Pharmacophysiological insights into the paradoxical increase in endogenous glucose production with metformin treatment.

Metformin, the first-line therapy for type 2 diabetes, has long been recognized for its ability to lower blood glucose primarily by suppressing endogenous glucose production through mechanisms involving mitochondrial complex I inhibition-a process that mildly reduce cellular energy charge by elevating the AMP/ATP ratio. This alteration in energy charge may directly impair ATP-dependent steps in gluconeogenesis and anatagonize glucagon signaling by inhibiting adenylate cyclase, thus reducing cyclic AMP levels and downstream signaling. Increased AMP/ATP ratio also activates AMP-activated protein kinase, which drives transcriptional repression of gluconeogenic enzymes. Paradoxically, in the last decade several reports documented that metformin may increase EGP. This unexpected rise in EGP has been tentatively attributed to compensatory hyperglucagonemia, consistent with a pronounced counter-regulatory response that offsets metformin's antigluconeogenic and glucagon antagonistic effects. Yet, this explanation fails to reconcile metformin's concurrent suppression of other glucagon-mediated catabolic effects, such as increased energy expenditure and reduced protein synthesis. Furthermore, elevated levels of gluconeogenic precursors, including branched-chain amino acids and alanine, may suggest a broader metabolic interplay, involving a liver alpha-cell cross talk, analogous to the liver-alpha cell axis. In this review, we re-examine metformin's mechanisms of action. We, then, provide a mechanistic perspective to this latter possibility, based on mechanistic studies of metformin's transcriptional targets; retaining thus its glucagon antagonistic and presumably anti-gluconeogenic effects on liver and attributing the increase in EGP to renal gluconeogenesis. We finally explore whether the increased EGP might reflect an adverse response to metformin treatment, drawing insights from available clinical evidence.