Rapid Endothelial CelleCSelective Loading of Connexin 40 Antibody Blocks Endothelium-Derived Hyperpolarizing Factor Dilation in Rat Small Me
http://www.100md.com
循环研究杂志 2005年第8期
the Department of Pharmacy and Pharmacology (S.M., K.A.D., S.L.S., P.W., C.J.G.), University of Bath, UK
the Department of Physiology and Pharmacology (S.L.S.), University of New South Wales, Sydney, Australia.
Abstract
In resistance arteries, spread of hyperpolarization from the endothelium to the adjacent smooth muscle is suggested to be a crucial component of dilation resulting from endothelium-derived hyperpolarizing factor (EDHF). To probe the role of endothelial gap junctions in EDHF-mediated dilation, we developed a method, which was originally used to load membrane impermeant molecules into cells in culture, to load connexin (Cx)-specific inhibitory molecules rapidly (15 minutes) into endothelial cells within isolated, pressurized mesenteric arteries of the rat. Validation was achieved by luminally loading cell-impermeant fluorescent dyes selectively into virtually all the arterial endothelial cells, without affecting either tissue morphology or function. The endothelial monolayer served as an effective barrier, preventing macromolecules from entering the underlying smooth muscle cells. Using this technique, endothelial cell loading either with antibodies to the intracellular carboxyl-terminal region of Cx40 (residues 340 to 358) or mimetic peptide for the cytoplasmic loop (Cx40; residues 130 to 140) each markedly depressed EDHF-mediated dilation. In contrast, multiple antibodies directed against different intracellular regions of Cx37 and Cx43, and mimetic peptide for the intracellular loop region of Cx37, were each without effect. Furthermore, simultaneous intra- and extraluminal incubation of pressurized arteries with inhibitory peptides targeted against extracellular regions of endothelial cell Cxs (43Gap 26, 40Gap 27, and 37,43Gap 27; 300 eol/L each) for 2 hours also failed to modify the EDHF response. High-resolution immunohistochemistry localized Cx40 to the end of endothelial cell projections at myoendothelial gap junctions. These data directly demonstrate a critical role for Cx40 in EDHF-mediated dilation of rat mesenteric arteries.
Key Words: endothelium-derived hyperpolarizing factor myoendothelial gap junctions endothelium-dependent dilation acetylcholine connexin 40
Introduction
In arterioles and some arteries, gap junctions between endothelial and smooth muscle cells (myoendothelial gap junctions [MEGJs]) enable changes in membrane potential to spread over considerable distances and, as a consequence, regulate blood flow by coordinating diameter change through the microcirculation.1eC3 For example, injection of hyperpolarizing current into a single endothelial cell can evoke extensive relaxation involving many smooth muscle cells throughout an isolated arteriole.4,5 An important aspect of this response is that MEGJs may provide a crucial route for endothelial cell hyperpolarization to spread radially to the adjacent smooth muscle and evoke the dilation attributed to endothelium-derived hyperpolarizing factor (EDHF).6eC8 In many arteries, there is more than 1 underlying mechanism for endothelium-dependent hyperpolarization of smooth muscle cells. In the rat mesenteric artery, during submaximal contraction to phenylephrine (PE), EDHF dilation appears to reflect hyperpolarizing current spread through MEGJs and an action of endothelium-derived K+.9,10 However, with maximal PE contraction, the K+ component is blocked; therefore, the unidentified mechanism predominates.11
Although ultrastructural studies have demonstrated the presence of MEGJs in a number of resistance arteries, these structures do not appear to be ubiquitous.12 Furthermore, when MEGJs are present, showing that they are both patent and necessary for an aspect of smooth muscle cell function has not proved to be straightforward. Studies that have attempted to correlate the heterocellular spread of dyes and current have been confounded by a dissociation between these parameters, while the use of gap junction blockers, such as glycyrrhetinic acid and its derivatives, have been complicated by the variety of indirect effects exerted by these agents on vascular cells.13eC15 Synthetic peptides with homology to conserved extracellular regions of vascular connexins (Cxs) (Cx37, Cx40, and Cx43) have been suggested to provide selective block.7 However, how these agents might block gap junctions is not known, and the possibility that they may exert indirect effects has not been satisfactorily excluded. In spite of these persistent concerns, experiments using combinations of Cx-mimetic peptides in rabbit iliac artery have suggested a role for Cx37 and Cx40 in the radial spread of endothelial hyperpolarization.16 This contrasts with studies in another large vessel, the porcine coronary artery. Gap peptides failed to modify muscle hyperpolarization immediately below the internal elastic lamina, but spread through the media was markedly reduced, suggesting the peptides blocked homocellular smooth muscle gap junctions, rather than MEGJs.17
The potential to target MEGJs directly would be increased considerably were it possible to load putative blockers selectively into the endothelium of intact vessels, minimizing simultaneous effects within the smooth muscle layers. Recent work suggests that a method originally used with cultured cells18 may enable rapid and selective loading of molecules into endothelial cells in situ. By altering osmolarity using a commercially available kit, macromolecules were introduced into endothelial cells in rabbit resistance arteries.19 However, in extensive preliminary experiments using a similar kit, we were unable to maintain resistance arteries (diameter, 350 e) in a viable state; therefore, we modified the original method for loading cells in culture for use in isolated resistance-size arteries under physiological pressures.
We validated the technique using electron and fluorescence microscopy and using functional studies to show that EDHF-mediated dilation in rat pressurized mesenteric arteries maximally stimulated with PE can be significantly disrupted with antibodies and mimetic peptides directed against intracellular regions of Cx40 but not Cx37 or Cx43.
Materials and Methods
Measurement of Diameter and Endothelial Cell Calcium in Pressurized Arteries
Male Wistar rats (Charles River; 200 to 250 g) were anesthetized and euthanized following procedures defined under the UK Animals (Scientific Procedures) Act 1986. Segments (2- to 3-mm long) of a third-order branch (ID, 250 to 350 e) of superior mesenteric artery were dissected at 4°C and cannulated with glass pipettes (OD, 150 e) in a myograph (Danish Myotechnology 120CP).20 To ensure optimal reactivity, the artery was pressurized (50 mm Hg) and stretched longitudinally by 20%.21 After 30 minutes superfusion at 37°C, reactivity was assessed by contraction with 3 eol/L PE, followed by endothelium-dependent relaxation to 1 eol/L acetylcholine (ACh). Only vessels relaxing 90%, reflecting undamaged endothelium, were subsequently used. Brightfield images (outer diameter) were captured with a laser scanning confocal microscope (FV500-SU, x10 objective; Olympus) and recorded (Tiempo software, Olympus) at 1 Hz. For Ca2+ imaging, pressurized arteries were luminally perfused with 3-(N-morpholino) propanesulfonic acid (MOPS) solution containing fluo 4-acetoxymethyl ester (23 eol/L) and 0.02% Pluronic F-127 selectively to load the endothelial cells.22 Fluorescence from endothelial cells at the bottom of the artery (>50 cells) were visualized (512x360 pixels) and recorded at 2.4 Hz (x20, 0.5 numerical aperture objective; Olympus). Data were analyzed with Metamorph software (version 6.1; Universal Imaging). The nitric-oxide synthase (NOS) inhibitor NG-nitro-L-arginine methyl ester (100 eol/L) was present throughout all experiments.
Electron Microscopy
Conventional
Pressurized arteries were fixed in 1% paraformaldehyde, 3% glutaraldehyde in 0.1 mol/L sodium cacodylate, with 35 mmol/L betaine, 0.2 mmol/L CaCl, and 0.15 mol/L sucrose (pH 7.4) and processed using standard procedures.12 From a section cut perpendicular to the longitudinal axis of each preparation, each of four 15-e long regions of endothelium 90o apart were photographed at x20 000 and vesicle (50- to 200-nm diameter) counts (excluding surface caveolae) made from prints.
