Regulation by glucose and cyclic nucleotides of the glucagon and insulin release induced by amino acids.

The glucagon and insulin release induced by amino acids was studied in the presence of glucose, dibutyryl cyclic AMP (dbcAMP) or theophylline on the splenic part of the pancreas of new born rats (48 to 64 hours). The results were compared to the literature data. Arginine or a mixture of the three amino acids (A.A.), arginine, lysine or alanine, stimulate glucagon secretion at 1.6 mM glucose. This stimulation is suppressed by 16.7 mM glucose. On the other hand, 16.7 mM glucose potentiates the effect of arginine or of the 3 A.A. on insulin release. At 1.6 mM glucose, theophylline potentiates the effect of 3 A.A. (10 mM each) on glucagon and insulin release : this effect reaches a maximum at 5 mM of theophylline; dbcAMP also potentiates the effect of 3 A.A. on glucagon and insulin release, and the effect of arginine, alanine or lysine on glucagon release. On the beta cell, the lack of potentiation observed between dbcAMP and arginine, lysine or alanine indicates that these A.A. interact positively when mixed together. In the presence of arginine or of the three A.A., the percentage stimulation of glucagon and insulin release depends on the dbcAMP dose and does not vary with the glucose concentration. The increase of glucagon and insulin release observed when the NaCl concentration in the incubation medium decreases cannot account for our results. Cyclic GMP (4 mM) does not modify the glucagon or insulin secretion induced by different concentrations of glucose or by the mixture of A.A. (10 mM each). The stimulating effect of acetylcholine on insulin release would not be related to the cyclic GMP molecule. In conclusion, instead of modifying the specificity of substrate, theophylline or dbcAMP accentuate it: glucose stimulates specifically the beta cell whereas 3 A.A. are more effective on the alpha2 cell than the beta cell. Cyclic AMP suppresses the glucose effect on glucagon release induced by the amino acids. Because of its interaction with glucose and amino acids, cyclic AMP seems to be a very important element in the regulation of the release of these pancreatic hormones.     

Does nitric oxide play a role in stimulus-secretion and/or synthesis-secretion coupling of insulin?

By using the perfused rat pancreas model and the nitric oxide synthase inhibitor, nitro-arginine methyl ester (NAME), we examined the hypothesis that first-phase potentiation or second-phase inhibition of insulin release due to the amino acid arginine (or both) are the result of its conversion to nitric oxide (NO). In the presence of 16.7 mM glucose, 20 mM arginine caused a first-phase potentiation of insulin release when compared with glucose controls, while inhibiting the second-phase insulin release. When 20 mM NAME was added in addition to 20 mM arginine and 16.7 mM glucose, the total insulin released during the first secretory phase was not significantly different from that of the glucose plus arginine group, suggesting that inhibition of NO production does not affect arginine-potentiated first-phase insulin release. Similarly, the presence of NAME failed to reverse the arginine inhibition of second-phase insulin release. The presence of NAME resulted in a more pronounced inhibition of insulin secretion. Correspondingly, compared with the glucose-only controls, the presence of 20 mM NAME plus 16.7 mM glucose resulted in a significant decrease in insulin release during the second phase, whereas the presence of NAME did not affect first-phase glucose-stimulated insulin release. Thus we conclude that the conversion of arginine to nitric oxide does not play a significant role in glucose-stimulated first-phase potentiation or second-phase inhibition of insulin release due to arginine.     

Somatostatin secretion from monolayer cultures of neonatal rat pancreas.

Monolayer cultures of neonatal rat pancreas have been characterized as an in vitro system for studying SRIF secretion. Marked 12- and 6-fold potentiation of SRIF release occurred with N-2-O-dibutyryl cAMP monosodium salt and theophylline, respectively. High glucose (300 mg/dl) stimulated SRIF release, whereas galactose was without effect. Exogenous insulin did not alter SRIF release, and the SRIF responses to theophylline and glucose were unaffected by the addition of antiinsulin serum to neutralize the insulin released by these agents. Arginine evoked a significant 2-fold increase in SRIF release. Exogenous glucagon produced slight but not significant stimulation of SRIF release. However, after exposure of the cultures to antiglucagon serum to diminish the concentration of glucagon in contact with the SRIF cells, exogenous glucagon produced a marked enhancement of SRIF secretion. These data suggest that glucose, arginine, glucagon, N-2-O-dibutyryl cAMP monosodium salt, and theophylline stimulate SRIF secretion, probably by direct effects on D cells or through mechanisms other than increased insulin secretion. Monolayer cultures of rat pancreas should provide a powerful in vitro system for studying pancreatic SRIF physiology.     

