The effect of adenylate cyclase inhibitor (ACI) on guanylate cyclase, phosphodiesterase and other enzymes in heart.

The effect of an inhibitor of adenylate cyclase (ACI) was measured on some enzymes associated with cyclic nucleotide-regulated metabolism. Soluble guanylate cyclase was inhibited; both soluble and particulate cyclic GMP-phosphodiesterases were stimulated. Cyclic AMP phosphodiesterases were unaffected. In contrast, the activities of Na, K-ATPase, protein kinase, phosphorylase kinase, glycogen synthetase and a number of glycosidases were not altered by equipotent amounts of the inhibitor. It is concluded that this substance acts as a modulator of both cyclic AMP and cyclic GMP metabolism in heart and other tissues.     

Mode of Action of the Crustacean Hyperglycemic Hormone

Glucose levels in crayfish hemolymph are enhanced by the crustaceanhyperglycemic hormone (CHH); at present there is some evidencethat this action is mediated by cyclic nucleotides. CHH is capableof stimulating adenylate cyclase in the abdominal muscle. Thereis an increase in cyclic AMP and cyclic GMP contents in hepatopancreasand abdominal muscle after CHH injection. Cyclic nucleotidesare able to evoke the same reaction as CHH in vivo and in vitro.Cyclic nucleotide-dependent protein kinases are activated bythe hormone, which leads to a phosphorylation and thereforeinhibition of glycogen synthase. So far, an effect of purifiedhormone on phosphorylase and phosphorylase kinase has not beendemonstrated in the abdominal muscle.     

Cyclic AMP does not mediate the action of synthetic hypertrehalosemic peptides from the corpus cardiacum of Periplaneta americana

Two synthetic peptides identical to those present in the corpus cardiacum of the American cockroach, Periplaneta americana, were tested for their effect on the production of cyclic AMP and the activation of glycogen phosphorylase in cockroach fat body. The peptides activate glycogen phosphorylase and promote trehalose production in incubated tissue when calcium is included in the incubation medium, but have no obvious effect on cyclic AMP levels. The lack of effect of the peptides on cyclic AMP production was confirmed in a fragmented membrane preparation. By contrast, an aqueous extract of corpus cardiacum activates glycogen phosphorylase, promotes trehalose production and elevates cyclic AMP levels in incubated tissue; the extract also enhances cyclic AMP production in the fragmented cell membrane preparation. Observations on the nature of cyclic AMP production in cockroach fat body indicate that the adenylate cyclase has a requirement for GTP and magnesium ions, is stimulated by fluoride and forskolin and, therefore, is similar to the adenylate cyclase complex of other eukaryotes.The results suggest that increases in intracellular calcium concentrations may mediate the expression of hypertrehalosemic effects by the synthetic peptides.     

Postnatal development of glycogen- and cyclic AMP-metabolizing enzymes in mammalian skeletal muscle

Glycogen and cyclic AMP-metabolizing enzymes of rabbit skeletal muscle were examined during postnatal development. Glycogen synthase I, glycogen phosporylase and lactate dehydrogenase activity increased 7-fold by the 6th--8th postnatal week while glycogen synthase D was present in the neonate at one-half adult levels. Cyclic AMP phosphodiesterase decreased; adenylate cyclase increased 10-fold for both the epinephrine and NaF-stimulated states of the enzyme.     

ACTIVATION OF GLYCOGEN PHOSPHORYLASE IN RAT CAUDATE NUCLEUS SLICES BY l-ISOPROPYLNOREPINEPHERINE AND DIBUTYRYL CYCLIC AMP

Abstract— l -Isopropylnorepinepherine (IPNE), 3-isobutyl-1-methylxanthine (IBMX) and N⁶,O²'-dibutyryl cyclic AMP have been found to stimulate the conversion of glycogen phosphorylase (GPase) from b to a forms in rat caudate nucleus slices. The average percentage of total GPase in the a form in control incubations was 32%. The percentage of total GPase in the a form was increased to 1.5 times the control value in the presence of 1 mM-IBMX, to twice the control value in the presence of 0.05 mM-IPINE and to 2.5 times the control in the presence of 0.05 mM-IPNE and 1 mM-IBMX in combination. The increase in GPase activation correlated well with the elevation of cyclic AMP levels by these agents in caudate slices. The percentage of total GPase in the a form was also increased to 2.5 times the control by 1 mM dibutyryl cyclic AMP. Dopamine (DA) and 2-chloroadenosine (2-CI-Ado), which also elevate cyclic AMP levels in rat caudate slices, were without significant effect on GPase. The results indicate that the β-adrenergc adenylate cyclase in the rat caudate nucleus plays a role in the regulation of glycogen metabolism, while the DA-stimulated adenylate cyclase is not significantly involved. 2-CI-Ado does have effects on glycogen metabolism in the caudate, but these effects do not appear to be mediated by GPase activation.     