Immunolocalization
Short segments of artery were isolated from urethane-anesthetized (3 mg/kg) rats, rapidly dissected, and then frozen in liquid nitrogen at 2100 bar (BAL TEC, HPM 010). Segments were freeze substituted (Leica, AFS) at eC90°C in 0.1% uranyl acetate in acetone over 4 days and embedded in LR gold polymerized under UV light at eC25°C. Thin sections were cut and placed on formvar- and carbon-coated slot grids and sequentially incubated in PBS containing 1% normal donkey serum and 0.2% Tween 20 (blocking buffer; 30 minutes) with antibody raised against mouse Cx40 (1:100; Chemicon) in blocking buffer for 18 hours at 37°C and then rinsed and incubated in secondary antibodies conjugated to 5-nm colloidal gold (1:40; British Biocell) in 0.01% Tween 20. Sections were subsequently treated with 1% glutaraldehyde in PBS and stained conventionally.
Selective Endothelial Cell Loading of Cell-Impermeant Macromolecules
The effectiveness of macromolecule loading was determined by incorporating fluorescent markers 5(6)-carboxyfluorescein (1 mg/mL), Alexa Fluor 633 goat anti-rabbit IgG (0.2 mg/mL), or fluorescein isothiocyanate (FITC)-conjugated dextran (3 kDa, 2 mg/mL) into the endothelium of pressurized mesenteric arteries. During loading, luminal pressure was reduced to 5 mm Hg. A microperistaltic pump (LKB, Bromma; speed 100 e蘈/min) was used luminally to perfuse 100 e蘈 of hypertonic solution (800 milliosmolar [mosM]; 10% polyethylene glycol (PEG) 1000 plus impermeant macromolecule). The solution remained in the lumen for periods specified in Results (10 to 40 minutes) and was then flushed with 200 e蘈 of hypotonic solution (180 mosM, 40% H2O added) and remained present for a further 2 minutes, before flushing through 250 e蘈 of isotonic MOPS. After 2 to 10 minutes, arteries were reequilibrated at 50 mm Hg. Endothelial cells were visualized (x20, 0.5 numerical aperture objective; Olympus) at 5- to 10-minute intervals to determine the extent and stability of loading. The efficiency of loading was estimated to be 1:1000 by comparing cellular fluorescence against a standard curve constructed from serial dilutions of carboxyfluorescein in vitro (n=5).
Drugs and Solutions
MOPS-buffer of the following composition (mmol/L) was used: 145 NaCl, 4.7 KCl, 2.0 CaCl2·2H2O, 1.17 MgSO4·7H2O, 2.00 MOPS, 1.2 NaH2PO4·H2O, 5.0 glucose, 2.0 pyruvate (Na salt), 0.02 EDTA free acid, 2.75 NaOH. For the hypertonic solution (800 mosM), sucrose (500 mmol/L), and 10% volume/volume PEG 1000 was added to isotonic MOPS. Hypotonic solution (180 mosM) was 60% MOPS, 40% H2O. Drugs were all from Sigma except carboxyfluorescein, FITC-dextran (3 kDa), fluo 4-acetoxymethyl ester, and Pluoronic F-127 (Molecular Probes); PEG 1000 (Calbiochem). Levcromakalim was a gift from GlaxoSmithKline (Stevenage, UK).
Cx40 antibody against amino acids 340 to 358 (carboxyl-terminus; Chemicon); Cx37 antibody against amino acids 318 to 333 (carboxyl-terminus; Alpha Diagnostics) and 99 to 147 (cytoplasmic loop; Diatheva); Cx43 antibody against amino acids 363 to 382 and 130 to 143 (carboxyl-terminus and cytoplasmic loop; both from Sigma). These antibodies have been previously characterized in both transfected Rin and HeLa cells and arterial tissue.23,24 Antibodies were diluted (1:10) to give a luminal concentration of 50 e/mL. All stock solutions were prepared using distilled water.
Cx-mimetic peptides 43Gap 26 (VCYDKSFPISHVR), 40Gap 27 (SRPTEKNVFIV), and 37,43Gap 27 (SRPTEKTIFII) with homology to the first (Gap 26) and second (Gap 27) extracellular Cx-loops were from Severn Biotech. The pH of peptide containing MOPS was carefully adjusted to 7.4 on each occasion. For endothelial cell loading, peptides with homology to the cytoplasmic loop amino acids 130 to 140 of rCx37 (EHQMAKISVAE) and mCx40 (AELSCWKEVDG) were obtained from Auspep (Parkville, Australia; 95% purity). Peptides targeting this region have been used previously as controls in studies examining the potential blocking action of extracellularly applied Gap 26 and Gap 27 peptides.25,26
Statistical Analysis
Peak increases in outer diameter at each concentration of ACh were used to construct cumulative concentration-response curves (10 to 3 eol/L) and EC50 derived using a curve-fitting program (Prism 4.0). All data are means±SEM of n replicates, with n representing individual rats. Fluorescence data are summarized as average fluorescence from 16 randomly chosen endothelial cells (F), sampled at 5-second intervals relative to basal intensity (F0). Analysis used either paired Student t test or 1-way ANOVA with Bonferroni post test and rejection of the null hypothesis at P0.05.
Results
Hyperosmolarity Increases Endothelial Cell Pinocytic Vesicles Without Changing Cell Integrity
Pinocytic loading solutions did not alter the ultrastructural integrity of the endothelium (online Figure I in the data supplement available at http://circres.ahajournals.org). Tissue fixed in hypertonic loading solution showed an 4-fold increase in intracellular vesicles within endothelial cells (online Figure I) and a reduction in cell volume. In contrast, after completing the loading protocol, the number of vesicles and the cell volume returned to control levels. Neither treatment altered ultrastructural morphology in the adjacent smooth muscle cells. Furthermore, the loading protocol damaged <5% of endothelial cells, assessed by staining dead cells with propidium iodide (online Figure IIA).
Changing Luminal Osmolarity Selectively Loads Endothelial Cells With Fluorescent Markers
A 10-minute hypertonic exposure was optimal for selective cellular loading with carboxyfluorescein (Figure 1A). Increasing intraluminal exposure to 20 minutes caused substantive loading of the smooth muscle (Figure 1B). Abluminal exposure of pressurized arteries to similar solutions directly loaded the smooth muscle and perivascular nerves with carboxyfluorescein (online Figure IIA). Selective, extensive endothelial cell loading also followed 10-minute luminal incubation with hypertonic solution containing Alexa Fluor 633 IgG or FITC-labeled dextran and without significant smooth muscle loading (Figure 1A). Unlike carboxyfluorescein, increasing luminal exposure to 40 minutes did not load the smooth muscle with these larger molecular-mass dyes, apart from the occasional cell.
After endothelial loading with carboxyfluorescein (for 10 minutes), subsequent images at 5-minute intervals revealed gradual loss of fluorescence (Figure 1B). By 30 minutes, most dye fluorescence was gone. This did not appear to reflect bleaching, as similar loss was apparent if fluorescence was assessed only once, at 30 minutes postloading. Following the loss of carboxyfluorescein fluorescence from the endothelium, the same cells could be "reloaded" with dye by simply repeating the osmotic loading protocol, with vascular reactivity retained. As IgG or dextran loading gave a stable fluorescent signal over time, carboxyfluorescein may be actively extruded by the cells.
Changing Luminal Osmolarity Does Not Alter Functionality in Pressurized Arteries
Endothelium-dependent relaxation to ACh (10 nmol/L to 3 eol/L; EC50, 177 nmol/L; maximal relaxation [Rmax], 93±3%; n=8) was not altered by the osmotic loading protocol (EC50, 170 nmol/L; Rmax, 90±2%; n=8; Figure 2A and 2B). Likewise, neither the initial contraction to 3 eol/L PE (77.4±1.9%; n=13) nor dilation to 3 eol/L levcromakalim (90.3±1.9%, n=12) or 15 eol/L cyclopiazonic acid (84.2±3.3%, n=13) was altered.
As an increase in endothelial cell [Ca2+]i to ACh is critical for endothelial and subsequent smooth muscle cell hyperpolarization,22 endothelial cell viability was assessed as time course of rises in [Ca2+]i in individual endothelial cells in pressurized arteries (Figure 2C and 2D). Before and after the osmotic loading protocol, ACh stimulated sustained and oscillating rises in endothelial cell [Ca2+]i (Figure 2C), the oscillations remaining asynchronous between cells, with repetitive waves of Ca2+ moving longitudinally across endothelial cells. Therefore, the loading protocol alone did not modify the ability of the endothelium to increase [Ca2+]i and evoke EDHF responses.
