Heart and Circulatory Physiology


Adenosine is a key myocardial metabolite that elicits coronary vasodilation in a variety of pathophysiological conditions. We examined the mechanism of adenosine-induced vasodilation in coronary arterioles from patients with heart disease. Human coronary arterioles (HCAs) were dissected from pieces of the atrial appendage obtained at the time of cardiac surgery and cannulated for the measurement of internal diameter with videomicroscopy. Adenosine-induced vasodilation was not inhibited by endothelial denudation, but A2 receptor antagonism with 3,7-dimethyl-1-propargylxanthine and adenylate cyclase (AC) inhibition with SQ22536 significantly attenuated the dilation. In contrast, A1 receptor antagonism with 8-cyclopentyl-1,3-dipropylxanthine significantly augmented the sensitivity to adenosine. Moreover, dilation to A2a receptor activation with 2-p-(2-carboxyethyl)phenethylamino-5′-N-ethylcarboxamido-adenosine hydrochloride was reduced by the A1 receptor agonist (2S)-N6-(2-endo-norbornyl)adenosine. The nonspecific calcium-activated potassium (KCa) channel blocker tetrabutylammonium attenuated adenosine-induced dilation, as did the intermediate-conductance KCa blocker clotrimazole. Neither the large-conductance KCa blocker iberiotoxin nor small-conductance KCa blocker apamin altered the dilation. In conclusion, adenosine endothelium independently dilates HCAs from patients with heart disease through a receptor-mediated mechanism that involves the activation of intermediate-conductance KCa channels via an AC signaling pathway. The roles of A1 and A2 receptor subtypes are opposing, with the former being inhibitory to AC-mediated dilator actions of the latter. These observations identify unique fundamental physiological characteristics of the human coronary circulation and may help to target the use of novel adenosine analogs for vasodilation in perfusion imaging or suggest new strategies for myocardial preconditioning.

  • receptors
  • potassium channels
  • microcirculation

adenosine is produced by myocardial metabolism under physiological or pathological conditions including myocardial ischemia. Adenosine regulates myocardial perfusion in both animals (9, 17) and humans (10). Some studies have shown that adenosine dilates coronary arteries through specific endothelial receptors linked to the release of nitric oxide (NO) (19), whereas others have suggested that adenosine-induced dilation is dependent on an adenylate cyclase (AC)-cAMP pathway (48). Still others have indicated no role for the endothelium (54). In some cases, adenosine activates potassium channels in vascular smooth muscle cells (VSMCs) (6); however, the specific channel involved varies depending on the species and vessel studied. Both ATP-sensitive potassium (KATP) channels (1) and calcium-activated potassium (KCa) channels (42) have been implicated in adenosine-induced dilation in coronary vessels. In contrast to studies in animals, much less is known about the mechanisms by which adenosine dilates human coronary arterioles (HCAs), especially in patients with heart diseases. Such knowledge should help therapies aimed at restoring impaired coronary dilation in disease states such as diabetes, where adenosine-induced dilation may be reduced (8). The present study examined the mechanisms of adenosine-induced vasodilation in coronary microvessels from patients with heart disease.


Tissue acquisition and general protocol.

All protocols were approved by the Medical College of Wisconsin committees on the use of human subjects in research. Human right atrial appendages, removed for cannulation during cardiopulmonary bypass, were obtained at the time of cardiac surgery and placed in HEPES buffer solution (4°C) consisting of (mmol/l) 138.0 NaCl, 4 KCl, 1.2 MgSO4, 1.6 CaCl2, 1.2 KH2PO4, 0.026 EDTA, 6 glucose, and 10 HEPES acid. Arterioles were dissected from the appendage in the same day and prepared for continuous measurements of diameter as described previously (29). Briefly, in a 20-ml tissue chamber, both ends of the arteriole were secured to glass pipettes using 10-0 Ethilon monofilament nylon sutures (Ethicon). Vessels were bathed continuously with Krebs bicarbonate buffer consisting of (in mmol/l) 123.0 NaCl, 4.7 KCl, 2.5 CaCl2, 1.2 MgSO4, 20 NaHCO3, 1.2 KH2PO4, 0.026 EDTA, and 11 glucose. The preparation was then transferred to the stage of an inverted microscope. Attached to the microscope were a video camera, a video monitor, and a calibrated video measurement system. Internal diameter (resolution of 0.4 μm) was measured by manually adjusting the video micrometer. Vessels were pressurized to a predetermined level by simultaneously adjusting the height of each reservoir attached to the pipettes. Vessels were incubated in Krebs buffer oxygenated with 21% O2-5% CO2-74% N2 for 30 min at 20 mmHg and 37°C. Intraluminal pressure was increased to 60 mmHg with a subsequent 30-min incubation period.