The cardiac sarcoplasmic reticulum-glycogenolytic complex. A possible effector site for cyclic AMP.

Cardiac sarcoplasmic reticulum-glycogenolytic complex, isolated as a single peak on sucrose density gradient, may function as a "compartmented" effector site for cyclic AMP resulting in modulation of both glycogenolysis and calcium transport. The conversion of phosphorylase b to a is stimulated by ATP and inhibited by protein kinase inhibitor. Cyclic AMP alone stimulated neither phosphorylase b to a conversion nor calcium uptake. An inhibitor of adenylate cyclase depressed both calcium uptake and phosphorylase activation and both functions were subsequently stimulated by micromolar concentrations of cyclic AMP. Endogenous phosphorylation of sarcoplasmic reticulum was also inhibited by adenylate cyclase inhibitor and the inhibition was reversed by cyclic AMP. These results suggest that the sarcoplasmic reticulum of cardiac muscle is an internal effector site for cyclic AMP which may regulate both calcium and metabolism. It appears that cyclic AMP formation in vitro is not the rate-controlling step in the activation sequence.     

Changes in brain cyclic AMP metabolism and acetylcholine and dopamine during narcotic dependence and withdrawal.

Effects of morphine administration were studied on cyclic AMP metabolism in several regions of rat brain. In the cortex, cerebellum and thalamus-hypothalamus, morphine dependence did not alter the activity of either adenylate cyclase or phosphodiesterase. However, during withdrawal from the opiate treatment, adenylate cyclase activity declined in all three regions studied. In contrast, the striatal cyclic AMP metabolism was enhanced during morphine treatment as reflected by elevated endogenous cyclic AMP and increased adenylate cyclase. Furthermore, narcotic dependence produced significant increases in acetylcholinesterase activity of rat striatum. Whereas morphine withdrawal reversed the changes in striatal acetylcholine levels and acetylcholinesterase activity, the enhanced striatal dopamine remained unaltered. Although the activity of striatal adenylate cyclase was significantly reduced when compared to the morphine-dependent rats, the drop in cyclic AMP levels was not significant. Methadone replacement did not affect the changes in striatal dopamine seen in morphine-withdrawn rats. Whereas dopamine stimulated equally well the striatal adenylate cyclase from control or morphine-dependent animals, it failed to stimulate the striatal enzyme from rats undergoing withdrawal. The crude synaptosomal fraction of the whole brain from morphine-dependent rats exhibited an increase in cyclic AMP which was accompanied by elevated adenylate cyclase and protein kinase activity. Naloxone administration suppressed this rise in cyclic AMP and reversed the morphine-stimulated increases in the activities of adenylate cyclase and protein kinase. Following the withdrawal of morphine treatment, alterations in cyclic AMP metabolism were similar to those noted in morphine-naloxone group. Furthermore, substitution of morphine with methadone antagonized the observed alterations in cyclic nucleotide metabolism during withdrawal.     

STIMULATION OF GLYCOGENOLYSIS IN RAT CAUDATE NUCLEUS SLICES BY L-ISOPROPYLNOREPINEPHERINE, DIBUTYRYL CYCLIC AMP AND 2-CHLOROADENOSINE

