Heart and Circulatory Physiology

Cardiac myocytes control release of endothelin-1 in coronary vasculature

Daphne Merkus, Anna K. Brzezinska, Cuihua Zhang, Shuichi Saito, William M. Chilian


α-Adrenergic vasoconstriction in the coronary circulation is mediated through α-adrenoceptors on cardiac myocytes and subsequent release of endothelin, a very potent, long-lasting vasoconstrictor. Recent studies found that adult cardiac myocytes do not express the preproendothelin gene. Thus we hypothesized that α-adrenoceptor stimulation on the cardiac myocytes results in the production of an endothelin-releasing factor, which stimulates the coronary vasculature to produce endothelin. We tested this hypothesis by using an in vitro model in which isolated adult rat cardiac myocytes can be stimulated with an α-adrenoceptor agonist (phenylephrine). Their bathing fluid is then transferred to isolated coronary arterioles, and vasoactive responses are measured. To identify the source of endothelin, the endothelin-converting enzyme inhibitor phosphoramidon was added to either the myocytes or the isolated arterioles. Phenylephrine enhanced the vasoconstrictor properties of the myocyte bathing fluid. Administration of phosphoramidon (in either the presence or the absence of phenylephrine) to the myocytes had no effect on the vasoactive properties of the bathing fluid. In contrast, administration of phosphoramidon to the isolated arteriole before administration of the bathing fluid converted vasoconstriction to vasodilation, similar to the effect of the endothelin A receptor antagonist JKC-301, indicating that the endothelin is indeed produced by the coronary vasculature. Administration of the angiotensin type 1 receptor antagonist losartan to the vessel bath enhanced vasodilation to the bathing fluid of the phenylephrine-treated but not control myocytes. In conclusion, during α-adrenergic activation cardiac myocytes release a factor, probably angiotensin II, that stimulates the vascular production of endothelin. Although the physiological implications of this mechanism are not obvious, this may represent a protective mechanism that integrates neuronal vasoconstrictor mechanisms with myocardial metabolism, which minimizes periods of both coronary underperfusion and overperfusion.

  • coronary circulation
  • coronary microcirculation
  • coronary blood flow
  • vasoconstriction

endothelin is an endogenous factor that produces robust and long-lasting coronary vasoconstriction (5, 6, 22). Therefore, it is imperative that the release and/or production of endothelin by coronary and myocardial cells is carefully regulated. Although the principal driving force for the control of coronary blood flow is myocardial metabolism, this intimate relationship between cardiac metabolism and coronary blood flow may be modified by a variety of conditions and stimuli (2, 9, 13, 15). This is especially important for vasoconstrictor influences, because they can potentially compromise myocardial perfusion and thereby impair cardiac function. In fact, endothelin-induced coronary constriction can be so potent that myocardial ischemia occurs.

The sympathetic nervous system may play an important role in matching myocardial oxygen supply to myocardial oxygen demand through the influence of norepinephrine release from the terminal nerve endings on the coronary vasculature. Norepinephrine can cause α-adrenergic vasoconstriction as well as β-adrenergic vasodilation in various species (1, 3, 7, 8, 14). In previous studies (20), we showed that, unlike in other organs, α1-adrenergic constriction in the coronary circulation is not mediated via activation of α1-adrenoceptors on the vascular myocytes. Rather, the α1-adrenoceptor agonists activate the receptor on cardiac myocytes, and the resulting coronary constriction is mediated through endothelin-1 (1, 20). On the basis of these observations, we formulated the following concept: activation of α1-adrenoceptors on the cardiac myocytes causes release of endothelin-1, which diffuses to the vasculature and causes constriction.

However, this conclusion is challenged by the observation that adult cardiac myocytes do not express the gene for preproendothelin, the precursor peptide for endothelin (17). Moreover, direct measurement of endothelin-1 concentration in the supernatant of phenylephrine-treated myocytes yielded concentrations below the vasoactive threshold (20).

