Effects of Insulin Deficiency or Excess on Hepatic Gluconeogenic Flux During Glycogenolytic Inhibition in the Conscious Dog

Experiments were conducted on 18 overnight-fasted conscious mongrel dogs (20–28 kg) of either sex. Housing and diet have been described previously (1). The surgical facility met the standards published by the American Association for the Accreditation of Laboratory Animal Care, and the protocols were approved by the Vanderbilt University Medical Center Animal Care Committee. All dogs underwent a laparotomy 2 weeks before the experiment to implant infusion catheters into the jejunal and splenic veins, sampling catheters into the portal and hepatic veins and the femoral artery, and Doppler flow probes around the hepatic artery and portal vein, as described elsewhere (1). Each dog was used for only one experiment. All dogs studied had 1) leukocyte count <18,000/mm³, 2) a hematocrit >35%, 3) a good appetite, and 4) normal stools.

Intraportal catheters (splenic and jejunal) were used for the infusion of insulin (Lilly, Indianapolis, IN) and glucagon (Lilly). Angiocaths (Deseret Medical, Becton-Dickinson, Sandy, UT) were inserted percutaneously into leg veins for [3-³H]glucose (DuPont NEN, Boston, MA), indocyanine green (Sigma, St. Louis, MO), and peripheral glucose (50% Dextrose; Baxter Healthcare Corporation, Deerfield, IL) infusion. Animals were allowed to rest quietly in a Pavlov harness for 30 min before the experiments started. Each of the three protocols consisted of an equilibration period (−120 to −40 min), a basal period (−40 to 0 min), and an experimental period (0–300 min). At −120 min, a priming dose of [3-³H]glucose (50 μCi) was given, and constant infusions of [3-³H]glucose (0.4 μCi/min) and indocyanine green (0.07 mg/min) were started. At the same time, a constant infusion of somatostatin (SRIF; 0.8 μg · kg⁻¹ · min⁻¹) was started in a peripheral vein to inhibit endogenous pancreatic hormone secretion, and constant intraportal glucagon (0.5 ng · kg⁻¹ · min⁻¹) and insulin (300 μU · kg⁻¹ · min⁻¹) infusions were started to replace basal secretion of these hormones. Also at −120 min, a glycogen phosphorylase inhibitor (BAY R 3401; 10 mg/kg in a 0.5% methyl cellulose/water solution) was given orally or intragastrically, depending on whether a stomach catheter had been implanted during surgery (there was no difference in the effectiveness of the drug with one route or the other). Shortly thereafter, a variable glucose infusion was begun to maintain euglycemia in the presence of the glycogen phosphorylase inhibitor. [3-³H]glucose infusion rates were modified in accordance with the cold glucose infusion rates to minimize changes in the specific activity. During the experimental period three insulin infusion rates were used: protocol 1 (basal insulin [control, n = 6]), no change was made in the insulin infusion rate over the course the experiment; protocol 2 (insulin deficiency [n = 6]), beginning at 0 min, the intraportal insulin infusion was stopped, resulting in 5 h of hypoinsulinemia; and protocol 3 (insulin excess [n = 6]), beginning at 0 min, the intraportal insulin infusion rate was increased to (1.2 mU · kg⁻¹ · min⁻¹), resulting in 5 h of hyperinsulinemia.

Immediately after the final sampling time, each animal was anesthetized and sections of three liver lobes were freeze-clamped in situ and stored at −70°C as previously described (1). All animals were then killed, and the correct position of the catheter tips was confirmed.

Hematocrit; plasma glucose, [³H]glucose, glucagon, insulin, cortisol, and nonesterified free fatty acids (NEFAs); and blood alanine, glycine, serine, threonine, lactate, glutamine, glutamate, glycerol, β-hydroxybutyrate, and hepatic glycogen concentrations were determined as previously described (1). Hepatic glycogen content was measured as previously described (15). Hepatic tissue glucose, G6P, and F6P were measured fluorometrically by standard enzymatic methods (16,17) after deproteinization with 3% perchloric acid.

