Endothelial β3-Adrenoceptors Mediate Vasorelaxation of Human Coronary Microarteries Through Nitric Oxide and Endothelium-Dependent Hyperpolarization
Background— Coronary vessel tone is modulated in part by β-adrenergic relaxation. However, the implication of specific β-adrenoceptor subtypes and their downstream vasorelaxing mechanism(s) in human coronary resistance arteries is poorly defined. β3-Adrenoceptors were recently shown to vasodilate animal vessels and are expressed in human hearts.
Methods and Results— We examined the expression and functional role of β3-adrenoceptors in human coronary microarteries and their coupling to vasodilating nitric oxide (NO) and/or hyperpolarization mechanisms. The expression of β3-adrenoceptor mRNA and protein was demonstrated in extracts of human coronary microarteries. Immunohistochemical analysis revealed their exclusive localization in the endothelium, with no staining of vascular smooth muscle. In contractility experiments in which videomicroscopy was used, the nonspecific β-agonist isoproterenol and the β3-preferential agonist BRL37344 evoked an ≈50% relaxation of endothelin-1-preconstricted human coronary microarteries. Relaxations were blocked by the β1/β2/β3-adrenoceptor antagonist bupranolol but were insensitive to the β1/β2-adrenoceptor antagonist nadolol, confirming a β3-adrenoceptor-mediated pathway. Relaxation in response to BRL37344 was absent in human coronary microarteries devoid of functional endothelium. When human coronary microarteries were precontracted with KCl (thereby preventing vessel hyperpolarization), the relaxation to BRL37344 was reduced to 15.5% and totally abrogated by the NO synthase inhibitor l-ω-nitroarginine, confirming the participation of a NO synthase-mediated relaxation. The NO synthase-independent relaxation was completely inhibited by the Ca2+-activated K+ channel inhibitors apamin and charybdotoxin, consistent with an additional endothelium-derived hyperpolarizing factor-like response. Accordingly, membrane potential recordings demonstrated vessel hyperpolarization in response to β3-adrenoceptor stimulation.
Conclusions— β3-adrenoceptors are expressed in the endothelium of human coronary resistance arteries and mediate adrenergic vasodilatation through both NO and vessel hyperpolarization.
Received June 16, 2003; de novo received December 26, 2003; revision received May 6, 2004; accepted May 10, 2004.
The existence of a rich sympathetic innervation of the coronary arteries raises the importance of catecholamines as endogenous modulators of myocardial perfusion, in addition to their direct inotropic effects. The increased oxygen demand after the latter is believed to be compensated by coronary vasodilation, in part through direct β-adrenergic relaxation of vascular smooth muscle. Classically, this has been ascribed to β1- and β2-adrenoceptor stimulation coupled to increases in smooth muscle cAMP.1–3 However, in other large conductance or small resistance arteries, an additional vasorelaxing mechanism may operate through β2-adrenoceptor activation of endothelial production of nitric oxide (NO).4,5 Although this pathway was shown to mediate the vasorelaxation of human forearm arteries with β-adrenergic agonists in vivo,6 its existence in human coronary resistance arteries has received little attention thus far.
In nonvascular smooth muscle, such as in the colon or stomach, β-adrenergic agonists also elicit a relaxation through activation of an “atypical” β-adrenoceptor, more recently identified as the β3-adrenoceptor.7 Since its molecular identification,8 the distribution and functional role of this receptor have been extended from the regulation of lipolysis in fat tissue 9 to modulation of cardiac contraction,10 including in human ventricle.11 In addition, we12 and others13,14 observed a profound peripheral vasodilation after infusion of β3-preferential agonists in the conscious dog, and recently, similar agonists were shown to elicit a NO-mediated relaxation of the rat aorta.15,16 This raised the possibility of a similar vasorelaxing mechanism in human coronary arteries. Because we previously observed an upregulation of β3-adrenoceptor proteins in extracts of ischemic and dilated human hearts,11 such a possibility would potentially pertain to our understanding of coronary vasoregulation in the diseased heart.
Endothelial cells modulate the vascular tone through both shear stress- and agonist-evoked release of vasorelaxants such as NO, prostacyclin, and (still incompletely characterized) endothelium-derived hyperpolarizing factor(s) (EDHF). The functional importance of the latter was suggested to be inversely correlated with vessel diameter 17,18 and to be more prominent in circumstances of impaired NO-mediated vasorelaxation, such as that associated with risk factors for atherosclerosis and ischemic diseases.19,20 An EDHF-mediated relaxation was observed in human resistance (including coronary) arteries in response to arachidonic acid21 and adrenomedulin,22 but the implication of EDHF in β-adrenergic vasodilation remained elusive.