We therefore proposed a revised hypothesis: cardiac myocytes produce an endothelin-releasing factor, and the release of this factor occurs during α1-adrenergic stimulation of cardiac myocytes, which induces endothelin release by vascular endothelial cells. The aim of this study was to test these hypotheses by determining the source of endothelin released during α1-adrenergic stimulation.


Studies were performed in accordance with the “Guiding principles in the Care and Use of Laboratory Animals” as approved by the Council of the American Physiological Society and with the approval of the Animal Care Committee of our institutions. Male Wistar rats (175–250 g, Harlan Sprague Dawley) were anesthetized with pentobarbital sodium (75 mg/kg ip) and decapitated. The thorax was opened, and the heart was removed and placed in ice-cold physiological saline solution (PSS). The hearts were then used for either dissection of coronary arterioles or isolation of cardiac myocytes.

Isolation of cardiac myocytes.

Myocytes were isolated with a modified Langendorff setup (16). In brief, the heart was perfused with buffer containing (in mM) 123 NaCl, 2.6 KCl, 1.2 KH2PO4, 1.2 MgSO4·7H2O, 25 HEPES, 11 glucose, 20 taurine, 20 creatine, and 1 CaCl2 (pH 7.4) for 3 min, after which the buffer was replaced by a buffer with the same composition except without the calcium for 6 min. Unless otherwise mentioned, all drugs were obtained from Sigma. Collagenase (type II, 0.6 mg/ml; Worthington) and CaCl2 (30 μM) were then added, and the heart was perfused for another 15 min. The heart was then cut into small pieces and resuspended in perfusion buffer to which BSA (1%) was added. After 5 min, during which the suspension was gently triturated, the buffer containing the myocytes was filtered through surgical gauze to remove big clumps of myocytes and gently spun down. Calcium was reintroduced to the myocytes in a stepwise manner (200 μM, 500 μM, and 1 mM CaCl2). The myocytes were allowed to settle under gravity in between the CaCl2 steps. The supernatant was removed, and the pellet was resuspended in buffer containing more calcium. Because rod-shaped living cells settle faster than dead cells, this increased the percentage of viable cells in the suspension. Dead cells were identified as cells that were round in shape, whereas healthy cells were rod shaped. Only preparations containing at least 70% rod-shaped cells were used.

Dissection of coronary arterioles.

Single arterioles (40- to 130-μm passive diameter) were dissected from the left ventricular free wall or the septum of the rat heart, as previously described for other species (10), and placed in ice-cold PSS of the following composition (in mM): 145.0 NaCl, 4.7 KCl, 2.0 CaCl2, 1.17 MgSO4, 1.2 NaH2PO4, 5.0 glucose, 2.0 pyruvate, 0.02 EDTA, and 3.0 MOPS, buffered to pH 7.4 at 4°C and filtered (dissection buffer). The heart was placed under a dissection microscope in a 4°C chamber, and vessels were carefully dissected free from the surrounding myocardial tissue and placed in dissection buffer containing 1% BSA (USB-Amersham). The vessels were cannulated on both ends with micropipettes (∼20- to 60-μm outer diameter, depending on the size of the vessel) connected to pressurized reservoirs filled with PSS buffered to pH 7.4 at 37°C. The height of these reservoirs was set to obtain the desired intraluminal pressure (60 mmHg). Vessels that failed to maintain pressure were excluded from analysis. Internal diameter of coronary microvessels was measured under a charge-coupled device camera (Sony CCD-IRIS) with a videocaliper system. The vessel was slowly warmed up to 37°C and allowed to develop spontaneous tone.

Detection of mRNA transcripts with RT-PCR.