Dopamine (DA), L-isópropylnorepinepherine (IPNE), 2-chloroadenosine (2-Cl-Ado), 3-isobutyl, I-methylxanthine (IBMX) and N⁶,O^2′-dibutyryl cyclic AMP have been investigated for their effects on glycogen levels in rat caudate nucleus slices. Incubation of slices with 1 mM-dibutyryl cyclic AMP, or with 50 μM-IPNE in the presence of 1mM-IBMX, or with 5-500 μM-2-Cl-Ado reduced glycogen levels to about 50% of control. Incubation of slices with 1 mM-IBMX alone, or with 50μM-IPNE alone, or with 50μM-DA either alone or in the presence of 1 mM-IBMX was without significant effect on glycogen levels. The effect of IPNE + IBMX was completely abolished by the prior addition of 10 μM-propranolol. The effect of 10 μM-2-Cl-Ado was not effectively prevented by either 100 μM-theophylline or 100 μM-cordycepin. The results indicate that the β-adrenergic adenylate cyclase in the rat caudate nucleus plays a role in the regulation of glycogen metabolism, while the DA-stimulated adenylate cyclase is not significantly involved. The 2-Cl-Ado-stimulated adenylate cyclase may be involved in the control of glycogen metabolism, but other mechanisms for the 2-CI-Ado action, such as interference with allosteric regulatory sites on glycogen phosphorylase, have not been ruled out.     

The cardiac sarcoplasmic reticulum-glycogenolytic complex A possible effector site for cyclic

Cardiac sarcoplasmic reticulum-glycogenolytic complex, isolated as a single peak on sucrose density gradient, may function as a “compartmented” effector site for cyclic AMP resulting in modulation of both glycogenolysis and calcium transport. The conversion of phosphorylase b to a is stimulated by ATP and inhibited by protein kinase inhibitor. Cyclic AMP alone stimulated neither phosphorylase b to a conversion nor calcium uptake. An inhibitor of adenylate cyclase depressed both calcium uptake and phosphorylase activation and both functions were subsequently stimulated by micromolar concentrations of cyclic AMP. Endogenous phosphorylation of sarcoplasmic reticulum was also inhibited by adenylate cyclase inhibitor and the inhibition was reversed by cyclic AMP. These results suggest that the sarcoplasmic reticulum of cardiac muscle is an internal effector site for cyclic AMP which may regulate both calcium and metabolism. It appears that cyclic AMP formation in vitro is not the rate-controlling step in the activation sequence.     

Studies on adenylate cyclase-cyclic AMP system of the myopathic hamster (UM-X7.1) skeletal and cardiac muscles.

Cyclic AMP content, adenylate cyclase (EC 4.6.1.1) activity and phosphodiesterase I (EC 3.1.4.1) activity of the hind leg skeletal muscle and cardiac muscle in 60- and 150-day-old normal and myopathic (UM-X7.1) hamsters were examined. In 60-day-old myopathic animals, cardiac cyclic AMP levels were higher and phosphodiesterase I activity was lower, without any changes in the basal adenylate cyclase activity, whereas in 150-day-old myopathic hamsters, cardiac cyclic AMP and basal adenylate cyclase activity were lower, without any changes in the homogenate phosphodiesterase I activity. On the other hand, basal adenylate cyclase and phosphodiesterase I activities in the skeletal muscle homogenate from 60- and 150-day-old myopathic animals were not different from the normal values but the skeletal muscle cyclic AMP levels were significantly less in 60-day-old myopathic hamsters only. The plasma cyclic AMP levels in 60-day-old myopathic hamsters, unlike 150-day-old myopathic animals, were higher than the normal. Although these results reveal differences in myopathic cardiac and skeletal muscles, it is concluded that changes in adenylate cyclase-cyclic AMP system in myopathy are dependent upon the degree of disease.     

Attenuation of epinephrine-induced increase in liver cyclic AMP by endogeneous insulin in vivo.

1. Epinephrine-induced increase in rat liver cyclic AMP in vivo was potentiated when the circulating insulin was suppressed by injection of anti-insulin serum or by induction of diabetes. Consequently, phosphorylase was activated, glycogen synthetase was inactivated and glycogen accumulation induced by glucose load was prevented by epinephrine in the insulin-deficient rats to a much larger extent than in normal rats. 2. Insulin lack was effective in potentiating epinephrine-induced increase in liver and muscule cyclic AMP even after the treatment of rats with theophylline; the potentiation could not be solely accounted for by the inhibition of cyclic AMP phosphodiesterase. Thus, it is likely that insulin lack enhaces epinephrine activation of adenylate cyclase. 3. Unlike epinephrine, glucagon increased liver cyclic AMP to essentially the same extent whether the rat was treated with anti-insulin serum or not. 4. Based on the difference in dose-response curves between normal and insulin-deficient rats, a possibility is discussed that there are two adenylate cylase in the liver with higher and lower affinities for epinephrine and that circulating insulin blocks the high affinity enzyme selectively.     