Experimental protocols.

After 30 min of stabilization, endothelin-1 (10−10–10−9 mol/l) was, if needed, added to adjust the vasomotor tone to a level between 30 and 60% of the “expected” passive diameter (a diameter observed just after vessels were pressurized to 60 mmHg) (29, 35), because HCAs develop varying degrees of spontaneous myogenic tone (10–60%) (35). Multivariate analysis revealed that the amount of baseline myogenic tone and use of endothelin-1 as a preconstrictor did not alter any of the results reported in this study.

Vascular responses to increasing concentrations of adenosine (10−9–10−5 mol/l), 2-p-(2-carboxyethyl)phenethylamino-5′-N-ethylcarboxamido-adenosine hydrochloride (CGS21680; 10−10–10−5 mol/l, an adenosine A2a receptor agonist), and sodium nitroprusside (SNP; 10−10–10−5 mol/l) were examined. In some experiments, the endothelium was removed by injecting 2 ml of air through the lumen of the cannulated vessel as described previously (35). Endothelial denudation was confirmed by the loss of dilation to bradykinin (10−6 mol/l), an endothelium-dependent vasodilator (31) (%maximal dilation: 98 ± 1% before, P < 0.05 vs. −1 ± 9% after). In separate studies, we examined the effects of Nω-nitro-l-arginine methyl ester [l-NAME; 10−4 M; a NO synthase inhibitor (50)], indomethacin [Indo; 10−5 mol/l, a cyclooxygenase inhibitor (35)], and KCl [3 × 10−2 mol/l (24)] on the dilation to adenosine. In some cases, 8-(p-sulfophenyl)-theophylline [8-SPT; 10−4 mol/l, a nonselective antagonist of adenosine receptors (24, 35)], 8-cyclopentyl-1,3-dipropylxanthine [CPX; 10−8 mol/l, an adenosine A1 receptor antagonist (16, 24)], 3,7-dimethyl-1-propargylxanthine [DMPX; 10−4 mol/l, an adenosine A2 receptor antagonist (5, 24)], (2S)-N6-(2-endo-norbornyl)adenosine [ENBA; 10−8 mol/l, an adenosine A1 agonist (21)], SQ22536 [10−4 mol/l, a specific inhibitor of AC (4)], glibenclamide [10−6 mol/l, a selective blocker of KATP channels (35)], 4-aminopyridine [4-AP; 3 × 10−3 mol/l, a selective blocker of voltage-gated potassium channels (27)], tetrabutylammonium chloride [TBA; 10−3 mol/l, a nonselective blocker of KCa channels (3, 53)], charybdotoxin [10−8 mol/l, a blocker of large-conductance (BKCa) and intermediate-conductance KCa (IKCa) channels (38)], iberiotoxin [10−7 mol/l, a selective blocker of BKCa channels (15)], apamin [10−8 mol/l, a selective blocker of small-conductance KCa (SKCa) channels (2)], clotrimazole [10−6 mol/l, a selective blocker of IKCa channels (22) with an inhibitory effect on cytochrome P-450 activity (44)], or 17-octadecynoic acid [17-ODYA; 10−5 mol/l, a selective cytochrome P-450 inhibitor with no effect on potassium channel activity (11)] were used. All studies using potassium channel blockers (TBA, glibenclamide, 4-AP, apamin, iberiotoxin, charybdotoxin, and clotrimazole) were performed in the presence of l-NAME and Indo. Inhibitors were extraluminally added to the chamber 30 min before dose responses were tested. The same vessel was used to obtain dose responses for control and inhibitor(s).

