Glucose-6-phosphatase flux and the hepatic glucose balance model.

Mice were studied with the euglycemic hyperinsulinemic and the hyperglycemic clamp techniques after a 6-h fast: 1) euglycemic (6.7 +/- 0.2 mM) hyperinsulinemia (approximately 800 microU/ml); 2) hyperglycemic (15.3 +/- 0.4 mM) hyperinsulinemia (approximately 800 microU/ml). All mice received an infusion of [3-3H]glucose and [U-14C]lactate. Basal hepatic glucose production (HGP) averaged approximately 170 mumol.kg-1.min-1 in both groups. During euglycemic and hyperglycemic hyperinsulinemia, HGP decreased by 53% (to 76.7 +/- 11.1 mumol.kg-1.min-1; P < 0.01) and 74% (to 43.3 +/- 7.2 mumol.kg-1.min-1; P < 0.01), respectively. Hyperglycemia increased glucose cycling (by 2.1-fold; P < 0.01) and the contribution of gluconeogenesis to HGP (88 vs. 43%; P < 0.01) while decreasing that of glycogenolysis (12 vs. 57%; P < 0.01). The percentage of neosynthetized hepatic glycogen formed via the direct pathway was markedly increased during hyperglycemia (53 +/- 2% vs. 23 +/- 3%; P < 0.01). These data indicate that the assessment of hepatic glucose fluxes can be accomplished in conscious unrestrained mice and that, in the presence of hyperinsulinemia, hyperglycemia causes 1) a further inhibition of HGP mainly via inhibition of glycogenolysis and increase in hepatic glucose cycling; and 2) about a fivefold stimulation in the direct pathway of hepatic glycogen formation.

Dicarboxylic acid fluxes during gluconeogenesis. No channelling of mitochondrial oxalacetate.

A rather complete model of the gluconeogenic pathway was used, with the known separate pools of mitochondrial and cytosolic oxalacetate, malate and aspartate. The fumarase, malate dehydrogenase and glutamate oxalacetate transaminase reactions were assumed to be isotopically actively reversible, but none at isotopic equilibrium. Malate was assumed to exchange actively between the mitochondria and cytosol, while aspartate exchange was more limited, in agreement with the known electrogenic nature of aspartate export from the mitochondria. This model was fit to 14C data obtained in hepatocyte studies, and to the whole rat 14C data obtained by Heath and Rose (Biochem J. 227, 851-876, 1985). The latter data were easily fit to our model, when a single mitochondrial oxalacetate pool was assumed. However, invoking two mitochondrial oxalacetate pools, as proposed by Heath and Rose, with the oxalacetate formed via pyruvate carboxylase preferentially channelled to gluconeogenesis, could not be fit with the known differences in scrambling in glucose and glutamate produced from L[3-14C]lactate.

On the estimation of alternative pathways of fatty acid oxidation in the liver in vivo.

The relative contributions of mitochondrial beta-oxidation and peroxisomal beta-oxidation and peroxisomal omega-oxidation to the oxidation of a given fatty acid in vivo can be quantitated by an isotopic method. The approach requires infusion of a fatty acid labelled on two specific carbon atoms (e.g. [1-14C] and [11-14C] palmitate) to an isotopic steady state, with subsequent isolation and degradation of an acetylated conjugate as a product of the liver cytosolic acetyl CoA pool and of ketone bodies as a product of the liver mitochondrial acetyl CoA pool.

Models of the liver pentose cycle.

Simple and complete models of the classical liver pentose cycle, and a model of Williams' proposed "L-type" pentose cycle, are compared. All extant experimental data on well-oxygenated whole cell systems can be fitted to the predicted output of the complete classical pentose cycle model; however, there are gross discrepancies between key experimental data and Williams' proposed scheme. The complete classical model allows isotopic reversibility in the non-oxidative segment of the cycle, but none of the reversible enzymes are extremely close to isotopic equilibrium. General approaches are presented to estimate the isotopic reversibility of most enzymic steps, without requiring isolation of the intermediates, present in some cases at very low concentrations. The isotopic reversibility of the non-oxidative pathway causes only minor errors in the equations used to estimate liver pentose cycle flux, which were based on simple unidirectional models.

Tests of the liver specificity of drug glucuronidation.

Hellerstein and Landau and their coworkers have developed the glucuronide conjugate approach to aid in the analysis of pathways of liver carbohydrate metabolism. This approach requires that the liver is essentially the sole site of glucuronidation of the given drug. Since UDPglucuronyl transferases are present also in other tissues, most notably the kidney and intestines, we need to test the liver specificity of this process. We develop isotopic approaches to do this, based upon a comparison of the specific activity of the conjugate with that of plasma glucose and liver glucose-6P.

Isotopic estimation of the hepatic glucose balance in vivo.