In the present study, we therefore examined the putative expression and functional role of β3-adrenoceptors in human coronary arterioles. We identified transcripts and proteins specific for β3-adrenoceptors in the endothelium of these vessels, and their activation mediated an endothelium-dependent relaxation that involves both NO and vessel hyperpolarization.
Human right atrial and left ventricular tissue specimens obtained from patients undergoing cardiac surgery were placed in physiological saline solution (PSS) containing the following (mmol/L): NaCl 120, KCl 5.9, NaHCO3 25, dextrose 17.5, CaCl2 2.5, MgCl2 1.2, NaH2PO4 1.2 (pH 7.4), maintained at 0°C to 4°C, and carefully dissected to isolate microarteries. Coronary microarteries (70 to 170 μm in diameter, 1 to 2 mm in length), rapidly cleaned of adherent connective tissue, were either kept at −80°C until homogenization or used for functional experiments.
Reverse Transcription-Polymerase Chain Reaction for mRNA Amplification
Human coronary microvessels were homogenized in GTC buffer (Tripure, Roche). After reverse transcription (RT), TaqMan polymerase chain reaction (PCR) was performed as described previously.23 The Ct (threshold cycle) was defined as the cycle number at which the reporter fluorescence generated by cleavage of the probe crossed a fixed threshold above baseline. In the absence of template, its value was 39.9±0.08 (n=3), indicating negligible background fluorescence from the probes.
Specific primer and TaqMan probe sequences for the human β3-adrenoceptor and human NOS3 (endothelial NO synthase [eNOS]) were designed as previously published.23,24
Microdissected vessels pooled from 6 to 9 atrial or ventricular specimens were ground in liquid nitrogen. Extracts were homogenized in 60 μL of buffer (mmol/L: Tris-HCl 20 [pH 7.4], EDTA 2.5, NaCl 100, NaF 10, Na3VO4 1, Na-deoxycholate 1%, SDS 0.1%, Triton X-100 1%, protease inhibitor cocktail). Protein samples were subjected to electrophoresis, transferred onto polyvinylidene difluoride membranes, and immunoblotted as described previously,11 with antibodies directed against human β3-adrenoceptors and eNOS.
Immunostaining of β3-Adrenoceptors
Pieces of atrial appendages were embedded in TissueTek OCT compound (Sakura) and snap-frozen in precooled isopentane at −80°C. Prewashed fixed cryosections (5 μm; 3.5% formaldehyde) were incubated with monoclonal anti-human β3-adrenoceptor antibodies (1/200 in PBS with 1% BSA), then rinsed (PBS with 0.1% BSA) and incubated with secondary polyclonal rabbit anti-rat IgG (1/200) coupled to peroxidase. After they were washed, sections were counterstained with Mayer’s hematoxylin and mounted for optical microscopy. Negative controls were performed in parallel where the primary antibody was omitted.
Videomotion Analysis of Vessel Contraction
Vessels were cannulated with dual glass micropipettes and secured with 10-0 nylon monofilament sutures in a Plexiglas isolated organ chamber circulated with oxygenated PSS (37°C) and placed on an inverted microscope (Axiovert S100, Zeiss) connected to a CCD camera. Microvessels were pressurized with a PSS-filled burette manometer at 60 mm Hg (a pressure chosen to avoid stretch-dependent effects typically manifested at higher pressures) in a no-flow state. Digitized imaging (IONOPTIX Corporation) allowed continuous monitoring of vessel external diameter. All experiments were performed in the presence of a cyclooxygenase inhibitor (indomethacin, 10 μmol/L). After 30 to 45 minutes of equilibration, vessels were contracted with high KCl solution (PSS, 50 mmol/L KCl replacing NaCl stoichiometrically). At the maximum contraction, vasorelaxation with substance P (100 nmol/L) was systematically tested to assess the presence of a functional endothelium. In some experiments the endothelium was selectively destroyed by an air bolus. All reagents were added in the bathing solution.
Measurement of Vessel Membrane Potential
Microvessels were mounted in an organ bath continuously circulated (6 mL/min) with oxygenated PSS at 37°C. All experiments were performed in the presence of a NOS inhibitor (l-ω-nitroarginine, 100 μmol/L) and indomethacin (10 μmol/L).