Secretory regulation of endocrine pancreas: Cyclic AMP and glucagon secretion.

Activation of adrenergic beta receptors has been found to stimulate insulin release in vitro that may be mediated through the augmentation of cyclic AMP in the beta cell. The activation of adrenergic alpha receptors in the beta cell inhibits the insulin release. The present studies have shown that isoproterenol (0.62 mug/ml) and sodium dibutyryl cyclic AMP (50 mug/ml) stimulate the insulin secretion and inhibit the glucagon secretion in the presence of 50 mg/100 ml glucose by the isolated pancreatic perfusion of the rat, while norepinephrine (0.5 mug/ml) inhibits the insulin secretion induced by 150 mg/100 ml glucose and stimulates the glucaton secretion. Theophylline (50 mug/ml) does not stimulate the insulin and the glucagon secretion. When norepinephrine is added to theophylline, the output of glucagon does not occur. From these results it can be deduced that the pancreatic alpha cell function may be inhibited by elevation of intracellular cyclic AMP, in contrast to the beta cell function which is stimulated by an increment of cyclic AMP.     

In vitro insulin secretion from the penguin (Pygoscelis papua) pancreas.

In experiments performed at the Almirante Brown Scientific Station (64° 53′ South; 62° 53′ West), arginine induced in penguins significant release of insulin from incubated pieces of pancreas, while glucose acted as stimulus only at very high concentrations. Substances active on the adenyl cyclase system, such as glucagon, theophylline, isoproterenol, and even 3′,5′ cyclic AMP, stimulated insulin release.l-Propranolol inhibited isoproterenol-induced insulin release. Phentolamine, in the doses used, had no effect on either basal or isoproterenol-induced insulin secretion.Ouabain and glibenclamide failed to promote any change in the amount of insulin secreted.The results suggest that (a) in the penguin pancreas arginine rather than glucose may be a prominent stimulus for insulin secretion; (b) the penguin beta cell shows decreased sensitivity to glucose; (c) the adenyl cyclase system plays an important role in the mechanism(s) regulating insulin secretion; (d) under our experimental conditions, adrenergic modulation of insulin release is mediated chiefly through beta receptors.     

Effect of cholera toxin on insulin release in monolayer cultures of the endocrine pancreas

Summary The effect of cholera toxin on insulin release by monolayer cultures of endocrine pancreas has been studied. The toxin has a marked stimulatory effect upon insulin release at concentrations as low as 10^–10M. The toxin had a small effect at low glucose concentrations, but was strongly stimulatory at high glucose concentrations and in the presence of arginine. Its effect could be detected within 30 min of application and only two minutes' exposure to the toxin was required for it to subsequently stimulate release. In comparative studies on insulin release the toxin was equal to, or slightly more potent than, 1.5 M glucagon and significantly more potent than cyclic AMP (10 mM) or theophylline (5 mM).     

Glucagon secretion in rats with non-insulin-dependent diabetes: an in vivo and in vitro study

Non-insulin-dependent diabetes ( NIDD ) was obtained in adult rats following a neonatal streptozotocin injection. Rats with NIDD exhibited a chronic low-insulin response to glucose in vivo, slightly elevated basal plasma glucose values (less than 2 g/l) and low pancreatic insulin stores (50% of the controls). Glucagon secretion was studied in this model, in vivo and in vitro using the isolated perfused pancreas technique. Normal basal plasma glucagon levels were observed in the fed state and were in accordance with normal basal glucagon release in vitro. The pancreatic glucagon stores were normal in the diabetics. In experiments with the perfused pancreas, the increased glucose concentration suppressed glucagon release as readily in the diabetics as in the controls. Moreover 5.5 mM glucose suppressed glucagon release stimulated by 19 mM arginine to the same extent in both groups. These data indicate that the suppression of A cell function by glucose is normal in rats with NIDD . Theophylline and isoproterenol also produced normal glucagon release in diabetics. By contrast, the glucagon secretion in response to arginine was lower in the diabetics. This was observed either in vivo (arginine infusion) or in vitro in the presence or the absence of glucose in the perfusate. But in the presence of theophylline the response to arginine was normalized in the diabetics. Impairment of A cell function of the diabetics is not limited to recognition of amino-acids, since acetylcholine evoked a lower glucagon response in the diabetics than in the controls. These defects are different from those described in their B cells.     

