Portal Venous 5-Aminoimidazole-4-Carboxamide-1-β-d-Ribofuranoside Infusion Overcomes Hyperinsulinemic Suppression of Endogenous Glucose Output

Experiments were performed on 23 overnight-fasted mongrel dogs (mean wt 23.2 ± 0.4 kg) of either sex that had been fed a standard diet (Pedigree beef dinner and Wayne Lab Blox; 51% carbohydrate, 31% protein, 11% fat, and 7% fiber based on dry wt) and given free access to water. Animals were housed under controlled temperature (21°C) and lighting conditions, with 12 h of light (6 a.m. to 6 p.m.) and 12 h of dark (6 p.m. to 6 a.m.). Protocols were approved by the Vanderbilt University Animal Care Committee.

At least 16 days before each experiment, a laparotomy was performed under general anesthesia (0.04 mg/kg of atropine and 15 mg/kg pentobarbital sodium before surgery and 1.0% isofluorane inhalation anesthetic during surgery). Catheters were inserted in portal and hepatic veins for blood sampling. An incision was made over the femoral artery into which a silastic catheter (inside diameter 0.04 in) was inserted and advanced to the level of the distal aorta (∼16 cm) for sampling and hemodynamic measurements during experiments. Silastic catheters (inside diameter 0.03 in) were inserted into the vena cava for infusions. A silastic catheter (inside diameter 0.03 in) was inserted into the splenic vein and positioned so that the catheter tip rested just beyond the point where the splenic and portal veins coalesce. This catheter was used for the intraportal infusions of glucagon, insulin, and AICAR. Catheters were filled with heparinized saline, and the free ends were knotted. Ultrasonic transit time flow probes were fitted and secured to the portal vein and hepatic artery (Transonic Systems, Ithaca, NY). The knotted catheter ends and flow probe leads were stored in a subcutaneous pocket in the abdominal region, so that the complete closure of skin incisions was possible.

Blood samples were drawn 3 days before the experiment to determine the leukocyte count and hematocrit of the animal. Only animals with 1) a leukocyte count <18,000/mm³, 2) a hematocrit >36%, 3) a good appetite (consumption of daily food ration), and 4) normal stools were used.

All studies were conducted in dogs fasted for 18 h. Free catheter ends and flow probe leads were accessed through small skin incisions made under local anesthesia in the abdominal and leg regions the morning of the experiment. The contents of the catheters were then aspirated and flushed with saline.

Experiments consisted of a tracer equilibration period (−130 to −30 min), a basal period (−30 to 0 min), and a hyperinsulinemic-hypoglycemic or -euglycemic clamp period (0−150 min). A primed (40 μCi) infusion (0.2 μCi/min) of [3-³H]glucose purified by high-performance liquid chromatography (HPLC) was initiated at t = −130 min. The tracer infusion rate was tripled at t = 0 min in euglycemic groups. A constant-rate indocyanine green infusion (0.1 mg · m⁻² · min⁻¹) was also started at t = −130 min and continued throughout the study. Indocyanine green was used as a backup method of blood flow measurement (21) if the Doppler probes did not provide a clear signal and as confirmation of hepatic vein catheter placement. A peripheral infusion of SRIF (0.8 μg · kg⁻¹ · min⁻¹) was started at t = 0 min to inhibit endogenous insulin and glucagon secretion. Concurrent with the infusion of SRIF was basal intraportal glucagon replacement (0.5 ng · kg⁻¹ · min⁻¹). An intraportal hyperinsulinemic infusion (2 mU · kg⁻¹ · min⁻¹) was also initiated at this time. Arterial glucose was clamped at euglycemic levels (∼105 mg/dl) or allowed to fall to hypoglycemic levels (∼50 mg/dl). In one hyperinsulinemic-hypoglycemic (H-AIC, n = 6) and hyperinsulinemic-euglycemic (E-AIC, n = 6) protocol, a portal venous infusion of AICAR (1 mg · kg⁻¹ · min⁻¹) was initiated at t = 60 min (Fig. 1). In a second hyperinsulinemic-hypoglycemic (H-SAL, n = 6) and hyperinsulinemic-euglycemic (E-SAL, n = 5) protocol, a portal venous infusion of saline was initiated at t = 60 min. Two initial H-AIC studies were performed to determine target glycemic responses for glucose clamping. All other experiments were randomized. In the H-SAL group, glucose was infused into the vena cava to match the increase in glucose seen in the H-AIC protocol. Portal vein and hepatic artery blood flows were monitored on-line throughout the experiments. At the end of the experiment, animals were killed with sodium pentobarbital and an autopsy was performed to confirm catheter placement. Liver biopsies were frozen in liquid nitrogen for later analyses. Blood samples were taken to determine circulating liver enzymes (alanine aminotransferase, aspartate aminotransferase, γ-glutamyl transpeptidase, and lactate dehydrogenase).