Adenosine 3',5'-monophosphate formation by preparations of rat liver soluble guanylate cyclase activated with nitric oxide, nitrosyl ferroheme, S-nitrosothiols, and other nitroso compounds

Cyclic AMP formation from ATP was stimulated by unpurified and partially purified soluble hepatic guanylate cyclase in the presence of nitric oxide (NO) or compounds containing a nitroso moiety such as nitroprusside, N-methyl-N-nitro-N-nitrosoguanidine (MNNG), nitrosyl ferroheme, and S-nitrosothiols. Cyclic AMP formation was undetectable in the absence of NO or nitroso compounds and was not stimulated by fluoride or glucagon, indicating the absence of adenylate cyclase activity. The nitroso compounds failed to activate, whereas fluoride or glucagon activated, adenylate cyclase in washed rat liver membrane fractions. Cyclic GMP formation from GTP was markedly stimulated by the soluble hepatic fraction in the presence of NO or nitroso compounds. Cyclic AMP formation by partially purified guanylate cyclase was competitively inhibited by GTP and cyclic GMP formation is well-known to be competitively inhibited by ATP. Therefore, it appears that activated guanylate cyclase, rather than adenylate cyclase, was responsible for the formation of cyclic AMP from ATP. Formation of cyclic AMP of cyclic GMP was enhanced by thiols, inhibited by hemoproteins and oxidants, and required the addition of either Mg2+ or Mn2+. Further, several nitrosyl ferroheme compounds and S-nitrosothiols stimulated the formation of both cyclic AMP and cyclic GMP by the soluble hepatic fraction. These observations support the view that soluble guanylate cyclase is capable, under certain well-defined conditions, of catalyzing the conversion of ATP to cyclic AMP.     

The control of glycogen metabolism in yeast. 1. Interconversion in vivo of glycogen synthase and glycogen phosphorylase induced by glucose, a nitrogen source or uncouplers

The addition of glucose to a suspension of yeast initiated glycogen synthesis and ethanol formation. Other effects of the glucose addition were a transient rise in the concentration of cyclic AMP and a more prolonged increase in the concentration of hexose 6-monophosphate and of fructose 2,6-bisphosphate. The activity of glycogen synthase increased about 4-fold and that of glycogen phosphorylase decreased 3-5-fold. These changes could be reversed by the removal of glucose from the medium and induced again by a new addition of the sugar. These effects of glucose were also obtained with glucose derivatives known to form the corresponding 6-phosphoester. Similar changes in glycogen synthase and glycogen phosphorylase activity were induced by glucose in a thermosensitive mutant deficient in adenylate cyclase (cdc35) when incubated at the permissive temperature of 26 degrees C, but were much more pronounced at the nonpermissive temperature of 35 degrees C. Under the latter condition, glycogen synthase was nearly fully activated and glycogen phosphorylase fully inactivated. Such large effects of glucose were, however, not seen in another adenylate-cyclase-deficient mutant (cyr1), able to incorporate exogenous cyclic AMP. When a nitrogen source or uncouplers were added to the incubation medium after glucose, they had effects on glycogen metabolism and on the activity of glycogen synthase and glycogen phosphorylase which were directly opposite to those of glucose. By contrast, like glucose, these agents also caused, under most experimental conditions, a detectable rise in cyclic AMP concentration and a series of cyclic-AMP-dependent effects such as an activation of phosphofructokinase 2 and of trehalase and an increase in the concentration of fructose 2,6-bisphosphate and in the rate of glycolysis. Under all experimental conditions, the rate of glycolysis was proportional to the concentration of fructose 2,6-bisphosphate. Uncouplers, but not a nitrogen source, also induced an activation of glycogen phosphorylase and an inactivation of glycogen synthase when added to the cdc35 mutant incubated at the restrictive temperature of 35 degrees C without affecting cyclic AMP concentration.     

