Enhanced glucose cycling and suppressed de novo synthesis of glucose-6-phosphate result in a net unchanged hepatic glucose output in ob/ob mice
- First Online:
- Cite this article as:
- Bandsma, R.H.J., Grefhorst, A., van Dijk, T.H. et al. Diabetologia (2004) 47: 2022. doi:10.1007/s00125-004-1571-8
- 257 Downloads
Leptin-deficient ob/ob mice are hyperinsulinaemic and hyperglycaemic; however, the cause of hyperglycaemia remains largely unknown.
Glucose metabolism in vivo in 9-h fasted ob/ob mice and lean littermates was studied by infusing [U-13C]-glucose, [2-13C]-glycerol, [1-2H]-galactose and paracetamol for 6 h, applying mass isotopomer distribution analysis on blood glucose and urinary paracetamol-glucuronide.
When expressed on the basis of body weight, endogenous glucose production (109±23 vs 152±27 µmol·kg–1·min–1, obese versus lean mice, p<0.01) and de novo synthesis of glucose-6-phosphate (122±13 vs 160±6 µmol·kg–1·min–1, obese versus lean mice, p<0.001) were lower in ob/ob mice than in lean littermates. In contrast, glucose cycling was greatly increased in obese mice (56±13 vs 26±4 µmol·kg–1·min–1, obese versus lean mice, p<0.001). As a result, total hepatic glucose output remained unaffected (165±31 vs 178±28 µmol·kg–1·min–1, obese vs lean mice, NS). The metabolic clearance rate of glucose was significantly lower in obese mice (8±2 vs 18±2 ml·kg–1·min–1, obese versus lean mice, p<0.001). Hepatic mRNA levels of genes encoding for glucokinase and pyruvate kinase were markedly increased in ob/ob mice.
Unaffected total hepatic glucose output in the presence of hyperinsulinaemia reflects hepatic insulin resistance in ob/ob mice, which is associated with markedly increased rates of glucose cycling. Hyperglycaemia in ob/ob mice primarily results from a decreased metabolic clearance rate of glucose.
KeywordsGlucose-6-phosphate Gluconeogenesis Glycogen Glycogenolysis Hepatic glucose production Mass isotopomer distribution analysis ob/ob mice Stable isotopes
mass isotopomer distribution analysis
peroxisome proliferator-activated receptor gamma
sterol regulatory element-binding protein-1c
Hyperinsulinaemia and fasting hyperglycaemia are hallmarks of type 2 diabetes. Insulin resistance of peripheral organs (muscle and adipocytes), as well as of the liver, may contribute to fasting hyperglycaemia. Peripheral insulin resistance reduces the ability of peripheral organs to clear glucose from the circulation. Hepatic insulin resistance develops in two stages. During the early stages in the development of type 2 diabetes, characterised by hyperinsulinaemia and normoglycaemia, hepatic glucose production is still normal under fasting conditions. However, during absorptive phases when insulin concentrations are elevated, hepatic glucose production remains inappropriately high. At later stages in the development of type 2 diabetes in humans, hepatic glucose production starts to increase even under fasting conditions .
Both gluconeogenesis and glycogenolysis may contribute to elevated hepatic glucose production. Furthermore, data indicates that cycling of glucose, the process of sequential glucose uptake and subsequent phosphorylation by glucokinase and dephosphorylation by glucose-6-phosphatase, occurs at increased rates in humans with type 2 diabetes [2, 3]. Little is known about the quantitative role of glucose cycling in the increased production of hepatic glucose in type 2 diabetes. Depending on the methodologies used for quantification of hepatic glucose fluxes, increased glucose cycling may affect the estimation of rates of gluconeogenesis and glycogenolysis.
Leptin-deficient ob/ob mice suffer from severe obesity and diabetes due to leptin deficiency, and provide a model for type 2 diabetes. These mice exhibit age-dependent hyperglycaemia and hyperinsulinaemia. Quantitative data on the perturbations of glucose metabolism in these mice in vivo are scarce. In vitro studies on perfused isolated livers of ob/ob mice have shown that glycogen turnover is increased . In addition, glucose cycling rates have been shown to be greatly increased in hepatocytes isolated from 24-h fasted ob/ob mice .
Novel methodologies using multiple stable isotopes in vivo now allow the determination of flux rates through the separate metabolic pathways involved in hepatic carbohydrate metabolism [6, 7, 8]. In the current study, we used these methods to evaluate the quantitative role of gluconeogenesis, glycogenolysis and glucose cycling in hyperglycaemia in modestly fasted ob/ob mice.
