Metformin has been the mainstay of therapy for diabetes mellitus for many years; however, the mechanistic aspects of metformin action remained ill-defined. Recent advances revealed that this drug, in addition to its glucose-lowering action, might be promising for specifically targeting metabolic differences between normal and abnormal metabolic signalling. The knowledge gained from dissecting the principal mechanisms by which metformin works can help us to develop novel treatments. The centre of metformin's mechanism of action is the alteration of the energy metabolism of the cell. Metformin exerts its prevailing, glucose-lowering effect by inhibiting hepatic gluconeogenesis and opposing the action of glucagon. The inhibition of mitochondrial complex I results in defective cAMP and protein kinase A signalling in response to glucagon. Stimulation of 5'-AMP-activated protein kinase, although dispensable for the glucose-lowering effect of metformin, confers insulin sensitivity, mainly by modulating lipid metabolism. Metformin might influence tumourigenesis, both indirectly, through the systemic reduction of insulin levels, and directly, via the induction of energetic stress; however, these effects require further investigation. Here, we discuss the updated understanding of the antigluconeogenic action of metformin in the liver and the implications of the discoveries of metformin targets for the treatment of diabetes mellitus and cancer.

Background: There is an urgent need to limit the metabolic side-effects of glucocorticoid overexposure as these can lead to Cushing's syndrome, associated with high morbidity. We have explored the potential for metformin to ameliorate such effects whilst sparing the anti-inflammatory benefits of glucocorticoids. Methods: In this double-blind, phase 2 proof-of-concept trial, 53 patients without known diabetes established on mid-to-high doses of glucocorticoids, administered as treatment for a chronic inflammatory disease, were randomised to receive 2550mg/day metformin (n=26) or an identical placebo (n=27) for 12 weeks. The primary endpoint was the change in visceral to truncal subcutaneous fat ratio assessed by computed tomography; secondary endpoints involved metabolic, bone, cardiovascular and inflammatory parameters. Findings: Nineteen patients on metformin and 21 on placebo completed the study. The groups received equivalent cumulative dose of glucocorticoids (1860mg (IQR 1060 to 2810) vs. 1770mg (IQR 1020-2356) prednisolone equivalent; p=0.76). There was no change in the visceral-to-subcutaneous fat ratio (0.11 (95%CI-0.02 to 0.24); p=0.09) between the treatment groups but metformin-treated patients lost truncal subcutaneous fat (-3835mm 2 (95%CI-6781 to-888); p=0.01) compared to placebo. Improvements in markers of carbohydrate, lipid, liver and bone metabolism were observed on metformin. Additionally, metformin-treated patients had improved fibrinolysis, carotid intima-media thickness, inflammatory parameters and clinical markers of disease activity. The frequency of pneumonia (1 vs. 7 events; p=0.01), overall rate of moderate-to-severe infections (2 vs. 11; p=0.001), and all-cause hospital admissions due to adverse events (1 vs. 9; p=0.001) were lower in the metformin group compared with placebo. Metformin-treated patients experienced more diarrhoea initially. Interpretation: Metformin administration improved the metabolic profile of glucocorticoid-treated patients with inflammatory disease, favourably modifying cardiovascular risk surrogates, reducing inflammation and hospitalisation.