Net hepatic balances (NHBs) were calculated with the arteriovenous (A-V) difference method using the following formula: NHB = Load_out − Load_in where Load_out = [H] · HF and Load_in = [A] · AF + [P] · PF and where [H], [A], and [P] were the substrate concentrations in hepatic vein, femoral artery, and portal vein blood or plasma, respectively, and HF, AF, and PF were the blood flow in the hepatic vein, hepatic artery, and portal vein, as determined by the ultrasonic flow probes. With this calculation, a positive value represents net output by the liver, whereas a negative value represents net hepatic uptake. Plasma glucose and [³H]glucose values were multiplied by 0.73 to convert them to blood glucose values as validated elsewhere (15). Net fractional substrate extraction by the liver was calculated by NHB/Load_in. The approximate insulin and glucagon levels in plasma entering the liver sinusoids were calculated using the formula [A] · %AF + [P] · %PF, where [A] and [P] were arterial and portal vein hormone concentrations, respectively, and %AF and %PF were the respective percentage contributions of arterial and portal flow to total hepatic blood flow.

Tracer-determined whole-body glucose production (R_a) was measured using a primed, constant infusion of [3-³H]glucose. Data calculation was carried out using the two-compartment model described by Mari (18), using canine parameters reported by Dobbins et al. (19).

For calculation of unidirectional hepatic glucose uptake (HGU), the net [³H]glucose uptake was divided by the arterial [³H]glucose specific activity. The above calculation is based on the assumption that intrahepatic uptake of glucose occurs before glucose production so that the plasma glucose specific activity is not diluted. Even if this assumption is not correct, the drop in specific activity across the liver is very small; thus, the assumption is of little consequence. Unidirectional hepatic glucose release (HGR) was determined by adding HGU to net hepatic glucose output.

Gluconeogenesis, as classically defined, is the synthesis and subsequent release of glucose from noncarbohydrate precursors. Carbon produced from flux through the gluconeogenic pathway does not necessarily have to be released as glucose; it can also be stored in glycogen. Therefore, we make a distinction between gluconeogenic flux to G6P (conversion of precursors to G6P: G6P neogenesis) and gluconeogenesis (release of glucose derived from gluconeogenic flux).

Gluconeogenic flux to G6P was determined by summing the net hepatic uptake rates of the gluconeogenic precursors (alanine, glycine, serine, threonine, glutamine, glutamate, glycerol, lactate, and pyruvate) and dividing by 2 to account for the incorporation of three carbon precursors into the six-carbon glucose molecule. Net hepatic gluconeogenic flux was determined by subtracting the summed net hepatic output rates (when such occurred) of the substrates noted above (in glucose equivalents) and glucose oxidation (GO) from gluconeogenic flux to G6P. GO was assumed to be 0.2 mg · kg⁻¹ · min⁻¹ throughout each experiment. In other experiments, GO varied from this estimate by no more than 0.1 mg · kg⁻¹ · min⁻¹ under widely varying conditions (S. Satake, M.C. Moore, A.D.C., unpublished observations); thus, any changes in GO that might have occurred in the present experiments should have been small and of little consequence to our estimates. A positive number represents net gluconeogenic flux to G6P, whereas a negative number indicates net glycolytic flux.

NHGLY was estimated by subtracting net hepatic gluconeogenic flux from net hepatic glucose balance (NHGB). A positive number therefore represents net glycogen breakdown, whereas a negative number indicates net glycogen synthesis.

Ideally, the gluconeogenic flux rate would be calculated using unidirectional hepatic uptake and output rates for each substrate, but this would be difficult, as it would require the simultaneous use of multiple stable isotopes, which could themselves induce a mild perturbation of the metabolic state. Therefore, NHB was used instead, necessitating consideration of the limits of this approach. There is little or no production of gluconeogenic amino acids or glycerol by the liver, so in this case the compromise is of little consequence. Such is not the case, however, for lactate. Our estimate of the rate of gluconeogenic flux to G6P will be quantitatively accurate only if we assume that lactate flux is unidirectional at a given moment (i.e., either in or out of the liver). In a given cell, this does not seem like an unreasonable assumption in light of the reciprocal control of gluconeogenesis or glycogenolysis (8). Katz and Jungermann suggested, however, that there is spatial separation of metabolic pathways so that gluconeogenic periportal hepatocytes primarily synthesize glucose and glycogen from lactate and other noncarbohydrate precursors, whereas glycolytic perivenous hepatocytes predominantly consume plasma glucose, which is incorporated into glycogen, oxidized, or released as lactate or other glycolytic substrates (20,21). Therefore, it is possible that under normal nutritional conditions in a net sense, hepatic gluconeogenic and glycolytic flux occurs simultaneously, with lactate output or uptake occurring in different cells. To the extent that flux occurs in both directions simultaneously, the A-V difference method will underestimate gluconeogenic flux to G6P. It should be noted, however, that net hepatic gluconeogenic flux and NHGLY can be calculated accurately without concern for the assumptions related to whether simultaneous gluconeogenic and glycolytic substrate flux occurs.