Arterial, portal vein, and hepatic vein blood samples were drawn at t = −30, −15, 0, 15, 30, 45, 60, 75, 90, 120, and 150 min in all groups. Arterial blood samples were taken every 10 min to determine plasma glucose−specific activity. Plasma glucose concentrations and radioactivity were determined as previously described (22). Blood samples were deproteinized (1 ml blood in 3 ml of 4% perchloric acid), and whole blood alanine, β-hydroxybutyrate, glycerol, and lactate concentrations were determined using standard enzymatic methods (23). Free fatty acids, cortisol, catecholamines, and immunoreactive insulin and glucagon were measured as previously described (22). Hepatic concentrations of glycogen were measured using enzymatic methods (after digestion of precipitated glycogen with amyloglucosidase) by the method of Chan and Exton (24). Hepatic purine nucleotide concentrations were determined from liver samples homogenized in 0.4 mol/l perchloric acid with 0.5 mmol/l EGTA and neutralized with 0.5 mol/l potassium carbonate (pH 6.8). HPLC analysis and ultraviolet detection were performed to assess nucleotide levels in the samples as previously described (25,26).

Liver samples (∼50 mg) were homogenized in a solution containing 10% glycerol, 20 mmol/l Na-pyrophosphate, 150 mmol/l NaCl, 50 mmol/l HEPES (pH 7.5), 1% NP-40, 20 mmol/l β-glycerophosphate, 10 mmol/l NaF, 2 mmol/l EDTA (pH 8.0), 2 mmol/l phenylmethylsulfonyl fluoride, 1 mmol/l CaCl₂, 1 mmol/l MgCl₂, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 2 mmol/l Na₂VO₃, and 3 mmol/l benzamide. Homogenized samples were assayed for protein concentration using a Pierce BCA protein assay kit (Rockford, IL). Next, 30 μg of protein were run on a SDS-PAGE gel and transferred to a polyvinylidine fluoride membrane. Membranes were treated with rabbit α-phospho-AMPK (Thr¹⁷²) or α-phospho-acetyl-CoA carboxylase (Ser⁷⁹) (Cell Signaling, Beverly, MA) and then incubated with anti-rabbit horseradish peroxidase (Promega, Madison, WI). Chemiluminescence was detected with a WesternBreeze kit (Invitrogen, Carlsbad, CA). Densitometry was performed using Scion Image software (Frederick, MD).

Net hepatic uptake of alanine, β-hydroxybutyrate, free fatty acids, glycerol, and lactate and NHGO were determined as [HAF × (H − A)] + [PVF × (H − P)], in which A, P, and H are the arterial, portal vein, and hepatic vein concentrations, and HAF and PVF are the hepatic artery and portal vein blood flows. The difference in glycogen due to AICAR was calculated as Gly(H-SAL) − Gly(H-AIC) and Gly(E-SAL) − Gly(E-AIC), where Gly(H-SAL) and Gly(E-SAL) are the average hepatic glycogen levels in the respective groups, and Gly(H-AIC) and Gly(E-AIC) are the individual hepatic glycogen levels in AICAR-treated dogs in the presence of hypoglycemia and euglycemia, respectively. Endogenous glucose production (R_a) and glucose utilization (R_d) were determined by isotope dilution (27).

