The effects of the antidiabetic drug metformin on glucose transport were investigated in freshly isolated heart muscle cells from healthy and streptozotocin-diabetic rats. In vivo treatment of diabetic rats with metformin failed to affect the basal and insulin-stimulated rate of glucose transport measured in isolated cells. In vitro exposure to therapeutic concentrations (< or = 10(-4) M) of metformin did not influence glucose transport, even upon incubation times up to 5 h or in the presence of high glucose (20 nM). In contrast, higher metformin concentrations produced an 8- to 12-fold increase in glucose uptake (with a lag of 90 min, and a maximum at 180 min and approximately 5 mM). In the presence of submaximal insulin concentrations (< or = 3.10(-10) M), the effects of metformin (5 mM) and of insulin were more than additive, whereas, at saturating insulin concentrations (10(-8) M), partial additivity was observed. Like insulin, metformin caused an approximately 1.6-fold increase in the content of both glucose transporter isoforms GLUT1 and GLUT4 in the plasma membrane of cardiac myocytes, with a corresponding decrease in an intracellular membrane fraction. cAMP-elevating treatments depressed the metformin-, but not the insulin-dependent glucose uptake, by 20-30%. In myocytes from diabetic rats, the rate of metformin-activated glucose transport was similar to that of cells from control animals, whereas basal and insulin-stimulated transport were substantially diminished. Finally, metformin (5 mM) induced a slight depression of oxygen consumption and energy metabolism of myocytes (as determined by measuring their level of energy-rich phosphates) comparable to the effects of hypoxia in rat hearts. In conclusion, these data do not provide evidence in favor of the hypothesis that glucose uptake by muscle tissue represents the site of metformin's therapeutic action in vivo. On the other hand, the large, insulin-independent effect of metformin at high concentrations (approximately mM) in vitro may be related to the action of hypoxia and occurs through a redistribution of glucose carriers from an intracellular locus to the plasma membrane. The mechanism (or signal) involved in metformin's action is likely to differ from that triggered by insulin and is not impaired in the diabetic state.