The total net contribution (TNC) of carbon to hepatic glycogen was determined using the average net hepatic glycogen synthetic rate during the experimental period to predict the total net accumulation of glycogen during that period (300 min) using the following formula: TNC (mg glycogen/g liver) = net hepatic glycogen synthesis (mg · kg⁻¹ · min⁻¹) · time (min)/liver weight (g/kg body wt).

The data were analyzed for differences from the basal period and for differences from the control group. Statistical comparisons were carried out using two-way ANOVA (JMP IN Software, SAS Institute Inc.). One-way ANOVA comparison tests were used post hoc when significant F ratios were obtained. The level of significance was P < 0.05 (two-sided test).

In the control group, the arterial and hepatic sinusoidal plasma insulin levels remained basal and unchanged throughout the study (5 and 15 μU/ml, respectively; Fig. 2). In the insulin deficiency group, the arterial and sinusoidal insulin levels fell to 1 ± 0 and 1 ± 1 μU/ml, respectively (P < 0.05), whereas in the insulin excess group these levels rose to 20 ± 1 and 66 ± 3 μU/ml, respectively (P < 0.05). The arterial and hepatic sinusoidal glucagon levels remained basal and unchanged in each group (Fig. 2).

Glucose infusion was required to maintain euglycemia during the basal period in each group as a result of the inhibition of glycogen breakdown (Table 1). Whereas the glucose infusion rate did not change significantly over time in the control group, it fell to 0 during insulin deficiency and rose to an average of 11.71 ± 1.17 mg · kg⁻¹ · min⁻¹ during the last hour of hyperinsulinemia. Thus, while the control and insulin excess groups remained euglycemic throughout, the plasma glucose level gradually rose during insulin deficiency from 106 ± 1 during the basal period to 135 ± 7 mg/dl at the end of the experimental period (P < 0.05; Fig. 3).

NHGB was close to 0 throughout the control protocol (Fig. 3). During insulin deficiency, it increased, averaging an output of 2.10 ± 0.20 mg · kg⁻¹ · min⁻¹ during the last hour of the study (P < 0.05). In the insulin excess group, it fell, averaging an uptake of 1.58 ± 0.31 mg · kg⁻¹ · min⁻¹ during the last hour of the study (P < 0.05). Hepatic artery and portal vein blood flows were similar in each group and did not change over time (26 ± 2, 29 ± 2, and 27 ± 2 ml/min average total hepatic blood flow in the control, insulin deficiency, and insulin excess groups, respectively). HGU did not change over time in the control group (∼1 mg · kg⁻¹ · min⁻¹) but fell to 0.25 ± 0.16 mg · kg⁻¹ · min⁻¹ during insulin deficiency (P < 0.05) and rose to 1.80 ± 0.16 during insulin excess (P < 0.05; Table 1). HGR tended to decline over the course of the experiment in the control and insulin excess groups (1.23 ± 0.21 to 0.80 ± 0.16 and 0.42 ± 0.19 to 0.22 ± 0.30, respectively) and rose from 0.98 ± 0.47 to 2.26 ± 0.17 mg · kg⁻¹ · min⁻¹ (P < 0.05) during insulin deficiency (Table 1).

Endogenous glucose appearance (R_a) in the control group was reduced relative to normal as a result of phosphorylase inhibition but did not change significantly over time. During insulin deficiency, R_a increased to 2.85 ± 0.13 mg · kg⁻¹ · min⁻¹ (P < 0.05), whereas during insulin excess it was completely inhibited (P < 0.05; Table 1). The rate of glucose disappearance (R_d) did not change in the control or insulin deficiency groups, whereas during insulin excess it increased significantly from 2.48 ± 0.42 in the basal period to 10.55 ± 2.05 mg · kg⁻¹ · min⁻¹ during the last hour of the study (P < 0.05; Table 1). Glucose clearance, an index of the avidity with which tissues take up glucose, was unchanged in the control group, tended to fall during insulin deficiency, and rose almost fivefold (P < 0.05) during insulin excess (Table 1).