EDHF-Mediated Dilatation to ACh: Effect of Endothelial Cell Loading With Cx Antibodies and Mimetic Peptides
Endothelial cell loading with antibodies directed against the intracellular carboxy-terminus of Cx40 did not alter EDHF-mediated dilation in arteries submaximally contracted to 0.9 eol/L PE (control: EC50, 130 nmol/L; Rmax, 97.0±0.8%; after antibody: EC50, 99 nmol/L; Rmax, 97.6±0.8%; n=4; Figure 3A and 3B) or in the additional presence of ouabain (online Figure IIIA). In marked contrast, raising the concentration of PE to evoke maximal contraction (and inhibit the simultaneous influence of diffusible EDHF apparent during submaximal stimulation11,20) was associated with block of EDHF dilation (Figure 3A and 3B and online Figure IIIA). In vessels contracted with 5 eol/L PE, the Rmax was only 26.7±8.8% (n=11), whereas subsequent addition of 3 eol/L levcromakalim evoked complete dilation (96.9±0.9%, n=4; Figure 3A). Cx40 antibody did not alter either the increase or oscillations in endothelial cell [Ca2+]i to ACh despite block of dilation (Figure 3C), and staining with propidium iodide indicated damage to <8% of cells. In contrast, endothelial loading with antibodies against both the carboxy-terminus and the cytoplasmic loop of Cx37 and Cx43 failed to modify EDHF dilation significantly (Figure 4A and 4B). Furthermore, if Cx40 antibody was loaded before block of NOS, dilation was sustained until NG-nitro-L-arginine methyl ester was applied (online Figure IIIB).
Overall, this profile was mirrored by endothelial cell loading with intracellular Cx-mimetic peptides. Cx40 peptide suppressed EDHF dilation in arteries contracted with 5 eol/L PE, whereas the Cx37 peptide had no significant effect (Figure 4C).
Effect of Extracellular Cx-Mimetic Peptides and Carbenoxolone Against EDHF Dilation
Simultaneous intra- and extraluminal incubation of pressurized arteries (for 2 hours; 3x40 minutes exposures) with a combination of 43Gap 26, 40Gap 27, and 37,43Gap 27 (each 300 eol/L) failed to modify EDHF dilation to ACh (Figure 5A). In contrast, whereas the gap junction blocker carbenoxolone (100 eol/L) did not alter the amplitude of, or oscillations in, endothelial increases in [Ca2+]i (Figure 5C and 5D), it completely blocked EDHF dilation (Figure 5B).
Ultrastructural Localization of Cx40
Cx40 was localized to points of close association between endothelial and smooth muscle cells that also exhibited small regions of pentalaminar membrane, characteristic of MEGJs (Figure 6).
Discussion
There are 2 major observations in this study. First, we show that it is possible to rapidly and selectively load large membrane impermeant molecules, including antibodies, into the endothelium of arteries under physiological pressures, without disrupting function. Second, using this approach combined with high-resolution ultrastructural methods, we provide direct evidence that MEGJs play a central role in EDHF dilation and highlight a primary role for Cx40 in the radial transfer of this signal from the endothelium.
The basis for cell loading is the normal process of endocytosis, describing the nondisruptive passage of macromolecules across the plasma membrane by phagacytosis or "cell eating" and pinocytosis or "cell drinking."27 Only a few specialized cells, such as macrophages and neutrophils, use phagacytosis, whereas virtually all cells can carry out some form of pinocytosis. An ability to increase the formation of pinocytic vesicles in cells was first described in L929 cells.18 Vesicles were then ruptured (by altering the osmolarity), enabling intracellular release of horseradish peroxidase, anti-ricin antibodies, or dextran (70 kDa), an effect termed osmotic lysis of pinosomes. A similar approach in whole tissue (rabbit mesenteric artery) used a commercially available kit,19 but, although we could load rat mesenteric smooth muscle cells with fluorescently-labeled dextran using kit reagents, endothelium-dependent vasodilation was lost after only 5 minutes of exposure. We, therefore, developed the original protocol for use in pressurized arteries.
Selective endothelial cell incorporation of macromolecules was facilitated by using artery segments cannulated at both ends. Luminal solution was changed independently of the bathing solution, and a cycle of 10 minutes hypertonic, 2 minutes hypotonic, before returning to isosmotic conditions, proved optimal to load the endothelium with fluorescent markers of differing structure and size: carboxyfluorescein, FITC-dextran, and Alexa Fluor 633 IgG. If the hyperosmotic phase was prolonged, carboxyfluorescein appeared in the smooth muscle, which might reflect movement through the myoendothelial gap junctions present in this artery.28 Molecules up to 1 kDa can pass through gap junctions; therefore, this seems a likely explanation for carboxyfluorescein (376 Da),29,30 supported by the observation that equivalent muscle loading was not evident after endothelial cell loading with either IgG (150 kDa) or dextran (3 kDa). In contrast to these markers, carboxyfluorescein-fluorescence within the endothelium decreased rapidly, presumably reflecting, in part, the activity of ATP-binding cassette transporters within the endothelial membrane, which extrude a range of molecules, including carboxyfluorescein and calcein.31,32
Following technique validation, the role of MEGJs in EDHF dilation was probed. Agonist activation of endothelial cells in the mesenteric artery increases [Ca2+]i22,33 and stimulates hyperpolarization,3,15,34,35 followed by hyperpolarization and relaxation in the adjacent smooth muscle.36,37 The smooth muscle effects are suggested to reflect the action of a diffusible EDHF, possibly K+, and the spread of hyperpolarization from the endothelium through MEGJs.9,38 That radial heterocellular coupling is important in local, endothelium-driven dilation is supported in the rat mesenteric artery by the presence of heterocellular membrane associations of pentalaminar structure, characteristic of MEGJs.12,28 However, although rat mesenteric artery endothelial cells express Cx37, Cx40, and possibly Cx43 (but not Cx4539), the precise profile in the smooth muscle is not clear.23,39 Thus, it is not known which Cxs are expressed at the MEGJs.
Endothelial cell hyperpolarization reflects the opening of Ca2+-activated K+ channels22,34,35 amplified by K+ efflux through the inwardly rectifying K+ channels focused within the endothelium of the mesenteric artery.40,41 Consequent extracellular K+ accumulation then stimulates smooth muscle hyperpolarization and relaxation by activating Na+/K+-ATPase.11 However, if background contraction with PE is increased, K+ loses the ability to serve as an EDHF, whereas smooth muscle cell hyperpolarization and relaxation to ACh is sustained by a persistent parallel mechanism.11 Loss of K+ as an EDHF appears to be caused by BKCa-channel activation in the smooth muscle (by PE-mediated depolarization and rises in [Ca2+]i with consequent K+ efflux eventually maximally activating the Na+/K+-ATPase), as it is reversed in the presence of the BKCa blocker iberiotoxin.20,42,43 In the present study, possible input from K+ as an EDHF was blocked by high levels of background arterial contraction to 5 eol/L PE,11 isolating EDHF dilation, which may be suggested to be attributable to MEGJs in this artery.9
The present study provides direct evidence that the MEGJs mediate the persistent dilation. Endothelial cell loading with antibody to the intracellular carboxy-terminal region of Cx40 completely abolished EDHF dilation, without altering the increase in [Ca2+]i necessary to initiate hyperpolarization, when arteries were stimulated with higher but not lower concentrations of PE. In these experiments, direct muscle dilation to levcromakalim3,35 persisted, and in separate experiments, EDHF dilation was maintained after endothelial loading with inhibitory molecules against different intracellular regions of Cx37 or Cx43. Interestingly, when arteries were contracted with low concentrations of PE, Cx40 antibody-insensitive ACh relaxation appeared not solely attributable to release of K+ acting on smooth muscle Na+/K+-ATPase. Although ouabain alone shifted the ACh-dilation curve, it did not abolish dilation in the presence of Cx40 antibody. This suggests compensatory pathways can still operate under low tone, possibly via electrical coupling through MEGJs comprised of Cx37.