After 60 minutes of equilibration, measurement of the smooth muscle membrane potential (Em) was made with a glass microelectrode (Clark, Electromedical Instruments, type GC 120F-15) filled with 1.5 mol/L KCl. The input resistance of the microelectrodes varied between 50 and 80 mol/LΩ. Differences in electric potential were measured with a Dagan amplifier (model 8100) and recorded. Criteria for a successful impalement were (1) an abrupt drop in voltage on entry of microelectrode into the cell, (2) stable membrane potential for at least 2 minutes, and (3) a sharp return to zero on withdrawal of the electrode.
All results are expressed as mean±SEM. Statistical comparisons were performed by use of Student t test or 1-way ANOVA when appropriate. Probability values <0.05 were considered significant.
Clinical patient characteristics are summarized in the Table. Samples were obtained from patients undergoing cardiac transplantation (n=4) or other cardiac surgical procedures (n=60). Most patients suffered from ischemic cardiac diseases (76%). All were treated with a variable combination of drugs as detailed in the Table. The local ethics committee approved all tissue collections.
Endothelial-Restricted Expression of β3-Adrenoceptors in Human Coronary Arterioles
An analysis of β3-adrenoceptor and NOS3 mRNAs was performed by RT-PCR with the use of dissected arterioles from right auricular appendages. Amplimers for both transcripts were detected from 3 different preparations, with a mean Ct of 34.0±0.15 (P<0.0001 versus background; n=3) for β3-adrenoceptors and 36.0±1.1 for NOS3 (P=0.022 versus background; n=3). By comparison, the highly expressed housekeeping gene GAPDH generated detectable signals at a mean Ct of 26.0±0.26 when amplified from the same cDNAs.
Figure 1 illustrates the identification of specific proteins with immunoaffinity for anti-human β3-adrenoceptor antibodies in both Western blotting (Figure 1A) and immunohistochemistry (Figure 1B). Bands corresponding to β3-adrenoceptors and eNOS were detected both in whole cardiac extracts from left ventricle and atria and in extracts of arterioles microdissected from the same tissues. In both cases, signals for β3-adrenoceptors and eNOS are more intense in atrial versus ventricular extracts. To identify the specific cell type(s) expressing β3-adrenoceptors, immunohistochemical analysis was performed in sections of human atrial myocardium. A positive staining was observed in cardiomyocytes, as previously described by us.11 In addition, sections of arterioles stained positively (Figure 1B, top left). Higher magnification revealed exclusive staining of endothelial cells of microarteries. No staining was observed in capillary endothelial cells (closely apposed to cardiomyocytes) or vascular smooth muscle cells (Figure 1B, bottom).
β3-Adrenoceptors Mediate Relaxation of Human Coronary Arterioles
To assess the function of β3-adrenoceptor signaling in the same vessels, variations of external diameter of preconstricted, pressurized human coronary microarterioles were studied by videomicroscopy with different β-adrenoceptor agonists. As illustrated by the typical experiment presented on Figure 2A, in vessels with an intact endothelium, the nonspecific β-agonist isoproterenol relaxed endothelin-1 (ET-1) preconstricted microarterioles by half. This relaxation was not affected on pretreatment with the β1/β2-antagonist nadolol, thereby ruling out a β1/β2-adrenoceptor-mediated effect, but was fully abrogated by the β1/β2/β3-antagonist bupranolol, suggesting a β3-adrenoceptor-mediated effect (Figure 2B). In support of the latter, the preferential β3-adrenoceptor agonist BRL37344 produced a dose-dependent relaxation of the same amplitude (Figure 2C), and the maximum relaxation to BRL37344 was 52.3±13.2% of the ET-1 contraction (n=6). This relaxation was also resistant to nadolol (54.9±16.8%; n=5) but was fully abrogated by bupranolol (n=4; Figure 2D).
In coronary microarterioles with an intact endothelium pretreated with nadolol and phentolamine (combining α1/α2- and β1/β2-adrenoceptor blockade), norepinephrine (1 μmol/L) also evoked a relaxation amounting to 41.4±7% (n=3) of ET-1 contraction (Figure 3).
Of note, the relaxation to BRL37344 was not observed in vessels that failed to relax to the endothelium-specific agonist substance P (not shown) or in which the endothelium was selectively destroyed, despite their full relaxation with sodium nitroprusside (a NO donor acting on the smooth muscle), confirming that the β3-adrenoceptor response is dependent on a functional endothelium (Figure 4A). Conversely, these deendothelialized vessels exhibited a residual relaxation to isoproterenol (21.0±6.3%; n=6), which was inhibited by nadolol (1.4±0.5%; n=4) (Figure 4B). This identified an additional endothelium-independent, β1/β2-adrenergic response on the vascular smooth muscle.
