- Email: [email protected]
Gap Junction-Dependent Increases in Smooth Muscle cAMP Underpin the EDHF Phenomenon in Rabbit Arteries
Biochemical and Biophysical Research Communications 283, 583–589 (2001) doi:10.1006/bbrc.2001.4791, available online at http://www.idealibrary.com on
Gap Junction-Dependent Increases in Smooth Muscle cAMP Underpin the EDHF Phenomenon in Rabbit Arteries Hannah J. Taylor, Andrew T. Chaytor, David H. Edwards, and Tudor M. Griffith 1 Department of Diagnostic Radiology, Wales Heart Research Institute, University of Wales College of Medicine, Cardiff CF14 4XN, United Kingdom
Received March 28, 2001
We have investigated the role of cAMP in nitric oxide (NO)- and prostanoid-independent vascular relaxations evoked by acetylcholine (ACh) in isolated arteries and perfused ear preparations from the rabbit. These EDHF-type responses are shown to be associated with elevated cAMP levels specifically in smooth muscle and are attenuated by blocking adenylyl cyclase or protein kinase A (PKA). Relaxations are amplified by 3-isobutyl-1-methylxanthine, which prevents cAMP hydrolysis, while remaining susceptible to inhibition by the combination of two K Ca channel blockers, apamin and charybdotoxin. Analogous endothelium- and cAMP-dependent relaxations were evoked by cyclopiazonic acid (CPA) which stimulates Ca 2ⴙ influx via channels linked to the depletion of Ca 2ⴙ stores. Responses to ACh and CPA were both inhibited by interrupting cell-to-cell coupling via gap junctions with 18␣-glycyrrhetinic acid and a connexin-specific Gap 27 peptide. The findings suggest that EDHF-type responses are initiated by capacitative Ca 2ⴙ influx into the endothelium and propagated by direct intercellular communication to effect relaxation via cAMP/PKAdependent phosphorylation events in smooth muscle. © 2001 Academic Press
Key Words: endothelium-derived hyperpolarizing factor (EDHF); gap junctions; cyclic AMP; epoxyeicosatrienoic acids; cyclopiazonic acid.
Nitric oxide (NO)- and prostanoid- independent vascular relaxations evoked by agonists such as acetylcholine (ACh) are closely associated with endothelial hyperpolarization mediated by the opening of Ca 2⫹activated K ⫹ channels (K Ca) (1–5). Recent evidence indicates that the mechanisms subsequently leading to relaxation involve electrotonic spread of endothelial hyperpolarization into the smooth muscle and/or the To whom correspondence should be addressed. Fax: ⫹44 2920744726. E-mail: [email protected]. 1
diffusion of an unidentified chemical mediator via direct intercellular communication pathways. Specific synthetic peptides that interrupt cell– cell coupling via gap junctions present in the vascular wall have thus been shown to inhibit ACh-evoked NO- and prostanoidindependent smooth muscle relaxations and hyperpolarizations in rabbit, rat, porcine, and guinea pig vessels (6 –14). Alternatively, it has been suggested that an endothelium-derived hyperpolarizing factor (EDHF) transfers from the endothelium to subjacent smooth muscle across the extracellular space, putative candidates for which include cytochrome P 450 epoxygenase metabolites of arachidonic acid (epoxyeicosatrienoic acids [EETs]) which activate smooth muscle K Ca channels, and K ⫹ ions which activate inwardly rectifying K ⫹ channels (K ir) and the membrane Na ⫹-K ⫹ ATPase (1, 15, 16). EETs and elevations in extracellular [K ⫹] however do not universally induce direct smooth muscle relaxation, and in some vessels may evoke endothelium-dependent responses that are susceptible to blockade of gap junctions (9, 17). While NOand prostanoid-independent relaxations have been attributed to reduced Ca 2⫹ influx via voltage-gated Ca 2⫹ channels following smooth muscle hyperpolarization (18), relaxation may not always be wholly dependent on changes in membrane potential. In the rat mesenteric artery, for example, blockade of smooth muscle hyperpolarization by simultaneous inhibition of K ir channels (by Ba 2⫹) and the Na ⫹-K ⫹ ATPase (by ouabain) causes a rightward shift in the concentrationrelaxation curve to ACh but does not alter the maximum response (1). In the present study we have used rabbit artery and vascular bed preparations to test the hypothesis that NO- and prostanoid-independent relaxations involve synthesis of cAMP, whose multiple actions encompass several characteristics of the “EDHF phenomenon” noted above. These include (i) hyperpolarization via the stimulation of K Ca channels (19) and the Na ⫹-K ⫹
583
0006-291X/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.
Vol. 283, No. 3, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
ATPase (20), (ii) enhanced electrical and chemical signaling via gap junctions (21, 22), and (iii) direct depression of smooth muscle contraction secondary to changes in the phosphorylation of myosin light chain kinase, phospholamban, and heat shock proteins (23). By exploiting the dependence of EDHF-type responses on gap junctional communication it has proved possible to dissociate the contribution of endothelial and smooth muscle cAMP synthesis to relaxation. The findings provide the first evidence that the endothelium can modulate contraction by elevating smooth muscle cAMP levels through a prostanoid-independent pathway that involves direct cell– cell communication in the vascular wall.
Intact ear preparations. Ears from NZW rabbits were cannulated via the central artery and perfused at 2 ml/min with oxygenated (95% O 2–5% CO 2) Holmans buffer containing L-NAME (300 M) and indomethacin (10 M) at 35°C. ACh was administered 20 min after administration of phenylephrine when perfusion pressure and cAMP efflux were stable. Cyclic AMP concentrations in the effluent were determined by collecting aliquots over 5 s which were stored at ⫺70°C before radioimmunoassay. Endothelial denudation was achieved by including 0.1% Triton X-100 in the buffer for 30 s (24). Statistical analysis. All data are given as mean ⫾ sem, where n denotes the number of animals studied for each data point. Concentration-response curves were assessed by one-way analysis of variance (Anova) followed by the Bonferroni multiple comparisons test. Maximal responses were compared by the Student’s t test for unpaired data. P ⬍ 0.05 was considered as significant.
RESULTS AND DISCUSSION MATERIALS AND METHODS Isolated ring preparations. Male NZW rabbits (2–2.5 kg) were sacrificed by injection of sodium pentobarbitone (120 mg/kg; i.v.) and the iliac artery (IA) and superior mesenteric artery (SMA) removed and transferred to Holmans buffer of the following composition (mM): 120 NaCl, 5 KCl, 2.5 CaCl 2, 1.3 NaH 2PO 4, 25 NaHCO 3, 11 glucose, and 10 sucrose. The vessels were stripped of adherent tissue, and rings 2–3 mm wide cut and suspended in organ baths containing gassed (95% O 2, 5% CO 2, pH 7.4) buffer at 37°C. IA and SMA rings were placed under 0.25 g and 0.6 g tension, respectively, and during an equilibration period of 1 h the tissues were repeatedly washed with fresh buffer and tension readjusted following stress relaxation. Endothelium-denuded rings were prepared by gentle abrasion of the intimal surface, and successful denudation subsequently confirmed by lack of response to ACh. The tissues were equilibrated for 40 – 60 min in the presence of N G-nitro-L-arginine methyl ester (L-NAME, 300 M) and indomethacin (10 M) before administration of phenylephrine (IA 1 M, SMA 10 M), which induced constrictor responses that stabilized within 3 min. Following constriction, concentration-response curves to either ACh, 5,6-epoxyeicosatrienoic acid (5,6-EET) or cyclopiazonic acid (CPA) were constructed. Concentration-response curves were also produced in the combined presence of L-NAME (300 M), indomethacin (10 M), and either 18␣-glycyrrhetinic acid (18␣-GA, 100 M), Gap 27 peptide (Gap 27, amino acid sequence SRPTEKTIFII), 2⬘,5⬘dideoxyadenosine (2⬘,5⬘-DDA, 50 M), 9-(tetrahydro-2-furyl) adenine (SQ22536, 100 M), Rp-8-bromoadenosine-3⬘,5⬘-cyclic monophosphorothioate (Rp-8BR-cAMPS, 50 M) or glibenclamide (10 M). In a further series of experiments with IA rings the effects of incubation with 3-isobutyl-1-methylxanthine (IBMX, 20 M) on responses to ACh were examined in the presence and absence of apamin (300 nM) plus charybdotoxin (CTX, 100 nM), which respectively block small (SK Ca) and large (BK Ca) K Ca channels. In these experiments with IBMX the concentration of phenylephrine used to induce tone was increased to 3 M to return initial tension to control levels. Analogous studies could not be performed in the SMA as tone became oscillatory in the presence of IBMX, thus precluding the construction of concentration-response curves. Gap 27 peptide (purity ⬎95%) was synthesized by Sigma Genosys, UK. With the exception of 5,6-EET (Affiniti, UK), all other agents were obtained from Sigma, UK. Radioimmunoassay. The protocol followed 40 min incubation in oxygenated Holmans buffer at 37°C containing L-NAME (300 M) plus indomethacin (10 M) and in some experiments 2⬘,5⬘-DDA (50 M), 18␣-GA (100 M) or Gap 27 (300 M). Phenylephrine (1 M) was added 3 min before the initial control point. Rings were frozen in liquid N 2 following addition of ACh and stored at ⫺70°C. cAMP and cGMP were subsequently extracted in 6% trichloroacetic acid and neutralized with water saturated diethyl ether followed by radioimmunoassay (Amersham UK Ltd.).