A role for Cx40 in heterocellular signaling is supported by ultrastructural localization of this Cx to the end of endothelial-derived projections associated with MEGJs.12,28 There is only 1 other equivalent ultrastructural report, localizing Cx40 within the membrane of gap junctions between endothelial cells and smooth muscle cells modified to secrete renin in the terminal region of kidney afferent arterioles.44 Whether or not this Cx is the primary Cx for myoendothelial signaling remains to be seen.
The Cx-mimetic peptides loaded into the endothelium were homologous to segments of the intracellular cytoplasmic loop sequences (amino acids 130 to 140) of Cx37 and Cx40. Patch-clamp experiments suggest this region may represent a Cx-gating element,45 making it a potentially useful target, particularly because there is no sequence homology here between Cx37 and Cx40. An ability to block EDHF dilation with Cx40 intracellular-mimetic peptide contrasts with the lack of effect of prolonged extracellular application of a triple combination of Gap peptides targeting extracellular Cx regions. This triple combination has been suggested as the most effective and reliable way to block vascular gap junctions.7 Previous use of Gap peptides against EDHF responses in the mesenteric artery provide inconsistent data. Gap 27 failed significantly to reduce EDHF hyperpolarization in 1 study,46 reduced it by 65% in another,12 and almost abolished EDHF dilation in pressurized mesenteric arteries.38 Interestingly, in the latter, Gap peptide was applied via the lumen of pressurized arteries, so it would not be predicted to even reach the MEGJs, due to the tight junctions between the endothelial cells. In any event, the reason such discrepancies exist is not clear. One possibility may be an alteration in pH caused by the peptides. In our experiments, where the triple peptide combination was applied within the lumen and around the artery at the same time, the reduction in pH caused by the peptides was corrected. Lack of effect of Gap peptides against EDHF dilation is similar to the lack of effect of 37,40Gap 26 in mouse mesenteric arteries, where MEGJs are abundant.47
Data linking cell-specific Cx isoforms to functional responses in vascular tissue are limited. A major physiological role for Cx40 is indicated by studies with knockout (Cx40eC/eC) mice. These mice are hypertensive, and, even without NOS block, cremaster arteriolar dilation to 10 eol/L ACh was attenuated.48 Furthermore, the ability of arterioles to conduct vasodilatation longitudinally through the vessel wall was markedly attenuated in these animals,48,49 a response normally ascribed to the endothelium.3,50 However, it remains possible that long-term compensatory mechanisms may underlie these changes in genetically modified mice, whereas short-term block in the present study circumvents this potential problem. The only equivalent long-term data come from a recent study with Gap peptides in the rabbit iliac artery (10 smooth muscle layers), suggesting both Cx37 and Cx40 are essential for hyperpolarization to spread through MEGJs.16 However, in this particular artery, definitive identification of MEGJs, and the constitutive Cxs, is required.
In summary, by selectively loading inhibitory Cx antibodies and peptides into endothelial cells in pressurized rat mesenteric arteries, we have demonstrated directly a major functional role for the MEGJs known to occur in this artery. Furthermore, our data suggest that Cx40 is the critical Cx isoform involved in heterocellular signaling.
Acknowledgments
This work was supported by the Wellcome Trust (UK) and the British Heart Foundation.
References
Duling BR, Berne RM. Propagated vasodilation in the microcirculation of the hamster cheek pouch. Circ Res. 1970; 26: 163eC170.
Gustafsson F, Holstein-Rathlou N. Conducted vasomotor responses in arterioles: characteristics, mechanisms and physiological significance. Acta Physiol Scand. 1999; 167: 11eC21.
Takano H, Dora KA, Spitaler MM, Garland CJ. Spreading dilatation in rat mesenteric arteries associated with calcium-independent endothelial cell hyperpolarization. J Physiol. 2004; 556: 887eC903.
Emerson GG, Segal SS. Electrical coupling between endothelial cells and smooth muscle cells in hamster feed arteries: role in vasomotor control. Circ Res. 2000; 87: 474eC479.
Emerson GG, Segal SS. Electrical activation of endothelium evokes vasodilation and hyperpolarization along hamster feed arteries. Am J Physiol Heart Circ Physiol. 2001; 280: H160eCH167.
Busse R, Edwards G, Feletou M, Fleming I, Vanhoutte PM, Weston AH. EDHF: bringing the concepts together. Trends Pharmacol Sci. 2002; 23: 374eC380.
Griffith TM. Endothelium-dependent smooth muscle hyperpolarization: do gap junctions provide a unifying hypothesis Br J Pharmacol. 2004; 141: 881eC903.
Sandow SL. Factors, fiction and endothelium-derived hyperpolarizing factor. Clin Exp Pharmacol Physiol. 2004; 31: 563eC570.
Edwards G, Dora KA, Gardener MJ, Garland CJ, Weston AH. K+ is an endothelium-derived hyperpolarizing factor in rat arteries. Nature. 1998; 396: 269eC272.
Dora KA, McEvoy L, King M, Ings NT, Garland CJ. The intensity of agonist-stimulation influences the mechanism for relaxation in rat mesenteric arteries. In: Vanhoutte PM, ed. EDHF 2002. London: Taylor & Francis; 2003: 256eC260.
Dora KA, Garland CJ. Properties of smooth muscle hyperpolarization and relaxation to K+ in the rat isolated mesenteric artery. Am J Physiol Heart Circ Physiol. 2001; 280: H2424eCH2429.
Sandow SL, Tare M, Coleman HA, Hill CE, Parkington HC. Involvement of myoendothelial gap junctions in the actions of endothelium-derived hyperpolarizing factor. Circ Res. 2002; 90: 1108eC1113.
Beny J. Endothelial and smooth muscle cells hyperpolarized by bradykinin are not dye coupled. Am J Physiol. 1990; 258: H836eCH841.
Coleman HA, Tare M, Parkington HC. K+ currents underlying the action of endothelium-derived hyperpolarizing factor in guinea-pig, rat and human blood vessels. J Physiol (Lond). 2001; 531: 359eC373.
Tare M, Coleman HA, Parkington HC. Glycyrrhetinic derivatives inhibit hyperpolarization in endothelial cells of guinea pig and rat arteries. Am J Physiol Heart Circ Physiol. 2002; 282: H335eCH341.
Chaytor AT, Bakker LM, Edwards DH, Griffith TM. Connexin-mimetic peptides dissociate electrotonic EDHF-type signalling via myoendothelial and smooth muscle gap junctions in the rabbit iliac artery. Br J Pharmacol. 2005; 144: 108eC114.
Edwards G, Thollon C, Gardener MJ, Feletou M, Vilaine J, Vanhoutte PM, Weston AH. Role of gap junctions and EETs in endothelium-dependent hyperpolarization of porcine coronary artery. Br J Pharmacol. 2000; 129: 1145eC1154.
Okada CY, Rechsteiner M. Introduction of macromolecules into cultured mammalian cells by osmotic lysis of pinocytic vesicles. Cell. 1982; 29: 33eC41.
Vequaud P, Thorin E. Endothelial G protein -subunits trigger nitric oxide- but not endothelium-derived hyperpolarizing factor-dependent dilation in rabbit resistance arteries. Circ Res. 2001; 89: 716eC722.
Dora KA, Ings NT, Garland CJ. KCa channel blockers reveal hyperpolarization and relaxation to K+ in the rat isolated mesenteric artery. Am J Physiol Heart Circ Physiol. 2002; 283: H606eCH614.
Coats P, Hillier C. Determination of an optimal axial-length tension for the study of isolated resistance arteries on a pressure myograph. Exp Physiol. 1999; 84: 1085eC1094.
McSherry IN, Spitaler MM, Takano H, Dora KA. Endothelial cell Ca2+ increases are independent of membrane potential in pressurized rat mesenteric arteries. Cell Calcium. 2005; 38: 23eC33.
Gustafsson F, Mikkelsen HB, Arensbak B, Thuneberg L, Neve S, Jensen LJ, Holstein-Rathlou NH. Expression of connexin 37, 40 and 43 in rat mesenteric arterioles and resistance arteries. Histochem Cell Biol. 2003; 119: 139eC148.
Sandow SL, Looft-Wilson R, Doran B, Grayson TH, Segal SS, Hill CE. Expression of homocellular and heterocellular gap junctions in hamster arterioles and feed arteries. Cardiovasc Res. 2003; 60: 643eC653.