The arterial blood lactate level did not change significantly in any group, although there was a tendency for it to fall slightly and then rise modestly (Fig. 4). Net hepatic lactate uptake (NHLU) did not change significantly over time in the control group but increased more than sixfold in the insulin deficiency group (rising from 1.42 ± 1.61 to 8.84 ± 1.11 μmol · kg⁻¹ · min⁻¹; P < 0.05) and rose modestly in the insulin excess group (2.06 ± 1.22 to 5.04 ± 0.62 μmol · kg⁻¹ · min⁻¹; P < 0.05; Fig. 4).

In the control and insulin deficiency groups, the summed arterial gluconeogenic amino acid levels tended to drift down slightly (Fig. 5). This was associated with a significant but small increase in the hepatic gluconeogenic amino acid fractional extraction and a tendency for net hepatic gluconeogenic amino acid uptake to increase (Fig. 5). In the insulin excess group, there was a decrease in the arterial blood level of all of the gluconeogenic amino acids (P < 0.05), although it did not reach significance in the cases of glutamate and glutamine (Table 2). Consequently, the summed gluconeogenic amino acid level decreased 36% (P < 0.05). At the same time, there was a 35% increase (P < 0.05) in the gluconeogenic amino acid fractional extraction by the liver (0.13 ± 0.03 to 0.20 ± 0.05; Fig. 5). Because the changes in fractional extraction occurred before the fall in amino acid levels, there was a tendency for gluconeogenic amino acid uptake to rise transiently in response to increased insulin (Table 3).

Despite an initial drift down, there was no significant change in blood glycerol levels during the course of the experiment in the control group (Fig. 6). During insulin deficiency, the glycerol level nearly doubled (P < 0.05), whereas during insulin excess it decreased >60% (P < 0.05). Similarly, net hepatic glycerol uptake did not change in the control group but increased more than twofold and decreased almost 60% in the insulin deficiency and excess groups, respectively (Fig. 6).

The arterial plasma NEFA levels tended to decrease in the control group, but as with glycerol, the change was not significant (Fig. 7). The NEFA level doubled during insulin deficiency (P < 0.05) and decreased 85% in the insulin excess group (P < 0.05). Net hepatic NEFA uptake showed a similar pattern, with no change in the control, a 2.5-fold increase during insulin deficiency (P < 0.05), and a 90% decrease during insulin excess (P < 0.05; Fig. 7).

In the control group, there was no significant change in the arterial blood β-hydroxybutyrate (β-OHB) level (35 ± 10 and 27 ± 9 μmol/l during the two periods) or in net hepatic β-OHB output (1.8 ± 0.6 and 1.3 ± 0.4 μmol · kg⁻¹ · min⁻¹; Table 4). In the insulin deficiency and insulin excess groups, the basal arterial β-OHB levels were not different from those evident in the control group but rose to 103 ± 21 (P < 0.05) and fell to 3 ± 1 μmol/l (P < 0.05), respectively. Net hepatic β-OHB output was also not different in the three groups during the basal period but increased to 3.2 ± 0.7 (P < 0.05) and decreased to 0.2 ± 0.1 μmol · kg⁻¹ · min⁻¹ (P < 0.05) in the two groups, respectively, during the last hour of the experimental period.

NHGLY decreased over time in the control group (−0.30 ± 0.32 to −1.20 ± 0.18 mg · kg⁻¹ · min⁻¹; P < 0.05), whereas it increased from −0.46 ± 0.39 to 0.76 ± 0.21 (P < 0.05) and decreased from −0.81 ± 0.28 to −2.44 ± 0.35 mg · kg⁻¹ · min⁻¹ (P < 0.05) during insulin deficiency and insulin excess, respectively. The changes in NHGLY between basal and the last hour of the experimental period were −0.90 ± 0.43, 1.22 ± 0.53, and −1.63 ± 0.36 mg · kg⁻¹ · min⁻¹ in the control, insulin deficiency, and insulin excess groups, respectively (P < 0.05; Fig. 8). The TNC of carbon to hepatic glycogen during the 5-h experimental period, calculated from NHGLY, was 11.4 ± 1.6 mg/g in the control group and 20.7 ± 3.0 mg/g during hyperinsulinemia (P < 0.05). Because there was net hepatic glycogen breakdown (1.8 ± 2.6 mg/g) during insulin deficiency, there was no net contribution of carbon to glycogen when insulin was lacking. Hepatic glycogen content was not different in the insulin deficiency (71 ± 6 mg/g) or excess groups (85 ± 6 mg/g) relative to the control (76 ± 6 mg/g) at the end of the studies; however, it tended to be lower after insulin deficiency and higher after hyperinsulinemia (Table 5). In fact, after insulin excess, the hepatic glycogen content was 9 mg/g higher than the control, similar to the independently estimated difference in TNC (9.3 mg/g) between these two groups. After insulin deficiency, the glycogen content was 5 mg/g less than the control, compared with a difference in TNC of 14.1 mg/g.