Immunohistochemistry was performed to localize adenosine receptors in the human coronary microcirculation as described previously (35). Briefly, small pieces of pectinate muscle were fixed with 4% paraformaldehyde in phosphate-buffered saline, infiltrated with 20% sucrose-HEPES buffer solution, and frozen in OCT compound. Sections (thickness, 8 μm) were immunolabeled with rabbit anti-human polyclonal antibodies against adenosine receptor proteins (dilutions 1:100, A1 and A2a, Affinity BioReagents; A2b, Chemicon International). Immunostains were visualized using avidin-biotin horseradish peroxidase visualization systems (Vectastain Universal Quick Kit, Vector Laboratories). As a control for nonspecific binding, the primary antibody was omitted (35).


Endothelin-1 was obtained from Peninsula Laboratories. 8-SPT, SQ22536, iberiotoxin, and charybdotoxin were obtained from Research Biochemicals. Other chemicals were obtained from Sigma.

Statistical analysis.

All data are expressed as means ± SE. The percent dilation was calculated as the percent change from the preconstricted diameter to the maximal diameter (maximal diameter in the experiment at 60-mmHg luminal pressure), which was generally the diameter after papaverine (10−4 mol/l). The percent constriction or basal tone was determined by calculating the percent reduction in maximal diameter. Statistical comparisons of maximal percent vasodilation, ED50 values (−log ED50), and basal tone under different treatments were performed by paired or unpaired Student's t-test. To compare dose-response relationships between treatment groups, two-way ANOVA supported by a Bonferroni post hoc test when appropriate was used. Multiple stepwise regression analyses were used to detect the influence of underlying diseases, age, and gender on vasodilation at various dosages (35). All procedures were done using “proc mixed” or “proc glm” programs of SAS for Windows version 8.2. Statistical significance was defined as a value of P < 0.05.


Right atrial appendages were obtained from 106 patients. One arteriole was used from each patient (mean maximal diameter = 141 ± 6 μm at 60 mmHg). Patient demographic information is summarized in Table 1.

View this table:
Table 1.

Demographics of the patients

As shown in Fig. 1, adenosine produced a potent concentration-dependent dilation of HCAs. The dilation was significantly reduced by treatment with the nonspecific adenosine receptor antagonist 8-SPT (Fig. 1A). Neither endothelial denudation (Fig. 1B) nor pretreatment with Indo plus l-NAME [%maximal dilation: 96 ± 1% vs. 89 ± 2% in control, P = not significant (NS); ED50: 6.8 ± 0.2 vs. 6.8 ± 0.1 in control, P = NS, n = 5] affected this dilation. As shown in Fig. 1C, SQ22536, an AC inhibitor, reduced the dilation to adenosine. Neither 8-SPT, endothelial denudation, nor SQ22536 altered the resting tones and responses to papaverine (data not shown). These findings suggest that dilation to adenosine is not dependent on the endothelium but does involve the activation of adenosine receptors and AC, most likely on VSMCs in HCAs.

Fig. 1.

A: 8-(p-sulfophenyl)-theophylline (8-SPT), a nonselective adenosine receptor antagonist, significantly attenuated dilation to adenosine [%maximal dilation: 53 ± 6% vs. 90 ± 3% in control, P < 0.05; ED50, 6.3 ± 0.3 vs. 6.8 ± 0.4 in control, P = not significant (NS); n = 6]. B: endothelial denudation did not alter dilation to adenosine (%maximal dilation: 92 ± 3% vs. 95 ± 1% in control, P = NS; ED50: 6.8 ± 0.3 vs. 6.8 ± 0.3 in control, P = NS; n = 5). C: dilation to adenosine was significantly attenuated by SQ22536, an inhibitor of adenylate cyclase (AC) (%maximal dilation: 53 ± 8% vs. 97 ± 1% in control, P < 0.05; ED50: 5.7 ± 0.3 vs. 6.6 ± 0.1 in control, P < 0.05; n = 6). #P < 0.05 vs. control.