Chaytor AT, Evans WH, Griffith TM. Peptides homologous to extracellular loop motifs of connexin 43 reversibly abolish rhythmic contractile activity in rabbit arteries. J Physiol. 1997; 503: 99eC110.
Earley S, Resta TC, Walker BR. Disruption of smooth muscle gap junctions attenuates myogenic vasoconstriction of mesenteric resistance arteries. Am J Physiol Heart Circ Physiol. 2004; 287: H2677eCH2686.
Conner SD, Schmid SL. Regulated portals of entry into the cell. Nature. 2003; 422: 37eC44.
Sandow SL, Hill CE. Incidence of myoendothelial gap junctions in the proximal and distal mesenteric arteries of the rat is suggestive of a role in endothelium-derived hyperpolarizing factor-mediated responses. Circ Res. 2000; 86: 341eC346.
Evans WH, Martin PE. Gap junctions: structure and function. Mol Membr Biol. 2002; 19: 121eC136.
Unger VM, Kumar NM, Gilula NB, Yeager M. Three-dimensional structure of a recombinant gap junction membrane channel. Science. 1999; 283: 1176eC1180.
Rychlik B, Balcerczyk A, Klimczak A, Bartosz G. The role of multidrug resistance protein 1 (MRP1) in transport of fluorescent anions across the human erythrocyte membrane. J Membr Biol. 2003; 193: 79eC90.
Bachmeier CJ, Trickler WJ, Miller DW. Drug efflux transport properties of 2',7'-bis(2-carboxyethyl)-5(6)-carboxyfluorescein acetoxymethyl ester (BCECF-AM) and its fluorescent free acid, BCECF. J Pharm Sci. 2004; 93: 932eC942.
Dora KA, Hinton JM, Walker SD, Garland CJ. An indirect influence of phenylephrine on the release of endothelium-derived vasodilators in rat small mesenteric artery. Br J Pharmacol. 2000; 129: 381eC387.
Hinton JM, Langton PD. Inhibition of EDHF by two new combinations of K+-channel inhibitors in rat isolated mesenteric arteries. Br J Pharmacol. 2003; 138: 1031eC1035.
White R, Hiley CR. Hyperpolarisation of rat mesenteric endothelial cells by ATP-sensitive K+ channel openers. Eur J Pharmacol. 2000; 397: 279eC290.
Garland CJ, McPherson GA. Evidence that nitric oxide does not mediate the hyperpolarization and relaxation to acetylcholine in the rat small mesenteric artery. Br J Pharmacol. 1992; 105: 429eC435.
Crane GJ, Gallagher NT, Dora KA, Garland CJ. Small and intermediate calcium-dependent K+ channels provide different facets of endothelium-dependent hyperpolarization in rat mesenteric artery. J Physiol. 2003; 553: 183eC189.
Doughty JM, Boyle JP, Langton PD. Potassium does not mimic EDHF in rat mesenteric arteries. Br J Pharmacol. 2000; 130: 1174eC1182.
Kansui Y, Fujii K, Nakamura K, Goto K, Oniki H, Abe I, Shibata Y, Iida M. Angiotensin II receptor blockade corrects altered expression of gap junctions in endothelial cells from hypertensive rats. Am J Physiol Heart Circ Physiol. 2004; 287: H216eCH224.
Doughty JM, Boyle JP, Langton PD. Blockade of chloride channels reveals relaxations of rat small mesenteric arteries to raised potassium. Br J Pharmacol. 2001; 132: 293eC301.
Crane GJ, Walker SD, Dora KA, Garland CJ. Evidence for a differential cellular distribution of inward rectifier K channels in the rat isolated mesenteric artery. J Vasc Res. 2003; 40: 159eC168.
Richards GR, Weston AH, Burnham MP, Feletou M, Vanhoutte PM, Edwards G. Suppression of K+-induced hyperpolarization by phenylephrine in rat mesenteric artery: relevance to studies of endothelium-derived hyperpolarizing factor. Br J Pharmacol. 2001; 134: 1eC5.
Walker SD, Dora KA, Ings NT, Crane GJ, Garland CJ. Activation of endothelial cell IKCa and 1-ethyl-2-benzimidazolinone evokes smooth muscle hyperpolarization in rat isolated mesenteric artery. Br J Pharmacol. 2001; 134: 1548eC1554.
Haefliger JA, Demotz S, Braissant O, Suter E, Waeber B, Nicod P, Meda P. Connexins 40 and 43 are differentially regulated within the kidneys of rats with renovascular hypertension. Kidney Int. 2001; 60: 190eC201.
Seki A, Duffy HS, Coombs W, Spray DC, Taffet SM, Delmar M. Modifications in the biophysical properties of connexin43 channels by a peptide of the cytoplasmic loop region. Circ Res. 2004; 95: e22eCe28.
Edwards G, Feletou M, Gardener MJ, Thollon C, Vanhoutte PM, Weston AH. Role of gap junctions in the responses to EDHF in rat and guinea-pig small arteries. Br J Pharmacol. 1999; 128: 1788eC1794.
Dora KA, Sandow SL, Gallagher NT, Takano H, Rummery NM, Hill CE, Garland CJ. Myoendothelial gap junctions may provide the pathway for EDHF in mouse mesenteric artery. J Vasc Res. 2003; 40: 480eC490.
de Wit C, Roos F, Bolz SS, Kirchhoff S, Kruger O, Willecke K, Pohl U. Impaired conduction of vasodilation along arterioles in connexin40-deficient mice. Circ Res. 2000; 86: 649eC655.
Figueroa XF, Paul DL, Simon AM, Goodenough DA, Day KH, Damon DN, Duling BR. Central role of connexin40 in the propagation of electrically activated vasodilation in mouse cremasteric arterioles in vivo. Circ Res. 2003; 92: 793eC800.
Emerson GG, Segal SS. Endothelial cell pathway for conduction of hyperpolarization and vasodilation along hamster feed artery. Circ Res. 2000; 86: 94eC100.(Simon Mather, Kim A. Dora)
the Department of Physiology and Pharmacology (S.L.S.), University of New South Wales, Sydney, Australia.
Abstract
In resistance arteries, spread of hyperpolarization from the endothelium to the adjacent smooth muscle is suggested to be a crucial component of dilation resulting from endothelium-derived hyperpolarizing factor (EDHF). To probe the role of endothelial gap junctions in EDHF-mediated dilation, we developed a method, which was originally used to load membrane impermeant molecules into cells in culture, to load connexin (Cx)-specific inhibitory molecules rapidly (15 minutes) into endothelial cells within isolated, pressurized mesenteric arteries of the rat. Validation was achieved by luminally loading cell-impermeant fluorescent dyes selectively into virtually all the arterial endothelial cells, without affecting either tissue morphology or function. The endothelial monolayer served as an effective barrier, preventing macromolecules from entering the underlying smooth muscle cells. Using this technique, endothelial cell loading either with antibodies to the intracellular carboxyl-terminal region of Cx40 (residues 340 to 358) or mimetic peptide for the cytoplasmic loop (Cx40; residues 130 to 140) each markedly depressed EDHF-mediated dilation. In contrast, multiple antibodies directed against different intracellular regions of Cx37 and Cx43, and mimetic peptide for the intracellular loop region of Cx37, were each without effect. Furthermore, simultaneous intra- and extraluminal incubation of pressurized arteries with inhibitory peptides targeted against extracellular regions of endothelial cell Cxs (43Gap 26, 40Gap 27, and 37,43Gap 27; 300 eol/L each) for 2 hours also failed to modify the EDHF response. High-resolution immunohistochemistry localized Cx40 to the end of endothelial cell projections at myoendothelial gap junctions. These data directly demonstrate a critical role for Cx40 in EDHF-mediated dilation of rat mesenteric arteries.
