Nurjahan Nurjhan

Metabolic Effects of Metformin in Non-Insulin-Dependent Diabetes Mellitus

BACKGROUND The metabolic effects and mechanism of action of metformin are still poorly understood, despite the fact that it has been used to treat patients with non-insulin-dependent diabetes mellitus (NIDDM) for more than 30 years. METHODS In 10 obese patients with NIDDM, we used a combination of isotope dilution, indirect calorimetry, bioimpedance, and tissue-balance techniques to assess the effects of metformin on systemic lactate, glucose, and free-fatty-acid turnover; lactate oxidation and the conversion of lactate to glucose; skeletal-muscle glucose and lactate metabolism; body composition; and energy expenditure before and after four months of treatment. RESULTS Metformin treatment decreased the mean (+/- SD) glycosylated hemoglobin value from 13.2 +/- 2.2 percent to 10.5 +/- 1.6 percent (P < 0.001) and reduced fasting plasma glucose concentrations from 220 +/- 41 to 155 +/- 28 mg per deciliter (12.2 +/- 0.7 to 8.6 +/- 0.5 mmol per liter) (P < 0.001). Although resting energy expenditure did not change, the patients lost 2.7 +/- 1.3 kg of weight (P < 0.001), 88 percent of which was adipose tissue. The mean (+/- SE) rate of plasma glucose turnover (hepatic glucose output and systemic glucose disposal) decreased from 2.8 +/- 0.2 to 2.0 +/- 0.2 mg per kilogram of body weight per minute (15.3 +/- 0.9 to 10.8 +/- 0.9 mumol per kilogram per minute) (P < 0.001), as a result of a decrease in hepatic glucose output; systemic glucose clearance did not change. The rate of conversion of lactate to glucose (gluconeogenesis) decreased by 37 percent (P < 0.001), whereas lactate oxidation increased by 25 percent (P < 0.001). There were no changes in the plasma lactate concentration, plasma lactate turnover, muscle lactate release, plasma free-fatty-acid turnover, or uptake of glucose by muscle. CONCLUSIONS Metformin acts primarily by decreasing hepatic glucose output, largely by inhibiting gluconeogenesis. It also seems to induce weight loss, preferentially involving adipose tissue.

Predominant Role of Gluconeogenesis in Increased Hepatic Glucose Production in NIDDM

Excessive hepatic glucose output is an important factor in the fasting hyperglycemia of non-insulindependent diabetes mellitus (NIDDM). To determine the relative contributions of gluconeogenesis and glycogenolysis in a quantitative manner, we applied a new isotopic approach, using infusions of [6-3H]glucose and [2-14C]acetate to trace overall hepatic glucose output and phosphoenolpyruvate gluconeogenesis in 14 postabsorptive NIDDM subjects and in 9 nondiabetic volunteers of similar age and weight. Overall hepatic glucose output was increased nearly twofold in the NIDDM subjects (22.7 ± 1.0 vs. 12.0 ± 0.6 μmol · kg−1 · min−1 in the nondiabetic volunteers, P < .001); phosphoenolpyruvate gluconeogenesis was increased more than threefold in the NIDDM subjects (12.7 ± 1.4 vs. 3.6 ± 0.4 μmol kg−1 min−1 in the nondiabetic subjects, P < .001) and was accompanied by increased plasma lactate, alanine, and glucagon concentrations (all P < .05). The increased phosphoenolpyruvate gluconeogenesis accounted for 89 ± 6% of the increase in overall hepatic glucose output in the NIDDM subjects and was significantly correlated with the fasting plasma glucose concentrations (r = .67, P < .01). Glycogenolysis, calculated as the difference between overall hepatic glucose output and phosphoenolpyruvate gluconeogenesis, was not significantly different in the NIDDM subjects (9.9 ± 0.06 μmol · kg−1 · min−1) and the nondiabetic volunteers (8.4 ± 0.3 μmol kg−1 · min−1). We conclude that increased gluconeogenesis is the predominant mechanism responsible for increased hepatic glucose output in NIDDM.