Next, we tested which adenosine receptor subtypes were involved in this dilation. First, expression of adenosine receptors in the human coronary microcirculation was immunohistochemically examined. As shown in Fig. 2, positive staining for adenosine receptors was observed by immunohistochemistry in HCAs. Both the endothelium and VSMCs showed strong immunostaining for proteins of A1, A2a, and A2b receptors (Fig. 2, A–C, respectively). This same pattern was observed in each of the three subjects.

Fig. 2.

Representative photomicrograph of atrial tissue immunostained with polyclonal antibodies against adenosine receptors (A: A1; B: A2a; C: A2b). Strong positive staining (brown tint) was seen in endothelium (arrow) and vascular smooth muscle cells (arrowhead) of human coronary arterioles (HCAs). Magnification, ×60 (scale bars = 50 μm). L, lumen.

Figure 3A shows that adenosine-induced dilation was significantly reduced by the A2 receptor antagonist DMPX. In contrast, the A1 receptor antagonist CPX surprisingly augmented the dilator sensitivity to adenosine (Fig. 3B). As shown in Fig. 3C, selective activation of A2a receptors, an A2 receptor subtype, with CGS21680 induced vasodilation in a dose-dependent manner with potency similar to that of adenosine. This A2a receptor-mediated dilation was reduced by A1 receptor activation with ENBA, confirming an inhibitory effect of A1 receptors on the vasodilator responses to adenosine. Resting tone and vasodilation to papaverine were not changed after treatments with these antagonists or agonist (data not shown).

Fig. 3.

A: vasodilation to adenosine was attenuated by 3,7-dimethyl-1-propargylxanthine (DMPX), an A2 receptor antagonist (%maximal dilation: 37 ± 8% vs. 91 ± 2% in control, P < 0.05; ED50: 7.3 ± 0.1 vs. 7.1 ± 0.4 in control, P = NS; n = 6). B: 8-cyclopentyl-1,3-dipropylxanthine (CPX), an A1 receptor antagonist, significantly augmented sensitivity to adenosine (%maximal dilation: 93 ± 1% vs. 92 ± 1% in control, P = NS; ED50: 8.4 ± 0.3 vs. 6.9 ± 0.2 in control, P < 0.05; n = 8). C: CGS21680, an A2a receptor agonist, dose dependently dilated HCAs (%maximal dilation: 92 ± 4%; ED50: 7.9 ± 0.3; n = 6). When vessels were treated with (2S)-N6-(2-endo-norbornyl)adenosine (ENBA), an A1 receptor agonist, the dilation to CGS21680 was significantly reduced (%maximal dilation: 38 ± 6% vs. control, P < 0.05; ED50, 6.8 ± 0.2 vs. control, P < 0.05). D: sodium nitroprusside (SNP)-induced dilation was not altered by ENBA (%maximal dilation: 98 ± 1% vs. 100 ± 0% in control, P = NS; ED50: 7.5 ± 0.3 vs. 7.0 ± 0.5 in control, P = NS; n = 5). #P < 0.05 vs. control.

Because A2a receptors stimulate AC by activating G proteins of the Gs subfamily (41), whereas A1 receptors inhibit AC by activating Gi/Go (37), we determined the role of A1 receptors in AC-mediated vasodilation. It is unlikely that this inhibitory effect of A1 receptor activation is nonspecific, because A1 receptor activation with ENBA did not alter the dilation to SNP, which is mediated by the activation of guanylate cyclase (Fig. 3D). These findings suggest that the stimulatory effect of A2 receptors, possibly via the A2a receptor subtype, and the inhibitory effect of A1 receptors underlie adenosine-induced vasodilation through AC signaling pathways in HCAs.