Key Words: endothelium-derived hyperpolarizing factor myoendothelial gap junctions endothelium-dependent dilation acetylcholine connexin 40
Introduction
In arterioles and some arteries, gap junctions between endothelial and smooth muscle cells (myoendothelial gap junctions [MEGJs]) enable changes in membrane potential to spread over considerable distances and, as a consequence, regulate blood flow by coordinating diameter change through the microcirculation.1eC3 For example, injection of hyperpolarizing current into a single endothelial cell can evoke extensive relaxation involving many smooth muscle cells throughout an isolated arteriole.4,5 An important aspect of this response is that MEGJs may provide a crucial route for endothelial cell hyperpolarization to spread radially to the adjacent smooth muscle and evoke the dilation attributed to endothelium-derived hyperpolarizing factor (EDHF).6eC8 In many arteries, there is more than 1 underlying mechanism for endothelium-dependent hyperpolarization of smooth muscle cells. In the rat mesenteric artery, during submaximal contraction to phenylephrine (PE), EDHF dilation appears to reflect hyperpolarizing current spread through MEGJs and an action of endothelium-derived K+.9,10 However, with maximal PE contraction, the K+ component is blocked; therefore, the unidentified mechanism predominates.11
Although ultrastructural studies have demonstrated the presence of MEGJs in a number of resistance arteries, these structures do not appear to be ubiquitous.12 Furthermore, when MEGJs are present, showing that they are both patent and necessary for an aspect of smooth muscle cell function has not proved to be straightforward. Studies that have attempted to correlate the heterocellular spread of dyes and current have been confounded by a dissociation between these parameters, while the use of gap junction blockers, such as glycyrrhetinic acid and its derivatives, have been complicated by the variety of indirect effects exerted by these agents on vascular cells.13eC15 Synthetic peptides with homology to conserved extracellular regions of vascular connexins (Cxs) (Cx37, Cx40, and Cx43) have been suggested to provide selective block.7 However, how these agents might block gap junctions is not known, and the possibility that they may exert indirect effects has not been satisfactorily excluded. In spite of these persistent concerns, experiments using combinations of Cx-mimetic peptides in rabbit iliac artery have suggested a role for Cx37 and Cx40 in the radial spread of endothelial hyperpolarization.16 This contrasts with studies in another large vessel, the porcine coronary artery. Gap peptides failed to modify muscle hyperpolarization immediately below the internal elastic lamina, but spread through the media was markedly reduced, suggesting the peptides blocked homocellular smooth muscle gap junctions, rather than MEGJs.17
The potential to target MEGJs directly would be increased considerably were it possible to load putative blockers selectively into the endothelium of intact vessels, minimizing simultaneous effects within the smooth muscle layers. Recent work suggests that a method originally used with cultured cells18 may enable rapid and selective loading of molecules into endothelial cells in situ. By altering osmolarity using a commercially available kit, macromolecules were introduced into endothelial cells in rabbit resistance arteries.19 However, in extensive preliminary experiments using a similar kit, we were unable to maintain resistance arteries (diameter, 350 e) in a viable state; therefore, we modified the original method for loading cells in culture for use in isolated resistance-size arteries under physiological pressures.
We validated the technique using electron and fluorescence microscopy and using functional studies to show that EDHF-mediated dilation in rat pressurized mesenteric arteries maximally stimulated with PE can be significantly disrupted with antibodies and mimetic peptides directed against intracellular regions of Cx40 but not Cx37 or Cx43.
Materials and Methods
Measurement of Diameter and Endothelial Cell Calcium in Pressurized Arteries
Male Wistar rats (Charles River; 200 to 250 g) were anesthetized and euthanized following procedures defined under the UK Animals (Scientific Procedures) Act 1986. Segments (2- to 3-mm long) of a third-order branch (ID, 250 to 350 e) of superior mesenteric artery were dissected at 4°C and cannulated with glass pipettes (OD, 150 e) in a myograph (Danish Myotechnology 120CP).20 To ensure optimal reactivity, the artery was pressurized (50 mm Hg) and stretched longitudinally by 20%.21 After 30 minutes superfusion at 37°C, reactivity was assessed by contraction with 3 eol/L PE, followed by endothelium-dependent relaxation to 1 eol/L acetylcholine (ACh). Only vessels relaxing 90%, reflecting undamaged endothelium, were subsequently used. Brightfield images (outer diameter) were captured with a laser scanning confocal microscope (FV500-SU, x10 objective; Olympus) and recorded (Tiempo software, Olympus) at 1 Hz. For Ca2+ imaging, pressurized arteries were luminally perfused with 3-(N-morpholino) propanesulfonic acid (MOPS) solution containing fluo 4-acetoxymethyl ester (23 eol/L) and 0.02% Pluronic F-127 selectively to load the endothelial cells.22 Fluorescence from endothelial cells at the bottom of the artery (>50 cells) were visualized (512x360 pixels) and recorded at 2.4 Hz (x20, 0.5 numerical aperture objective; Olympus). Data were analyzed with Metamorph software (version 6.1; Universal Imaging). The nitric-oxide synthase (NOS) inhibitor NG-nitro-L-arginine methyl ester (100 eol/L) was present throughout all experiments.
Electron Microscopy
Conventional
Pressurized arteries were fixed in 1% paraformaldehyde, 3% glutaraldehyde in 0.1 mol/L sodium cacodylate, with 35 mmol/L betaine, 0.2 mmol/L CaCl, and 0.15 mol/L sucrose (pH 7.4) and processed using standard procedures.12 From a section cut perpendicular to the longitudinal axis of each preparation, each of four 15-e long regions of endothelium 90o apart were photographed at x20 000 and vesicle (50- to 200-nm diameter) counts (excluding surface caveolae) made from prints.
Immunolocalization
Short segments of artery were isolated from urethane-anesthetized (3 mg/kg) rats, rapidly dissected, and then frozen in liquid nitrogen at 2100 bar (BAL TEC, HPM 010). Segments were freeze substituted (Leica, AFS) at eC90°C in 0.1% uranyl acetate in acetone over 4 days and embedded in LR gold polymerized under UV light at eC25°C. Thin sections were cut and placed on formvar- and carbon-coated slot grids and sequentially incubated in PBS containing 1% normal donkey serum and 0.2% Tween 20 (blocking buffer; 30 minutes) with antibody raised against mouse Cx40 (1:100; Chemicon) in blocking buffer for 18 hours at 37°C and then rinsed and incubated in secondary antibodies conjugated to 5-nm colloidal gold (1:40; British Biocell) in 0.01% Tween 20. Sections were subsequently treated with 1% glutaraldehyde in PBS and stained conventionally.
Selective Endothelial Cell Loading of Cell-Impermeant Macromolecules
The effectiveness of macromolecule loading was determined by incorporating fluorescent markers 5(6)-carboxyfluorescein (1 mg/mL), Alexa Fluor 633 goat anti-rabbit IgG (0.2 mg/mL), or fluorescein isothiocyanate (FITC)-conjugated dextran (3 kDa, 2 mg/mL) into the endothelium of pressurized mesenteric arteries. During loading, luminal pressure was reduced to 5 mm Hg. A microperistaltic pump (LKB, Bromma; speed 100 e蘈/min) was used luminally to perfuse 100 e蘈 of hypertonic solution (800 milliosmolar [mosM]; 10% polyethylene glycol (PEG) 1000 plus impermeant macromolecule). The solution remained in the lumen for periods specified in Results (10 to 40 minutes) and was then flushed with 200 e蘈 of hypotonic solution (180 mosM, 40% H2O added) and remained present for a further 2 minutes, before flushing through 250 e蘈 of isotonic MOPS. After 2 to 10 minutes, arteries were reequilibrated at 50 mm Hg. Endothelial cells were visualized (x20, 0.5 numerical aperture objective; Olympus) at 5- to 10-minute intervals to determine the extent and stability of loading. The efficiency of loading was estimated to be 1:1000 by comparing cellular fluorescence against a standard curve constructed from serial dilutions of carboxyfluorescein in vitro (n=5).
Drugs and Solutions
MOPS-buffer of the following composition (mmol/L) was used: 145 NaCl, 4.7 KCl, 2.0 CaCl2·2H2O, 1.17 MgSO4·7H2O, 2.00 MOPS, 1.2 NaH2PO4·H2O, 5.0 glucose, 2.0 pyruvate (Na salt), 0.02 EDTA free acid, 2.75 NaOH. For the hypertonic solution (800 mosM), sucrose (500 mmol/L), and 10% volume/volume PEG 1000 was added to isotonic MOPS. Hypotonic solution (180 mosM) was 60% MOPS, 40% H2O. Drugs were all from Sigma except carboxyfluorescein, FITC-dextran (3 kDa), fluo 4-acetoxymethyl ester, and Pluoronic F-127 (Molecular Probes); PEG 1000 (Calbiochem). Levcromakalim was a gift from GlaxoSmithKline (Stevenage, UK).
Cx40 antibody against amino acids 340 to 358 (carboxyl-terminus; Chemicon); Cx37 antibody against amino acids 318 to 333 (carboxyl-terminus; Alpha Diagnostics) and 99 to 147 (cytoplasmic loop; Diatheva); Cx43 antibody against amino acids 363 to 382 and 130 to 143 (carboxyl-terminus and cytoplasmic loop; both from Sigma). These antibodies have been previously characterized in both transfected Rin and HeLa cells and arterial tissue.23,24 Antibodies were diluted (1:10) to give a luminal concentration of 50 e/mL. All stock solutions were prepared using distilled water.
Cx-mimetic peptides 43Gap 26 (VCYDKSFPISHVR), 40Gap 27 (SRPTEKNVFIV), and 37,43Gap 27 (SRPTEKTIFII) with homology to the first (Gap 26) and second (Gap 27) extracellular Cx-loops were from Severn Biotech. The pH of peptide containing MOPS was carefully adjusted to 7.4 on each occasion. For endothelial cell loading, peptides with homology to the cytoplasmic loop amino acids 130 to 140 of rCx37 (EHQMAKISVAE) and mCx40 (AELSCWKEVDG) were obtained from Auspep (Parkville, Australia; 95% purity). Peptides targeting this region have been used previously as controls in studies examining the potential blocking action of extracellularly applied Gap 26 and Gap 27 peptides.25,26
Statistical Analysis
Peak increases in outer diameter at each concentration of ACh were used to construct cumulative concentration-response curves (10 to 3 eol/L) and EC50 derived using a curve-fitting program (Prism 4.0). All data are means±SEM of n replicates, with n representing individual rats. Fluorescence data are summarized as average fluorescence from 16 randomly chosen endothelial cells (F), sampled at 5-second intervals relative to basal intensity (F0). Analysis used either paired Student t test or 1-way ANOVA with Bonferroni post test and rejection of the null hypothesis at P0.05.
Results
Hyperosmolarity Increases Endothelial Cell Pinocytic Vesicles Without Changing Cell Integrity
Pinocytic loading solutions did not alter the ultrastructural integrity of the endothelium (online Figure I in the data supplement available at http://circres.ahajournals.org). Tissue fixed in hypertonic loading solution showed an 4-fold increase in intracellular vesicles within endothelial cells (online Figure I) and a reduction in cell volume. In contrast, after completing the loading protocol, the number of vesicles and the cell volume returned to control levels. Neither treatment altered ultrastructural morphology in the adjacent smooth muscle cells. Furthermore, the loading protocol damaged <5% of endothelial cells, assessed by staining dead cells with propidium iodide (online Figure IIA).
Changing Luminal Osmolarity Selectively Loads Endothelial Cells With Fluorescent Markers
A 10-minute hypertonic exposure was optimal for selective cellular loading with carboxyfluorescein (Figure 1A). Increasing intraluminal exposure to 20 minutes caused substantive loading of the smooth muscle (Figure 1B). Abluminal exposure of pressurized arteries to similar solutions directly loaded the smooth muscle and perivascular nerves with carboxyfluorescein (online Figure IIA). Selective, extensive endothelial cell loading also followed 10-minute luminal incubation with hypertonic solution containing Alexa Fluor 633 IgG or FITC-labeled dextran and without significant smooth muscle loading (Figure 1A). Unlike carboxyfluorescein, increasing luminal exposure to 40 minutes did not load the smooth muscle with these larger molecular-mass dyes, apart from the occasional cell.
After endothelial loading with carboxyfluorescein (for 10 minutes), subsequent images at 5-minute intervals revealed gradual loss of fluorescence (Figure 1B). By 30 minutes, most dye fluorescence was gone. This did not appear to reflect bleaching, as similar loss was apparent if fluorescence was assessed only once, at 30 minutes postloading. Following the loss of carboxyfluorescein fluorescence from the endothelium, the same cells could be "reloaded" with dye by simply repeating the osmotic loading protocol, with vascular reactivity retained. As IgG or dextran loading gave a stable fluorescent signal over time, carboxyfluorescein may be actively extruded by the cells.
Changing Luminal Osmolarity Does Not Alter Functionality in Pressurized Arteries
Endothelium-dependent relaxation to ACh (10 nmol/L to 3 eol/L; EC50, 177 nmol/L; maximal relaxation [Rmax], 93±3%; n=8) was not altered by the osmotic loading protocol (EC50, 170 nmol/L; Rmax, 90±2%; n=8; Figure 2A and 2B). Likewise, neither the initial contraction to 3 eol/L PE (77.4±1.9%; n=13) nor dilation to 3 eol/L levcromakalim (90.3±1.9%, n=12) or 15 eol/L cyclopiazonic acid (84.2±3.3%, n=13) was altered.
As an increase in endothelial cell [Ca2+]i to ACh is critical for endothelial and subsequent smooth muscle cell hyperpolarization,22 endothelial cell viability was assessed as time course of rises in [Ca2+]i in individual endothelial cells in pressurized arteries (Figure 2C and 2D). Before and after the osmotic loading protocol, ACh stimulated sustained and oscillating rises in endothelial cell [Ca2+]i (Figure 2C), the oscillations remaining asynchronous between cells, with repetitive waves of Ca2+ moving longitudinally across endothelial cells. Therefore, the loading protocol alone did not modify the ability of the endothelium to increase [Ca2+]i and evoke EDHF responses.
EDHF-Mediated Dilatation to ACh: Effect of Endothelial Cell Loading With Cx Antibodies and Mimetic Peptides
Endothelial cell loading with antibodies directed against the intracellular carboxy-terminus of Cx40 did not alter EDHF-mediated dilation in arteries submaximally contracted to 0.9 eol/L PE (control: EC50, 130 nmol/L; Rmax, 97.0±0.8%; after antibody: EC50, 99 nmol/L; Rmax, 97.6±0.8%; n=4; Figure 3A and 3B) or in the additional presence of ouabain (online Figure IIIA). In marked contrast, raising the concentration of PE to evoke maximal contraction (and inhibit the simultaneous influence of diffusible EDHF apparent during submaximal stimulation11,20) was associated with block of EDHF dilation (Figure 3A and 3B and online Figure IIIA). In vessels contracted with 5 eol/L PE, the Rmax was only 26.7±8.8% (n=11), whereas subsequent addition of 3 eol/L levcromakalim evoked complete dilation (96.9±0.9%, n=4; Figure 3A). Cx40 antibody did not alter either the increase or oscillations in endothelial cell [Ca2+]i to ACh despite block of dilation (Figure 3C), and staining with propidium iodide indicated damage to <8% of cells. In contrast, endothelial loading with antibodies against both the carboxy-terminus and the cytoplasmic loop of Cx37 and Cx43 failed to modify EDHF dilation significantly (Figure 4A and 4B). Furthermore, if Cx40 antibody was loaded before block of NOS, dilation was sustained until NG-nitro-L-arginine methyl ester was applied (online Figure IIIB).
Overall, this profile was mirrored by endothelial cell loading with intracellular Cx-mimetic peptides. Cx40 peptide suppressed EDHF dilation in arteries contracted with 5 eol/L PE, whereas the Cx37 peptide had no significant effect (Figure 4C).
Effect of Extracellular Cx-Mimetic Peptides and Carbenoxolone Against EDHF Dilation
Simultaneous intra- and extraluminal incubation of pressurized arteries (for 2 hours; 3x40 minutes exposures) with a combination of 43Gap 26, 40Gap 27, and 37,43Gap 27 (each 300 eol/L) failed to modify EDHF dilation to ACh (Figure 5A). In contrast, whereas the gap junction blocker carbenoxolone (100 eol/L) did not alter the amplitude of, or oscillations in, endothelial increases in [Ca2+]i (Figure 5C and 5D), it completely blocked EDHF dilation (Figure 5B).