Statistical comparisons among groups and over time were made using ANOVA designed to account for repeated measures. Specific time points were examined for significance using contrasts solved by univariate repeated measures. Liver purine nucleotides were compared using ANOVA. Liver glycogen was compared using t tests. AMPK and ACC phosphorylation were compared using one-tailed t tests. Statistics are reported in tables and figures. Data are presented as means ± SE, and statistical significance was defined as P < 0.05.

Arterial plasma glucose levels were similar during the basal period in all groups and were clamped at euglycemic levels (104 ± 1 and 107 ± 1 mg/dl in the E-SAL and E-AIC groups, respectively) or allowed to fall to similar hypoglycemic levels (49 ± 2 and 51 ± 2 mg/dl in the H-SAL and H-AIC groups, respectively) (Fig. 2). AICAR infusion caused arterial plasma glucose to rise to 61 ± 5 mg/dl by t = 150 min in the H-AIC group. This increase in glucose was simulated in the H-SAL animals to 62 ± 1 mg/dl by t = 150 min. The mean glucose infusion rate required to simulate this rise in H-SAL animals was 1.6 ± 0.1 mg · kg⁻¹ · min⁻¹ (Fig. 3B). The glucose infusion rates required to clamp glucose at euglycemic levels during the last 90 min were similar in the presence and absence of AICAR (11.6 ± 0.4 and 11.5 ± 0.4 mg · kg⁻¹ · min⁻¹ in the E-SAL and E-AIC groups, respectively) (Fig. 3A). Arterial and portal venous plasma insulin levels were similar during the basal period in all groups and rose similarly with clamps in all groups (Table 1, Fig. 2B).

Arterial plasma glucagon was constant throughout the experiments at concentrations that were similar among groups (Table 1). Basal arterial plasma cortisol, epinephrine, and norepinephrine were similar in all groups. Epinephrine and norepinephrine levels rose similarly in hypoglycemic groups so that they were greater than in the euglycemic groups (P < 0.05). Cortisol was higher in the hypoglycemic compared with the euglycemic groups. Cortisol was greater in the H-SAL than in the H-AIC groups (P < 0.05) at t = 120 and 150 min.

NHGO was similar during the basal period among groups and fell similarly during the first 60 min of hyperinsulinemia in each group (Fig. 4). NHGO remained suppressed in the E-SAL and H-SAL groups but was significantly increased in the E-AIC (from −0.3 ± 0.2 at t = 60 min to 1.4 ± 0.4 mg · kg⁻¹ · min⁻¹ at t = 150 min) and H-AIC (from 1.6 ± 0.5 at t = 60 min to 3.4 ± 0.3 mg · kg⁻¹ · min⁻¹ at t = 150 min) groups.

Endogenous R_a was similar during the basal period and fell similarly during the first 60 min of hyperinsulinemia in each group (Fig. 5). R_a remained suppressed in the E-SAL and H-SAL groups but was significantly increased in the E-AIC (from 2.2 ± 0.3 at t = 60 min to 4.6 ± 1.0 mg · kg⁻¹ · min⁻¹ at t = 150 min) and H-AIC (from 1.9 ± 0.5 at t = 60 min to 5.2 ± 0.5 mg · kg⁻¹ · min⁻¹ at t = 150 min) groups. R_d was similar during the basal period among groups and rose similarly during the first 60 min of hyperinsulinemia in the euglycemic (Fig. 6A) and hypoglycemic (Fig. 6B) groups. R_d continued to rise in the E-SAL and E-AIC groups but was significantly greater in the E-AIC compared with the E-SAL animals at t = 120–140 min. R_d fell in the H-SAL animals after t = 60 min. R_d initially fell in the H-AIC group after t = 60 min but then rose and was significantly greater than in the H-SAL group for the final 30 min. Metabolic clearance rates showed similar results (data not shown).