Materials and methods
Female ob/ob mice (n=7) and lean littermates (n=7), 8 weeks of age and on a C57Bl/6 genetic background, were purchased from Harlan (Zeist, The Netherlands). The mice were housed in a temperature-controlled (21 °C) room with a dark–light cycle of 12 h each. Experimental procedures were approved by the Ethics Committee for Animal Experiments of the State University Groningen. Mice were fitted with a permanent catheter in the right atrium via the right jugular vein, as described previously . Mice were allowed to recover from surgery for at least 4 days.
The following isotopes were used: [2-13C]-glycerol (99% 13C atom percent excess), [1-2H]-galactose (98% 2H atom percent excess) (Isotec, Miamisburg, Ohio, USA), [U-13C]-glucose (99% 13C atom percent excess) (Cambridge Isotope Laboratories, Andover, Mass., USA). All chemicals used were reagent pro analysis grade. Blood spots and urine were collected on Schleicher and Schuell No. 2992 filter paper (Schleicher and Schuells’, Hertogenbosch, The Netherlands). Infusates were freshly prepared and sterilised by the Hospital Pharmacy at the day before the experiment.
Experiments were performed in awake, chronically catheterised mice, essentially as described previously . Mice were fasted for 9 h, after which they were placed in metabolic cages to allow frequent collection of blood spots and urine. Mice were infused with a sterile solution containing [U-13C]-glucose (13.9 µmol/ml), [2-13C]-glycerol (160 µmol/ml), [1-2H]-galactose (33 µmol/ml) and paracetamol (1.0 mg/ml) at a rate of 0.6 ml/h. During the experiment, blood glucose was measured using EuroFlash test strips (LifeScan Benelux, Beerse, Belgium). Blood spots were collected on filter paper before the start of the infusion and hourly afterwards until 6 h after the start of the infusion. Blood spots were air-dried and stored at room temperature until analysis. Timed urine samples were collected on filter paper strips at hourly intervals. Strips were air-dried and stored at room temperature until analysis. At the end of the experiment, animals were anaesthetised with isofurane, and a large blood sample was collected in heparin-containing tubes by heart puncture. The sample was centrifuged immediately and stored at −20 °C until analysis. The liver was quickly excised, weighed and immediately frozen in liquid nitrogen.
Determination of metabolite concentrations
Plasma was isolated from blood by centrifugation, and liver tissue was homogenised. Commercially available kits were use to determine plasma levels of β-hydroxybutyrate, lactate (Roche Diagnostics, Mannheim, Germany) and NEFA (Wako Chemicals, Neuss, Germany). Plasma insulin levels were determined by RIA (RI-13K; Linco Research, St. Charles, Mo., USA). Total liver protein content was determined according to the method of Lowry et al. .
Hepatic glycogen was determined by sonication after extraction with 1 mol/l KOH. The extract was incubated at 90 °C for 30 min, cooled and then adjusted to pH 4.5 by the addition of 3 mol/l acetic acid. Precipitated protein was removed by centrifugation. Glycogen was converted to glucose by treating the samples with amyloglucosidase. A glucose assay was then performed at pH 7.4 with ATP, NADP+, hexokinase and G6P dehydrogenase.
Liver samples for the determination of G6P were treated by sonication in a 5% (w/v) HClO4 solution. Precipitated protein was removed by rapid centrifugation at 20000 g for 1 min in a cold microcentrifuge, and the supernatant was neutralised to pH 7 by the addition of small amounts of a solution containing 2 mol/l KOH and 0.3 mol/l MOPS. Levels of G6P were determined fluorimetrically with NADP+ and G6P dehydrogenase.