Hyperglucagonemia and altered responsiveness of hepatic adenylate cyclase-adenosine 3′,3′-monophosphate system to hormonal stimulation during chronic ingestion of dl-ethionine

Basal activity and hormonal responsiveness of the adenylate cyclase-adenosine 3′,5′-monophosphate system were examined in premalignant liver from rat chronically fed the hepatic carcinogen DL-ethionine, and these data were correlated with endogenous levels of plasma glucagon. By 2 weeks basal hepatic cyclic AMP levels, determined in tissue quick-frozen in situ, were 2-fold higher in rats ingesting ethionine than in the pair-fed control. Enhanced tissue cyclic AMP content was associated with an increase in the adenylate cyclase activity of whole homogenates of fresh liver from rats fed ethionine (68 ± 5 pmol cyclic AMP/10 min per mg protein) compared to control (48 ± 4). Cyclic AMP-dependent protein kinase activity ratios were also significantly higher (control, 0.38 ± 0.04; ethionine 0.55 ± 0.05) and the percent glycogen synthetase activity in the glucose 6-phosphate-independent form was markedly reduced (control, 52 ± 7%; ethionine, 15 ± 1.5 %) in the livers of ethionine-fed rats compared to the controls, suggesting that the high total hepatic cyclic AMP which accompanied ethione ingestion was biologically effective. These changes persisted throughout the 38 weeks of drug ingestion. Immunoreactive glucagon levels, determined in portal venous plasma, were 8-fold higher than control after 2 weeks of the ethionine diet (contro, 185 ± 24 pg/ml; ethionine, 1532 ± 195). Analogous to the changes in hepatic parameters, plasma glucagon levels remained elevated during the entire period of drug ingestion until the development of hepatomas. The hepatic cyclic AMP response to a maximal stimulatory dose of injected glucagon was blunted in vivo in ethionine-fed rats (control, 14-fold increase over basal, to 8.63 ± 1.1 pmol/mg wet weight; ethionine, 4.6-fold rise over basal, to 5.42 ± 0.9). Reduced cyclic AMP responses to both maximal and submaximal glucagon stimulation were also evident in vitro in hepatic slices prepared from rats fed the drug, and the reduction was specific to glucagon. Absolute or relative hepatic cyclic AMP responses to maximally effective concentrations of prostaglandin E₁ or isoproterenol in hepatic slices from ethionine-fed rats were greater than or equal to those observed in control slices. Parallel alterations in hormonal responsiveness were observed in adenylate cyclase activity of whole homogenates of these livers, implying that the changes in cyclic AMP accumulation following hormone stimulation were related to an alteration in cyclic AMP generation in the premalignant tissue.In view of the recognized hepatic actions of glucagon and the desensitization of adenylate cyclase which can occur during sustained stimulation of the liver with this hormone, the endogenous hyperglucagonemia that accompanies ethionine ingestion could play a role in the pathogenesis of both the basal alterations in hepatic cyclic AMP metabolism and the reduced responsiveness to glucagon observed in liver from rats fed this carcinogen.     

Metabolic and cyclic nucleotide enzyme activities in muscle and nonmuscle cells of rat heart during perinatal development

Enzyme activities related to aerobic metabolism and cyclic nucleotides were evaluated in muscle and nonmuscle cells of rat heart. The perinatal period from weaning to adult was studied. Malate dehydrogenase, citrate synthase, and 3-hydroxyacyl-CoA dehydrogenase activities of nonmuscle cells equal or exceed muscle cell activities in the weanling heart. Aerobic enzymes remain unchanged in nonmuscle cells during growth; however, muscle cell activities are enhanced. Adenylate cyclase and guanylate cyclase activities are higher in heart homogenates of weanling than adult rats. Despite elevated adenylate cyclase activity, cyclic AMP levels are identical in weanling and adult rats. Cyclic GMP levels are twofold higher in weanling than in adult rats. Muscle cell metabolism and cyclic nucleotide levels are associated with growth-related changes in heart function and cellularity, respectively.     

Forskolin Stimulates Adenylate Cyclase Activity, Cyclic AMP Accumulation, and Adrenocorticotropin Secretion from Mouse Anterior Pituitary Tumor Cells

Abstract: The effects of forskolin, an adenylate cyclase activator, were investigated on adrenocorticotropin (ACTH) secretion from AtT-20/D16-16 mouse pituitary tumor cells. Forskolin increased adenylate cyclase activity in these cells in the absence of added guanyl nu-cleotide, an effect blocked by somatostatin. Cyclic AMP synthesis and ACTH secretion increased in a concentration-dependent manner, not only in the clonal cells, but in primary cultures of rat anterior pituitary as well. Somatostatin inhibited cyclic AMP synthesis and ACTH secretion in response to forskolin. When forskolin was coapplied with corticotropin releasing factor, cyclic AMP synthesis was potentiated and ACTH secretion additive. The calcium channel blocker, nifedipine, inhibited forskolin, and 8-bromocyclic AMP stimulated ACTH secretion. These data suggest that ACTH secretion may be regulated at the molecular level by changes in cyclic AMP formation, which in turn regulate a calcium gating mechanism.     

