NON-INVASIVE ASSESSMENT OF METFORMIN INDUCED CHANGES IN CARDIAC AND HEPATIC REDOX STATE USING HYPERPOLARIZED [1-13C] PYRUVATE

Background and methods Metformin’s mechanism of action is controversial though may involve suppression of hepatic gluconeogenesis through perturbation of cellular redox state. Hyperpolarized magnetic resonance with [1-13C] pyruvate exploits huge increases in tracer signal to non-invasively assess both the redox coupled lactate dehydrogenase reaction and also pyruvate dehydrogenase activity. We performed hyperpolarized [1-13C] pyruvate magnetic resonance spectroscopic and imaging assessments of heart and liver metabolism in Wistar rats following either a single infusion of metformin or longer term oral treatment. Results Metformin infusion increased both cardiac and hepatic hyperpolarized [1-13C] lactate:[1-13C] pyruvate ratio (0.22 ± 0.04 versus 0.11 ± 0.01, P = 0.004 and 0.39 ± 0.04 versus 0.30 ± 0.02, P = 0.04 respectively) without changing pyruvate dehydrogenase flux. Metformin also increased both the cardiac and hepatic [lactate]:[pyruvate] ratio, a surrogate of cytosolic redox state (50 ± 8 versus 25 ± 7, P = 0.02 and 72 ± 12 versus 16 ± 8, P = 0.002 respectively). Longer term Metformin treatment for 4 weeks also increased the cardiac and hepatic hyperpolarized [1-13C] lactate:[1-13C] pyruvate ratio (0.27 ± 0.06 versus 0.10 ± 0.01, P = 0.005 and 0.87 ± 0.2 versus 0.36 ± 0.04, P = 0.003 respectively) without significantly changing pyruvate dehydrogenase flux and also increased the cardiac and hepatic [lactate]:[pyruvate] ratio (46 ± 6 versus 30 ± 6, P = 0.04 and 60 ± 9 versus 27 ± 3, P = 0.002 respectively). Conclusions Both acute and chronic metformin treatment significantly increased both the cardiac and hepatic hyperpolarized [1-13C] lactate:[1-13C] pyruvate ratio, reflecting an increase in cytosolic redox state. These findings 1) identify a novel cardiac effect of metformin, 2) demonstrate the sensitivity of hyperpolarized [1-13C] pyruvate to metformin induced changes in redox biology and 3) have implications for the design of upcoming human studies using hyperpolarized MR to study cardiac metabolism in diabetes.