Tissue Iron Predicts Metformin Responsiveness in a Mouse Model and in Humans with Type 2 Diabetes

Purpose: Action of metformin, the preferred pharmacologic treatment for type 2 diabetes (T2D), remains incompletely understood. Metformin induces an iron starvation response in yeast, suggesting tissue iron might be related to metformin responsiveness. We therefore examined the effects of different levels of tissue iron on metformin action. Subjects and Methods: We examined glycemic responses to metformin and effects on downstream targets of metformin in mice on diabetogenic diets with different iron contents and performed a retrospective study of metformin effectiveness as a function of serum ferritin, a reliable marker of tissue iron, in human patients after initiation of metformin therapy. Results: Metformin provided the most glycemic benefit to mice on the normal-iron diet, compared to high or low iron. The low-iron diet itself provided glycemic benefit, and on low iron, metformin provided no additional benefit. These results were parallelled by effects of dietary iron on downstream targets of metformin action, including AMP-dependent kinase and glycerol-3-phosphate dehydrogenase 2. Increased protein modification by O-linked N-acetylglucosamine (O-GlcNAc) mimicked lower iron levels, abrogated the effect of high iron, and resulted in smaller relative metformin responses on glycemia and downstream reporters. Humans with higher or lower levels of the iron biomarker ferritin also exhibited a decreased HbA1c response to metformin. Conclusion: Dietary iron and serum ferritin predict the glycemic response to metformin in mouse models and humans with T2D. Protein O-GlcNAcylation has effects paralleling low iron and may play a role in mediating these interactions.