β3-Adrenoceptor-Mediated Relaxation Involves Both NO and an EDHF
To characterize the endothelial mediator for β3-adrenoceptor relaxation, the effect of BRL37344 was first compared in vessels preconstricted with ET-1 or a high (50 mmol/L) KCl solution. The latter is known to depolarize the vessel membrane and prevent the relaxing effect of a (EDHF-like) hyperpolarizing factor. As shown in Figure 5A, although BRL37344 relaxed vessels preconstricted with both ET-1 and KCl, its relaxing effect was substantially reduced in vessels contracted with KCl (ET-1: 52.3±13.2% [n=6] versus KCl: 15.5±5.3% [n=4]), suggesting the participation of an EDHF-like response.
To determine the involvement of NO production in the KCl-resistant relaxation, similar experiments were performed in vessels preincubated with the NOS inhibitor l-ω-nitroarginine. NOS inhibition abrogated the residual relaxation with BRL37344 in KCl-preconstricted vessels, confirming NOS involvement in the β3-adrenoceptor response. Of note, the relaxing effect of BRL37344 on ET-1-preconstricted vessels was unaffected by NOS inhibition, suggesting compensation by the EDHF-like response (Figure 5B).
Two additional approaches were used to confirm the involvement of an EDHF-like response in the NO-independent β3-adrenoceptor-mediated relaxation. First, we tested the effect of BRL37344 on the membrane potential of human arterioles mounted in the same conditions as for the relaxation assays. As illustrated in Figure 6A, acute application of BRL37344 in the presence of NOS and cyclooxygenase inhibition (to rule out confounding effects of NO and prostanoids) resulted in significant hyperpolarization. Second, the sensitivity of the BRL37344-mediated relaxation to the calcium-activated K+ channel inhibitors charybdotoxin and apamin was tested in similar vessels under videomicroscopy. As shown in Figure 6B, these inhibitors fully abrogated the residual relaxation, further confirming its mediation through an EDHF pathway.
We characterized a novel pathway for the adrenergic vasorelaxation of human coronary microarteries through activation of β3-adrenoceptors on endothelial cells. This is distinct from the vasodilation that follows activation of adenylyl cyclase and increases in cAMP, which are most often ascribed to β2-adrenoceptor (and perhaps β1-adrenoreceptor) activation in vascular smooth muscle cells. Indeed, our functional experiments with nonspecific β-adrenergic (as well as β3-preferential) agonists demonstrate a vasorelaxation of human coronary microarteries with slow kinetics, similar to the smooth muscle relaxation attributed to the activation of atypical β-adrenoceptors in other tissues. The fact that this vasorelaxation was insensitive to nadolol (a β1/β2-adrenoceptor antagonist) and was abrogated by bupranolol (a full β-antagonist) strongly supported a β3-adrenoceptor-mediated response.
Accordingly, we provide evidence for the expression of mRNA and proteins of β3-adrenoceptors in extracts of dissected cardiac microarterioles. We had successfully used the same molecular approaches to identify and quantify human β3-adrenoceptors in whole human myocardium.11 Using immunohistochemistry, we show that, in addition to cardiomyocytes, β3-adrenoceptor expression is restricted to the endothelium of microarteries, with no staining of vascular smooth muscle. This is consistent with similar endothelium-restricted expression in rat aorta.16 It would also account for the lack of β3-adrenoceptor-mediated relaxation in vessels with dysfunctional or destroyed endothelium in which we failed to obtain a typical endothelial-mediated relaxation with substance P. Likewise, such response may have been undetected in previous studies in which human coronary microarteries from patients with end-stage heart failure and endothelial dysfunction were used, leaving only a residual smooth muscle-mediated β2-adrenergic response.3
Aside from prostanoids, NO and EDHF(s) account for the prototypical endothelium-mediated vasorelaxation. Consistent with the expression of eNOS in our vessels, we found the β3-adrenoceptor relaxation to be mediated partly through NO production. This was evidenced by its complete abrogation by NOS inhibition under circumstances when both prostanoids and EDHF are inoperative (ie, after cyclooxygenase inhibition and preconstriction with high KCl, respectively). This also recapitulates our previous demonstration of a functional coupling of β3-adrenoceptor agonists to NO production in whole human ventricular muscle through Gαi proteins.11 However, NOS inhibition had little (if any) effect on the vasorelaxation of vessels preconstricted with ET-1. This cannot be explained by incomplete NOS inhibition because similar treatment with l-ω-nitroarginine abrogated the endothelium-dependent relaxation to substance P in the same vessels preconstricted with KCl (not shown). This strongly indicated the involvement of an alternative, EDHF-like response.