Arterial concentrations and net hepatic balances of alanine, β-hydroxybutyrate (data not shown), free fatty acids, glycerol, and lactate were similar between the E-SAL and E-AIC animals and the H-SAL and H-AIC animals, respectively, during the basal period and first 60 min of hyperinsulinemia (Table 2). Arterial free fatty acid levels were significantly greater in the H-SAL versus the H-AIC animals at t = 120 min. Arterial lactate was significantly greater in the H-AIC versus the H-SAL animals from t = 90 min onward, as the liver had switched from a net uptake to a net output mode.

Liver glycogen and purine nucleotides are shown in Table 3. Liver glycogen was significantly decreased in the H-AIC versus the H-SAL group (P < 0.05). The difference in liver glycogen due to AICAR was significantly greater in the presence of hypoglycemia versus euglycemia (Fig. 7). Liver ZMP was significantly greater in AICAR- versus saline-treated groups in the presence of euglycemia or hypoglycemia (P < 0.05). Phosphorylation of hepatic AMPK (Thr¹⁷²) was similar in all groups (data not shown). Phosphorylation of hepatic ACC (Ser⁷⁹) was significantly greater in AICAR- versus saline-treated groups (P < 0.05) (Fig. 8).

Circulating liver enzymes (alanine aminotransferase, aspartate aminotransferase, γ-glutamyl transpeptidase, and lactate dehydrogenase) were measured in a subset of AICAR-treated dogs. All were within normal ranges, as were circulating levels of albumin and bilirubin.

The present studies demonstrated that an intraportal infusion of AICAR at a rate of 1 mg · kg⁻¹ · min⁻¹ can stimulate hepatic glucose output in the presence of insulin levels that are high enough to suppress the actions of maximally effective glucagon (28). The increase in glucose output was due to a stimulation of net hepatic glycogenolysis. Moreover, these studies showed that intraportal AICAR’s potent stimulation of hepatic glycogenolysis is increased under hypoglycemic conditions. In the presence of an intraportal insulin infusion of 2 mU · kg⁻¹ · min⁻¹, basal intraportal glucagon replacement, and euglycemia (∼100 mg/dl), intraportal AICAR increased R_a by 2.5 ± 1.0 mg · kg⁻¹ · min⁻¹ (Fig. 5) and decreased hepatic glycogen by 5 ± 3 mg/g tissue⁻¹ (Fig. 7). Under hypoglycemic (∼50 mg/dl) conditions, AICAR increased R_a by 3.4 ± 0.4 mg · kg⁻¹ · min⁻¹ and decreased hepatic glycogen by 19 ± 5 mg/g tissue⁻¹. NHGO responses paralleled increases in R_a.