Mechanism for Underestimation of Isotopically Determined Glucose Disposal

Use of [3H]glucose and a one-compartment model to determine glucose kinetics frequently underestimates the rate of glucose production (Ra). To assess to what extent an isotope effect, a tracer contaminant, or inadequacy of the model was responsible, we measured glucose Ra and forearm clearance of tracer and unlabeled glucose at various concentrations of plasma insulin (∼50, ∼160, and ∼1800 μU/ml) and plasma glucose (∼90, ∼160, ∼250, and ∼400 mg/dl) under steady-state and non-steady-state conditions. Under isotopic steady-state conditions, the clearances of tracer and unlabeled glucose across the forearm were identical, and exogenous glucose infusion rates did not differ significantly from the isotopically determined glucose Ra (10.0 ±1.3 vs. 10.5 ± 1.0 mg ⋅ kg−1 fat-free mass ⋅ min−1, respectively). However, under isotopic non-steady-state conditions, the isotopically determined Ra was significantly lower than the glucose infusion rate (11.5 ± 1.3 vs. 13.7 ± 1.5 mg ⋅ kg−1 fat-free mass ⋅ min−1, respectively, P < .001), and the underestimation was related to the deviation from the isotopic steady state. When [3H]glucose specific activity of plasma samples from experiments with the greatest underestimation of Ra was determined by high-performance liquid chromatography, <7% of the underestimation could be accounted for by a contaminant. These results indicate that inadequacy of the one-compartment model is responsible for underestimation of glucose Ra under non-steady-state conditions and that there is no detectable isotopic effect or appreciable contaminant of [3-3H]glucose. We conclude that under isotopic steady-state conditions, [3-3H]glucose is a reliable tracer for glucose kinetic studies in humans.

Glutamine and Alanine Metabolism in NIDDM

Gluconeogenesis is increased in NIDDM. We therefore examined the metabolism of glutamine and alanine, the most important gluconeogenic amino acids, in 14 postabsorptive NIDDM subjects and 18 nondiabetic volunteers using a combination of isotopic ([6-3H] glucose (20 µCi, 0.2 µCi/min), [U-14C]glutamine (20 µCi, 0.2 µCi/min), [3-13C]alanine (99% 13C, 2 mmol, 20 µmol/min), [ring-2H5]phenylalanine (99% 2H, 2 µmol/kg, 0.03 µmol · kg-1 · min-1), and limb balance techniques. Alanine turnover (4.54 ± 0.24 vs. 5.64 ± 0.33 µmol · kg-1 · min-1), de novo, synthesis (3.00 ± 0.25 vs. 4.01 ± 0.33 µmol · kg-1 · min-1) were increased in NIDDM subjects (all P < 0.001), while its forearm release (0.45 ± 0.04 vs. 0.39 ± 0.04 µmol · kg-1 · min-1) was unaltered. Although glutamine turnover (4.81 ± 0.23 vs. 4.40 ± 0.31 µmol · kg-1 · min-1) was unaltered in NIDDM, its conversion to glucose (0.57 ± 0.04 vs. 1.08 ± 0.10 µmol · kg-1 · min-1) and to alanine (0.10 ± 0.01 vs. 0.34 ± 0.04 µmol · kg-1 · min-1) (both P = 0.001) was increased while its oxidation (2.84 ± 0.27 vs. 1.84 ± 0.15 µmol kg-1 · min-1 P = 0.03) and forearm release (0.77 ± 0.05 vs. 0.62 ± 0.09 µmol · kg-1 · min-1 P < 0.008) were both reduced. Our results thus demonstrate that there are substantial alterations of glutamine and alanine metabolism in NIDDM. Conversion of both amino acids to glucose and the proportion of their turnover used for gluconeogenesis are increased; release of both amino acids from tissues other than skeletal muscle seems to be increased. Finally, the reduction in glutamine oxidation, possibly the result of competition with glucose and free fatty acids as fuels, makes more glutamine available for gluconeogenesis without a change in its turnover.

An improved method to calculate adipose tissue interstitial substrate recovery for microdialysis studies.

We simultaneously compared the conventional, time-consuming point of no net flux method for calculation of interstitial substrate recovery necessary for in vivo microdialysis studies with a simple isotopic method using rat epididymal fat pads. The recovery (%) calculated with the conventional method and the isotopic method for glucose (7.4 +/- 1.1 vs. 6.6 +/- 0.6), glycerol (23 +/- 4 vs. 26 +/- 5) and lactate (40 +/- 8 vs. 38 +/- 5), respectively, were not significantly different. Moreover, the overall correlation coefficient (N = 25) between the methods was 0.87, p < 0.001. We therefore conclude that the methods yield comparable results, and the more convenient isotopic method should become the method of choice for determining adipose tissue interstitial recovery for glucose, lactate and glycerol.

Dose‐response effects of lactate infusions on gluconeogenesis from lactate in normal man

Abstract. Lactate is the predominant gluconeogenic precursor in man. To determine the dose‐response relationships between plasma lactate concentration and rates of lactate incorporation in plasma glucose (lactate gluconeogenesis, LGN), we infused 17 normal volunteers with sodium lactate for 180 min at rates ranging from 6 to 40 γmol kg‐1 min‐1 and measured [U‐14C]lactate incorporation into plasma glucose, as well as rates of lactate and glucose appearance in plasma. With the highest lactate infusions, plasma lactate increased up to 7 mM (compared to 1.1±0.13 mM during control sodium bicarbonate infusions, n=10) and LGN averaged 4.73 ± 0.23 μmol kg‐1 min‐1 (compared to 1.57 ± 0.26 μmol kg‐1 min‐l in bicarbonate control experiments, P< 0.001). The data relating plasma lactate concentration to LGN best fit a sigmoidal curve which plateaued at plasma lactate concentrations of approximately 6 mM and yielded an ED50 of 2.04 ± 0.20 (SD) mM and a Vmax (6.25±1.2) (SD) (mUmol kg‐1 min‐1). The sum of the basal rate of lactate appearance and the rate of lactate infusion was not significantly different from the overall rates of lactate appearance during the lactate infusions (35.8±2.2 vs. 34.8±2.9 μmol kg‐1 min‐1, P = 0.23). Thus, our results support the view that infusion of exogenous lactate does not suppress endogenous lactate appearance in plasma.

Contribution of Impaired Muscle Glucose Clearance to Reduced Postabsorptive Systemic Glucose Clearance in NIDDM

The reduced postabsorptive rates of systemic glucose clearance in non-insulin-dependent diabetes mellitus (NIDDM) are thought to be the consequence of insulin resistance in peripheral tissues. Although the peripheral tissues involved have not been identified, it is generally assumed to be primarily muscle, the major site of insulin-mediated glucose disposal. To test this hypothesis, we measured postabsorptive systemic and forearm glucose utilization and clearance in 15 volunteers with NIDDM and 15 age- and weightmatched nondiabetic volunteers. Although systemic glucose utilization was increased in NIDDM subjects (14.5 ± 0.5 vs. 11.2 ± 0.2 μmol · kg−1 · min−1, P < 0.001), systemic glucose clearance was reduced 1.40 ± 0.06 vs. 2.13 ± 0.05 ml · kg−1 · min−1, P < 0.01). Although forearm glucose utilization was increased in NIDDM subjects (0.663 ± 0.058 vs. 0.411 ± 0.019 μmol · dl−1 · min−1, P < 0.001), forearm glucose dl−1 clearance was reduced (0.628 ± 0.044 vs. 0.774 ± 0.037 ml · L−1 · min−1, P < 0.01). However, extrapolation of forearm data to total-body muscle indicated that impaired clearance reduced muscle glucose disposal by only 61 ± 21 μmol<min, whereas impaired systemic clearance reduced systemic glucose disposal by 662 ± 82 μmol<min. Thus, impaired muscle glucose clearance accounted for <10% of the reduced systemic glucose clearance in NIDDM subjects. Therefore, we conclude that muscle insulin resistance plays only a minor role in the reduced systemic glucose clearance found in NIDDM in the postabsorptive state and propose that reduced brain glucose clearance is largely responsible.