We studied the role of potassium channels in the adenosine-induced response. As shown in Fig. 4A, high concentrations of KCl, which depolarizes VSMC membranes by blockade of potassium channels, markedly reduced the vasodilation to adenosine. In the presence of AC inhibition with SQ22536, KCl had no additional inhibitory effect (Fig. 4B). This indicates that potassium channel activation via AC and cAMP pathways in VSMCs is involved in the receptor-mediated vasodilation to adenosine.

Fig. 4.

A: KCl reduced adenosine-induced dilation (%maximal dilation: 39 ± 11% vs. 96 ± 1% in control, P < 0.05; ED50: 6.5 ± 0.4 vs. 7.4 ± 0.2 in control, P = NS; n = 6). B: no additional inhibition by KCl (%maximal dilation: 36 ± 9%; ED50: 6.2 ± 0.2, P = NS vs. SQ22536 alone, respectively; n = 5) was seen in dilation to adenosine after inhibition of AC with SQ22536 (%maximal dilation: 40 ± 10%; ED50: 6.0 ± 0.2; n = 5). C: clotrimazole, a selective blocker of intermediate-conductance calcium-activated potassium channels and cytochrome P-450, significantly attenuated dilation to adenosine (%maximal dilation: 45 ± 9 vs. 91 ± 4 in control, P = NS; ED50, 5.8 ± 0.1 vs. 7.4 ± 0.4 in control, P = NS; n = 6). D: 17-octadecynoic acid (17-ODYA), a selective inhibitor of cytochrome P-450, had no effect on the dilation (%maximal dilation: 93 ± 5 vs. 99 ± 1 in control, P = NS; ED50: 6.8 ± 0.2 vs. 6.8 ± 0.1 in control, P = NS; n = 4). #P < 0.05 vs. control.

Next, we tested the involvement of specific potassium channels in the dilation to adenosine. As shown in Table 2, TBA, a nonspecific KCa channel blocker, reduced the dilation to adenosine, whereas neither 4-AP, a selective blocker of voltage-gated potassium channels, nor glibenclamide, a selective KATP channel blocker, had any effect. These findings suggest the involvement of KCa channels in dilation. Charybdotoxin but not iberiotoxin or apamin inhibited adenosine-induced dilation, indicating that opening of IKCa but not BKCa or SKCa channels is necessary. We tested the effect of clotrimazole, a chemically distinct IKCa channel blocker (22). As shown in Fig. 4C, clotrimazole also reduced the dilation to adenosine. Clotrimazole is also an inhibitor of cytochrome P-450. However, 17-ODYA, a specific inhibitor of cytochrome P-450 (11), did not affect dilation (Fig. 4D), suggesting that clotrimazole attenuated adenosine-induced dilation by blockade of IKCa channels. In the presence of A2 receptor antagonism with DMPX, clotrimazole failed to further reduce the dilation (%maximal dilation: 29 ± 6% with DMPX, P = NS vs. 33 ± 9% with DMPX + clotrimazole; n = 4). Pretreatment with KCl increased the vascular tone (−2 ± 3% with vehicle, P < 0.05 vs. −24 ± 8% with KCl), whereas other potassium channel blockers unchanged the tone (data not shown). The vasodilation to papaverine was not impaired by any potassium channel blockers (data not shown). These findings, together with reports of IKCa upregulation in VSMCs and its involvement in vasodilation in diseased vasculatures including the human coronary microcirculation (32, 45), suggest a role for IKCa channels in the adenosine-induced dilation of HCAs from patients with heart disease.

View this table:
Table 2.

Effects of potassium channel blockers on vasodilation to adenosine

Multiple stepwise regression analysis confirmed that the vasodilation to adenosine was not influenced by underlying diseases (coronary artery disease, hypertension, hypercholesterolemia, diabetes mellitus, myocardial infarction, or congestive heart failure), sex, or age.