Few studies have examined the hepatic effects of administering AICAR in vivo. Two recent studies have shown that intracerebroventricular infusion of AICAR increases endogenous R_a in mice (29) and rats (30). Two other studies by Bergeron and colleagues (18,19) have shown that primed (100 mg/kg) systemic AICAR doses of 7.5 (18) and 10 (19) mg · kg⁻¹ · min⁻¹ suppress glucose production in overnight-fasted rats. These findings were supported by the fact that injecting mice with adenoviruses expressing dominant negative AMPK (31) or treating hepatocytes with AICAR (32) increases and reduces, respectively, the expression of the key gluconeogenic enzyme, PEPCK. Moreover, AICAR has been shown to inhibit gluconeogenesis in rat hepatocytes (33). Because laboratory rats consume 80% of their total daily amount of food during nocturnal periods (34), these animals were effectively fasted for closer to 24 h, leading to a large reduction in glycogen stores in the rat. The disparities between the studies of Bergeron and colleagues and the current study showing an AICAR stimulation of glucose production can be attributed to differences in glycogen, as the liver of the 18-h−fasted dog model used in this study contained ∼65 mg glycogen/g liver (35) and gluconeogenesis is only 5–25% of NHGO (36). They are unlikely to be due to a species difference, as others have reported a dramatic increase in hepatic glycogen breakdown after administration of AICAR in 5-h−fasted rats (12). Consistent with the idea that variations in hepatic glycogen levels were responsible for differences between this study and others is the fact that the marked stimulation of NHGO from a 2 mg · kg⁻¹ · min⁻¹ intraportal AICAR infusion was lost when given to a 42-h−fasted dog in which liver glycogen was markedly reduced (R.R.P., R.C.C., D.H.W., unpublished observation). Further support of a glycogenolytic, and not a gluconeogenic, effect is evidenced by the lack of differences in hepatic uptake of the gluconeogenic precursors, alanine and glycerol, in the presence of AICAR. Finally, supplying glucose via glycogenolysis instead of through gluconeogenesis would limit hepatic ATP usage. To assess whether AICAR infusion at the dosages used in these experiments caused liver toxicity, circulating levels of liver enzymes were measured in a subgroup of dogs. All were within predicted ranges, suggesting that liver function was normal.

The intraportal infusion of AICAR was associated with a 5 ± 3 and 19 ± 5 mg/g tissue decrease (P < 0.05) in hepatic glycogen in the presence of euglycemia and hypoglycemia, respectively (Table 3). Thus, hypoglycemia amplified the net hepatic glycogenolytic response to AICAR by approximately fourfold. These differences in glycogen correspond to rates of net hepatic glycogenolysis of 1.9 and 4.5 mg · kg⁻¹ · min⁻¹. NHGO was 2.2 ± 0.4 mg · kg⁻¹ · min⁻¹ higher in the E-AIC compared with the E-SAL animals, an observation that is in line with the observed difference in net hepatic glycogenolysis (i.e., 1.9 mg · kg⁻¹ · min⁻¹). NHGO was 2.5 ± 0.3 mg · kg⁻¹ · min⁻¹ higher in the H-AIC compared with the H-SAL group, and net hepatic lactate output was ∼1.0 mg · kg⁻¹ · min⁻¹ greater in the H-AIC group. Together, these account for ∼80% of the net hepatic glycogenolysis under hypoglycemic conditions (i.e., 4.5 mg · kg⁻¹ · min⁻¹); the remaining 20% may be distributed among various other metabolic intermediates. Net hepatic lactate output was stimulated in the E-AIC group as well, although differences failed to reach significance. Stimulation of net hepatic lactate output is consistent with AMPK phosphorylation of phosphofructokinase and stimulation of glycolysis (11,37–40).

A glycogenolytic effect of AMPK is consistent with several lines of evidence. Glycogen synthase, a substrate that AMPK phosphorylates and inactivates (41), and glycogen phosphorylase, a substrate that AMPK phosphorylates and activates (15), co-immunoprecipitate with AMPK (42). A glycogen-binding domain exists on the AMPK-δ3 subunit in rat liver (43) as well as on the AMPK-β domain (43,44). A mutation in the δ3 subunit results in glycogen storage disease, at least in skeletal (45) and cardiac (46) muscle. When one considers these properties of AMPK and couples them with the effect of purine nucleotides as allosteric activators of glycogen phosphorylase (14,15,47,48), it becomes clear that there are cellular bases for the potent effect of AICAR on hepatic glucose output.

An interesting finding of this study was that AICAR reduced cortisol levels in response to hypoglycemia. AICAR was shown to have no effect on the counterregulatory hormone responses of glucagon and epinephrine to hypoglycemia when administered centrally (30), but cortisol was not measured. Our finding is consistent with the work of Fabian et al. (49), who showed that cortisol was reduced with AICAR administration to pigs after lipopolysaccharide-evoked pulmonary dysfunction after resuscitation from traumatic shock. It has been suggested that this reduction in cortisol plays a part in improving hemodynamic parameters to aid in response to stress. AICAR effects in changing cortisol (and by inference, adrenocorticotropic hormone) are consistent with this compound mediating central effects, as has been shown by other investigators (29,30,50).

In the presence of both euglycemia and hypoglycemia, intraportal AICAR significantly increased R_d (Fig. 6). This finding is consistent with those of previous studies showing that AICAR increases muscle glucose uptake (12,18–20,41,51,52). However, it is important to note that this peripheral stimulation of glucose uptake by AICAR occurred even though the dosage used in the present studies was considerably less than that in other studies and was administered in the portal vein.

Studies have shown that one way in which AMPK is activated is by phosphorylation of the Thr¹⁷² residue by an upstream kinase, AMPKK or LKB1 (53–55). However, another recent study has shown that AICAR does not activate AMPK by stimulating LKB1 (56). Given that AMPK activity is regulated by both allosteric interactions as well as phosphorylation (57), phosphorylation of ACC is a more sensitive marker of AMPK activity in vivo (58). Although there was no apparent increase in the phosphorylation of the Thr¹⁷² residue of AMPK, the downstream target of AMPK, ACC (Ser⁷⁹), showed a significant increase in phosphorylation in the presence of both euglycemia and hypoglycemia (Fig. 7), proof that AMPK activity was increased. This notion is supported by work in myocardial tissue in which AICAR treatment resulted in a clear increase in the phosphorylation of ACC (14). The lower dosage of AICAR infusion used in this study may not have been high enough to cause covalent (phosphorylation) modification of AMPK, yet may have been sufficient to cause allosteric activation (14). As expected, intraportal AICAR infusion significantly increased hepatic ZMP concentrations (Table 3). Both increased AMPK activity (11,13) and increased AMP/ZMP concentrations (14,15,17,47) have been shown to result in activation of glycogen phosphorylase. Thus, it is highly probable that the increases in net hepatic glycogen breakdown were due to increased hepatic AMPK activity and ZMP concentrations.

In summary, our results show that in the presence of basal glucagon and high physiological levels of insulin, increasing hepatic ZMP levels via an intraportal infusion of AICAR stimulates endogenous glucose output via an increase in hepatic glycogenolysis and increases R_d during euglycemic and hypoglycemic conditions. Hypoglycemia amplifies the net hepatic glycogenolytic response to AICAR by approximately fourfold. These effects on hepatic and peripheral glucose metabolism occurred using infusion rates that were 10% of those previously used in rats (18–20). These responses are consistent with the concept that increased AMP concentrations not only signal increased glucose metabolism within the muscle, but also coordinate a whole-body response that increases the availability of blood glucose by stimulating glucose output by the liver. However, effects on processes such as renal glucose production, adipocyte glucose uptake, and gut glucose absorption may play a role as well. The present results combined with the results of an earlier study (17) have three important implications. First, because purine nucleotides are a natural constituent of all cells, including hepatocytes, they may comprise part of an endogenous pathway involved in the physiological activation of hepatic glucose output. Second, understanding hepatic insulin resistance caused by AICAR may provide insight into basic mechanisms underlying this liver pathology. Third, mechanisms that increase hepatic purine nucleotide concentrations may be effective in countering insulin-induced hypoglycemia. Because hypoglycemia increases AICAR-stimulated net hepatic glycogenolysis and the liver of the AICAR-infused dog appears to be relatively blind to insulin adds further weight to this third point. This suggests that a compound with similar hepatic actions could effectively be exploited to combat insulin-induced hypoglycemia, which is the single most frequently occurring complication in diabetes.

This work was funded by National Institute of Diabetes and Digestive and Kidney Diseases Grant RO1-DK-50277 and Diabetes Research and Training Center Grant DK-20593.

We thank the Hormone Assay and Animal Resources Cores of the Vanderbilt Diabetes Research and Training Center for providing the hormone and catecholamine data provided.