The plot thickens.

The dilator function of the vascular endothelium is critical for optimal tissue perfusion, and this is brought into stark reality in a number of diseases, such as hypertension, diabetes, and preeclampsia, in which endothelium-dependent vasodilator function is impaired. Prostacyclin and nitric oxide (NO) were the earliest identified endothelium-dependent vasodilators, but it was soon found that in many vascular beds, particularly in resistance vessels, vasorelaxation persisted when production of these autacoids was suppressed. This residual vasorelaxation was invariably associated with hyperpolarization of the vascular smooth muscle and the entity was dubbed “endothelium-derived hyperpolarizing factor” (EDHF).1 From the outset, the identity of EDHF has been the subject of vigorous debate. Nonetheless, that the “EDHF effect” is blocked by agents that target calcium-activated K channels is among the few aspects of EDHF on which there is apparent consensus. There was also the debate on the significance, or otherwise, of EDHF. Is it just the poor cousin of NO, stepping into the breech when the latter is compromised, especially under conditions of enhanced oxidative stress? Then a beacon of light shone out from Japan, with the demonstration that EDHF assumed greater importance as vessel diameter decreased.2 This was a key observation, supporting the importance of EDHF in the region of the vascular system where resistance is primarily determined. This was hotly followed by the revelation that EDHF may be targeted in diseases, diabetes,3–5 hypertension,6 and apolipoprotein E-deficient mice.7 Thus, EDHF, suggested to provide vasodilation when NO was compromised, was also under attack in disease. Despite its Cinderella status, investigation into the nature of EDHF has continued with breakneck intensity, with the conviction that EDHF has an important role in tissue perfusion and blood flow. The fact that the acronym has been retained to describe the NOand prostacyclin-insensitive relaxation reflects the fact that the identity of and mechanism(s) mediating EDHF remain unresolved, and these can be different in the various vascular beds. Small-conductance, calcium-activated K channels (KCa2.3, blocked by apamin) and intermediate-conductance, calciumactivated K channels (KCa3.1, blocked by charybdotoxin or TRAM-34) are located on the endothelial cells (ECs) and play a pivotal role in the smooth muscle cell (SMC) hyperpolarization that underpins EDHF. A hotly debated issue surrounding the nature of EDHF is whether the hyperpolarization generated in ECs by activation of these channels spreads directly to subjacent SMCs via myoendothelial gap junctions.8,9 Myoendothelial gap junctions arise predominantly on EC projections passing through holes in the internal elastic lamina to contact medial SMCs. An alternative view is that K exiting ECs activates inwardly rectifying K channels and Na /K -ATPase pumps to induce hyperpolarization secondarily in the SMCs,10 although their contribution to the EDHF current, recorded under voltage clamp, are minimal in guinea pig submucosal arterioles.11 It has been demonstrated previously that KCa3.1 channels are localized to the holes in the internal elastic lamina and are therefore on EC projections.12 In this issue of Circulation Research, Dora et al show that Na /K -ATPase pumps are also localized to EC projections, suggesting an intimate link between the pump and the KCa3.1 channel, and they hypothesize that this amplifies EC hyperpolarization.13 In arteries that had been treated with carbenoxolone to block gap junctions, the remaining TRAM-34–sensitive relaxation was significantly reduced by ouabain, consistent with previous observations that K released from KCa3.1 on EC projections activated Na /K -ATPase pumps located on SMC.10 However, in the new scheme, would not transfer of the amplified hyperpolarization component in EC be expected to be blocked in carbenoxolone, leaving the ouabain-sensitive component that results from extracellular K stimulation of SMC pumps? Results with ouabain must be interpreted with caution, because it also has uncoupling effects at gap junctions.14,15 Progress in defining EDHF mechanisms is hampered by the nonselective actions of currently available pharmacological tools. The development of antibodies to proteins of interest (ion channels, connexins, etc) has provided valuable insights into their spatial distribution. An elegant study in rat small mesenteric arteries demonstrated that KCa2.3, colocated with EC–EC connexons around the periphery of these cells, are clearly spatially separate from KCa3.1, present on EC projections.12 Dora et al suggest that KCa2.3 are also localized on EC projections, along with the KCa3.1 channels.13 The high level of sequence homology between KCa2.3 and KCa3.1 channels16 makes it wise to use more than one antibody source, preferably made to separate epitopes targeting the same channel. Consistent with previous reports, not all internal elastic lamina holes are associated with KCa expression, and this reflects the absence of myoendothelial gap The opinions expressed in this editorial are not necessarily those of the editors or of the American Heart Association. From the Department of Physiology, Monash University, Clayton, Australia. Correspondence to Prof Helena C. Parkington, Department of Physiology, Monash University, Wellington Rd, Clayton, Victoria 3800, Australia. E-mail helena. [email protected] (Circ Res. 2008;102:1148-1150.) © 2008 American Heart Association, Inc.