The novel findings of this study are fourfold. First, in contrast to many animal models, adenosine is an endothelium-independent receptor-mediated dilator of HCAs in subjects with heart disease. Second, A1, A2a, and A2b receptors are expressed in both the endothelium and VSMCs of HCAs. Third, the vasomotor effects of adenosine are complex and include A2, possibly A2a, receptor-mediated dilation and A1 receptor-mediated inhibition of dilation through the modulation of AC activity. Finally, adenosine-induced dilation involves hyperpolarization of VSMCs through the opening of IKCa channels. The importance of these findings relates to the significant role of adenosine in regulating myocardial perfusion in the human heart.

Endothelium-independent vasodilation to adenosine.

We demonstrated in HCAs from subjects with heart disease that A1, A2a, and A2b receptors are expressed in the endothelium and VSMCs and that adenosine-induced dilation is not affected by endothelial denudation or by inhibitions of NO synthase and cyclooxygenase. These results support those from a previous study (24) of isolated human small coronary arteries, where adenosine-induced dilations were unaffected by the inhibition of NO synthase. However, they contrast with the results from an animal study (19) where adenosine-induced dilation was in part mediated by the endothelial release of NO. This discrepancy could be related to species differences or differences arising from studies of tissue from normal versus diseased subjects where endothelial function or adenosine receptor activity may be somewhat altered. In disease, NO release may be decreased, and non-endothelium-dependent mechanisms could compensate for the loss of endothelial dilator substances (56). However, previous studies (31, 34) using HCAs from patients with coronary artery disease demonstrated that vasodilations to bradykinin and shear stress are endothelium dependent, indicating that endothelial function in subjects with heart disease is relatively preserved in the human coronary microcirculation. Thus, unless an adenosine-specific defect in endothelial function is present, we speculate that the endothelium has minimal role in modulating the arteriolar dilation to adenosine. This prediction must be tempered by the fact that we have not examined the role of the endothelium in adenosine-induced vasodilation in normal subjects. For obvious reasons, obtaining fresh cardiac tissue from normal humans is not possible. Nevertheless, by examining tissues from subjects with coronary artery disease, we gain insight into the mechanisms of altered vascular function in a clinically relevant situation using a chronic “model,” the chronicity and complexity of which cannot be duplicated in animals.

Modulator effect of adenosine receptor subtypes on AC activity and vasodilation.

Adenosine receptors have been cloned as four subtypes: A1, A2a, A2b, and A3 (47). Adenosine induces vasodilation through A2a receptor activation in pig coronary arterioles (18) and through an A2b receptor mechanism in human small coronary arteries (24). In the present study, we observed A1, A2a, and A2b receptors expressed in the endothelium and VSMCs in HCAs, and adenosine-induced vasodilation in HCAs was inhibited by the nonselective adenosine receptor antagonist 8-SPT and by a selective A2 receptor antagonist. Furthermore, A2a receptor activation with CGS21680 produced a potent vasodilation of HCAs with an effectiveness similar to adenosine. In contrast, activation of A1 receptors reduced the dilation to adenosine, indicating an opposing effect of A2 and A1 receptors in HCAs. Previous studies have demonstrated that adenosine receptors are coupled to G proteins, which regulate AC activity. A2a receptors activate AC by stimulating G proteins in the Gs subfamily (41), whereas A1 receptors inhibit AC by stimulating Gi/Go (37). This is also supported by our findings in human vessels where 1) adenosine-induced dilation was reduced by SQ22536, an AC inhibitor; and 2) both adenosine-induced and A2a receptor agonist-induced dilations were reduced in the presence of an A1 receptor agonist. Meanwhile, it has been reported that adenosine concentration of 1,000 times higher is necessary to elicit a receptor response for A2 receptors compared with A1 receptors (46). Thus, in contrast to the nanomolar levels of adenosine required for the half-maximal effect on A1 receptors, the half-maximal concentration for stimulatory adenosine receptors (A2) falls in micromolar ranges. This clearly supports our data showing that the A1 receptor antagonist CPX potentiates the vasodilation to adenosine at lower concentrations (Fig. 3B), whereas the A1 receptor agonist ENBA reduces vasodilation to the A2 receptor agonist CGS21680 (Fig. 3C). On the basis of these data, it can be concluded that A1 and A2 (A2a) receptor-mediated signals modulate AC activity and the resultant vasodilation in opposing directions. Pathophysiological conditions where A2 receptor-mediated responses are preserved and A1 receptor function is impaired (14, 52) could augment adenosine-induced dilation.