Hepatic mRNA levels
Sequences of the primers and probes used in PCR measurements
ACC CAC ACT GTG CCC ATC TAC
GCT CGG TCA GGA TCT TCA TGA
AGG GCT ATG CTC TCC CTC ACG CCA
CGG CTA CCA CAT CCA AGG A
CCA ATT ACA GGG CCT CGA AA
CGC GCA AAT TAC CCA CTC CCG A
CTG CAA GGG AGA ACT CAG CAA
GAG GAC CAA GGA AGC CAC AAT
TGC TCC CAT TCC GCT TCG CCT
GAG GCC TTG TAG GAA GCA TTG
CCA TCC CAG CCA TCA TGA GTA
CTC TGT ATG GGA ACC CTC GCC ACG
CCT GGG CTT CAC CTT CTC CTT
GAG GCC TTG AAG CCC TTG GT
CAC GAA GAC ATA GAC AAG GGC ATC CTG CTC
GAA GGA GGC AAA CGG ATC AAC
TCA CGA TGT CCG AGT GGA TCT
CCT CTG CAT CGT GGG CTG CCA
GCT CTC CAG ACG ATT CTT GCA
GTG CGG TTC CTC TGA ATG ATC
CCT CTA CGG GTT TTG TAA ACA GTC ACG CC
CGT TTG TGC CAC ACA GAT GCT
CAT TGG CCA CAT CGC TTG TCT
AGC ATG ATC ACT AAG GCT CGA CCA ACT CGG
GTG TCA TCC GCA AGC TGA AG
CTT TCG ATC CTG GCC ACA TC
CAA CTG TTG GCT GGC TCT CAC TGA CCC
Mass isotopomer distribution analysis
Glucose and paracetamol-glucuronic acid (Par-GlcUA) were extracted from blood spot and urine filter paper strips respectively, derivatised, and measured by GC-MS, essentially as described previously [8, 10]. The fractional isotopomer distribution according to GC-MS (m0–m6) was corrected for the fractional distribution due to the natural abundance of 13C by multiple linear regression, as described by Lee et al.  to obtain the excess mole fraction of mass isotopomers M0–M6 due to incorporation of infused labelled compounds, i.e. [2-13C]-glycerol, [U-13C]-glucose and [1-2H]-galactose.
In contrast to blood glucose, the total rate of appearance of UDP-glucose is determined by: (i) gluconeogenic flux from G6P; (ii) glycogenolysis; and (iii) the flux of blood glucose into the UDP-glucose pool. The flux of glycogen into UDP-glucose is a measure of glycogen / glucose-1-phosphate cycling .
All values are means ± SD. Levels of significance of difference of metabolite concentrations, gene expression and the values of the individual time points during isotope infusion experiments were determined using the non-parametric Mann–Whitney test for unpaired data. Levels of significance of differences between the averages of the values of the fluxes at individual time points between 3 and 6 h during the experiment were estimated using repeated measures ANOVA. A p value of less than 0.05 was considered statistically significant.
Hepatic and plasma parameters in ob/ob mice and lean littermates
Lean mice (n=7)
ob/ob mice (n=7)
Body weight (g)
Liver weight (g)
2.6 ± 0.4a
Relative liver weight (% body weight)
5.0 ± 0.8
Total liver protein (mg)
Liver protein content (mg protein/g liver weight)
G6P (nmol/g liver weight)
Glycogen (µmol glucose/g liver weight)
Summary of the calculated values of the various flux rates, using the isotopic model shown in Fig. 1, normalised to either body weight or liver protein
Normalised to body weight (µmol·kg–1·min–1)
Normalised to liver protein (µmol·g protein–1·min–1)
De novo synthesis of G6P (phosphenolpyruvate carboxykinase)
Endogenous glucose production
Total endogenous glucose production (glucose-6-phosphatase)
The leptin-deficient ob/ob mouse is a commonly used mouse model of type 2 diabetes, but quantitative in vivo data on the disturbances that underlie hyperglycaemia in this model are sparse. In this study, we determined flux rates through various pathways relevant in hepatic carbohydrate metabolism in lean and ob/ob mice. When expressed per unit of body weight or liver protein, hepatic glucose metabolism activity was, in general, suppressed in obese mice compared with that in their lean littermates. However, glucose cycling was an exception to this, and was observed to be greatly increased in obese mice. Interestingly, the newly produced G6P was not preferentially directed towards plasma glucose in ob/ob mice, but instead was partitioned to glycogen stores to a similar extent as that observed in lean mice. Furthermore, the expression of genes of key enzymes involved in glucose metabolism were similar in livers of ob/ob and lean mice, apart from the expression of glucokinase and liver-type pyruvate kinase which was increased in the liver of ob/ob mice.
Before discussing the results, some methodological issues have to be addressed. In this study, a multiple isotope infusion protocol was used to calculate the relevant fluxes of glucose metabolism . The validity of the isotope model, with the application of glycoconjugates, and the mass isotopomer distribution analysis (MIDA) approach has been substantiated in various studies, although some controversy still remains [14, 15]. Since the contribution of glycolysis to intracellular G6P metabolism has not been included in this model, the calculated flux rate through glucokinase represents a minimal estimate. We have validated the application of MIDA in 9-h fasted C57Bl/6 mice in a separate study . In 24-h fasted mice, no stable isotopic steady-state could be obtained . In the current study, we compared hepatic glucose metabolism in groups of mice with strongly different body compositions. Our data show that, with respect to hepatic glucose metabolism, normalisation to body weight appeared to be appropriate, since the same conclusions could be drawn regardless of whether the data was normalised to body weight or liver protein.