Regulation of the synthesis of the major surfactant apoprotein in fetal rabbit lung tissue

Antibodies directed against the major apoprotein of rabbit lung surfactant, a 29-36-kDa glycoprotein, were used to study changes in the levels of translatable surfactant apoprotein mRNA in rabbit lung tissue during development, as well as the effects of cortisol and cyclic AMP analogues on the levels of surfactant apoprotein and its mRNA in fetal rabbit lung tissue in organ culture. The major surfactant apoprotein and its mRNA were undetectable in lung tissues of 21-day gestational age fetal rabbits. Translatable mRNA specific for the major surfactant apoprotein was first detectable in lung tissues of 26-day fetuses, increased 25-fold on day 28, reached peak levels at day 31, and declined after birth. Incubation of 21-day fetal rabbit lung explants with cortisol in serum-free medium resulted in an increase in the specific content of the 29-36-kDa apoprotein. Cyclic AMP analogues and forskolin, an activator of adenylate cyclase, also caused a marked increase in the accumulation of surfactant apoprotein. When fetal lung explants were incubated with cortisol and dibutyryl cyclic AMP in combination, the specific content of the surfactant apoprotein was increased to levels greater than that of explants treated with either cortisol or dibutyryl cyclic AMP alone. These effects of dibutyryl cyclic AMP and cortisol on surfactant apoprotein accumulation were associated with comparable changes in the levels of translatable surfactant apoprotein mRNA. Thus, we have shown for the first time that the induction of pulmonary surfactant apoprotein synthesis during differentiation in vitro and in vivo is associated with an increase in the level of translatable mRNA and that cortisol and cyclic AMP increase both the accumulation of the major surfactant apoprotein and the corresponding mRNA in fetal rabbit lung tissue in vitro.     

Developmental changes in cyclic AMP, protein kinase, phosphorylase kinase, and phosphorylase in liver, heart, and skeletal muscle of the rat

Changes in cyclic AMP, protein kinase, phosphorylase kinase, and phosphorylase levels were examined during development in the rat. In liver, cyclic AMP increased prenatally and for the first 10 postnatal days; protein kinase levels (both cyclic AMP-dependent and independent activities) were high prenatally and declined during the first 10 postnatal days. Both phosphorylase and phosphorylase kinase in liver increased rapidly prenatally and more slowly postnatally. In heart and skeletal muscle cyclic AMP increased prenatally and for the first 10 days after birth, then declined. Protein kinase in both these tissues was highest prenatally and declined perinatally. In heart and skeletal muscle phosphorylase and phosphorylase kinase activities were extremely low prenatally although both enzymes were largely in their activated forms. Postnatally the nonactive form of both enzymes increased greatly throughout 30 postnatal days. In all three tissues, particularly heart and skeletal muscle, these changes could not be correlated with levels of tissue glycogen.     

Glycogen metabolism and cyclic AMP levels in isolated islets of lean and genetically obese mice.

The levels of glycogen and cyclic AMP, incorporation of glucose into glycogen and activities of glycogen synthetase and phosphorylase were determined in pancreatic islets isolated from genetically obese mice and their lean litter-mates. Islets from obese mice had elevated glycogen levels, increased phosphorylase activity and an increased amount of glycogen synthetase in the physiologically more effective I-form, indicating an increased turnover of glycogen. There was no significant difference in cyclic AMP levels between islets of lean and obese mice, but inhibition of phosphodiesterase or stimulation of adenyl cyclase increased cyclic AMP levels more in obese than in lean mouse islets, indicating a more rapid turnover of cyclic AMP in the former. It is suggested that cyclic AMP stimulated phosphorolytic breakdown of glycogen may be one of the mechanisms responsible for the increased insulin secretory response to glucose observed in islets from genetically obese mice.