EDHF-type relaxations evoked by ACh in rings of rabbit IA and SMA artery were maximal at 3 M (R max 35.0 ⫾ 2.6% and 58.3 ⫾ 5.5%, EC 50 650 ⫾ 167 nM and 720 ⫾ 178 nM for IA, and SMA, respectively), and were attenuated by blockade of cAMP synthesis with the P-site adenylyl cyclase inhibitors (25) 2⬘,5⬘-DDA (50 M, Fig. 1) and SQ22536 (100 M, data not shown). In similarly incubated rings, radioimmunoassay showed that 3 M ACh induced a transient elevation of cAMP levels that was abolished by 2⬘,5⬘-DDA (50 M) and following endothelial denudation, whereas there was no associated increase in cGMP levels, confirming NO synthase inhibition (Fig. 2). Both the relaxation and the rise in cAMP evoked by ACh were inhibited by 18␣-GA (100 M), a steroidal aglycone that disrupts gap junction plaques, and by Gap 27 (300 M), a synthetic peptide inhibitor of gap junctional communication possessing sequence homology with the Gap 27 domain of the 2nd extracellular loop of connexin 43 (8, 26). Such agents inhibit EDHF-type responses through a receptor-independent mechanism that does not involve non-specific effects on smooth muscle contraction, relaxation or hyperpolarization (6, 8, 10). Furthermore, they are without effect on ACh-induced endothelial hyperpolarization (3) and NO release (6, 8, 9), precluding direct effects on endothelial synthetic pathways. In endothelium-denuded IA rings, the adenylyl cyclase activator forskolin reversed phenylephrine-induced constriction with equal potency in the presence or absence of these inhibitors (EC 50: 55.0 ⫾ 7.8 nM control, 54.1 ⫾ 19.9 nM 18␣-GA, 51.4 ⫾ 8.5 nM Gap 27, n ⫽ 4–7). This confirms that the 18␣-GA- and Gap 27-induced abolition of the ACh-evoked increase in cAMP levels observed reflects loss of intercellular communication via gap junctions, rather than inhibition of adenylyl cyclase per se. Any rise in cAMP within the relatively small endothelial monolayer induced by ACh thus appears to contribute negligibly to the overall cAMP content measured in intact rings in the presence of these gap junction inhibitors. Since ⬃1.5 fold elevations in cAMP levels, as evoked by ACh in
584
Vol. 283, No. 3, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Effects of pharmacological interventions on EDHF-type relaxations evoked by acetylcholine (ACh) in isolated rings of rabbit iliac (IA) and superior mesenteric (SMA) artery constricted by phenylephrine. (a) Representative IA traces. Control relaxations were amplified by the cAMP phosphodiesterase inhibitor IBMX (20 M) and these enhanced responses were almost abolished by the combination of apamin (300 nM) plus charybdotoxin (CTX, 100 nM). Control relaxations were attenuated by the adenylyl cyclase inhibitor 2⬘,5⬘-DDA (50 M), the protein kinase A inhibitor Rp-8Br-cAMPS (50 M), and the gap junction inhibitor Gap 27 (300 M). Ordinate represents isometric force development above baseline. (b) Cumulative concentration-relaxation curves to acetylcholine in the IA and SMA were depressed by ⬎60% in the presence of 2⬘,5⬘-DDA, Rp-8Br-cAMPS, Gap 27 and the additional gap junction inhibitor 18␣-GA (100 M). Relaxations of the IA were amplified by IBMX, with an associated leftward shift in the concentration-response curve, but suppressed in the combined presence of IBMX, apamin and CTX. The K ATP channel blocker glibenclamide (10 M) was without effect on control relaxations. Following endothelial denudation relaxations were abolished. n ⫽ 4–9 for each data point obtained in the presence of inhibitors; controls pooled.
endothelium-intact rings in the presence of L-NAME and indomethacin (Fig. 2), are sufficient to elicit nearmaximal biological responses (27), the findings indicate that elevations of cAMP in vascular smooth muscle underpin the EDHF phenomenon in rabbit conduit arteries. Cyclic AMP levels peaked ⬃15 s after exposure to ACh and declined to control by ⬃2 min, whereas the plateau of mechanical relaxation was attained some 40 –120 s after each successive dose (Figs. 1 and 2; see also [9]). This is consistent with the dependence of cAMP functional effects largely on subsequent phosphorylation events (27), as supported by the attenuation of ACh-evoked relaxation of IA and SMA by the protein kinase A (PKA) inhibitor Rp-8Br-cAMPS (50 M) (Fig. 1). Evidence that the elevation of cAMP levels reflects increased adenylyl cyclase activity,
rather than decreased nucleotide hydrolysis, is provided by observations that the cAMP phosphodiesterase inhibitor IBMX (20 M) amplified rather than attenuated EDHF-type responses (Fig. 1). The ability of the K ⫹ channel blockers apamin and charybdotoxin to suppress relaxation in combination has become the hallmark of the EDHF phenomenon (28), and in the iliac artery their coadministration effectively abolished both control relaxations (data not shown) and the enhanced responses observed in the presence of IBMX (Fig. 1). Consistent with previous findings in rabbit arteries (29), the lack of effect of glibenclamide (10 M) excludes the participation of cAMP/PKA-mediated activation of K ATP channels in the EDHF response to ACh (Fig. 1). The key mechanism that sustains agonist-induced endothelial hyperpolarization and NO-mediated and
585
Vol. 283, No. 3, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Cyclic nucleotide levels in IA rings. In endotheliumintact preparations incubated with L-NAME and indomethacin, acetylcholine (3 M) stimulated a transient peak in cAMP accumulation at ⬃15 s that was abolished by 2⬘,5⬘-DDA (50 M), 18␣-GA (100 M), Gap 27 (300 M), and endothelial denudation. In endothelium-intact preparations basal cAMP levels were not significantly affected by the combination of L-NAME and indomethacin, whereas basal cGMP levels were significantly reduced. Inhibition of NO synthase was confirmed by the lack of effect of acetylcholine on cGMP content. (n ⫽ 4–9, *P ⬍ 0.05 vs control).
EDHF-type relaxations is thought to be capacitative Ca 2⫹ influx into the endothelial cell, whereby depletion of Ca 2⫹ stores in the endoplasmic reticulum results in K Ca channel activation, hyperpolarization, and enhanced electrochemical Ca 2⫹ entry (30, 31). The role of these events in cAMP-mediated relaxation of intact SMA rings was evaluated using CPA (which depletes the endoplasmic reticulum Ca 2⫹ store by blocking its Ca 2⫹-ATPase pump) and 5,6-EET (which activates store-operated Ca 2⫹ channels directly without the necessity for store depletion) (6, 8, 30, 32). These agents were inactive in endothelium-denuded preparations, in contrast to intact preparations where both evoked concentration-dependent EDHF-type relaxations (R max 33.5 ⫾ 2.9% and 52.0 ⫾ 4.5%, EC 50 850 ⫾ 90 nM and 460 ⫾ 80 nM for CPA and 5,6-EET, respectively), which like those evoked by ACh, were attenuated by inhibition of adenylyl cyclase, PKA or gap junctional communication (Fig. 3). This observation that 5,6-EET stimulates cAMP-mediated relaxations of rabbit SMA by acting on endothelium, and not on smooth muscle, is consistent with electrophysiological measurements in porcine arteries with the 11,12-EET regioisomer (12), and does not support the idea that EDHF is an EET in rabbit arteries. It may, however, reflect the role of arachidonic metabolism in the maintenance of capacitative Ca 2⫹ entry (32–34) in addition to direct stimulatory effects of EETs on endothelial K Ca channels (35). Indeed, EDHF-type relaxations of rabbit SMA are dependent on the mobilization of arachidonic acid in the
endothelium by a Ca 2⫹ sensitive phospholipase (9, 36), and depletion of the endoplasmic reticulum Ca 2⫹ store mobilizes arachidonic acid by activating this same enzyme (34). We next investigated the contribution of endothelial cAMP synthesis to the EDHF phenomenon by correlating constrictor tone with nucleotide efflux from isolated, perfused rabbit ear preparations where NO synthase and cyclooxygenase were pharmacologically inhibited. In ears with intact endothelium, preconstricted with 1 M phenylephrine, ACh caused dilatation and induced sustained ⬃2-fold elevations in cAMP efflux; both were abolished by endothelial denudation but inhibition of gap junctional communication with 18␣-GA (100 M) blocked dilatation without diminishing the magnitude of the ACh-induced increment in cAMP efflux (Fig. 4a, right panel). This confirms that ACh-induced cAMP efflux originates from the endothelium rather than smooth muscle (24) and that 18␣-GA does not directly impair the activation of endothelial adenylyl cyclase. Constriction of rabbit ear preparations by phenylephrine itself induced a biphasic cAMP response, with an initial transient efflux from smooth muscle being followed by sustained release (also ⬃2fold above baseline) that was abolished by endothelial denudation and by 18␣-GA (Fig. 4a, left panel). Since endothelial cells are devoid of ␣ 1 receptors and not directly affected by phenylephrine (37), this implies that the formation of cAMP within the endothelium can be stimulated by a signal transmitted from activated smooth muscle via gap junctions (an analogous mechanism promotes Ca 2⫹-dependent stimulation of NO synthase [37]). Importantly, however, neither the time course nor the magnitude of the pressor response to phenylephrine were affected by endothelial denudation or 18␣-GA (Fig. 4b, left panel), indicating that the endothelium-dependent component of the cAMP response to phenylephrine does not modulate constriction in the rabbit ear through negative feedback, despite its equivalence in magnitude with the sustained increment in cAMP efflux induced by ACh. Cyclic AMP generated within the endothelium could in theory contribute to relaxation by diffusing to smooth muscle via gap junctions and/or promoting a conducted endothelial hyperpolarization (7). Indeed, we have previously shown that 18␣-GA causes a small rightward shift in concentration-relaxation curves to forskolin in endothelium-intact rings of rabbit jugular vein (7). In analogous experiments with endothelium-intact IA and ear preparations, however, forskolin was equally potent in the presence and absence of 18␣-GA (EC 50 for IA 77⫾ 24 nM control, 115 ⫾ 33 nM 18␣-GA, n ⫽ 12–18; EC 50 for rabbit ear 67⫾ 11 nM control, 85 ⫾ 16 nM 18␣-GA, n ⫽ 4). That endothelial and endothelium-dependent smooth muscle cAMP synthesis evoked by ACh can be dissociated is exemplified also by the transience of the elevation of smooth muscle
586
Vol. 283, No. 3, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. EDHF-type relaxations to cyclopiazonic acid (CPA) and 5,6-EET in SMA rings constricted by 10 M phenylephrine in the presence of L-NAME and indomethacin. (a) Representative traces demonstrating that responses to CPA and 5,6-EET were attenuated by 2⬘,5⬘-DDA (50 M) and 18␣-GA (100 M). (b) Cumulative concentration-relaxation curves to CPA and 5,6-EET were depressed by ⬎60% in the presence of 2⬘, 5⬘-DDA, 18␣-GA and Gap 27 (300 M). In endothelium-denuded rings relaxations to CPA were ⬍10% of developed force, whereas 5,6-EET was inactive.
cAMP in rabbit IA rings (Fig. 2), which contrasts with the sustained efflux of endothelial cAMP observed in intact ears (Fig. 4) and reported also in perfused rat mesentery (24). In conclusion, we have provided evidence that elevations in smooth muscle cAMP levels and associated PKA-mediated phosphorylations underpin the EDHF phenomenon in rabbit arteries. The underlying cellular events may be initiated by capacitative Ca 2⫹ influx into endothelium and are propagated by direct cell– cell communication through gap junctions. Whether the signal activating smooth muscle adenylyl cyclase is carried by a specific and as yet unidentified endogenous agent, or by electrotonic spread of the associated endothelial hyperpolarization remains unknown. There is no evidence that membrane hyperpolarization can
activate adenylyl cyclase directly (38, 39), however, hyperpolarization-induced reductions in intracellular [Ca 2⫹] (18) might disinhibit Type V and VI adenylyl cyclase isoforms to promote cAMP synthesis in smooth muscle (40). In any event, elevations of cAMP could sustain regenerative hyperpolarization not only by activating K ⫹ channels (19) and the Na ⫹-K ⫹ ATPase (20), but also by promoting cAMP-dependent increases in gap junction permeability/conductance that enhance the diffusion of signaling molecules and electrotonic spread of current (21, 22). It remains to be determined if cross-inhibition in the closely-linked pathways that regulate intracellular cGMP and cAMP homeostasis contribute to the inverse relationship between the amplitude of NO-mediated and EDHF-type relaxations previously reported in rabbit arteries (9).
587
Vol. 283, No. 3, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
REFERENCES
FIG. 4. Effects of phenylephrine and acetylcholine on endothelial cAMP efflux and perfusion pressure in isolated rabbit ears. (a) Left: Under control conditions phenylephrine (1 M) induced a sustained twofold increase in cAMP release into the effluent (from 2008 ⫾ 100 fmol/min [normalized to 100%] to 4240 ⫾ 520 fmol/ min). Following endothelial denudation increments in cAMP efflux declined to baseline before maximal constriction was attained. An equivalent transient cAMP efflux was observed in the presence of 18␣-GA (100 M). Right: In endothelium-intact preparations stably constricted by phenylephrine (breaks in abscissa reflect data from separate experiments), acetylcholine (3 M) evoked similar increments in cAMP efflux in the presence and absence of 18␣-GA (with compensation being made for the initial loss of phenylephrine-evoked cAMP production), confirming that 18␣-GA does not inhibit adenylyl cyclase directly. No cAMP response to acetylcholine was evident in endothelium-denuded preparations. (b) Left: Time course and magnitude of the pressor response to phenylephrine was unaffected by 18␣-GA or endothelial denudation. Right: Subsequent depressor effects of acetylcholine were abolished by 18␣-GA and endothelial denudation.
ACKNOWLEDGMENTS The work was supported by the Medical Research Council. The authors thank Professor A.H. Henderson for helpful comments during preparation of the manuscript.
1. Edwards, G., Dora, K. A., Gardener, M. J., Garland, C. J., and Weston, A. H. (1998) K ⫹ is an endothelium-derived hyperpolarizing factor in rat arteries. Nature 96, 269 –272. 2. Zygmunt, P. M., and Hogestatt, E. D. (1996) Role of potassium channels in endothelium-dependent relaxation resistant to nitroarginine in the rat hepatic artery Br. J. Pharmacol. 117, 1600 – 1606. 3. Yamamoto, Y., Fukuta, H., Nakahira, Y., and Suzuki, H. (1998) Blockage by 18-glycyrrhetinic acid of intercellular electrical coupling in guinea-pig arterioles. J. Physiol. 511, 501–508. 4. Doughty, J. M., Plane, F., and Langton, P. D. (1999) Charybdotoxin and apamin block EDHF in rat mesenteric artery if selectively applied to the endothelium. Am. J. Physiol. 276, H1107– H1112. 5. Coleman, H. A., Tare, M., and Parkington, H. C. (2001) K ⫹ currents underlying the action of endothelium-derived hyperpolarizing factor in guinea-pig, rat and human blood vessels. J. Physiol. 531, 359 –373. 6. Chaytor, A. T., Evans, W. H., and Griffith, T. M. (1998) Central role of heterocellular gap junctional communication in endothelium-dependent relaxations of rabbit arteries. J. Physiol. 508, 561–573. 7. Griffith, T. M., and Taylor, H. J. (1999) Cyclic AMP mediates EDHF-type relaxations of rabbit jugular vein. Biochem. Biophys. Res. Commun. 263, 52–57. 8. Taylor, H. J., Chaytor, A. T., Evans, W. H., and Griffith, T. M. (1998) Inhibition of the gap junctional component of endothelium-dependent relaxations in rabbit iliac artery by 18␣glycyrrhetinic acid. Br. J. Pharmacol. 125, 1– 4. 9. Hutcheson, I. R., Chaytor, A. T., Evans, W. H., and Griffith, T. M. (1999) Nitric oxide-independent relaxations to acetylcholine and A23187 involve different routes of heterocellular communication: Role of gap junctions and phospholipase A 2. Circ. Res. 84, 53– 63. 10. Dora, K. A., Martin, P. E. M., Chaytor, A. T., Evans, W. H., Garland, C. J., and Griffith, T. M. (1999) Role of heterocellular gap junctional communication in endothelium-dependent smooth muscle hyperpolarization: Inhibition by a connexin mimetic peptide. Biochem. Biophys. Res. Commun. 254, 27–31. 11. Yamamoto, Y., Imaeda, K., and Suzuki, H. (1999) Endotheliumdependent hyperpolarization and intercellular electrical coupling in guinea-pig mesenteric arterioles. J. Physiol. 514, 505– 513. 12. Edwards, G., Thollon, C., Gardener, F. M., Vilaine, J-P., Vanhoutte, P. M., and Weston, A. H. (2000) Role of gap junctions and EETs in endothelium-dependent hyperpolarization of porcine coronary artery. Br. J. Pharmacol. 129, 1145–1154. 13. Chaytor, A. T., Martin, P. E. M., Edwards, D. H., and Griffith, T. M. (2001) Gap junctional communication underpins EDHFtype relaxations evoked by acetylcholine in the rat hepatic artery. Am. J. Physiol. 280, in press. 14. Doughty, J. M., Boyle, J. P., and Langton, P. D. (2000) Potassium does not mimic EDHF in rat mesenteric arteries. Br. J. Pharmacol. 130, 1174 –1182. 15. Fisslthaler, B., Popp, R., Kiss, L., Potente, M., Harder, D. R., Fleming, I., and Busse, R. (1999) Cytochrome P450 2C is an EDHF synthase in coronary arteries. Nature 401, 493– 497. 16. Campbell, W. B., and Harder, D. R. (1999) Endothelium-derived hyperpolarizing factors and vascular cytochrome P450 metabolites of arachidonic acid in the regulation of tone. Circ. Res. 84, 484 – 488. 17. Doughty, J. M., Boyle, J. P., and Langton, P. D. (2001) Blockade of chloride channels reveals relaxations of rat small mesenteric arteries to raised potassium. Br. J. Pharmacol. 132, 293–301.
588
Vol. 283, No. 3, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
18. Bolz, S. S., de Wit, C., and Pohl, U. (1999) Endothelium-derived hyperpolarizing factor but not NO reduces smooth muscle Ca 2⫹ during acetylcholine-induced dilation of microvessels. Br. J. Pharmacol. 128, 124 –134. 19. Jaggar, J. H., Wellman, G. C., Heppner, T. J., Porter, V. A., Perez, G. J., Gollasch, M., Kleppisch, T., Rubort, M., Stevenson, A. S., Lederer, W. J., Knot, H. J., Bonev, A. D., and Nelson, M. T. (1998) Ca 2⫹ channels, ryanodine receptors and Ca 2⫹-activated K ⫹ channels: A functional unit for regulating arterial tone. Acta Physiol. Scand. 164, 577–587. 20. Therien, A. G., and Blostein, R. (2000) Mechanisms of sodium pump regulation. Am. J. Physiol. 279, C541–C566. 21. Abudara, V., Eyzaguirre, C., and Saez, J. C. (2000) Short- and long-term regulation of rat carotid body gap junctions by cAMP. Identification of connexin43, a gap junction subunit. Adv. Exp. Med. Biol. 475, 359 –369. 22. Gladwell, S. J., and Jefferys, J. G. (2001) Second messenger modulation of electrotonic coupling between region CA3 pyramidal cell axons in the rat hippocampus. Neurosci. Lett. 300, 1– 4. 23. Woodrum, D. A., Brophy, C. M., Wingard, C. J., Beall, A., and Rasmussen, H. (1999) Phosphorylation events associated with cyclic nucleotide-dependent inhibition of smooth muscle contraction. Am. J. Physiol. 277, H931–H939. 24. Abiru, T., Watanabe, Y., Kamata, K. & Kasuya, Y. (1993) Simultaneous measurement of vasodilation and changes in cyclic nucleotides in the perfused mesenteric arterial bed of the rat. Eur. J. Pharmacol. 242, 15–22. 25. Johnson, R. A., Desaubry, L., Bianchi, G., Shoshani, I., Lyons,E., Taussig, R., Watson, P. A., Cali, J. J., Krupinski, J., Pieroni, J. P., and Iyengar, R. (1997) Isozyme-dependent sensitivity of adenylyl cyclases to P-site-mediated inhibition by adenine nucleosides and nucleoside 3⬘-polyphosphates. J. Biol. Chem. 272, 8962– 8966. 26. Chaytor, A. T., Evans, W. H., and Griffith, T. M. (1997) Peptides homologous to extracellular loop motifs of connexin 43 reversibly abolish rhythmic contractile activity in rabbit arteries. J. Physiol. 503, 99 –110. 27. Francis, S. H., and Corbin, J. D. (1994) Structure and function of cyclic nucleotide-dependent protein kinases. Ann. Rev. Physiol. 56, 237–272. 28. Edwards, G., and Weston, A. H. (2001) EDHF-are there gaps in the pathway? J. Physiol. 531, 2.
29. Murphy, M. E., and Brayden, J. E. (1995) Apamin-sensitive K ⫹ channels mediate an endothelium-dependent hyperpolarization in rabbit mesenteric arteries. J. Physiol. 489, 723–734. 30. Fukao, M., Hattori, Y., Kanno, M., Sakuma, I., and Kitabatake, A. (1997) Sources of Ca 2⫹ in relation to generation of acetylcholine-induced endothelium-dependent hyperpolarization in rat mesenteric artery. Br. J. Pharmacol. 120, 1328 –1334. 31. Davis, M. J., and Sharma, N. R. (1997) Calcium-releaseactivated calcium influx in endothelium. J. Vasc. Res. 34, 186 – 195. 32. Rzigalinski, B. A., Willoughby, K. A., Hoffman, S. W., Falck, J. R., and Ellis, E. F. (1999) Calcium influx factor, further evidence it is 5,6-epoxyeicosatrienoic acid. J. Biol. Chem. 274, 175–182. 33. Hoebel, B. G., Kostner, G. M., and Graier, W. F. (1997) Activation of microsomal cytochrome P450 mono-oxygenase by Ca 2⫹ store depletion and its contribution to Ca 2⫹ entry in porcine aortic endothelial cells. Br. J. Pharmacol. 121, 1579 –1588. 34. Gailly, P. (1998) Ca 2⫹ entry in CHO cells, after Ca 2⫹ stores depletion, is mediated by arachidonic acid. Cell Calcium. 24, 293–304. 35. Baron, A., Frieden, M., and Beny, J. L. (1997) Epoxyeicosatrienoic acids activate a high-conductance, Ca 2⫹-dependent K ⫹ channel on pig coronary artery endothelial cells. J. Physiol. 504, 537–543. 36. Hutcheson, I. R., and Griffith, T. M. (2000) Role of phospholipase A 2 and myoendothelial gap junctions in melittin-induced arterial relaxation. Eur. J. Pharmacol. 406, 239 –245. 37. Dora, K. A., Doyle, M. P., and Duling, B. R. (1997) Elevation of intracellular calcium in smooth muscle causes endothelial cell generation of NO in arterioles. Proc. Natl. Acad. Sci. USA 94, 6529 – 6534. 38. Cooper, D. M., Schell, M. J., Thorn, P., and Irvine, R. F. (1998) Regulation of adenylyl cyclase by membrane potential. J. Biol. Chem. 273, 27703–27707. 39. Reddy, R., Smith, D., Wayman, G., Wu, Z., Villacres, E. C., and Storm, D. R. (1995) Voltage-sensitive adenylyl cyclase activity in cultured neurons. J. Biol. Chem. 270, 14340 –14346. 40. Murthy, K. S., and Makhlouf, G. M. (1998) Regulation of adenylyl cyclase type V/VI in smooth muscle: Interplay of inhibitory G protein and Ca 2⫹ influx. Mol. Pharmacol. 54, 122–128.
589
Gap Junction-Dependent Increases in Smooth Muscle cAMP Underpin the EDHF Phenomenon in Rabbit Arteries Hannah J. Taylor, Andrew T. Chaytor, David H. Edwards, and Tudor M. Griffith 1 Department of Diagnostic Radiology, Wales Heart Research Institute, University of Wales College of Medicine, Cardiff CF14 4XN, United Kingdom
Received March 28, 2001
We have investigated the role of cAMP in nitric oxide (NO)- and prostanoid-independent vascular relaxations evoked by acetylcholine (ACh) in isolated arteries and perfused ear preparations from the rabbit. These EDHF-type responses are shown to be associated with elevated cAMP levels specifically in smooth muscle and are attenuated by blocking adenylyl cyclase or protein kinase A (PKA). Relaxations are amplified by 3-isobutyl-1-methylxanthine, which prevents cAMP hydrolysis, while remaining susceptible to inhibition by the combination of two K Ca channel blockers, apamin and charybdotoxin. Analogous endothelium- and cAMP-dependent relaxations were evoked by cyclopiazonic acid (CPA) which stimulates Ca 2ⴙ influx via channels linked to the depletion of Ca 2ⴙ stores. Responses to ACh and CPA were both inhibited by interrupting cell-to-cell coupling via gap junctions with 18␣-glycyrrhetinic acid and a connexin-specific Gap 27 peptide. The findings suggest that EDHF-type responses are initiated by capacitative Ca 2ⴙ influx into the endothelium and propagated by direct intercellular communication to effect relaxation via cAMP/PKAdependent phosphorylation events in smooth muscle. © 2001 Academic Press
Key Words: endothelium-derived hyperpolarizing factor (EDHF); gap junctions; cyclic AMP; epoxyeicosatrienoic acids; cyclopiazonic acid.
Nitric oxide (NO)- and prostanoid- independent vascular relaxations evoked by agonists such as acetylcholine (ACh) are closely associated with endothelial hyperpolarization mediated by the opening of Ca 2⫹activated K ⫹ channels (K Ca) (1–5). Recent evidence indicates that the mechanisms subsequently leading to relaxation involve electrotonic spread of endothelial hyperpolarization into the smooth muscle and/or the To whom correspondence should be addressed. Fax: ⫹44 2920744726. E-mail: [email protected]. 1
diffusion of an unidentified chemical mediator via direct intercellular communication pathways. Specific synthetic peptides that interrupt cell– cell coupling via gap junctions present in the vascular wall have thus been shown to inhibit ACh-evoked NO- and prostanoidindependent smooth muscle relaxations and hyperpolarizations in rabbit, rat, porcine, and guinea pig vessels (6 –14). Alternatively, it has been suggested that an endothelium-derived hyperpolarizing factor (EDHF) transfers from the endothelium to subjacent smooth muscle across the extracellular space, putative candidates for which include cytochrome P 450 epoxygenase metabolites of arachidonic acid (epoxyeicosatrienoic acids [EETs]) which activate smooth muscle K Ca channels, and K ⫹ ions which activate inwardly rectifying K ⫹ channels (K ir) and the membrane Na ⫹-K ⫹ ATPase (1, 15, 16). EETs and elevations in extracellular [K ⫹] however do not universally induce direct smooth muscle relaxation, and in some vessels may evoke endothelium-dependent responses that are susceptible to blockade of gap junctions (9, 17). While NOand prostanoid-independent relaxations have been attributed to reduced Ca 2⫹ influx via voltage-gated Ca 2⫹ channels following smooth muscle hyperpolarization (18), relaxation may not always be wholly dependent on changes in membrane potential. In the rat mesenteric artery, for example, blockade of smooth muscle hyperpolarization by simultaneous inhibition of K ir channels (by Ba 2⫹) and the Na ⫹-K ⫹ ATPase (by ouabain) causes a rightward shift in the concentrationrelaxation curve to ACh but does not alter the maximum response (1). In the present study we have used rabbit artery and vascular bed preparations to test the hypothesis that NO- and prostanoid-independent relaxations involve synthesis of cAMP, whose multiple actions encompass several characteristics of the “EDHF phenomenon” noted above. These include (i) hyperpolarization via the stimulation of K Ca channels (19) and the Na ⫹-K ⫹
583
0006-291X/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.
Vol. 283, No. 3, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
ATPase (20), (ii) enhanced electrical and chemical signaling via gap junctions (21, 22), and (iii) direct depression of smooth muscle contraction secondary to changes in the phosphorylation of myosin light chain kinase, phospholamban, and heat shock proteins (23). By exploiting the dependence of EDHF-type responses on gap junctional communication it has proved possible to dissociate the contribution of endothelial and smooth muscle cAMP synthesis to relaxation. The findings provide the first evidence that the endothelium can modulate contraction by elevating smooth muscle cAMP levels through a prostanoid-independent pathway that involves direct cell– cell communication in the vascular wall.
Intact ear preparations. Ears from NZW rabbits were cannulated via the central artery and perfused at 2 ml/min with oxygenated (95% O 2–5% CO 2) Holmans buffer containing L-NAME (300 M) and indomethacin (10 M) at 35°C. ACh was administered 20 min after administration of phenylephrine when perfusion pressure and cAMP efflux were stable. Cyclic AMP concentrations in the effluent were determined by collecting aliquots over 5 s which were stored at ⫺70°C before radioimmunoassay. Endothelial denudation was achieved by including 0.1% Triton X-100 in the buffer for 30 s (24). Statistical analysis. All data are given as mean ⫾ sem, where n denotes the number of animals studied for each data point. Concentration-response curves were assessed by one-way analysis of variance (Anova) followed by the Bonferroni multiple comparisons test. Maximal responses were compared by the Student’s t test for unpaired data. P ⬍ 0.05 was considered as significant.
RESULTS AND DISCUSSION MATERIALS AND METHODS Isolated ring preparations. Male NZW rabbits (2–2.5 kg) were sacrificed by injection of sodium pentobarbitone (120 mg/kg; i.v.) and the iliac artery (IA) and superior mesenteric artery (SMA) removed and transferred to Holmans buffer of the following composition (mM): 120 NaCl, 5 KCl, 2.5 CaCl 2, 1.3 NaH 2PO 4, 25 NaHCO 3, 11 glucose, and 10 sucrose. The vessels were stripped of adherent tissue, and rings 2–3 mm wide cut and suspended in organ baths containing gassed (95% O 2, 5% CO 2, pH 7.4) buffer at 37°C. IA and SMA rings were placed under 0.25 g and 0.6 g tension, respectively, and during an equilibration period of 1 h the tissues were repeatedly washed with fresh buffer and tension readjusted following stress relaxation. Endothelium-denuded rings were prepared by gentle abrasion of the intimal surface, and successful denudation subsequently confirmed by lack of response to ACh. The tissues were equilibrated for 40 – 60 min in the presence of N G-nitro-L-arginine methyl ester (L-NAME, 300 M) and indomethacin (10 M) before administration of phenylephrine (IA 1 M, SMA 10 M), which induced constrictor responses that stabilized within 3 min. Following constriction, concentration-response curves to either ACh, 5,6-epoxyeicosatrienoic acid (5,6-EET) or cyclopiazonic acid (CPA) were constructed. Concentration-response curves were also produced in the combined presence of L-NAME (300 M), indomethacin (10 M), and either 18␣-glycyrrhetinic acid (18␣-GA, 100 M), Gap 27 peptide (Gap 27, amino acid sequence SRPTEKTIFII), 2⬘,5⬘dideoxyadenosine (2⬘,5⬘-DDA, 50 M), 9-(tetrahydro-2-furyl) adenine (SQ22536, 100 M), Rp-8-bromoadenosine-3⬘,5⬘-cyclic monophosphorothioate (Rp-8BR-cAMPS, 50 M) or glibenclamide (10 M). In a further series of experiments with IA rings the effects of incubation with 3-isobutyl-1-methylxanthine (IBMX, 20 M) on responses to ACh were examined in the presence and absence of apamin (300 nM) plus charybdotoxin (CTX, 100 nM), which respectively block small (SK Ca) and large (BK Ca) K Ca channels. In these experiments with IBMX the concentration of phenylephrine used to induce tone was increased to 3 M to return initial tension to control levels. Analogous studies could not be performed in the SMA as tone became oscillatory in the presence of IBMX, thus precluding the construction of concentration-response curves. Gap 27 peptide (purity ⬎95%) was synthesized by Sigma Genosys, UK. With the exception of 5,6-EET (Affiniti, UK), all other agents were obtained from Sigma, UK. Radioimmunoassay. The protocol followed 40 min incubation in oxygenated Holmans buffer at 37°C containing L-NAME (300 M) plus indomethacin (10 M) and in some experiments 2⬘,5⬘-DDA (50 M), 18␣-GA (100 M) or Gap 27 (300 M). Phenylephrine (1 M) was added 3 min before the initial control point. Rings were frozen in liquid N 2 following addition of ACh and stored at ⫺70°C. cAMP and cGMP were subsequently extracted in 6% trichloroacetic acid and neutralized with water saturated diethyl ether followed by radioimmunoassay (Amersham UK Ltd.).
EDHF-type relaxations evoked by ACh in rings of rabbit IA and SMA artery were maximal at 3 M (R max 35.0 ⫾ 2.6% and 58.3 ⫾ 5.5%, EC 50 650 ⫾ 167 nM and 720 ⫾ 178 nM for IA, and SMA, respectively), and were attenuated by blockade of cAMP synthesis with the P-site adenylyl cyclase inhibitors (25) 2⬘,5⬘-DDA (50 M, Fig. 1) and SQ22536 (100 M, data not shown). In similarly incubated rings, radioimmunoassay showed that 3 M ACh induced a transient elevation of cAMP levels that was abolished by 2⬘,5⬘-DDA (50 M) and following endothelial denudation, whereas there was no associated increase in cGMP levels, confirming NO synthase inhibition (Fig. 2). Both the relaxation and the rise in cAMP evoked by ACh were inhibited by 18␣-GA (100 M), a steroidal aglycone that disrupts gap junction plaques, and by Gap 27 (300 M), a synthetic peptide inhibitor of gap junctional communication possessing sequence homology with the Gap 27 domain of the 2nd extracellular loop of connexin 43 (8, 26). Such agents inhibit EDHF-type responses through a receptor-independent mechanism that does not involve non-specific effects on smooth muscle contraction, relaxation or hyperpolarization (6, 8, 10). Furthermore, they are without effect on ACh-induced endothelial hyperpolarization (3) and NO release (6, 8, 9), precluding direct effects on endothelial synthetic pathways. In endothelium-denuded IA rings, the adenylyl cyclase activator forskolin reversed phenylephrine-induced constriction with equal potency in the presence or absence of these inhibitors (EC 50: 55.0 ⫾ 7.8 nM control, 54.1 ⫾ 19.9 nM 18␣-GA, 51.4 ⫾ 8.5 nM Gap 27, n ⫽ 4–7). This confirms that the 18␣-GA- and Gap 27-induced abolition of the ACh-evoked increase in cAMP levels observed reflects loss of intercellular communication via gap junctions, rather than inhibition of adenylyl cyclase per se. Any rise in cAMP within the relatively small endothelial monolayer induced by ACh thus appears to contribute negligibly to the overall cAMP content measured in intact rings in the presence of these gap junction inhibitors. Since ⬃1.5 fold elevations in cAMP levels, as evoked by ACh in
584
Vol. 283, No. 3, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Effects of pharmacological interventions on EDHF-type relaxations evoked by acetylcholine (ACh) in isolated rings of rabbit iliac (IA) and superior mesenteric (SMA) artery constricted by phenylephrine. (a) Representative IA traces. Control relaxations were amplified by the cAMP phosphodiesterase inhibitor IBMX (20 M) and these enhanced responses were almost abolished by the combination of apamin (300 nM) plus charybdotoxin (CTX, 100 nM). Control relaxations were attenuated by the adenylyl cyclase inhibitor 2⬘,5⬘-DDA (50 M), the protein kinase A inhibitor Rp-8Br-cAMPS (50 M), and the gap junction inhibitor Gap 27 (300 M). Ordinate represents isometric force development above baseline. (b) Cumulative concentration-relaxation curves to acetylcholine in the IA and SMA were depressed by ⬎60% in the presence of 2⬘,5⬘-DDA, Rp-8Br-cAMPS, Gap 27 and the additional gap junction inhibitor 18␣-GA (100 M). Relaxations of the IA were amplified by IBMX, with an associated leftward shift in the concentration-response curve, but suppressed in the combined presence of IBMX, apamin and CTX. The K ATP channel blocker glibenclamide (10 M) was without effect on control relaxations. Following endothelial denudation relaxations were abolished. n ⫽ 4–9 for each data point obtained in the presence of inhibitors; controls pooled.
endothelium-intact rings in the presence of L-NAME and indomethacin (Fig. 2), are sufficient to elicit nearmaximal biological responses (27), the findings indicate that elevations of cAMP in vascular smooth muscle underpin the EDHF phenomenon in rabbit conduit arteries. Cyclic AMP levels peaked ⬃15 s after exposure to ACh and declined to control by ⬃2 min, whereas the plateau of mechanical relaxation was attained some 40 –120 s after each successive dose (Figs. 1 and 2; see also [9]). This is consistent with the dependence of cAMP functional effects largely on subsequent phosphorylation events (27), as supported by the attenuation of ACh-evoked relaxation of IA and SMA by the protein kinase A (PKA) inhibitor Rp-8Br-cAMPS (50 M) (Fig. 1). Evidence that the elevation of cAMP levels reflects increased adenylyl cyclase activity,
rather than decreased nucleotide hydrolysis, is provided by observations that the cAMP phosphodiesterase inhibitor IBMX (20 M) amplified rather than attenuated EDHF-type responses (Fig. 1). The ability of the K ⫹ channel blockers apamin and charybdotoxin to suppress relaxation in combination has become the hallmark of the EDHF phenomenon (28), and in the iliac artery their coadministration effectively abolished both control relaxations (data not shown) and the enhanced responses observed in the presence of IBMX (Fig. 1). Consistent with previous findings in rabbit arteries (29), the lack of effect of glibenclamide (10 M) excludes the participation of cAMP/PKA-mediated activation of K ATP channels in the EDHF response to ACh (Fig. 1). The key mechanism that sustains agonist-induced endothelial hyperpolarization and NO-mediated and
585
Vol. 283, No. 3, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Cyclic nucleotide levels in IA rings. In endotheliumintact preparations incubated with L-NAME and indomethacin, acetylcholine (3 M) stimulated a transient peak in cAMP accumulation at ⬃15 s that was abolished by 2⬘,5⬘-DDA (50 M), 18␣-GA (100 M), Gap 27 (300 M), and endothelial denudation. In endothelium-intact preparations basal cAMP levels were not significantly affected by the combination of L-NAME and indomethacin, whereas basal cGMP levels were significantly reduced. Inhibition of NO synthase was confirmed by the lack of effect of acetylcholine on cGMP content. (n ⫽ 4–9, *P ⬍ 0.05 vs control).
EDHF-type relaxations is thought to be capacitative Ca 2⫹ influx into the endothelial cell, whereby depletion of Ca 2⫹ stores in the endoplasmic reticulum results in K Ca channel activation, hyperpolarization, and enhanced electrochemical Ca 2⫹ entry (30, 31). The role of these events in cAMP-mediated relaxation of intact SMA rings was evaluated using CPA (which depletes the endoplasmic reticulum Ca 2⫹ store by blocking its Ca 2⫹-ATPase pump) and 5,6-EET (which activates store-operated Ca 2⫹ channels directly without the necessity for store depletion) (6, 8, 30, 32). These agents were inactive in endothelium-denuded preparations, in contrast to intact preparations where both evoked concentration-dependent EDHF-type relaxations (R max 33.5 ⫾ 2.9% and 52.0 ⫾ 4.5%, EC 50 850 ⫾ 90 nM and 460 ⫾ 80 nM for CPA and 5,6-EET, respectively), which like those evoked by ACh, were attenuated by inhibition of adenylyl cyclase, PKA or gap junctional communication (Fig. 3). This observation that 5,6-EET stimulates cAMP-mediated relaxations of rabbit SMA by acting on endothelium, and not on smooth muscle, is consistent with electrophysiological measurements in porcine arteries with the 11,12-EET regioisomer (12), and does not support the idea that EDHF is an EET in rabbit arteries. It may, however, reflect the role of arachidonic metabolism in the maintenance of capacitative Ca 2⫹ entry (32–34) in addition to direct stimulatory effects of EETs on endothelial K Ca channels (35). Indeed, EDHF-type relaxations of rabbit SMA are dependent on the mobilization of arachidonic acid in the
endothelium by a Ca 2⫹ sensitive phospholipase (9, 36), and depletion of the endoplasmic reticulum Ca 2⫹ store mobilizes arachidonic acid by activating this same enzyme (34). We next investigated the contribution of endothelial cAMP synthesis to the EDHF phenomenon by correlating constrictor tone with nucleotide efflux from isolated, perfused rabbit ear preparations where NO synthase and cyclooxygenase were pharmacologically inhibited. In ears with intact endothelium, preconstricted with 1 M phenylephrine, ACh caused dilatation and induced sustained ⬃2-fold elevations in cAMP efflux; both were abolished by endothelial denudation but inhibition of gap junctional communication with 18␣-GA (100 M) blocked dilatation without diminishing the magnitude of the ACh-induced increment in cAMP efflux (Fig. 4a, right panel). This confirms that ACh-induced cAMP efflux originates from the endothelium rather than smooth muscle (24) and that 18␣-GA does not directly impair the activation of endothelial adenylyl cyclase. Constriction of rabbit ear preparations by phenylephrine itself induced a biphasic cAMP response, with an initial transient efflux from smooth muscle being followed by sustained release (also ⬃2fold above baseline) that was abolished by endothelial denudation and by 18␣-GA (Fig. 4a, left panel). Since endothelial cells are devoid of ␣ 1 receptors and not directly affected by phenylephrine (37), this implies that the formation of cAMP within the endothelium can be stimulated by a signal transmitted from activated smooth muscle via gap junctions (an analogous mechanism promotes Ca 2⫹-dependent stimulation of NO synthase [37]). Importantly, however, neither the time course nor the magnitude of the pressor response to phenylephrine were affected by endothelial denudation or 18␣-GA (Fig. 4b, left panel), indicating that the endothelium-dependent component of the cAMP response to phenylephrine does not modulate constriction in the rabbit ear through negative feedback, despite its equivalence in magnitude with the sustained increment in cAMP efflux induced by ACh. Cyclic AMP generated within the endothelium could in theory contribute to relaxation by diffusing to smooth muscle via gap junctions and/or promoting a conducted endothelial hyperpolarization (7). Indeed, we have previously shown that 18␣-GA causes a small rightward shift in concentration-relaxation curves to forskolin in endothelium-intact rings of rabbit jugular vein (7). In analogous experiments with endothelium-intact IA and ear preparations, however, forskolin was equally potent in the presence and absence of 18␣-GA (EC 50 for IA 77⫾ 24 nM control, 115 ⫾ 33 nM 18␣-GA, n ⫽ 12–18; EC 50 for rabbit ear 67⫾ 11 nM control, 85 ⫾ 16 nM 18␣-GA, n ⫽ 4). That endothelial and endothelium-dependent smooth muscle cAMP synthesis evoked by ACh can be dissociated is exemplified also by the transience of the elevation of smooth muscle
586
Vol. 283, No. 3, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. EDHF-type relaxations to cyclopiazonic acid (CPA) and 5,6-EET in SMA rings constricted by 10 M phenylephrine in the presence of L-NAME and indomethacin. (a) Representative traces demonstrating that responses to CPA and 5,6-EET were attenuated by 2⬘,5⬘-DDA (50 M) and 18␣-GA (100 M). (b) Cumulative concentration-relaxation curves to CPA and 5,6-EET were depressed by ⬎60% in the presence of 2⬘, 5⬘-DDA, 18␣-GA and Gap 27 (300 M). In endothelium-denuded rings relaxations to CPA were ⬍10% of developed force, whereas 5,6-EET was inactive.
cAMP in rabbit IA rings (Fig. 2), which contrasts with the sustained efflux of endothelial cAMP observed in intact ears (Fig. 4) and reported also in perfused rat mesentery (24). In conclusion, we have provided evidence that elevations in smooth muscle cAMP levels and associated PKA-mediated phosphorylations underpin the EDHF phenomenon in rabbit arteries. The underlying cellular events may be initiated by capacitative Ca 2⫹ influx into endothelium and are propagated by direct cell– cell communication through gap junctions. Whether the signal activating smooth muscle adenylyl cyclase is carried by a specific and as yet unidentified endogenous agent, or by electrotonic spread of the associated endothelial hyperpolarization remains unknown. There is no evidence that membrane hyperpolarization can
activate adenylyl cyclase directly (38, 39), however, hyperpolarization-induced reductions in intracellular [Ca 2⫹] (18) might disinhibit Type V and VI adenylyl cyclase isoforms to promote cAMP synthesis in smooth muscle (40). In any event, elevations of cAMP could sustain regenerative hyperpolarization not only by activating K ⫹ channels (19) and the Na ⫹-K ⫹ ATPase (20), but also by promoting cAMP-dependent increases in gap junction permeability/conductance that enhance the diffusion of signaling molecules and electrotonic spread of current (21, 22). It remains to be determined if cross-inhibition in the closely-linked pathways that regulate intracellular cGMP and cAMP homeostasis contribute to the inverse relationship between the amplitude of NO-mediated and EDHF-type relaxations previously reported in rabbit arteries (9).
587
Vol. 283, No. 3, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
REFERENCES
FIG. 4. Effects of phenylephrine and acetylcholine on endothelial cAMP efflux and perfusion pressure in isolated rabbit ears. (a) Left: Under control conditions phenylephrine (1 M) induced a sustained twofold increase in cAMP release into the effluent (from 2008 ⫾ 100 fmol/min [normalized to 100%] to 4240 ⫾ 520 fmol/ min). Following endothelial denudation increments in cAMP efflux declined to baseline before maximal constriction was attained. An equivalent transient cAMP efflux was observed in the presence of 18␣-GA (100 M). Right: In endothelium-intact preparations stably constricted by phenylephrine (breaks in abscissa reflect data from separate experiments), acetylcholine (3 M) evoked similar increments in cAMP efflux in the presence and absence of 18␣-GA (with compensation being made for the initial loss of phenylephrine-evoked cAMP production), confirming that 18␣-GA does not inhibit adenylyl cyclase directly. No cAMP response to acetylcholine was evident in endothelium-denuded preparations. (b) Left: Time course and magnitude of the pressor response to phenylephrine was unaffected by 18␣-GA or endothelial denudation. Right: Subsequent depressor effects of acetylcholine were abolished by 18␣-GA and endothelial denudation.
ACKNOWLEDGMENTS The work was supported by the Medical Research Council. The authors thank Professor A.H. Henderson for helpful comments during preparation of the manuscript.
1. Edwards, G., Dora, K. A., Gardener, M. J., Garland, C. J., and Weston, A. H. (1998) K ⫹ is an endothelium-derived hyperpolarizing factor in rat arteries. Nature 96, 269 –272. 2. Zygmunt, P. M., and Hogestatt, E. D. (1996) Role of potassium channels in endothelium-dependent relaxation resistant to nitroarginine in the rat hepatic artery Br. J. Pharmacol. 117, 1600 – 1606. 3. Yamamoto, Y., Fukuta, H., Nakahira, Y., and Suzuki, H. (1998) Blockage by 18-glycyrrhetinic acid of intercellular electrical coupling in guinea-pig arterioles. J. Physiol. 511, 501–508. 4. Doughty, J. M., Plane, F., and Langton, P. D. (1999) Charybdotoxin and apamin block EDHF in rat mesenteric artery if selectively applied to the endothelium. Am. J. Physiol. 276, H1107– H1112. 5. Coleman, H. A., Tare, M., and Parkington, H. C. (2001) K ⫹ currents underlying the action of endothelium-derived hyperpolarizing factor in guinea-pig, rat and human blood vessels. J. Physiol. 531, 359 –373. 6. Chaytor, A. T., Evans, W. H., and Griffith, T. M. (1998) Central role of heterocellular gap junctional communication in endothelium-dependent relaxations of rabbit arteries. J. Physiol. 508, 561–573. 7. Griffith, T. M., and Taylor, H. J. (1999) Cyclic AMP mediates EDHF-type relaxations of rabbit jugular vein. Biochem. Biophys. Res. Commun. 263, 52–57. 8. Taylor, H. J., Chaytor, A. T., Evans, W. H., and Griffith, T. M. (1998) Inhibition of the gap junctional component of endothelium-dependent relaxations in rabbit iliac artery by 18␣glycyrrhetinic acid. Br. J. Pharmacol. 125, 1– 4. 9. Hutcheson, I. R., Chaytor, A. T., Evans, W. H., and Griffith, T. M. (1999) Nitric oxide-independent relaxations to acetylcholine and A23187 involve different routes of heterocellular communication: Role of gap junctions and phospholipase A 2. Circ. Res. 84, 53– 63. 10. Dora, K. A., Martin, P. E. M., Chaytor, A. T., Evans, W. H., Garland, C. J., and Griffith, T. M. (1999) Role of heterocellular gap junctional communication in endothelium-dependent smooth muscle hyperpolarization: Inhibition by a connexin mimetic peptide. Biochem. Biophys. Res. Commun. 254, 27–31. 11. Yamamoto, Y., Imaeda, K., and Suzuki, H. (1999) Endotheliumdependent hyperpolarization and intercellular electrical coupling in guinea-pig mesenteric arterioles. J. Physiol. 514, 505– 513. 12. Edwards, G., Thollon, C., Gardener, F. M., Vilaine, J-P., Vanhoutte, P. M., and Weston, A. H. (2000) Role of gap junctions and EETs in endothelium-dependent hyperpolarization of porcine coronary artery. Br. J. Pharmacol. 129, 1145–1154. 13. Chaytor, A. T., Martin, P. E. M., Edwards, D. H., and Griffith, T. M. (2001) Gap junctional communication underpins EDHFtype relaxations evoked by acetylcholine in the rat hepatic artery. Am. J. Physiol. 280, in press. 14. Doughty, J. M., Boyle, J. P., and Langton, P. D. (2000) Potassium does not mimic EDHF in rat mesenteric arteries. Br. J. Pharmacol. 130, 1174 –1182. 15. Fisslthaler, B., Popp, R., Kiss, L., Potente, M., Harder, D. R., Fleming, I., and Busse, R. (1999) Cytochrome P450 2C is an EDHF synthase in coronary arteries. Nature 401, 493– 497. 16. Campbell, W. B., and Harder, D. R. (1999) Endothelium-derived hyperpolarizing factors and vascular cytochrome P450 metabolites of arachidonic acid in the regulation of tone. Circ. Res. 84, 484 – 488. 17. Doughty, J. M., Boyle, J. P., and Langton, P. D. (2001) Blockade of chloride channels reveals relaxations of rat small mesenteric arteries to raised potassium. Br. J. Pharmacol. 132, 293–301.
588
Vol. 283, No. 3, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
18. Bolz, S. S., de Wit, C., and Pohl, U. (1999) Endothelium-derived hyperpolarizing factor but not NO reduces smooth muscle Ca 2⫹ during acetylcholine-induced dilation of microvessels. Br. J. Pharmacol. 128, 124 –134. 19. Jaggar, J. H., Wellman, G. C., Heppner, T. J., Porter, V. A., Perez, G. J., Gollasch, M., Kleppisch, T., Rubort, M., Stevenson, A. S., Lederer, W. J., Knot, H. J., Bonev, A. D., and Nelson, M. T. (1998) Ca 2⫹ channels, ryanodine receptors and Ca 2⫹-activated K ⫹ channels: A functional unit for regulating arterial tone. Acta Physiol. Scand. 164, 577–587. 20. Therien, A. G., and Blostein, R. (2000) Mechanisms of sodium pump regulation. Am. J. Physiol. 279, C541–C566. 21. Abudara, V., Eyzaguirre, C., and Saez, J. C. (2000) Short- and long-term regulation of rat carotid body gap junctions by cAMP. Identification of connexin43, a gap junction subunit. Adv. Exp. Med. Biol. 475, 359 –369. 22. Gladwell, S. J., and Jefferys, J. G. (2001) Second messenger modulation of electrotonic coupling between region CA3 pyramidal cell axons in the rat hippocampus. Neurosci. Lett. 300, 1– 4. 23. Woodrum, D. A., Brophy, C. M., Wingard, C. J., Beall, A., and Rasmussen, H. (1999) Phosphorylation events associated with cyclic nucleotide-dependent inhibition of smooth muscle contraction. Am. J. Physiol. 277, H931–H939. 24. Abiru, T., Watanabe, Y., Kamata, K. & Kasuya, Y. (1993) Simultaneous measurement of vasodilation and changes in cyclic nucleotides in the perfused mesenteric arterial bed of the rat. Eur. J. Pharmacol. 242, 15–22. 25. Johnson, R. A., Desaubry, L., Bianchi, G., Shoshani, I., Lyons,E., Taussig, R., Watson, P. A., Cali, J. J., Krupinski, J., Pieroni, J. P., and Iyengar, R. (1997) Isozyme-dependent sensitivity of adenylyl cyclases to P-site-mediated inhibition by adenine nucleosides and nucleoside 3⬘-polyphosphates. J. Biol. Chem. 272, 8962– 8966. 26. Chaytor, A. T., Evans, W. H., and Griffith, T. M. (1997) Peptides homologous to extracellular loop motifs of connexin 43 reversibly abolish rhythmic contractile activity in rabbit arteries. J. Physiol. 503, 99 –110. 27. Francis, S. H., and Corbin, J. D. (1994) Structure and function of cyclic nucleotide-dependent protein kinases. Ann. Rev. Physiol. 56, 237–272. 28. Edwards, G., and Weston, A. H. (2001) EDHF-are there gaps in the pathway? J. Physiol. 531, 2.
29. Murphy, M. E., and Brayden, J. E. (1995) Apamin-sensitive K ⫹ channels mediate an endothelium-dependent hyperpolarization in rabbit mesenteric arteries. J. Physiol. 489, 723–734. 30. Fukao, M., Hattori, Y., Kanno, M., Sakuma, I., and Kitabatake, A. (1997) Sources of Ca 2⫹ in relation to generation of acetylcholine-induced endothelium-dependent hyperpolarization in rat mesenteric artery. Br. J. Pharmacol. 120, 1328 –1334. 31. Davis, M. J., and Sharma, N. R. (1997) Calcium-releaseactivated calcium influx in endothelium. J. Vasc. Res. 34, 186 – 195. 32. Rzigalinski, B. A., Willoughby, K. A., Hoffman, S. W., Falck, J. R., and Ellis, E. F. (1999) Calcium influx factor, further evidence it is 5,6-epoxyeicosatrienoic acid. J. Biol. Chem. 274, 175–182. 33. Hoebel, B. G., Kostner, G. M., and Graier, W. F. (1997) Activation of microsomal cytochrome P450 mono-oxygenase by Ca 2⫹ store depletion and its contribution to Ca 2⫹ entry in porcine aortic endothelial cells. Br. J. Pharmacol. 121, 1579 –1588. 34. Gailly, P. (1998) Ca 2⫹ entry in CHO cells, after Ca 2⫹ stores depletion, is mediated by arachidonic acid. Cell Calcium. 24, 293–304. 35. Baron, A., Frieden, M., and Beny, J. L. (1997) Epoxyeicosatrienoic acids activate a high-conductance, Ca 2⫹-dependent K ⫹ channel on pig coronary artery endothelial cells. J. Physiol. 504, 537–543. 36. Hutcheson, I. R., and Griffith, T. M. (2000) Role of phospholipase A 2 and myoendothelial gap junctions in melittin-induced arterial relaxation. Eur. J. Pharmacol. 406, 239 –245. 37. Dora, K. A., Doyle, M. P., and Duling, B. R. (1997) Elevation of intracellular calcium in smooth muscle causes endothelial cell generation of NO in arterioles. Proc. Natl. Acad. Sci. USA 94, 6529 – 6534. 38. Cooper, D. M., Schell, M. J., Thorn, P., and Irvine, R. F. (1998) Regulation of adenylyl cyclase by membrane potential. J. Biol. Chem. 273, 27703–27707. 39. Reddy, R., Smith, D., Wayman, G., Wu, Z., Villacres, E. C., and Storm, D. R. (1995) Voltage-sensitive adenylyl cyclase activity in cultured neurons. J. Biol. Chem. 270, 14340 –14346. 40. Murthy, K. S., and Makhlouf, G. M. (1998) Regulation of adenylyl cyclase type V/VI in smooth muscle: Interplay of inhibitory G protein and Ca 2⫹ influx. Mol. Pharmacol. 54, 122–128.
589