To demonstrate the source of endothelin in cardiac and coronary arterioles, we used RT-PCR. Enzymatically isolated cardiac myocytes were selected from three different batches according to a recently described method (16), to ensure that only mRNA from these cells was analyzed. In addition, mRNA was extracted from isolated coronary arterioles. RT-PCR was performed to investigate the presence of mRNA for preproendothelin, troponin T, and GAPDH. mRNA was isolated with a kit from Dynal, according to the manufacturer's instructions. Briefly, samples were sonicated in a Tris·HCl-LiCl buffer in the presence of RNase inhibitors to liberate RNA. Samples were then incubated with 30-μm magnetic beads coated with polythymidine linkers, which bind specifically and with high avidity to the polyadenylate tail of mRNA. After 15 min of incubation to allow binding of mRNA, the beads were pelleted in a magnetic precipitator and the supernatant was removed. The beads and mRNA were resuspended in extraction buffer, and the mRNA was converted to cDNA with a First Strand Ready-To-Go Beads kit (Pharmacia Biotech) according to the manufacturer's instructions. To compare expression of individual genes, cDNA was incubated with specific primers (Operon) and GIBCO PCR Supermix and cycled in a PerkinElmer thermocycler (35 cycles of 30 s at 95°C, 30 s at 55°C, 45 s at 72°C, followed by 7-min elongation at 72°C). The PCR was performed in duplicate. PCR primers for endothelial NO synthase (eNOS, endothelial cell marker; Ref. 21), endothelin-1, troponin T (cardiac myocyte marker), and β-actin (marker for both endothelial cells and cardiac myocytes) were designed with Primer Picker from rat sequences obtained from GenBank. The sequences for the primer sets were β-actin: sense 5′-gtc gta cca ctg gca ttg tg-3′, antisense 5′-ctc tca gct gtg gtg gtg aa-3′; endothelin-1: sense 5′-gcg tcc ttg ttc caa aca tt-3′; antisense 5′-gcc tga gtc aga cac gaa ca-3′; eNOS: sense 5′-tga ccc tca ccg ata caa ca-3′, antisense 5′-ctg gcc ttc tgc tca ttt tc-3′; and troponin T: sense 5′-cct gca gga aaa gtt caa gc-3′, antisense 5′-gtg cct ggc aag acc tag ag-3′.


Cardiac myocytes were incubated in the presence and absence of the α1-adrenoceptor agonist phenylephrine (50 μM; Ref. 20) and the endothelin-converting enzyme inhibitor (phosphoramidon, 10 μM). We showed previously (20) that the response to this dose of phenylephrine can be blocked with prazosin and is therefore α1-adrenoceptor mediated. The production of vasoactive factors by the isolated myocytes was measured in a bioassay by withdrawing supernatant from the myocyte suspension, which was then added to the isolated arterioles to establish effects on diameter (12) . Aliquots of myocyte supernatant (100, 200, and 500 μl) were added to the vessel bath (vol 4 ml). Vascular diameter was measured 5 min after addition of the supernatant and after 10 min of washout. The addition of the supernatant was repeated in the presence of the endothelin A receptor antagonist JKC-301 (5 μM; American Peptide), the angiotensin AT-1 receptor antagonist losartan (1 μM), and the endothelin-converting enzyme inhibitor phosphoramidon (10 μM)

Data analysis and statistics.

All statistic analyses were performed with StatView software for the Macintosh (Abacus Concepts, Berkeley, CA). Data were compared with ANOVA for repeated measures (dose of supernatant) with Scheffé's test as post hoc multiple-comparison test. Vascular diameters are normalized to the diameter with tone, before administration of the supernatant. Data are presented as means ± SE. Significance was accepted at P < 0.05 in all experiments.


Vasoactive properties.

Supernatant of cardiac myocytes was added to isolated coronary arterioles, and their response was measured. Similar to results obtained in the canine circulation (20), treatment of myocytes with the α1-adrenoceptor agonist phenylephrine (50 μM) increased the vasoconstrictor properties of the myocyte supernatant (Fig. 1), whereas treatment of the coronary arterioles with phenylephrine in this concentration had no effect (data not shown). Administration of the endothelin A receptor antagonist JKC-301 to the coronary arterioles increased the vasodilator properties of the supernatant, indicating that dilation of the vessels was limited by the presence of endothelin-1.