Although the precise nature of EDHF is still elusive, a consensus view is that this (these) factor(s) released from endothelial cells produces a hyperpolarization leading to vascular muscle relaxation through activation of calcium-dependent K+ channels.25 Our results in vessels constricted with high KCl solution (which modifies the electrochemical gradient for K+ ions, thereby preventing hyperpolarization) are in agreement with the participation of an EDHF. In addition, we directly demonstrated vessel hyperpolarization in response to β3-adrenoceptor agonists and the abrogation of β3-adrenoceptor-mediated relaxation after vessel pretreatment with the K+ channel inhibitors charybdotoxin and apamin, 2 signatures of an EDHF response. These results are also in agreement with the recent proposition of β3-adrenoceptor-mediated relaxation through K+ channel activation in rat aorta.16 Of note, the apparent insensitivity of the β3 relaxation to NOS inhibition in vessels preconstricted with ET-1 suggests that this EDHF response fully compensates for the absence of NO. Indeed, previous reports have suggested that EDHF acts as a backup relaxing mechanism in circumstances of endothelial NO-dependent dysfunction.18,20,26
If our findings in isolated vessels truly reflect vasoregulation in vivo, our demonstration of a functional β3-adrenoceptor vasorelaxation mediated in part by EDHF in human coronary resistance arteries may have a major bearing on the understanding of the regulation of coronary perfusion in circumstances such as dyslipidemia, diabetes, and atherosclerosis, all associated with decreased NO production and/or bioavailability. Indeed, previous work in human arteries suggested EDHF to be preserved despite the presence of risk factors for atherosclerosis (as is the case in the present study, in which we used arterioles mostly from ischemic patients).18,27 Notably, the natural catecholamine norepinephrine elicited similar β3-adrenoceptor vasorelaxation, extending the relevance of our paradigm. Given the relative resistance of β3-adrenoceptors to homologous desensitization,28 their activation of such backup relaxation would seem particularly appropriate in circumstances of increased adrenergic tone, such as ischemia or heart failure, to preserve myocardial perfusion.
This study was supported by grants ARC 01/06-271 and SSTC-PAI-P5/02 to Dr Balligand. Dr Dessy is a research associate of the Fonds National de la Recherche Scientifique (FNRS).
Sun D, Huang A, Mital S, et al. Norepinephrine elicits beta2-receptor-mediated dilation of isolated human coronary arterioles. Circulation. 2002; 106: 550–555.
Emorine LJ, Marullo S, Briend-Sutren MM, et al. Molecular characterization of the human beta3-adrenergic receptor. Science. 1989; 245: 1118–1121.
Moniotte S, Kobzik L, Feron O, et al. Upregulation of beta(3)-adrenoceptors and altered contractile response to inotropic amines in human failing myocardium. Circulation. 2001; 103: 1649–1655.
Shen YT, Zhang H, Vatner SF. Peripheral vascular effects of beta-3 adrenergic receptor stimulation in conscious dogs. J Pharmacol Exp Ther. 1994; 268: 466–473.
Berlan M, Galitzky J, Bousquet-Melou A, et al. Beta-3 adrenoceptor-mediated increase in cutaneous blood flow in the dog. J Pharmacol Exp Ther. 1994; 268: 1444–1451.
Nishikawa Y, Stepp DW, Chilian WM. Nitric oxide exerts feedback inhibition on EDHF-induced coronary arteriolar dilation in vivo. Am J Physiol. 2000; 279: H459–H465.
Miura H, Gutterman DD. Human coronary arteriolar dilation to arachidonic acid depends on cytochrome P-450 monooxygenase and Ca2+-activated K+ channels. Circ Res. 1998; 83: 501–507.
Terata K, Miura H, Liu Y, et al. Human coronary arteriolar dilation to adrenomedullin: role of nitric oxide and K(+) channels. Am J Physiol. 2000; 279: H2620–H2626.
Sonveaux P, Brouet A, Havaux X, et al. Irradiation-induced angiogenesis through the up-regulation of the nitric oxide pathway: implications for tumor radiotherapy. Cancer Res. 2003; 63: 1012–1019.
Miura H, Liu Y, Gutterman DD. Human coronary arteriolar dilation to bradykinin depends on membrane hyperpolarization: contribution of nitric oxide and Ca2+-activated K+ channels. Circulation. 1999; 99: 3132–3138.
Liggett SB, Freedman NJ, Schwinn DA, et al. Structural basis for receptor subtype-specific regulation revealed by a chimeric beta 3/beta 2-adrenergic receptor. Proc Natl Acad Sci U S A. 1993; 90: 3665–3669.