The hepatic tissue glucose level at the end of the study was not different in the insulin excess group (0.21 ± 0.02 mg/g) relative to the control group (0.23 ± 6 mg/g); however, it was increased after insulin deficiency (0.30 ± 0.03 mg/g; P < 0.05; Table 5). Hepatic G6P and F6P levels were similar at the end of the control and hyperinsulinemic periods (72 ± 5 vs. 75 ± 6 and 37 ± 3 vs. 39 ± 3 nmol/g, respectively), whereas both tended to be higher after insulin deficiency (84 ± 9 and 44 ± 5 nmol/g; Table 5).

Hepatic gluconeogenic flux to G6P did not change significantly over time in the control group (1.05 ± 0.17 and 1.15 ± 0.17 mg · kg⁻¹ · min⁻¹ in the two periods). During insulin deficiency, however, it rose twofold, from 0.78 ± 0.09 to 1.59 ± 0.14 mg · kg⁻¹ · min⁻¹ (P < 0.05), whereas during insulin excess it did not change significantly. The changes in gluconeogenic flux to G6P between the basal period and the last hour of the experiment were 0.10 ± 0.10, 0.81 ± 0.16, and 0.21 ± 0.13 mg · kg⁻¹ · min⁻¹ in the control, insulin deficiency, and insulin excess groups, respectively (P < 0.05; Figs. 8 and 9).

Net hepatic gluconeogenic flux did not change significantly over time in the control group (0.58 ± 0.26 and 0.87 ± 0.19 mg · kg⁻¹ · min⁻¹ in the two periods). It rose more than fourfold during insulin deficiency, however, from 0.31 ± 0.17 to 1.35 ± 0.15 mg · kg⁻¹ · min⁻¹ (P < 0.05). During insulin excess, it rose from 0.47 ± 0.11 to 0.85 ± 0.11 mg · kg⁻¹ · min⁻¹ (P < 0.05), a rate not different, however, from that in the control group. The changes in net gluconeogenic flux between basal and the last hour of the experimental period were 0.29 ± 0.21, 1.04 ± 0.20, and 0.38 ± 0.14 mg · kg⁻¹ · min⁻¹ in the control, insulin deficiency, and insulin excess groups, respectively (P < 0.05; Figs. 8 and 9).

To examine insulin’s acute action on gluconeogenic flux to G6P in the absence of its effects on glycogenolytic flux, we inhibited glycogen breakdown using a glycogen phosphorylase inhibitor. The drug used in these studies, BAY R3401, was previously shown to effectively suppress glycogenolysis in hepatocytes (11), in perfused intact liver (11), and in the whole animal (rat and dog) (12–14). The active metabolite of BAY R3401 reduces glycogenolysis both by facilitating inactivation of the enzyme through allosteric inhibition and by dephosphorylation of glycogen phosphorylase a (11). In the dog, after administration of BAY R3401, inhibition of NHGLY resulted in a decrease in hepatic glucose production and net hepatic lactate balance (NHLB) but had no effect on the metabolism (level or NHB) of other gluconeogenic precursors, NEFAs, pancreatic or adrenal hormones, or hepatic blood flow (14). Whereas glucose R_a decreased 50%, glucose R_d and clearance were unaffected, indicating that the portion of the drug that passed the liver had little or no effect on muscle glycogen metabolism. These data suggest that BAY R3401 reduces hepatic glycogenolysis through a direct inhibitory action on glycogen phosphorylase and that it acts without producing other direct metabolic alterations.