Irrespective of hyperinsulinaemia and hyperglycaemia, total glucose output (glucose-6-phosphatase flux) was not affected, while endogenous glucose production was only modestly inhibited in ob/ob mice. This points to hepatic insulin resistance. Previously reported values of endogenous glucose production in C57Bl/6 mice are almost identical to those reported in this study in lean littermates of ob/ob mice [16, 17]. Furthermore, endogenous glucose production in C57Bl/6 mice could be suppressed almost completely during hyperinsulinaemic clamp at normal or increased glucose concentrations [16, 17] (and unpublished observations, A. Grefhorst et al.). The impaired suppression of endogenous glucose production was mainly due to the blunted response of de novo synthesis of G6P to the combined hyperinsulinaemia and hyperglycaemia in ob/ob mice. This indicates that, in the absence of leptin, insulin appears to be largely ineffective in suppressing hepatic de novo synthesis of G6P by an as yet unknown mechanism.
In the liver of ob/ob mice, the decreased contribution of endogenous glucose production to total glucose output appeared to be compensated by enhanced glucose cycling. Glucose cycling was increased by a factor of ~2.5 in the liver of ob/ob mice, due to an enhanced flux through glucokinase. A previous study reported a high rate of glucose cycling in hepatocytes isolated from the liver of ob/ob mice fasted for 24 h . Similarly, in an earlier publication, glucokinase activity was found to remain elevated in the liver of ob/ob mice throughout a 48-h fast . Collectively, these observations indicate that, independent of the duration of fasting, the liver of ob/ob mice maintains a high capacity to phosphorylate glucose.
Besides hepatic insulin resistance, peripheral organs were also found to be insulin resistant in ob/ob mice. Metabolic clearance of plasma glucose was decreased by a factor of ~2 at blood glucose concentrations that were almost double that in lean mice. This indicates that net glucose uptake by peripheral tissue was similar in ob/ob and lean mice, irrespective of the elevated insulin concentrations in the obese group. Thus, hyperglycaemia in ob/ob mice is due to peripheral insulin resistance. This finding is in agreement with an earlier study using different means to investigate peripheral insulin resistance in ob/ob mice, which reported that uptake of 2-deoxyglucose was severely inhibited in isolated skeletal muscle of obese mice compared with that in lean mice .
As discussed, we observed ‘normal’ rates of total glucose output and high rates of glucose cycling. In accordance with these observations, ‘normal’ mRNA levels of the gluconeogenic enzymes G6P hydrolase and phosphoenolpyruvate carboxykinase were observed, while mRNA levels of glucokinase and liver-type pyruvate kinase were significantly increased in the liver of obese mice compared to those in liver of lean littermates. Significant increases in mRNA levels of Srebp-1c and its target genes in lipogenesis, i.e. Fas and Acc1, in livers of fasted ob/ob mice have previously been reported [13, 20]. This indicates an enhanced glycolytic flux into lipogenesis. It should be realised that the glycolytic flux as part of the glucokinase flux cannot be assessed in the isotopic model applied. Recent data indicate that hyperglycaemia could directly induce increased expression of the genes encoding Srebp-1c and pyruvate kinase in an insulin-independent way .
In humans with type 2 diabetes, there is evidence for enhanced gluconeogenesis and glycogenolysis after an overnight fast, particularly in patients with severe fasting hyperglycaemia [1, 22, 23, 24, 25, 26, 27]. Although liver was insulin resistant in ob/ob mice, this did not result in enhanced hepatic glucose production. In the present study, only a moderate fasting hyperglycaemia was observed in ob/ob mice at 8 weeks of age. Apparently, in these mice the disease had not yet progressed to a more severe stage with (very) high fasting blood glucose concentrations. Furthermore, in most studies on (often obese) diabetic subjects, the data was normalised to lean body mass instead of body weight, which might have led to seemingly elevated values for gluconeogenic and glycogenolytic fluxes in these individuals compared with those in non-diabetic subjects. Until now, only very few studies have considered the role of glucose cycling in hepatic glucose production. There are indications that hepatic cycling of glucose is elevated in humans with type 2 diabetes [2, 3].
In conclusion, this study demonstrates that in ob/ob mice, de novo synthesis of G6P is diminished while glucose cycling is increased, resulting in an unaffected total glucose output by the liver. However, these observations were made where there was a background of hyperglycaemia and hyperinsulinaemia. This points to a co-existence of hepatic and peripheral insulin resistance, with peripheral insulin resistance as the cause of hyperglycaemia.
This work was supported by the Dutch Diabetes Foundation (grant 96.604). R.H.J. Bandsma is supported by the Dutch Organisation for Scientific Research (NWO). We thank T. Boer, P. Modderman and T. Jager for excellent technical assistance.