Role of IKCa channels in vasodilation to adenosine.

In the present study, adenosine-induced dilation was reduced by treatment with either an AC inhibitor or KCl. No additive effect was seen. This suggests a necessary role for VSMC membrane hyperpolarization and AC activity. The inhibitory effect of TBA on adenosine-induced dilation indicates the involvement of KCa channels. Charybdotoxin, a blocker of both BKCa and IKCa channels (38), also inhibited the dilation to adenosine. Selective blockade of BKCa or SKCa channels with iberiotoxin or apamin, respectively, failed to attenuate this dilation. We infer therefore that IKCa channels are involved, because IKCa channels are inhibited by TBA (39) and charybdotoxin (38) but not by iberiotoxin and apamin. This is consistent with a role for IKCa channels in VSMCs of rat coronary arterioles (45, 51), carotid arteries (26), and HCAs (32, 33). Clotrimazole, a selective IKCa channel blocker (22), was also effective in blocking the dilation to adenosine through a mechanism that did not involve cytochrome P-450, because 17-ODYA, a specific inhibitor of cytochrome P-450 (11), had no effect on the dilation. In addition, antagonism of A2 receptors abolished the inhibitory effect of IKCa channel blockade on adenosine-induced vasodilation. Our results are consistent with a role for IKCa channels in the dilation to adenosine in HCAs. Actions of adenosine are associated with the activities of AC and cAMP (23, 37, 41). cAMP and cAMP-dependent protein kinase play an important role in the regulation of IKCa channel activity by the phosphorylation of the channels (7). Taken together, these findings support the linkage of adenosine receptors and the AC signaling pathway to IKCa channel activity. A residual dilation to adenosine (30%) was observed in the presence of KCl, suggesting that a potassium channel-independent mechanism may also contribute to adenosine-induced dilation.

Clinical implications.

Adenosine is proposed as a mediator for ischemic preconditioning of myocardium and for metabolic regulation of the coronary microcirculation in hearts from several species including humans (28). However, the mechanism of the coronary microvascular dilation to adenosine has not been fully determined. For example, a recent human in vivo study (13) reported that an increase in coronary blood flow in response to an intracoronary infusion of adenosine is sensitive to KATP channel inhibition, whereas in vitro studies, including the present study, identify no role for KATP channels in the vasodilation to adenosine in small coronary arteries and arterioles in humans (24). It is also recognized that KATP channels play a critical role in adenosine-induced cardiac protection during ischemia and reperfusion (57). However, KATP channel function is impaired in diseases such as diabetes mellitus in humans (35), and KATP channels do not likely contribute to the cardiac protection during ischemia and reperfusion in those disease states (25). Recently, a cardioprotective role of KCa channels during ischemia and reperfusion was proposed. For example, the estrogen receptor-mediated cardioprotective effect is attributed to opening of KCa channels (40). It is also reported that hydrogen peroxide, an endogenous EDHF, which is a major vasodilator in the human coronary microcirculation (30) and which stimulates KCa channels, contributes to the myocardial protection during ischemia and reperfusion (55). Therefore, it is important to clarify the channel remodeling and the compensatory or altered mechanisms of actions of adenosine in diseased humans.

Potential limitations of the study.