Without the use of a phosphorylase inhibitor in the present study, complete insulin deficiency would have resulted in a large and rapid increase in glycogen breakdown. In a previous study in which hepatic sinusoidal insulin levels were selectively reduced by 75% (there was no decrease in arterial insulin level), NHGLY rose more than sixfold (to 5.39 mg · kg⁻¹ · min⁻¹) within 30 min (10). In the present study, when insulin was made completely deficient, net hepatic glycogen synthesis, which was evident during the basal period, slowly decreased until eventually a minimal rate of net glycogen breakdown was evident (0.71 ± mg · kg⁻¹ · min⁻¹ during the last hour of the experiment). As a result, insulin deficiency in the present study resulted in only a slight (∼30 mg/dl) rise in the plasma glucose level, rather than the two- to threefold increase that was expected during insulin deficiency in the absence of BAY R3401 (10). Glucose inhibits hepatic glucose production independent from insulin (22); thus, with the phosphorylase inhibitor present, the effect of insulin deficiency on gluconeogenic flux to G6P could be assessed, for the most part, without the confounding effects attributable to glycogenolysis and marked hyperglycemia, because carbon flux into G6P from plasma glucose or glycogen would reduce the ability of insulin deficiency to increase gluconeogenic flux.

With glycogenolytic flux inhibited, insulin deficiency caused significant increases in gluconeogenic flux to G6P and net hepatic gluconeogenic flux, which were evident within 3 h. During the basal period, there was no gluconeogenic contribution to NHGB (there was slight net glucose uptake); however, gluconeogenesis (the flux of gluconeogenically derived glucose into plasma) became evident during insulin deficiency, accounting for ∼65% of net hepatic glucose output by the end of the study. Gluconeogenic flux to G6P, as estimated using the A-V difference technique, was previously shown to decrease by 50% during selective hepatic insulin deficiency (the arterial insulin level was unchanged) (10). This finding was mainly due to a switch from NHLU to output, presumably reflecting the increase in glycolysis caused by augmented glycogen breakdown. In the present study, when glycogen breakdown was inhibited, insulin deficiency resulted in a twofold rise in hepatic gluconeogenic flux to G6P. This was due to a doubling of net hepatic glycerol uptake and a sixfold increase in NHLU. The increase in glycerol uptake was the result of an increase in substrate availability, which in turn resulted from increased lipolysis. The effect on lactate uptake was due to an increase in the fractional extraction of lactate by the liver and an increase in the supply of lactate from nonhepatic tissues. The latter is evident from the lack of a decrease in the blood lactate level despite an increase in NHLU. An effect on the fractional extraction of lactate by the liver could have resulted from the direct effect of insulin deficiency per se or from the rise in NEFA, which may itself have stimulated NHLU. Regardless of the mechanisms by which it occurred, there was more than a fourfold increase in net hepatic gluconeogenic flux when insulin was made deficient.

Gluconeogenic amino acid levels and uptake rates were unchanged during insulin deficiency and thus did not affect the carbon flux rates. Because insulin is generally considered to increase protein synthesis and decrease protein breakdown, insulin deficiency might have been expected to cause a rise in blood amino acid levels. Although whole-body protein breakdown seems to increase during insulin deficiency, some investigators have found that protein synthesis is normal or increased (23), possibly explaining why we observed no change in the gluconeogenic amino acid levels even when insulin was deficient.

In contrast to insulin deficiency, insulin excess did not seem to alter gluconeogenic flux to G6P or the net gluconeogenic carbon flux in the presence of the phosphorylase inhibitor. Previous studies have shown that NHGLY is quickly suppressed when increments in plasma insulin are created (1). Therefore, it was hoped that in the present study phosphorylase inhibition before elevation in insulin would prevent significant additional effects of the hormone on glycogenolysis. Indeed, NHGLY was already completely suppressed by the drug before the fourfold rise in plasma insulin; therefore, the hormone did not further inhibit glycogen breakdown.

In these studies, we found no inhibition of gluconeogenic flux to G6P after 5 h of a fourfold rise in plasma insulin level. In fact, there was a twofold rise in net gluconeogenic flux, an increase similar to that seen in the control group. In keeping with the anabolic effects of insulin on protein synthesis and breakdown (23), we observed an ∼40% decrease in arterial gluconeogenic amino acid levels. However, an equivalent increase in the fractional extraction of these amino acids by the liver prevented a fall in their net hepatic uptake rates (Fig. 5). Insulin has previously been reported to stimulate amino acid uptake by the liver (23). Net hepatic glycerol uptake tended to decrease during insulin excess as a result of the fall in blood glycerol secondary to the inhibition of lipolysis (Fig. 6). This was counterbalanced by a slight increase in NHLU (Fig. 4); thus, no net effect of the rise in insulin on gluconeogenic flux to G6P was observed even after 5 h of hyperinsulinemia.