Fig. 1.

Dilation of coronary arterioles to myocyte supernatant (n = 8) was dose dependent. Stimulation of the α1-adrenoceptor with phenylephrine (PE, 50 μM; n = 8) converted vasodilation to the supernatant into vasoconstriction. Administration of the endothelin A receptor antagonist JKC-301 (5 μM; n = 6) to the coronary arterioles increased the vasodilator properties of the myocyte supernatant, indicating that endothelin limits the dilation to the myocyte supernatant. *P < 0.05 vs. control.

Source of endothelin.

We showed in a previous study (20) that the amount of endothelin-1 in the supernatant was too low to have any vasoactive effects. However, the vasoactive effects of the supernatant increase with blockade of the endothelin A receptor (1, 20). Additionally, a recent study showed that the preproendothelin gene is not expressed in cardiac myocytes (17). We confirmed these results in our myocyte suspension. The mRNA levels for endothelin-1, troponin T, β-actin, and eNOS are shown in Fig. 2. Transcripts from whole heart revealed signals for all the genes, as would be expected from RNA from many cell types, e.g., cardiac myocytes, fibroblasts, endothelial cells, and smooth muscle cells. However, the analysis of transcripts from cardiac myocytes showed transcripts for troponin T and β-actin only, with signals from eNOS or endothelin-1 being absent. Thus these results support the contention that in the adult heart endothelin-1 is produced not by cardiac myocytes, but in the vascular wall.

Fig. 2.

Ethidium bromide signals of products from RT-PCR (35 cycles) for β-actin (β-act), cardiac troponin T (trop), endothelial NO synthase (eNOS), and endothelin-1 (ET-1) for transcripts isolated from enzymatically isolated cardiac myocytes (M) and whole heart (H). One hundred-base pair ladders were run on both sides of the gel. Note that products from β-actin and troponin were robust in both preparations but the signals from eNOS and ET-1 were observed only in transcripts from the whole heart and were absent in the myocyte preparations.

Inhibition of endothelin-converting enzyme (essential in the conversion of big endothelin to endothelin-1) with phosphoramidon (10 μM) in the myocyte suspension did not affect the vasoactive properties of the supernatant of either quiescent myocytes or phenylephrine-treated myocytes (Fig. 3). We therefore investigated whether the arterioles could be the source of endothelin-1. Inhibition of endothelin-converting enzyme in the isolated arterioles with phosphoramidon (10 μM) increased the vasodilator properties of the supernatant of quiescent and phenylephrine-treated myocytes (Fig. 4) to a similar extent as treatment with an endothelin receptor antagonist, indicating that coronary arterioles produce endothelin in response to a signal from the myocytes, which subsequently limits their dilation.

Fig. 3.

Administration of phosphoramidon (Phos, 10 μM), an inhibitor of endothelin-converting enzyme, to the myocytes did not affect dilation of the vessels to the supernatant of quiescent myocytes (n = 6 with and without Phos) or PE-treated myocytes (50 μM; n = 5 with and without Phos). *P < 0.05 vs. control.

Fig. 4.

Administration of Phos (10 μM), an inhibitor of endothelin-converting enzyme, to the vessels increased their dilation to the supernatant of both quiescent myocytes (n = 6, with and without Phos) and PE (50 μM; n = 5, with and without Phos)-treated myocytes. *P < 0.05 vs. control; †P < 0.05 vs. PE; ‡P < 0.05 effect of Phos is different from control.

Because Winegrad et al. (23) identified angiotensin as the endothelin-releasing signal in their study, we administered the AT-1 receptor antagonist losartan (1 μM) to the isolated arterioles and measured dilation to the highest dose of myocyte supernatant (Fig. 5). Losartan did not affect the vasodilator properties of the supernatant of untreated myocytes, whereas it increased the dilation to supernatant of myocytes treated with phenylephrine (Fig. 6).