Conventional determination of "hepatic glucose production" in the refed state is based on the 30-year-old Steele model, which omits glucose<-->glucose-6P cycling. Using the more complete model, conventional hepatic glucose production is shown to be a complex ratio of fluxes, and not a simple physiological flux. Our new model allows some prospective isotopic approaches to the estimation of glucose-6-phosphatase and glucokinase fluxes, and thus real net hepatic glucose production or uptake.

Estimation of peroxisomal and mitochondrial fatty acid oxidation in rat hepatocytes using tritiated substrates.

The pathways of peroxisomal and mitochondrial fatty acid oxidation were monitored with the use of substrates which produce NAD3H. I used as marker substrates: D-[3-3H]3-hydroxybutyrate for mitochondrial NAD3H production, [2-3H]glycerol for cytosolic NAD3H production, and [2-3H]acetate to measure carbon-bound 3H which was also generated by the metabolism of the commercial 9,10-3H-labelled fatty acids. The assumption that peroxisomal NAD3H can be considered to be equivalent to cytosolic NAD3H was supported using a specific inhibitor of mitochondrial fatty acid oxidation. The approach involves determination of the specific yields, and the relative distribution on carbons 4 and 6, of 3H in glucose from the marker substrates and the labelled fatty acids. In hepatocytes from clofibrate-treated rats, the amount of palmitate or oleate oxidation which starts in the peroxisomes is comparable with that which starts in the mitochondria.

Evidence against tight channelling of NADH in hepatocytes.

Tritiated substrates at tracer levels were incubated with rat hepatocytes plus 10 mM L-lactate, and the yields of tritium in glucose and water, as well as the tritium distribution on C-6 and C-4 of glucose, determined. Substrates of cytosolic type A NAD-linked dehydrogenases showed some preferential labeling of C-6 of glucose (the pathway involving type A malate dehydrogenase), whereas substrates of cytosolic type B NAD-linked dehydrogenases showed some preferential labeling of C-4 of glucose (the pathway involving type B glyceraldehyde-3P dehydrogenase). The results found are consistent with a classical diffusion model of NADH metabolism, and are at odds with the Srivastava hypothesis (based on isolated enzyme studies) which indicated that direct transfer of NADH can occur between many NAD-linked enzymes but only when they are of opposite (A or B) specificity.

Effects of salicylate on hepatocyte lactate metabolism.

We have examined the effects of salicylate on fluxes of lactate metabolism in rat hepatocytes using a steady state model. Salicylate produces an uncoupling effect, an inhibition of gluconeogenesis, a marked activation of pyruvate dehydrogenase flux, and an inhibition of endogenous fatty acid oxidation. Agents with known functions such as dinitrophenol and dichloroacetate were also compared in this system. The in vitro inhibition of gluconeogenesis caused by salicylate is not primarily related to the uncoupling effect. The fact that octanoate, but not palmitate, overcomes the salicylate inhibition of gluconeogenesis suggests that salicylyl CoA is involved in the inhibition. To relate in vitro studies to Reye's syndrome in vivo, in which medium chain dicarboxylic acids accumulate, we have also examined the effects of monomethyl suberate on liver lactate metabolism. This half ester is taken up by the hepatocytes, and causes inhibition of lactate gluconeogenesis, and uncoupling. Both salicylate and monomethyl suberate inhibit the oxidation of 0.2 mM octanoate by hepatocytes.

Errors in the isotopic estimations of hepatic glycogen synthesis and glucose output.

Ignoring the reality of glucose in equilibrium glucose-6P cycling in the liver in vivo produces large potential errors in the isotopic quantitation of percentage contributions of direct and indirect pathways of glycogen synthesis and of hepatic glucose output.

Possible role for carbamyl phosphate in the control of liver glycogen synthesis.

Isotopic data suggested a possible correlation between the activation of liver glycogen synthesis and some unstable intermediate. Carbamyl phosphate was one candidate considered. L-Norvaline, an inhibitor of ornithine transcarbamylase, is known to increase intracellular carbamyl phosphate levels. L-Norvaline augmented the increase in glycogen synthesis caused by L-glutamine and L-alanine, and also induced considerable glycogen synthesis with ammonia as the nitrogen source.

Inhibition of glycogen synthesis in rat hepatocytes by medium Zn2+.