We could not test the mechanisms for adenosine-induced vasodilation in “normal” human coronary arterioles, because the human right atrial appendages used in this study were removed for cannulation during cardiopulmonary bypass at the time of cardiac surgery (e.g., coronary artery bypass graft, valve replacement, and repair of congenital heart disease). Even those subjects who do not have coronary artery disease are not normal (control). There are no tissues available from normal subjects. Our goal was not to evaluate the properties of coronary vessels in normal subjects but to examine the influence of underlying diseases, age, and gender on adenosine-induced vasodilation of human coronary arterioles. Thus lack of a pure normal control group does not preclude our ability to define the influence of disease on coronary arteriolar responses to adenosine. To account for this limitation, we performed a multiple stepwise regression analysis to determine whether there were relationships between the dilation to adenosine and coronary risk factors (34, 35). Although risk factors were not found to influence the effect of inhibitors on dilation to adenosine, we cannot exclude the possibility that coronary risk factors influenced the specific potassium channels involved in adenosine-induced dilation. Preliminary data using HCAs from children, in which coronary atherosclerosis is negligible, showed an endothelial NO-dependent and KATP channel-dependent vasodilation to adenosine. We speculate that the mechanism of dilation to adenosine may be altered in HCAs from patients with heart disease as a result of age or presence of disease (3235). Because the present study examined adenosine-induced dilation in subjects with disease, the findings may be specific for patients with disease and may not apply to normals. Future studies using vessels from subjects without acquired heart diseases or its risk factors are needed to further examine this possibility.

We observed varied potencies of adenosine-induced dilations among patients (ED50 = 10−8 ∼ 10−5 M), although no factors were predicted to impair the dilation. It seems that this variety of potencies results in the inconsistent potency of dilations between figures in this study, although use of same vessel for dose-response curves with control or inhibitor(s) eliminates the possibility that different potencies of vessels affect the interpretation of results obtained.

We previously reported that acetylcholine constricts atrial microvessels but dilates ventricular microvessels, whereas other agonists (such as bradykinin and substance P) and shear stress elicit vasodilations of both vessels in a similar manner (29, 34). Hein et al. (20) reported the heterogeneous distribution of A1 and A2a receptors along the coronary vascular tree in porcine hearts. They showed that both A1 and A2a receptors express in large coronary arteries (e.g., the left anterior descending coronary artery), whereas only A2a receptors express in coronary microvessels. Thus it is possible that vascular response to adenosine and expression of the receptors may vary among regions of human hearts (e.g., atria and ventricles).

In this study, the important role of IKCa channel activity in adenosine-induced vasodilation was identified with pharmacological techniques. In VSMCs, three types of KCa channels are recognized dependent on their conductance: BKCa (49), IKCa (38), and SKCa channels (43). Interestingly, the VSMC phenotypic conversion from a contractile to proliferative type is concomitant with clonal transition of KCa channels from BKCa to IKCa (38). This transition occurs most frequently in disease states such as diabetes mellitus and atherosclerosis (12, 36). We and others also demonstrated the differential expression of KCa channels in diseased animal models such as rat hearts after ischemia and reperfusion (45) and after chronic inhibition of NO synthesis (51), in a rat restenosis model after balloon injuries (26), and in the human coronary microcirculation (32, 33), where IKCa channel expression is upregulated and BKCa channel expression is downregulated. Thus upregulation of IKCa channels may be compensatory in preserving the action of adenosine in the diseased human coronary microcirculation, suggesting the possibility of new therapies for treatment of these patients with adenosine or adenosine analogs.

In conclusion, the present study provides evidence that adenosine-induced vasodilation in HCAs from patients with heart disease does not require the endothelium but is receptor mediated in VSMCs, leading to activation of AC and membrane hyperpolarization, which involves activation of IKCa channels. These findings may be important for therapeutical strategies to improve myocardial perfusion in patients with heart disease.


This study was supported by a Postdoctoral Fellowship Grant from the American Heart Association, Northland Affiliate, grants from the National Institutes of Health, and a Veterans Administration Merit Award.


The authors acknowledge the contribution of tissue acquisition by participating investigators from Cardiothoracic Surgery at the Medical College of Wisconsin (Drs. A. C. Nicolosi and G. H. Almassi) and St. Luke's Medical Center (Drs. L. H. Kleinman and D. C. Kress).


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract