Responsiveness of glycogen breakdown to cyclic AMP in perfused liver from rats with insulin-induced hypoglycemia

! " # $ % " &! " # $ # " ' ( ( )* + * ( ( " "( # $ ( " " Correspondence R.B. Bazotte Departamento de Farmácia e Farmacologia Universidade Estadual de Maringá 87020-900 Maringá, PR Brasil Fax: +55-44-263-6231 E-mail: rbbazotte@uem.br Research supported by CNPq, CAPES and PRONEX (No. 168/97). S. Pagliarini e Silva and K.F. Nascimento are recipients of CNPq fellowships. Received December 17, 2001 Accepted September 4, 2002


Introduction
During insulin-induced hypoglycemia, hepatic glucose production must increase to match the energy demands of the brain.In the fed state, when liver glycogen stores are present, the hepatic response to insulininduced hypoglycemia operates primarily through glycogenolysis (1,2) rather than gluconeogenesis (3).Furthermore, there is considerable evidence that during insulininduced hypoglycemia glucagon and epinephrine are important counter-regulatory hormones involved in the activation of glycogen breakdown (4,5).
Recently we demonstrated (6) that the administration of insulin at pharmacological levels was capable of inhibiting isoproterenol-induced hepatic glycogen breakdown.Because the activation of hepatic glycogenolysis promoted by ß-adrenergic agonists is mediated by adenosine-3'-5'-cyclic monophosphate (cAMP), we decided to investigate the participation of cAMP in this effect (6).
Since the cellular levels of cAMP and analogues are proportional to the extracellular concentration used in the liver perfusion experiments (7), it is possible to investigate the hepatic responsiveness of glycogen break-down to cAMP during insulin-induced hypoglycemia.For this purpose, we used in situ perfused liver.This technique has the advantage of determining metabolic rates on the basis of the composition of the perfusate in the intact organ (8,9).Furthermore, in order to obtain information about the participation of insulin in hepatic responsiveness to cAMP, additional experiments measuring the direct effect of insulin on glycogen catabolism promoted by cAMP and analogues were performed.

Material and Methods
Male Wistar fed rats (200-220 g) were employed.Insulin rats received an intraperitoneal (ip) injection of 1 IU/kg of Neosulin ® R (regular insulin obtained from Biobrás, Montes Claros, MG, Brazil).Control rats were injected with the same volume of saline.One hour after the administration of insulin or saline the rats were anesthetized ip with 40 mg/kg sodium pentobarbital.After laparotomy, blood was collected from the vena cava for the measurement of glucose (10) and insulin (11).As expected, insulin rats showed increased (P<0.05)insulinemia (436.0 ± 13.2 µIU/ml, N = 23) compared to control rats (47.6 ± 3.8 µIU/ml, N = 23) and decreased (P<0.05)glycemia (42.1 ± 2.0 mg/dl, N = 23) compared to the control group (139 ± 2.4 mg/dl, N = 23).Since hypoglycemia was established 60 min after insulin injection, this period of time was chosen to carry out the first set of experiments with isolated perfused livers.Thus, 60 min after the administration of insulin or saline the rats were anesthetized ip with 40 mg/kg sodium pentobarbital.After laparotomy, the livers were perfused as previously described (6,8,9,12).
For N 6 ,2'-O-dibutyryl-adenosine-3'-5'-cyclic monophosphate (DB-cAMP) the concentrations needed for 50 and 100% catabolism were 0.10 and 0.15 µM, respectively (results not shown).Therefore, after a pre-infusion period (10 min) livers from insulin and control rats were perfused for 20 min (10-30 min) with cAMP or DB-cAMP dissolved in the perfusion fluid followed by a post-infusion period (20 min) to allow a return to basal levels.Thus, the activation of glucose (10), L-lactate (13) and pyruvate (14) production was taken to be the difference between the rates of release of these compounds during (10-30 min) and before (0-10 min) the infusion of cAMP or DB-cAMP.Glycogen catabolism was calculated to be the sum of glucose plus the half-sum of Llactate and pyruvate [glucose + 1/2 (L-lactate + pyruvate)].The release of these metabolites provides the rate of glycogenolysis because pyruvate oxidation, pentose-monophosphate shunt and recycling of pyruvate to glucose are minimal (15).All metabolic measurements are expressed as µmol .min -1 .g -1 .
In the second set of experiments the direct effect of insulin on the kinetics of the activation of glycogen catabolism promoted by 3 µM cAMP was investigated.Thus, after a pre-infusion period of 10 min, livers from control rats were perfused for 15 min (10-25 min) with 3 µM cAMP, followed by a combined infusion of 3 µM cAMP and 500 µIU/ ml insulin for 10 min (25-35 min).The concentration of insulin used was similar to that obtained for insulin rats 60 min after insulin administration.The effect of insulin on the activation of glycogen catabolism promoted by cAMP was measured as described above.
In the third set of experiments the direct effect of insulin (500 µIU/ml) on the activation of glycogen catabolism promoted by 3 and 15 µM cAMP, 0.10 and 0.15 µM DB-cAMP, 3 µM 8-bromo-adenosine-3'-5'cyclic monophosphate (8Br-cAMP) or 3 µM N 6 -monobutyryladenosine-3'-5'-cyclic mono-phosphate (6MB-cAMP) was investigated.Thus, after a pre-infusion period of 10 min, livers from control rats were perfused for 20 min (10-30 min) with a combined infusion of insulin (500 µIU/ml) and a cyclic nucleotide (cAMP, DB-cAMP, 8Br-cAMP or 6MB-cAMP) followed by a post-infusion period of 20 min to allow a return to basal levels.The activation of glycogen catabolism was measured as the difference between the rates of glycogenolysis during (10-30 min) and before (0-10 min) the infusion of the cyclic nucleotides.
The computer program GraphPad Prism (version 2.0) was used to calculate the area under the curve, expressed as µmol/g liver fresh weight.Data were analyzed statisti-cally by the unpaired Student t-test.A 95% level of confidence (P<0.05) was accepted for all comparisons.Results are reported as means ± SEM.

Results
In the first set of experiments the effect of cAMP and DB-cAMP on glycogen catabolism in livers from control and insulin rats was compared.As shown in Figure 1, the infusion of cAMP or DB-cAMP promoted a rapid increase in hepatic glycogenolysis.However, the values obtained for cAMP (3 µM) and DB-cAMP (0.10 µM) were lower (P<0.05) in the insulin group (Figure 1A,C).In contrast, the increase of glycogenolysis Because insulin showed a direct effect on the activation of glycogen catabolism promoted by cAMP (Figure 2), we assessed the effect of insulin perfusion on glycogen catabolism promoted by cAMP and analogues (DB-cAMP, 8Br-cAMP, 6MB-cAMP).As shown in Figure 3, activation of hepatic glycogenolysis with cAMP (3 and 15 µM) and DB-cAMP (0.1 and 0.15 µM) was significantly reduced (P<0.05)when insulin was present in the perfusion fluid (Figure 3).In contrast, the increase of glycogenolysis promoted by 3 µM 8Br-cAMP (Figure 4A) or 3 µM 6MB-cAMP (Figure 4B) was not influenced by the presence of insulin.