In previous studies conducted in isolated hepatocytes and perfused rat liver, BAY R3401 significantly reduced G6P levels (11). Because net hepatic glycogen synthesis increased 1.63 mg · kg⁻¹ · min⁻¹ during hyperinsulinemia whereas net HGU increased only 1.25 mg · kg⁻¹ · min⁻¹, an insulin-mediated decrease in net hepatic gluconeogenic flux would have resulted in a decrease in G6P levels. In fact, net gluconeogenic flux increased 0.38 mg · kg⁻¹ · min⁻¹ in this group, resulting in no decline in G6P levels compared with the control. It is possible, therefore, that the buffering capacity of the liver to prevent critically low G6P levels did not permit an insulin-mediated inhibition of gluconeogenic flux to G6P, because a decline in F2,6P₂, associated with low G6P levels, could counteract the direct effects of insulin and prevent a decrease in gluconeogenic flux to G6P. It is possible that inhibition of gluconeogenic flux to G6P by insulin would be unmasked if in conjunction with hyperinsulinemia and the phosphorylase inhibitor the glucose level was increased, providing an alternate means for supplying glycolytic intermediates. We can nevertheless conclude that the rise in insulin inhibited gluconeogenesis per se (as opposed to G6P neogenesis) because glucose production was fully suppressed (Table 1). In the present study, therefore, the inhibition of gluconeogenesis caused by the increase in insulin was secondary to activation of glycogen synthesis rather than a decrease in gluconeogenic flux to G6P per se. This is in keeping with previous studies that have shown that on refeeding after an extended fast, 40–60% of glycogen is synthesized from gluconeogenic substrates, despite a rise in insulin (21).

HGU occurs by flux through glucokinase, with insulin stimulating uptake by activating glycogen synthase (24) and thereby providing a “pull mechanism” (21). Thus, as expected, HGU, the net direct contribution of plasma glucose to glycogen, and net hepatic glycogen synthesis were greatest during hyperinsulinemia. In contrast, during insulin deficiency, the liver switched from net HGU to output and net glycogen synthesis to breakdown. HGR is dependent on glucose-6-phosphatase (G6Pase), and the G6Pase catalytic subunit is thought to be a major control point in the G6Pase reaction (25). Using a subgroup of four of the six animals from each group in these experiments, induction of G6Pase catalytic subunit mRNA expression was determined and reported by Hornbuckle et al. (25). After 5 h of insulin treatment, the G6Pase catalytic subunit mRNA was 30% of that in the control group, in agreement with the observed decrease in hepatic glucose production, and the 30% increase in hepatic tissue glucose level. Conversely, G6Pase catalytic subunit mRNA expression rose to fourfold that in the control group during insulin deficiency, and correspondingly HGR was 2.8 times greater than the control, although in the presence of BAY R3401 the majority of the glucose released was gluconeogenic rather than glycogenolytic.

Insulin can regulate the fate of hepatic G6P indirectly through its effects on lipolysis. Plasma free fatty acids have been shown to increase hepatic gluconeogenesis (26–29) and to decrease whole-body glucose utilization (30). Whereas some studies have demonstrated increased glucose production in response to elevated NEFAs (31,32), others have failed to do so, presumably as a result of a compensatory decrease in NHGLY (26,28,33,34). The mechanisms by which an increase in NEFAs affects gluconeogenesis, glycolysis, and NHLB most likely include a rise in NEFA oxidation and thus an increase in the mitochondrial ratios of [acetyl CoA]/[CoA] and [NADH]/[NAD⁺] (30,35,36). Such an increase would inhibit the pyruvate dehydrogenase complex, increase citrate (an inhibitor of PFK-1 and PFK-2), and thereby decrease F2,6P₂ and F1,6P₂ concentrations. This in turn would result in decreased glycolysis and increased gluconeogenic flux to G6P (7,8,30). Generation of acetyl CoA or long-chain fatty acetyl CoA could also increase PC activity, and a rise in free fatty acid oxidation would supply the reducing equivalents (NADH) and energy (ATP) necessary for gluconeogenic flux (30,35–39). In addition, free fatty acids have been reported to stimulate PEPCK gene transcription (40).

