SUMMARY1. The mechanisms underlying hypoxic dilatation of coronary arteries were studied in isolated guinea-pig hearts perfused with physiological salt solution at 37°C. The hearts were perfused at a constant rate of 3-10 ml min-'; coronary perfusion pressure (CPP) and isovolumetric left ventricular pressure (LVP) were measured with piezoresistive transducers.2. Addition of the K+ channel opener cromakalim (500 nM) to the perfusate caused a maximal vasodilatation in beating hearts, i.e. a decrease in CPP of about 50%.Switching from normal perfusate (partial pressure of 02 (Po2), 650-700 mmHg) to hypoxic perfusate (Po2, 9-10 mmHg) caused a similar vasodilatation. Both of these effects were prevented by 2 ,um-glibenclamide, a blocker of ATP-sensitive potassium channels. Hypoxic vasodilatation was accompanied by a marked decrease in LVP, which was reduced by 56 + 22 % (mean + S.D.) in the presence of glibenclamide.3. In hearts arrested by increasing the K+ concentration of the perfusate to 15 mm, the addition of the adenosine-uptake inhibitor dipyridamole evoked a maximal vasodilatation and this was inhibited by 76+7 % in the presence of glibenclamide.4. The adenosine antagonist 8-phenyltheophylline (8-PT; 5 /M) inhibited the vasodilatation induced by dipyridamole by 88 + 10 %. In contrast, hypoxic vasodilatation was unaffected by 5 /M 8-PT. This suggests that hypoxic dilatation of coronary arteries is not mediated by release of adenosine from cardiomyocytes.5. In order to test whether release of endothelium-derived relaxing factor (EDRF) contributed to hypoxic vasodilatation we blocked EDRF synthesis with Nd-nitro-Larginine (NNA). When applied at a perfusion rate of 10 ml min-' to arrested hearts, 10 #mM-NNA increased CPP by 35 % and prolonged the delay between application of hypoxic solution and half-maximal vasodilatation from 52 + 9 to 129 + 29 s.6. Under control conditions the relation between perfusion rate and the CPP measured in the steady state was linear. In the presence of 10 /sM-NNA coronary resistance was increased more than twofold at low perfusion rates; at perfusion rates between 4 and 10 ml min-' coronary resistance decreased progressively. This change * To whom correspondence should be addressed. MS 9042 N. VON BECKERATH AND OTHERSin the pressure-flow relationship may be responsible for the alterations in the time course of hypoxic vasodilatation induced by NNA.7. In order to test whether changes in energy metabolism in coronary smooth muscle cells were responsible for hypoxic vasodilatation we blocked glycolysis by replacing the glucose in the perfusate with deoxyglucose (DOG). DOG produced a near-maximal dilatation of coronary arteries that could be reversed by 2 /Mglibenclamide but was unaffected by 5 /tM-8-PT.8. Our results suggest that adenosine can induce the opening of ATP-sensitive potassium channels in smooth muscle cells of the coronary resistance vessels, but that this is not the principal mechanism underlying hypoxic vasodilatation. It is proposed that early hypoxic dilatation of coron...