulin release. This study shows that l‐glutamate stimulates insulin secretion in rat pancreas, by acting on an excitatory amino acid receptor of the AMPA subtype.

We have previously shown that the antidiabetic property of fenugreek seeds ( Trigonella foenum graecum L.) is associated with the defatted seed material which is rich in fibers, saponins, and proteins. In the present work this defatted preparation was divided into two subfractions: subfraction "a" which contains the testa and endosperm and is rich in fibers (79.6%); and subfraction "b" which contains the cotyledons and axes and is rich in saponins (7.2%) and proteins (52.8%). We investigated the effects of each subfraction on hyperglycemia and the levels of pancreatic hormones when chronically administered to alloxan-diabetic dogs. Each subfraction was studied separately and was given to the dogs per 0s (mixed with the two daily meals), in addition to the insulin treatment (which was kept the same throughout the experiment) for a period of 2 1 days. The addition of subfraction "a" to insulin treatment resulted in a clear decrease of hyperglycemia and glycosuria accompanied by a reduction of the high plasma glucagon and somatostatin levels in diabetic dogs. The treatment also decreased the hyperglycemic response to the oral glucose tolerance test. In contrast the chronic administration of subfraction "b" had no effect on hyperglycemia or on the levels of pancreatic hormones in diabetic dogs. Our results show that the antidiabetic properties of fenugreek seeds are contained in the testa and endosperm. Although this subfraction is rich in fibers (high viscosity; 1 15 cP), it is not possible to exclude the existence of one or more unknown active pharmacological compounds in this subfraction of the seed.