Effect of K+ deprivation and ouabain upon glucagon release

The effects of K+ depletion and ouabain (0.1, 0.5, and 1 mM) on glucagon and insulin release from the in vitro perfused rat pancreas were studied in the absence or presence of arginine (10 mM). In the absence of arginine, K+ depletion resulted in an inhibition of glucagon release, both at a low (2.8 mM) and a higher (8.3 mM) concentration of glucose, whereas ouabain enhanced glucagon secretion. In the presence of arginine, K+ deprivation induced limited and short lived oscillations in the rate of glucagon release at both low and high glucose levels. In the presence of arginine, ouabain induced first a short lived stimulation and, thereafter, a severe and sustained inhibition of glucagon release. The restoration of the normal K+ concentration or the removal of ouabain was followed by a marked and transient inhibition of glucagon release which was most prominent in the presence of arginine. In agreement with previous investigations, K+ deprivation or ouabain facilitated insulin release evoked by glucose and/or arginine. The results suggest that a primary alteration in K+ availability and/or transport affects in a dual manner the release of glucagon, depending on environmental factors such as the type of stimulus used to activate the secretory process.     

Insulin and glucagon secretory responses to arginine, glucagon, and theophylline during perifusion of human fetal islet-like cell clusters

The secretory responsiveness of human fetal pancreatic endocrine cells was studied by perifusion of cultured islet-like cell clusters (ICC). ICC were obtained from 7 fetuses at 13-15 weeks gestation and 21 fetuses at 17-22 weeks gestation. The ICC were challenged with glucose (20 mmol/L), arginine (10 mmol/L), glucagon (1.4 mumol/L), and theophylline (10 mmol/L) combined with zero, low (2 mmol/L), or high (20 mmol/L) glucose. At 13-15 weeks, glucose and arginine enhanced insulin release in some experiments, whereas glucagon and theophylline were always potent stimuli (mean response, 4-fold regardless of the glucose concentration). At 17-22 weeks, both glucose alone (20 mmol/L) and arginine (10 mmol/L, with 2 mmol/L glucose) induced a small (1.4- to 1.5-fold) increase in insulin release. When arginine was combined with 20 mmol/L glucose, the response was potentiated to become a 2.3-fold increase. In contrast, glucagon was equally effective in 2 and 20 mmol/L glucose (2.9- and 2.6-fold responses, respectively) and produced a half-maximal response even in the absence of glucose. In this age range the most potent stimulus for insulin release was clearly theophylline. The effect of theophylline was also remarkably independent of the glucose concentration of the perifusate (5.6-, 8.1-, and 8.6-fold responses at 0, 2, and 20 mmol/L glucose, respectively). Glucagon release from the ICC of the 17- to 22-week-old fetuses was low (mean basal glucagon release, 2.9; insulin, 24.8 fmol/100 ICC/min). Glucagon release was not affected by 20 mmol/L glucose, but was stimulated by arginine and theophylline. These findings suggest that in the fetal pancreas, in contrast to the adult organ, insulin release results from elevation of intracellular cAMP concentrations (by glucagon or theophylline) relatively independent of the exogenous glucose concentration. Therefore, glucagon may have an important role in regulating insulin release during the early development of human fetal B-cells.     

Glucose regulates proinsulin and prosomatostatin but not proglucagon messenger ribonucleic acid levels in rat pancreatic islets

Insulin and glucagon are the major hormones involved in the control of fuel metabolism and particularly of glucose homeostasis; in turn, nutrients tightly regulate insulin and glucagon secretion from the islets of Langerhans. Nutrients have clearly been shown to affect insulin secretion, as well as insulin biosynthesis and proinsulin gene expression; by contrast, the effects of nutrients on proglucagon gene expression have not been studied. We have investigated the effect of glucose, arginine, and palmitate on glucagon release, glucagon cell content, and proglucagon messenger RNA (mRNA) levels from isolated rat islets in 24-h incubations. We report here that concentrations of glucose that clearly regulate insulin and somatostatin release as well as proinsulin and prosomatostatin mRNA levels, do not significantly affect glucagon release, glucagon cell content or proglucagon mRNA levels. In addition, though both 10 mM arginine and 1 mM palmitate strongly stimulated glucagon release, they did not affect proglucagon mRNA levels. We conclude that, in contrast to insulin and somatostatin, glucose does not affect glucagon release and proglucagon mRNA levels, and arginine and palmitate do not coordinately regulate glucagon release and proglucagon mRNA levels.     

Investigations on isolated islets in vitro. IX. Influence of food deprivation and glucose refeeding in vitro on insulin secretion

Summary Starvation resulted in a progressive decrease of insulin secretion in response to 15 mM glucose, but it neither inhibited insulin release completely nor changed the shape of the glucose response curve. Long-term fasting caused the glucose threshold to rise to 30 mM glucose during an incubation time of 15 min, whereas fed animals were characterized by a sigmoidal insulin response and a threshold level of 5 mM glucose during short-term incubation, too. Prolongation of incubation time (60 or 120 min) or pretreatment of isletsin vitro with 20 mM glucose for 60 min caused progressive lowering of the glucose threshold, indicating that glucoseper se influenced the effective stimulating concentration. The effect of starvation is not specific for glucose, since hormone secretion in response to glibenclamide, nicotinic acid and theophylline was also reduced, whereas theophylline can partly counteract the diminished release. Although the content of insulin as well as glucagon in the islets was reduced by food deprivation it is assumed that the islet hormone content is not responsible for the rise of the glucose threshold caused by starvation.     

Insulin and glucagon release in the diabetic Chinese hamster: Differences among inbred sublines

Summary Release of insulin and glucagon from perfused pancreases in vitro of 40 normal male and female Chinese hamsters (from one inbred subline) and 110 male and female diabetic hamsters (from three inbred sublines) was measured in response to glucose plus arginine, theophylline alone, or potassium alone, in order to determine if differences in hormone secretion exist among different diabetic sublines. Glucose plus arginine and potassium produced subnormal insulin responses in all three diabetic sublines, whereas theophylline induced normal or above normal insulin responses. Excessive glucagon release was consistently seen in only one diabetic subline. The female normal animals showed greater insulin release than the male normal hamsters in response to glucose plus arginine. This sex difference was not seen in the diabetic animals.     

Selective impairment of pancreatic A cell suppression by glucose during acute alloxan-induced insulinopenia: in vitro study on isolated perfused rat pancreas

This study was performed in order to investigate the role of insulin in the modulation of pancreatic A cell response to glucose. The isolated perfused rat pancreas model was used: intraislet insulinopenia was induced in vitro by 0.56 mM alloxan infusion over 15 min. Alloxan caused a transitory insulin release but did not affect glucagon secretion. Exposure to alloxan completely abolished insulin response to 20 mM arginine, 1.6 mM glucose, and 11.1 mM glucose. Glucagon response to 20 mM arginine and 1.6 mM glucose was unchanged by alloxan pretreatment compared to control pancreata not treated with alloxan. However, the suppression of glucagon release by 11.1 mM glucose was abolished in the alloxan experiments. Twenty milliunits per ml of insulin infused during 11.1 mM glucose infusion restored glycemic suppression of glucagon release, but it produced only a slight inhibitory effect on A cell function in the presence of 3.9 mM glucose. Our study indicates that glucose is the physiological suppressor of the pancreatic A cell and that, in this regard, insulin exerts only a permissive effect.     

Hyperglucagonemia of the isolated perfused pancreas of diabetic mice (db/db)

Summary Diabetes mellitus is held to be accompanied by inappropriately high levels of plasma glucagon relative to blood glucose concentrations. This has been interpreted as indicating lack of insulin. To establish glucagon release in presence of high levels of endogenous insulin, the effects of both glucose and arginine were studied in the isolated perfused pancreas of genetically diabetic mice (db/db). Stimulation with glucose 2.75 mM or glucose plus arginine 8.25 mM exhibited a pronounced hyperglucagonemia. Following glucose 8.25 mM, however, there was no depression of glucagon secretion. Despite excessive high levels of endogenous insulin, there was a pattern of rather non-suppressible glucagon release. Lack of insulin per se, therefore, is unlikely to be the sole cause of hyperglucagonemia in this type of genetic animal diabetes mellitus.Supported by Deutsche Forschungsgemeinschaft, Bad Godesberg.     

Stimulatory and inhibitory effects of cyclic AMP on pancreatic glucagon release from monolayer cultures and the controlling role of calcium

Summary When glucagon release from monolayer cultures of newborn rat pancreas was measured over four hours in media containing 2.5 mM Ca⁺⁺, a significant cyclic AMP-related inhibition of release was observed. This was noted whether intracellular cyclic AMP levels were raised by the addition of exogenous cyclic AMP or dibutyryl cyclic AMP, by phosphodiesterase inhibition with theophylline, or by the stimulation of adenylate cyclase with cholera toxin. The inhibition was concentration dependent for cyclic AMP and could not be reproduced by the addition of AMP, ADP or ATP. Adenosine also inhibited glucagon release while ATP was stimulatory. From time course studies it appeared that the inhibitory effects of cyclic AMP and cholera toxin were progressive after two hours of incubation. With cholera toxin an early stimulation of glucagon release was observed. The effects of cyclic AMP and cholera toxin on argininestimulated glucagon release were to stimulate further the glucagon release during the first hour of the incubation. Thus, the effects of raising intracellular cyclic AMP levels were biphasic in that both an early stimulation and a late inhibition of glucagon release were observed. In examining the nature of these responses a remarkable controlling role for Ca⁺⁺ was uncovered: at Ca⁺⁺ concentrations of 0.3 mM and lower no effect of cyclic AMP on glucagon release was found. With 1 mM Ca⁺⁺ in the medium cyclic AMP stimulated glucagon release early (30 min) and thereafter had no further effect. In the presence of 2.5 mM Ca⁺⁺ cyclic AMP did not stimulate early but did cause the delayed inhibition of release. It is concluded that the effect of cyclic AMP on glucagon release can be either stimulatory or inhibitory depending upon the Ca⁺⁺ concentration of the medium and the duration of exposure to raised cyclic AMP levels.     

Defective early phase insulin release in perifused isolated pancreatic islets of spiny mice (acomys cahirinus)

In order to characterize pancreatic beta cell function in Geneva bred spiny mice (acomys cahirinus), the dynamics of immunoreactive insulin release were examined during perifusion of pancreatic islets isolated from normoglycemic acomys. The initial insulin response of acomys was slow: no clear-cut early (1 to 10 min) peak of insulin release was observed when glucose in the perifusion medium was abruptly raised from 2.8 mM to concentrations as high as 56 mM. This was true for islets of either young, or older more obese acomys. However, after 20 to 30 min of perifusion at the high glucose concentrations, the rate of insulin release from acomysislets became similar to that from islets of rats or mice. By contrast, glucose-induced insulin release responses observed with islets of Wistar-derived rats, Swiss albino mice, and inbred C57BL/6J lean or obese (ob/ob) mice, were clearly biphasic. Tolbutamide 1.5 mM, arginine 16 mM, and theophylline 10 mM were ineffective in stimulating insulin release from acomys islets in the presence of a substimulatory glucose concentration (2.8 mM), whereas these agents were effective in rat islets at the same substimulatory concentration of glucose. On the other hand, when these agents, as well as cyclic AMP 10 mM or cytochalasin B 10 mug/ml were applied in the presence of a stimulating concentration of glucose (16.8 mM), the glucose-stimulated insulin release from acomys islets was increased to the same or to a greater extent than from rat islets. It is suggested that the failure of all the agents tested to stimulate an early rapid phase of insulin release from acomys islets may be secondary to the observed initial insensitivity to glucose, which insensitivity may in turn reflect a selective impairment in the recognition of glucose as an insulinogenic signal in this species.     

Pancreastatin and islet hormone release.

The effect of pancreastatin on the release of insulin, glucagon, and somatostatin was studied in the isolated perfused rat pancreas. After an initial equilibration period (-20 to 0 min) with a basal glucose concentration (3.3 mM), the pancreata were perfused with either 16.7 mM glucose (0-40 min) or with 20 mM arginine (0-20 min). Pancreastatin was introduced 10 min prior to and throughout the administration of the high glucose and arginine and continued during their perfusion. As expected, the glucose and the arginine augmented insulin and somatostatin release. Pancreastatin (1 and 10 nM) markedly suppressed the first phase of insulin release with both insulinogogues used, while the early somatostatin secretion was not significantly decreased. However, the peak incremental somatostatin response to arginine was reduced by 50% (P less than 0.05). Conversely, the peptide (10 nM) tended to augment arginine-induced glucagon release. Pancreastatin (100 nM) also suppressed glucose-stimulated insulin release from isolated rat islets. These pancreastatin-mediated alterations in islet hormone release are reminiscent of those known to characterize non-insulin-dependent diabetes. Therefore, the significance of pancreastatin in islet physiology and pathophysiology deserves special consideration.     

Acute effects of alloxan- and streptozotocin-induced insulin deficiency on somatostatin and glucagon secretion by the perfused isolated rat pancreatico-duodenal preparation

Summary The secretion of somatostatin and glucagon by the perfused rat pancreatico-duodenal preparation was examined in situ under control conditions and after the induction of acute insulin deficiency by alloxan or streptozotocin. A 10 min 0.625 mmol/l alloxan perfusion resulted in an immediate and transient increase in basal insulin and glucagon release and a slightly delayed and persistent increase in basal somatostatin secretion. The insulin responses to 16.7 mmol/l glucose, 1 mmol/l theophylline, and 19 mmol/l arginine alone or in combination were virtually eliminated by alloxan treatment, Somatostatin secretion in response to the stimuli was completely inhibited or markedly attenuated. The glucagon-suppressive effect of glucose was unaltered by alloxan and the stimulatory effect of arginine was enhanced. Addition of 1 g/ml porcine insulin to the perfusion medium did not modify the alterations in somatostatin and glucagon responses to arginine. Streptozotocin treatment 90 min prior to the onset of perfusion resulted in changes in somatostatin, glucagon, and insulin responses to glucose and arginine similar to those of alloxan. The present results are consistent with an effect of alloxan and streptozotocin on the D cell similar to that on the B cell, namely, interference with a glucose-mediated effect on hormone secretion.     

Structure-function relationships in pancreatic islets: support for intraislet modulation of insulin secretion

Pancreatic islet B cell function was studied in vitro using three structurally different preparations of islet tissues: isolated, intact islets, dispersed islet cells attached singly to microcarrier beads, and reaggregated islet cells. Mechanisms of intercellular communication are eliminated with single cell preparations, whereas in aggregates cell to cell communications are reestablished and a defined microenvironment restored. Perifusion studies measured nonstimulated and glucose- and arginine-stimulated insulin release from the three islet tissues. Insulin secretion rates were expressed as a function of cellular DNA content, permitting direct comparison between tissues. During perifusion with low (2.8 or 5.5 mM) glucose concentrations, secretion rates of single islet cells were up to 6-fold greater (P less than 0.001) than those of intact islets. Perifusion of islet cells with 2.8 mM glucose and 100 or 500 pg glucagon/ml had no effect whereas GH-release-inhibiting factor (330 and 1000 pg/ml) decreased nonstimulated insulin secretion rates by 15% (P less than 0.05). After reaggregation, basal insulin secretion rates were restored toward those of intact islets. Glucose (5.5-30 mM) and L-arginine (5-20 mM) elicited first phase insulin responses from single islet cells that were not significantly different from those observed with intact islets; in contrast, second phase responses of single islets to glucose were approximately 50% those seen with intact islets, and their second phase responses to arginine were absent. Single islet cell first and second phase insulin responses to 5.5 mM glucose were enhanced 2.2-fold (P less than 0.01) and 2.8-fold (P less than 0.05), respectively, in the presence of exogenous glucagon, resulting in secretory profiles characteristic of intact islets. Reaggregation of single islet cells was associated with markedly increased first and second phase insulin responses to both glucose and arginine stimulation. These data show that disruption of the islet microanatomy results in alteration of insulin secretory responses and that these effects can be reversed, in part by exogenous glucagon and GH-release-inhibiting factor, and by reaggregation. Although different mechanisms appear important for nonstimulated, first and second phase insulin release, the findings support a role for both direct intercellular communication and hormonal secretion by islet A and D cells in the modulation of B cell function.     

Interaction of cyclic AMP and alloxan on insulin secretion in isolated rat islets perifused in vitro.

The interaction of alloxan and cyclic AMP on glucose-induced insulin secretion was studied with the use of isolated rat islet perifusion. Simultaneous perifusion with cyclic AMP and alloxan did not protect the islets against the effect of alloxan. However, addition of dibutyryl cyclic AMP to the alloxan solution produced 40% protection of glucose-induced insulin secretion. Partial protection was obtained with either theophylline (41%) or caffeine alone (54%), and the addition of 1 mg/ml glucose to the theophylline or caffeine solution provided greater than 68% protection. The levels of islet tissue cyclic AMP were more than 2.1 times that of islets not protected against the alloxan effect, when partial or nearly complete reversal of the inhibitory action of alloxan on glucose-induced insulin secretion was effected by theophylline or caffeine. These results suggest that cyclic AMP affords partial protection against the effect of alloxan on glucose-induced insulin secretion.