Discussion
Hepatic glycogen breakdown in the liver is regulated in an opposite manner by insulin and cAMP-elevating agents.Recent studies from our laboratory have shown decreased responsiveness of glycogen breakdown to isoproterenol in livers from rats with insulininduced hypoglycemia (6).Because postreceptor mechanisms mediated by cAMP are important to activate hepatic glycogenolysis during insulin-induced hypoglycemia, we investigated this second messenger.Thus, by using cAMP and DB-cAMP at levels that produce about 50% of the maximal effect on glycogen catabolism (3 and 0.10 µM, respectively), we found a decreased hepatic responsiveness of glycogen catabolism during insulin-induced hypoglycemia (Figure 1A,C).However, no difference (Figure 1B,D) was observed when the concentrations of cAMP and DB-cAMP which produce the maximal glycogen catabolism were employed (15 and 0.15 µM, respectively).
To overcome the influence of counterregulatory factors which might have been released during insulin-induced hypoglycemia we examined the direct effect of insulin on hepatic glycogenolysis promoted by cAMP (3 µM).As shown in Figure 2, insulin decreased (P<0.05) the hepatic responsiveness to cAMP, with the effect occurring as early as 3 min after the beginning of insulin infusion.
Since insulin showed a direct effect on hepatic responsiveness to cAMP we decided to study the direct effect of insulin on glycogenolysis promoted by cAMP and DB-cAMP at the concentrations used in the first set of experiments.Thus, in contrast to insulin rats, we observed decreased activation of glycogenolysis (Figure 3) not only with 3 µM cAMP or 0.1 µM DB-cAMP but also with 15 µM cAMP or 0.15 µM DB-cAMP.Thus, the direct effect of insulin was more intense than that observed in the livers from rats with insulin-induced hypoglycemia (Figure 1).
The direct effect of insulin on the activation of glycogen catabolism promoted by 15 µM cAMP or 0.1 µM DB-cAMP can be obtained with 0.3 µM 8Br-cAMP (data not shown).Since 8Br-cAMP is more resistant than cAMP to hydrolysis by phosphodiesterases, it is possible to overcome the influence of these enzymes by employing higher concentrations of 8Br-cAMP.Therefore, for comparison with the results obtained with 3 µM cAMP we employed 3 µM 8Br-cAMP.
In view of the fact that the activation of glycogen catabolism promoted by 3 µM 8Br-cAMP was not influenced by insulin infusion (Figure 4A) and considering that this cAMP analogue operates as a protein kinase agonist (16), we may suggest that insulin affects the amount of cAMP in the liver.
Whereas 8Br-cAMP is susceptible to hydrolysis by phosphodiesterase 3B, 6MB-cAMP is resistant to hydrolysis by this enzyme due to the covalent modification at the N 6 position (17).Thus, using 6MB-cAMP at a concentration which permits comparison with cAMP and 8Br-cAMP (i.e., 3 µM), we that the activation of glycogen catabolism promoted by 3 µM 6MB-cAMP was not influenced by insulin infusion (Figure 4B).Consequently, the direct effect of insulin appears to be mediated mainly by the phosphodiesterase present in the liver, i.e., phosphodiesterase 3B.
Recently we reported that the decreased hepatic responsiveness to isoproterenol occurred at least 30 min after insulin administration (1).In contrast, we show here that the direct effect of insulin on hepatic responsiveness to cAMP occurred as early as 3 min after the beginning of insulin infusion (Figure 2).Since insulin acutely decreases the activation of glycogenolysis promoted by cAMP, this suggests the activation of phosphodiesterase 3B (18) and inhibition of adenylate cyclase (19).
Thus, our previous report (1) and the results obtained here suggest that the decreased hepatic responsiveness to isoproterenol is mediated not only by the adrenergic desensitization promoted by the release of catecholamines during insulin-induced hypoglycemia (6,20), but also by a direct effect of insulin decreasing the cellular levels of cAMP.Additionally, we suggest that both mechanisms operated simultaneously in the liver, but that the decreased responsiveness of glycogenolysis mediated by insulin occurred prior to the adrenergic desensitization mediated by catecholamines released during insulin-induced hypoglycemia.

Figure 1 .
Figure 1.Effect of adenosine-3'-5'-cyclic monophosphate (cAMP) (panels A and B) and dibutyryl cAMP (DB-cAMP) (panels C and D) on glycogenolysis.The data are reported as the means of 4-6 perfused livers obtained from rats killed 60 min after the injection of saline (control, squares) or insulin (1 IU/kg, lozenges).The area under the curve (AUC) data (µmol/g liver fresh weight ± SEM) are presented in the figure.The statistical test used was the unpaired Student t-test.Glycogenolysis (µmol min -1 g -1) 4