In hepatocytes from fasted rats, Zn2+ in the range from 0 to 500 microM has relatively minor effects on gluconeogenesis from most substrates, or on ureagenesis from NH3. In hepatocytes from fed rats, Zn2+ does not affect glycogenolysis. In hepatocytes from fasted rats, in which glycogen is being actively synthesized using the substrate combination (Katz et al. (1976) Proc. Natl.Acad.Sci.USA 73,3433-3437) of glucose, lactate and glutamine (all 10mM), Zn2+ markedly inhibits glycogen synthesis, with total inhibition at 500 microM, and a half maximal effect in the range from 50 to 100 microM. Dipicolinate (pyridine 2,6-dicarboxylate), a zinc chelator, is about as effective as L-glutamine in activating glycogen synthesis with the substrate combination of dihydroxyacetone, lactate and glucose (all 10mM). This suggests the possible hypothesis that endogenous Zn2+ might control the rate of glycogen synthesis in vivo. However, alternate explanations such as metabolite accumulation are also possible, since dipicolinate causes inhibition of gluconeogenesis from L-lactate.

Gluconeogenesis in rat hepatocytes from monomethyl succinate and other esters.

Although little glucose is formed from succinate in rat hepatocytes, the rate of gluconeogenesis from monomethyl succinate approaches that from L-lactate. Dimethyl succinate is as good as monomethyl succinate at 5 mM, but not at 20 mM. Monoethyl fumarate and 4-methyl malate are only fair glucogenic substrates, but 1-methyl malate is another good substrate at high concentrations. The esters are apparently taken up either directly through the cell membrane, or by monocarboxylate transporters, and then hydrolyzed intracellularly by some esterase(s). This approach may permit the use of a wider range of substrates and inhibitors for the study of liver cell metabolism.

The role of mitochondrial pyruvate transport in the control of lactate gluconeogenesis.

At low L-lactate concentrations, plots of the reciprocal of the gluconeogenic rate vs hydroxycyanocinnamate concentration were linear up to high inhibitor concentrations, indicating that the pyruvate transporter was a rate-limiting step and alternate pathways were at best minor. At 10 mM L-lactate and in the absence of added acids, the 1/V vs I plots became sigmoidal, indicating both some excess capacity of the transporter, and a significant alternative pathway. Use of transaminase inhibitors suggests that the alternate pathway does not primarily involve a dual glutamate-pyruvate transaminase mechanism.

Further evidence for the classical pentose phosphate cycle in the liver.

Isolated rat hepatocytes were incubated with [3-(14)C]xylitol or d-[3-(14)C]xylulose plus xylitol or glucose at substrate concentrations. The glucose formed was isolated and degraded to give the relative specific radioactivities in each carbon atom. C-4 of glucose had the highest specific radioactivity, followed by C-3, with half to one-fifth that of C-4. Only about 1% of the total radioactivity was in C-1. The data are compared with the predictions of the classical pentose phosphate cycle [Horecker, Gibbs, Klenow & Smyrniotis (1954) J. Biol. Chem.207, 393-403], and the proposed new version of the pentose phosphate cycle in liver [Longenecker & Williams (1980) Biochem. J.188, 847-857], which they denoted as the ;L-type pentose cycle'. The Williams pathway predicts that the specific radioactivity of C-1 of glucose should be half that of C-4 (after correction for approximately equal labelling on C-3 and C-4 of hexose phosphate in the pathway involving fructose 1,6-bisphosphatase). The actual labelling in C-1 is 20-350-fold less than this. When the hepatocytes are incubated with phenazine methosulphate, to stimulate the oxidative branch of the pentose phosphate cycle, the predicted relationship between (C-2/C-3) and (C-1/C-3) ratios of specific radio-activities is nearly exactly in accord with the classical pentose phosphate cycle. Glucose and glucose 6-phosphate were isolated and degraded from an incubation of hepatocytes from starved/re-fed rats with [3-(14)C]xylitol. Although the patterns were of the classical type, there was more randomization of (14)C into C-2 and C-1 in the glucose 6-phosphate isolated at the end of the incubation than in the glucose which was continuously produced.

14CO2 fixation by phosphoenolpyruvate carboxykinase during gluconeogenesis in the intact rat liver cell.

During gluconeogenesis from L-glutamine, 14CO2 is fixed into glucose. Inhibitors of pyruvate transport or pyruvate carboxylase only slightly decrease the 14CO2 incorporation, indicating that a pathway of formation of pyruvate, followed by pyruvate carboxylation, is not primarily involved. These results suggest that 14CO2 fixation is effected by a reverse (exchange) reaction of P-enolpyruvate carboxykinase. MnCl2 (0.5 mM) stimulates the 14CO2 fixation in glucose from L-glutamine by nearly 50%. This result is in accord with a recent study (Colombo, G., Carlson, G. M., and Lardy, H. A. (1981) Biochemistry 20, 2749-2757) showing that Mn2+ greatly stimulates the reverse reaction (P-enolpyruvate leads to oxalacetate) of purified rat liver P-enolpyruvate carboxykinase. Preliminary calculations suggest that 14CO2 is also fixed by reversible P-enolpyruvate carboxykinase activity during gluconeogenesis from L-lactate, in addition to the fixation of H14CO3(-) in the pyruvate carboxylase forward reaction.