Contribution of Gluconeogenesis to Overall Glucose Output in Diabetic and Nondiabetic Men

Increased hepatic glucose output is the main cause of fasting hyperglycemia in non-insulin dependent diabetes mellitus. Due to difficulties in obtaining a quantitative estimate of gluconeogenesis in vivo, the relative contribution of gluconeogenesis and glycogenolysis to this increased hepatic glucose output was unknown. The application in vivo of a new isotopic approach based on a mathematical model of the Krebs cycle enabled us to obtain a quantitative estimate of gluconeogenesis in vivo. Using this approach, gluconeogenesis was found to account for approximately 28% and approximately 97% of overall hepatic glucose output in healthy volunteers in the postabsorptive and in the fasted state respectively. When this technique was used to compare gluconeogenesis rates in non-insulin dependent diabetes mellitus and nondiabetic patients, gluconeogenesis was found to be increased threefold in the patients with non-insulin dependent diabetes mellitus (12.7 +/- 1.6 mu vs 3.6 +/- 0.6 mumol/Kg/min) and to be significantly correlated with fasting plasma glucose. Furthermore, the increase in gluconeogenesis could explain more than 80% of the increase in overall hepatic glucose output in patients with non-insulin dependent diabetes mellitus. In conclusion, in non-insulin dependent diabetes mellitus, gluconeogenesis, as measured by a new isotopic technique, is increased and this increase represents the main cause for increased overall hepatic glucose output and fasting hyperglycemia.

Limitations in the Use of [2-14C]Acetate for Measuring Gluconeogenesis In Vivo

This study was undertaken to test two assumptions critical for use of [2-14C]acetate to measure gluconeogenesis in vivo. For the assumption that incorporation into glucose of products of [14C]acetate metabolism does not affect the distribution of label within the glucose molecule, we infused [2-14C]acetate in 17 healthy subjects and [3-14C]lactate in 10 healthy subjects and compared the ratio of the resultant specific activities of plasma glucose carbons 1, 2, 5, 6, and 3, 4 obtained with each tracer. The ratio obtained with [2-14C]acetate (2.99 ± 0.07) was significantly different from the ratio obtained with [3-14C]lactate, (3.82 ± 0.2, P < 0.01). Because the model predicts that these ratios should be identical, these results indicate that either the model is incorrect or that metabolism of [14C]acetate to other compounds affects the distribution of the label within the glucose molecule. To test the assumption that plasma 3-OH-butyrate specific activity approximates the specific activity of hepatic intramitochondrial acetyl CoA, we compared the ratio of specific activities of plasma glucose and 3-OH-butyrate obtained in 7 healthy subjects infused with [2-14C]acetate and [2-14C]octanoate. The ratio obtained with [2-14C]acetate (0.18 ± 0.03) was significantly different from that obtained with [2-14C]octanoate, (0.10 ± 0.02), P < 0.001. These results suggest compartmentalization of acetyl CoA within liver mitochondria and indicate that plasma 3-OH-butyrate specific activity may not necessarily approximate intramitochondrial acetyl CoA specific activity during [2-14C]acetate infusion. We conclude that assumptions critical for use of [2-14C]acetate to measure gluconeogenesis in vivo are not valid.

Gluconeogenesis in Type 2 Diabetes

There is now considerable evidence that increased hepatic glucose output rather than reduced peripheral glucose uptake is the primary factor responsible for both fasting and postprandial hyperglycemia in type 2 diabetes1,2. It is, therefore, appropriate to consider the mechanisms that may be involved in permitting and promoting this excessive hepatic output of glucose.