Given the potential effects of free fatty acids on glycolytic and gluconeogenic fluxes, a rise in NEFAs, such as occurs with insulin deficiency, would be expected to reduce NHLB and stimulate net gluconeogenic flux. Previously, when selective hepatic insulin deficiency was created so that arterial insulin and thus NEFA levels did not change, NHGLY was stimulated and net hepatic lactate output increased to ∼9 μmol · kg⁻¹ · min⁻¹ (10). Conversely, during insulin deficiency in the present study when the NEFA level doubled and glycogenolysis was inhibited, lactate was taken up by the liver at a net of ∼9 μmol · kg⁻¹ · min⁻¹. This was associated with an increase in fatty acid oxidation as indicated by a threefold rise in β-OHB level and increased net hepatic β-OHB output. Additional investigation will be required to determine whether the effect of insulin deficiency on gluconeogenic flux to G6P was due to direct hepatic insulin action or was the result of insulin’s indirect peripheral effects, because the increase in gluconeogenic flux to G6P was the result of increased hepatic glycerol and lactate uptake, and both occurred in conjunction with an increase in lipolysis and NEFA level.

Previous studies have shown that a fall in NEFAs results in increased net hepatic lactate output, decreased glucose production (to the extent that G6P-derived carbon exits the liver as lactate instead of glucose), and decreased gluconeogenic flux to G6P (to the extent that lactate uptake no longer contributes to the process). This was demonstrated by Sindelar et al. (9), who found that a selective increase in arterial insulin (no change in portal vein insulin) led to a decrease in NEFA levels and resulted in the redirection of glycogenolytically derived carbon to lactate rather than glucose. However, when free fatty acid levels were prevented from falling during a selective increase in arterial insulin (41) or insulin was increased selectively in the liver so that NEFA levels did not change (9), an increase in net hepatic lactate output did not occur. In the present study, although during hyperinsulinemia the net hepatic uptakes and levels of NEFA and β-OHB decreased by ∼90%, there was no increase in net hepatic lactate output. This supports the conclusion of Sindelar et al. that the change in lactate balance caused by a change in NEFAs was the result of an effect on the routing of glycogenolytically derived carbon (41). Although there was a moderate increase in net HGU during hyperinsulinemia, this was probably directed to glycogen. Even though a fall in NEFA level accompanied the increase in insulin, we did not observe a decrease in hepatic gluconeogenic flux to G6P or net gluconeogenic flux.

Insulin has been demonstrated to acutely inhibit gluconeogenic flux to G6P in vitro (7,8), and the process is known to be chronically increased by fasting (42) and diabetes (43,44). Conversely, acute effects of physiological changes in insulin on gluconeogenic flux to G6P in vivo have been difficult to demonstrate. The present experiments were designed, therefore, to investigate whether inhibiting glycogen breakdown in the presence of changes in insulin would allow an acute effect of the hormone on gluconeogenic flux to G6P to be observed. This was found to be the case during insulin deficiency, when an increase in gluconeogenic flux to G6P (and gluconeogenesis) was observed when the offsetting increase in glycogen breakdown was prevented. Whether this was due to low hepatic insulin receptor activation or an increase in plasma NEFA concentration or both remains unclear. During hyperinsulinemia, inhibition of gluconeogenic flux to G6P was not observed, however, despite inhibition of glycogen breakdown before insulin treatment. It is possible that the basal rate of gluconeogenic flux in the overnight-fasted euglycemic state cannot be acutely reduced by physiological levels of insulin. Gluconeogenesis itself was inhibited by the rise in insulin, but this was the result of redirection of gluconeogenically derived carbon to glycogen instead of plasma glucose. These data demonstrate that in the acute setting, the effects of insulin on glycogenolytic flux dominate changes in gluconeogenic flux to G6P.

This research was supported in part by National Institutes of Health Grants R37-DK18243 and DK14507 and the Clinical Nutrition Research Unit Grant P30-DK26657. D.S.E. was supported by the Vanderbilt Molecular Endocrinology Training Program (5 T 32 DK07563-12).

This work was presented in part at the 59th Annual Meeting of the American Diabetes Association, Chicago, IL, June 1999.

We thank Jon Hastings, Angelina Penaloza, Wanda Snead, Patrick Donahue, and Yang Ying for excellent technical support.