Fig. 5.

Angiotensin type 1 receptor blockade (losartan, 1 μM; n = 4 in all protocols) on the coronary arterioles did not affect the vasodilator properties of the myocyte supernatant, whereas endothelin receptor blockade did increase the dilation, indicating that angiotensin II is not the endothelin-releasing factor in our preparation. *P < 0.05 vs. control.

Fig. 6.

Angiotensin type 1 receptor blockade (losartan, 1 μM) on the coronary arterioles did not increase vasodilation to control myocyte supernatant (500 μl; n = 4, paired data, left), whereas it increased vasodilation to supernatant of myocytes treated with PE (50 μM; n = 5, paired data, right). *P < 0.05 vs. supernatant without PE; P < 0.05 vs. supernatant without losartan.


In our study, we have shown that adult rat cardiac myocytes do not produce endothelin-1. These myocytes do not transcribe the gene for preproendothelin, the precursor peptide of endothelin-1, and administration of the endothelin-converting enzyme inhibitor phosphoramidon to cardiac myocytes did not affect vasodilator properties of the myocyte supernatant. However, phosphoramidon, when administered to coronary arterioles, enhanced the vasodilator properties of the myocyte supernatant. Therefore, we conclude that endothelin-1 is produced in coronary microvessels in response to a chemical signal from cardiac myocytes. Interestingly, this signal is modified by α1-adrenergic stimulation of cardiac myocytes, which implies that these cells produce a substance or substances that stimulate the production of endothelin by coronary arterioles. Our results also imply that regulation of coronary vascular tone depends on a tightly regulated balance in the production and/or release of constrictors and vasodilators by cardiac myocytes, which builds upon our previous study (12) in which we found that the production of endothelin is dependent on myocardial oxygen consumption. Before discussing the implications of our findings, we will first evaluate the limitations and benefits of our methodological approaches.

Methodological considerations.

We used an in vitro system to investigate the control of endothelin production by studying enzymatically isolated cardiac myocytes and isolated coronary arterioles (12). A potential problem is that enzymatic isolation may damage or even kill the cells. We minimized the problem of dead cells by using only preparations with >70% rod-shaped cells. There is the possibility that dead cells may release vasoactive substances, which could complicate our findings. However, supernatant from preparations with mainly dead cells (≥80%) was not vasoactive (data not shown). Therefore, we believe that the responses to the supernatant involve vasoactive substances released by the viable cardiac myocytes.

The main advantage of our in vitro system is that we can distinguish between production of factors by cardiac myocytes and vascular cells. This is important for unequivocal identification of the source of endothelin. Within this context, two separate experiments identify the vasculature as the source of endothelin. First, the RT-PCR results show that the gene for preproendothelin, the precursor of endothelin, is not expressed in adult cardiac myocytes, which is in accordance with a study from Preisig-Muller et al. (17). Second, administration of the endothelin-converting enzyme inhibitor separately to both cardiac myocytes and coronary arterioles allowed us to identify the vasculature as the source of endothelin. Interestingly, the cardiac myocytes released a factor, which may be angiotensin II, that stimulates the vascular production of endothelin. Additionally, many substances are both vasoactive and cardioactive, because their receptors are present both on vascular and cardiac myocytes. With our system, we can block the receptors on the vasculature without affecting the myocytes and vice versa. This is especially important for the endothelin A receptor, which, when stimulated, causes vasoconstriction as well as an increase in inotropy of the myocytes (in both humans and rats; Ref. 4), which in turn may increase the production of vasodilators and increase extravascular compression. Thus administration of an endothelin A receptor antagonist in vivo also has indirect effects on the vasculature and may therefore result in overestimation of the true vascular effects of endothelin-1 when tested in vivo.

Implications of findings.

