Mechanism of action of metformin on glucose 6-phosphate in hepatocytes

Abstract

The anti‐hyperglycaemic drug metformin is the first‐line oral therapy for Type 2 Diabetes, yet its mechanism of action remains contentious. We tested the hypothesis that metformin counteracts the effect of high glucose on gene regulation in hepatocytes by lowering cell glucose 6‐phosphate (G6P) and we investigated the possible candidate mechanisms for the lowering of G6P by metformin. Rat and mouse hepatocytes were treated with metformin (0.1‐1.0mM) in incubation conditions to achieve cellular loads of metformin within the therapeutic range and higher (1‐10 nmol / mg protein) as determined from cellular accumulation of [ 14C] metformin. Rates of flux of glucose metabolism were determined using 3H‐labelled (2,3 and 5 positions) or 14C‐labelled glucose (1,6 or uniformly labelled) in hepatocytes incubated with metformin, allosteric activators of AMPK (AMP‐activated protein kinase), and allosteric modulators of enzymes that alter flux and G6P levels in hepatocytes. Metformin, at cellular loads corresponding to the therapeutic range (1‐2 nmol/mg), counteracted the elevation in G6P caused by 15‐25mM glucose, dihydroxyacetone, xylitol and fructose in conditions where cell adenosine triphosphate (ATP) levels were maintained, but had little effect on cell G6P at basal glucose (5mM). In conditions of high glucose or gluconeogenic precursors the metformin efficacy in lowering G6P was greater at high cell G6P levels as occurs during compromised hydrolysis of G6P with a chlorogenic acid inhibitor of the G6P‐transporter encoded by Slc37a4. Metformin also counteracted the effects of high glucose on both gene induction (Pklr and G6pc) and Gck gene repression and the latter was not mimicked by AMPK activators. The G6P lowering effect of metformin was not mimicked by pharmacological activators of AMPK (A769662, C13 and 991) but was mimicked by inhibitors of complex 1, mitochondrial uncouplers, an inhibitor of nicotinamide nucleotide transhydrogenase (NNT) which is coupled to the mitochondrial proton gradient, and phenazine methosulphate which oxidizes NADPH. Inhibition of glucokinase and glucose phosphorylation were excluded as a mechanism for the lowering of G6P by therapeutic concentrations of metformin. However, inhibition of glucose phosphorylation, estimated from metabolism of [2‐ 3H] glucose, and glucokinase translocation occurred with diverse selective AMPK activators and also at higher metformin concentrations ≥1.0mM. We excluded the possibility that vi metformin lowers G6P by stimulating glycogen synthesis or inhibiting glycogen degradation from measurement of glycogen accumulation and correlation of glycogen synthesis with G6P using an allosteric inhibitor of phosphorylase. These studies showed inhibition of glycogen synthesis by metformin which paralleled the lowering of G6P suggesting that it is “secondary to” rather than “causative of” the G6P lowering. Metformin raised cell NADP indicating possible increased flux of G6P through the pentose phosphate pathway, and dehydroepiandrosterone (DHEA), an inhibitor of glucose 6‐phosphate dehydrogenase (G6PD), abolished the increase in NADP with 0.5mM metformin and partially counteracted the decrease in G6P consistent with a possible increased pentose phosphate pathway flux by metformin. However, this could not be confirmed by knock down of G6PD. We excluded the possibility that inhibition of NNT by metformin is a major mechanism in lowered G6P because metformin lowered G6P in hepatocytes from mice with a deletion in the NNT gene. With high glucose as substrate, low metformin increased production of lactate and pyruvate, and metabolism of [3‐ 3H] glucose but not [2‐ 3H] glucose in conditions of negligible inhibition of [U‐ 14C] glucose oxidation indicating increased glycolysis downstream of G6P. Targeting the first regulated site of glycolysis by allosteric inhibition of phosphofructokinase‐1 (with either aurintricarboxylic acid a potent inhibitor, or by selective lowering of fructose 2,6‐P2 by expressing a kinase‐deficient variant of PFKFB1) caused marked elevation in G6P for a small fractional inhibition of flux through glycolysis. This suggests the phosphofructokinase‐1 / fructose bisphosphatase‐1 site is a strong candidate target for the stimulation of glycolysis and lowering of G6P by metformin. Elevated inorganic phosphate was identified as one candidate allosteric activator of phosphofructokinase‐1 that is raised by metformin. In summary, metformin lowers G6P in hepatocytes at least in part by stimulation of glycolysis and most likely by altered allosteric control at phosphofrucokinase‐1 through raised inorganic phosphate or other effectors. The stimulation glycolysis by metformin is not mimicked by allosteric activators of AMPK. The lowering of hepatocyte G6P by metformin contributes to the counter‐effects of metformin on gene regulation by high glucose.