Background: Accumulating evidences showed metformin and berberine, well-known glucose-lowering agents, were able to activate adenosine monophosphate-activated protein kinase (AMPK). Strikingly, they were also found to be able to inhibit mitochondrial electron transport chain at complex I. Whether AMPK or complex I is the target of metformin and berberine remains in conflict.
Aims: In the present study, we aimed to explore the hypoglycemic effect of complex I inhibition and illuminate the role of AMPK in the beneficial effects of complex I inhibition.
Methods: Rotenone, amobarbital and gene silence of NDUFA13 were used to inhibit complex I. Intraperitoneal glucose tolerance test (IPGTT) and insulin tolerance test (ITT) were performed in db/db mice. Lactate release and glucose consumption were measured to investigate glucose metabolism in HepG2 hepatocytes and C2C12 myotubes. Glucose output was measured in primary hepatocytes. Compound C and recombinant adenoviruses expressing dominant negative forms of AMPKα1 (D159A) and AMPKα2 (K45R) were exploited to inactivate AMPK pathway. Western blotting was conducted to analyze protein expression. Cellular NAD+/NADH ratio was assayed to evaluate energy transforming and redox state. 
Results: Rotenone ameliorated hyperglycemia and insulin resistance in db/db mice. In HepG2 cells, at concentrations between 0.02 and 0.2 μmol/L, rotenone improved glucose consumption by 20.1% to 31.7% and lactate production by 30.2% to 56.1% in a dose-dependent manner. Similar effects were observed in C2C12 cells. In primary hepatocytes, rotenone significantly inhibited glucose output by 30.6% to 72.4% in a dose-dependent manner. Rotenone also activated AMP-activated protein kinase (AMPK). Furthermore, it remained effective with AMPK inactivation. The enhanced glycolysis and repressed gluconeogenesis correlated with a reduction in cellular NAD+/NADH ratio, which resulted from complex I suppression. Amobarbital, another representative complex I inhibitor, stimulated glucose consumption and decreased hepatic glucose output in vitro, too. Similar changes were observed while expression of NDUFA13, a subunit of complex I, was knockdown with gene silencing. 
Discussion: These findings reveal mitochondrial complex I emerges as a key drug target for diabetes treatment. Inhibition of complex I improves glucose homeostasis through reducing cellular NAD+/NADH ratio in an AMPK-independent way.

Inhibition of electron transport chain at complex I improves glucose metabolism through decreasing NAD+/NADH ratio