Quiescent cardiac myocytes produce factors that stimulate the production of endothelin in coronary arterioles, which increases coronary vascular tone. The production of these factors is controlled by cardiac metabolism (12) as well as by the sympathetic nervous system, through the α-adrenoceptor. α-Adrenergic vasoconstriction has been proposed to prevent excessive backflow from the coronary circulation, thereby promoting subendocardial perfusion. Through production of vasoconstrictors, cardiac myocytes are capable of increasing basal coronary resistance to prevent excess myocardial perfusion.

Conventionally, the dogma for coronary control is that the production of an unknown vasodilator(s) is directly coupled to myocardial oxygen consumption. Thus at low levels of myocardial oxygen consumption reduced production of vasodilator(s) is associated with augmented coronary tone. However, our results challenge this dogma by suggesting that the enhanced production of a constrictor at low levels of myocardial oxygen consumption increases coronary tone. Perhaps a system that utilizes two opposing control schemes (dilators vs. constrictors) can more quickly adjust to alterations in metabolism than one that simply titrates flow to altered production of vasodilators. In accordance with this theory, we recently showed (12) that, when an increase in myocardial oxygen consumption occurs, flow can be increased both by augmented production of vasodilators and by concomitant reduction in the production of endothelin. Moreover, recent experiments by Winegrad et al. showed that myocytes can act as oxygen sensors in the heart, producing a vasodilator when oxygen tension decreases and a constrictor when oxygen tension increases. In that study, the constrictor involved was identified as angiotensin, which caused endothelin-mediated constriction (23). In our experiments with supernatant from quiescent myocytes the endothelin-releasing factor did not seem to be angiotensin II, because administration of losartan to the coronary arterioles did not alter the vasoactive properties of the supernatant. Surprisingly, when the myocytes were stimulated with phenylephrine, the vasoconstriction of the coronary arterioles to the supernatant was reversed by losartan, indicating that phenylephrine-stimulated myocytes do release angiotensin II. Yet inhibition of endothelin-converting enzyme in the coronary arterioles converts vasoconstriction to supernatant of phenylephrine-treated myocytes to vasodilation. Thus, even after angiotensin receptor blockade, there may still be some vascular endothelin production. Apparently, there are two factors that can be released from cardiac myocytes, one of them being angiotensin II, that induce endothelin production in the coronary vasculature.

An advantage of endothelin production in the endothelium is that the endothelium directly senses changes in shear stress and thus can provide a feedback system if vasoconstriction is too intense. Vasoconstriction will result in an increase in shear stress, thereby increasing the production of NO, which then limits the vasoconstriction through its vasodilator action on vascular smooth muscle (19). Also, NO limits the production and release of endothelin by the endothelium (11). Thus NO provides a negative feedback mechanism to limit excess vasoconstriction induced by endothelin.

Clinically, our findings are important in patients with heart failure. Heart failure results in increased sympathetic nervous system activity (18), aimed at maintaining blood pressure, through an increase in heart rate and peripheral vasoconstriction, to ensure adequate perfusion of the brain. However, the enhanced sympathetic stimulation of the cardiac myocytes will also result in enhanced endothelin production by the coronary vasculature. Moreover, because heart failure is often accompanied by endothelial dysfunction, NO production is decreased, which may result in inadequate negative feedback on vascular endothelin production. Thus coronary blood supply may be limited by an increased endothelin-induced vasoconstriction, while myocardial oxygen demand is increased because of the increased workload of the heart (increased heart rate as well as increased peripheral resistance).

In conclusion, because endothelin is a very potent and long-lasting vasoconstrictor, its production and release need to be carefully regulated. The prime site of production of endothelin in the heart is the coronary vasculature, but endothelin production is regulated based on signals obtained from the cardiac myocytes.


We gratefully acknowledge support from the American Heart Association (Grant 9920433Z), the National Heart, Lung, and Blood Institute (Grants HL-32788 and HL-65203), as well as the Netherlands Heart Foundation (Grant 2000T042).


  • 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