Ultrastructural Localization of Cx40
Cx40 was localized to points of close association between endothelial and smooth muscle cells that also exhibited small regions of pentalaminar membrane, characteristic of MEGJs (Figure 6).
Discussion
There are 2 major observations in this study. First, we show that it is possible to rapidly and selectively load large membrane impermeant molecules, including antibodies, into the endothelium of arteries under physiological pressures, without disrupting function. Second, using this approach combined with high-resolution ultrastructural methods, we provide direct evidence that MEGJs play a central role in EDHF dilation and highlight a primary role for Cx40 in the radial transfer of this signal from the endothelium.
The basis for cell loading is the normal process of endocytosis, describing the nondisruptive passage of macromolecules across the plasma membrane by phagacytosis or "cell eating" and pinocytosis or "cell drinking."27 Only a few specialized cells, such as macrophages and neutrophils, use phagacytosis, whereas virtually all cells can carry out some form of pinocytosis. An ability to increase the formation of pinocytic vesicles in cells was first described in L929 cells.18 Vesicles were then ruptured (by altering the osmolarity), enabling intracellular release of horseradish peroxidase, anti-ricin antibodies, or dextran (70 kDa), an effect termed osmotic lysis of pinosomes. A similar approach in whole tissue (rabbit mesenteric artery) used a commercially available kit,19 but, although we could load rat mesenteric smooth muscle cells with fluorescently-labeled dextran using kit reagents, endothelium-dependent vasodilation was lost after only 5 minutes of exposure. We, therefore, developed the original protocol for use in pressurized arteries.
Selective endothelial cell incorporation of macromolecules was facilitated by using artery segments cannulated at both ends. Luminal solution was changed independently of the bathing solution, and a cycle of 10 minutes hypertonic, 2 minutes hypotonic, before returning to isosmotic conditions, proved optimal to load the endothelium with fluorescent markers of differing structure and size: carboxyfluorescein, FITC-dextran, and Alexa Fluor 633 IgG. If the hyperosmotic phase was prolonged, carboxyfluorescein appeared in the smooth muscle, which might reflect movement through the myoendothelial gap junctions present in this artery.28 Molecules up to 1 kDa can pass through gap junctions; therefore, this seems a likely explanation for carboxyfluorescein (376 Da),29,30 supported by the observation that equivalent muscle loading was not evident after endothelial cell loading with either IgG (150 kDa) or dextran (3 kDa). In contrast to these markers, carboxyfluorescein-fluorescence within the endothelium decreased rapidly, presumably reflecting, in part, the activity of ATP-binding cassette transporters within the endothelial membrane, which extrude a range of molecules, including carboxyfluorescein and calcein.31,32
Following technique validation, the role of MEGJs in EDHF dilation was probed. Agonist activation of endothelial cells in the mesenteric artery increases [Ca2+]i22,33 and stimulates hyperpolarization,3,15,34,35 followed by hyperpolarization and relaxation in the adjacent smooth muscle.36,37 The smooth muscle effects are suggested to reflect the action of a diffusible EDHF, possibly K+, and the spread of hyperpolarization from the endothelium through MEGJs.9,38 That radial heterocellular coupling is important in local, endothelium-driven dilation is supported in the rat mesenteric artery by the presence of heterocellular membrane associations of pentalaminar structure, characteristic of MEGJs.12,28 However, although rat mesenteric artery endothelial cells express Cx37, Cx40, and possibly Cx43 (but not Cx4539), the precise profile in the smooth muscle is not clear.23,39 Thus, it is not known which Cxs are expressed at the MEGJs.
Endothelial cell hyperpolarization reflects the opening of Ca2+-activated K+ channels22,34,35 amplified by K+ efflux through the inwardly rectifying K+ channels focused within the endothelium of the mesenteric artery.40,41 Consequent extracellular K+ accumulation then stimulates smooth muscle hyperpolarization and relaxation by activating Na+/K+-ATPase.11 However, if background contraction with PE is increased, K+ loses the ability to serve as an EDHF, whereas smooth muscle cell hyperpolarization and relaxation to ACh is sustained by a persistent parallel mechanism.11 Loss of K+ as an EDHF appears to be caused by BKCa-channel activation in the smooth muscle (by PE-mediated depolarization and rises in [Ca2+]i with consequent K+ efflux eventually maximally activating the Na+/K+-ATPase), as it is reversed in the presence of the BKCa blocker iberiotoxin.20,42,43 In the present study, possible input from K+ as an EDHF was blocked by high levels of background arterial contraction to 5 eol/L PE,11 isolating EDHF dilation, which may be suggested to be attributable to MEGJs in this artery.9
The present study provides direct evidence that the MEGJs mediate the persistent dilation. Endothelial cell loading with antibody to the intracellular carboxy-terminal region of Cx40 completely abolished EDHF dilation, without altering the increase in [Ca2+]i necessary to initiate hyperpolarization, when arteries were stimulated with higher but not lower concentrations of PE. In these experiments, direct muscle dilation to levcromakalim3,35 persisted, and in separate experiments, EDHF dilation was maintained after endothelial loading with inhibitory molecules against different intracellular regions of Cx37 or Cx43. Interestingly, when arteries were contracted with low concentrations of PE, Cx40 antibody-insensitive ACh relaxation appeared not solely attributable to release of K+ acting on smooth muscle Na+/K+-ATPase. Although ouabain alone shifted the ACh-dilation curve, it did not abolish dilation in the presence of Cx40 antibody. This suggests compensatory pathways can still operate under low tone, possibly via electrical coupling through MEGJs comprised of Cx37.
A role for Cx40 in heterocellular signaling is supported by ultrastructural localization of this Cx to the end of endothelial-derived projections associated with MEGJs.12,28 There is only 1 other equivalent ultrastructural report, localizing Cx40 within the membrane of gap junctions between endothelial cells and smooth muscle cells modified to secrete renin in the terminal region of kidney afferent arterioles.44 Whether or not this Cx is the primary Cx for myoendothelial signaling remains to be seen.
The Cx-mimetic peptides loaded into the endothelium were homologous to segments of the intracellular cytoplasmic loop sequences (amino acids 130 to 140) of Cx37 and Cx40. Patch-clamp experiments suggest this region may represent a Cx-gating element,45 making it a potentially useful target, particularly because there is no sequence homology here between Cx37 and Cx40. An ability to block EDHF dilation with Cx40 intracellular-mimetic peptide contrasts with the lack of effect of prolonged extracellular application of a triple combination of Gap peptides targeting extracellular Cx regions. This triple combination has been suggested as the most effective and reliable way to block vascular gap junctions.7 Previous use of Gap peptides against EDHF responses in the mesenteric artery provide inconsistent data. Gap 27 failed significantly to reduce EDHF hyperpolarization in 1 study,46 reduced it by 65% in another,12 and almost abolished EDHF dilation in pressurized mesenteric arteries.38 Interestingly, in the latter, Gap peptide was applied via the lumen of pressurized arteries, so it would not be predicted to even reach the MEGJs, due to the tight junctions between the endothelial cells. In any event, the reason such discrepancies exist is not clear. One possibility may be an alteration in pH caused by the peptides. In our experiments, where the triple peptide combination was applied within the lumen and around the artery at the same time, the reduction in pH caused by the peptides was corrected. Lack of effect of Gap peptides against EDHF dilation is similar to the lack of effect of 37,40Gap 26 in mouse mesenteric arteries, where MEGJs are abundant.47
Data linking cell-specific Cx isoforms to functional responses in vascular tissue are limited. A major physiological role for Cx40 is indicated by studies with knockout (Cx40eC/eC) mice. These mice are hypertensive, and, even without NOS block, cremaster arteriolar dilation to 10 eol/L ACh was attenuated.48 Furthermore, the ability of arterioles to conduct vasodilatation longitudinally through the vessel wall was markedly attenuated in these animals,48,49 a response normally ascribed to the endothelium.3,50 However, it remains possible that long-term compensatory mechanisms may underlie these changes in genetically modified mice, whereas short-term block in the present study circumvents this potential problem. The only equivalent long-term data come from a recent study with Gap peptides in the rabbit iliac artery (10 smooth muscle layers), suggesting both Cx37 and Cx40 are essential for hyperpolarization to spread through MEGJs.16 However, in this particular artery, definitive identification of MEGJs, and the constitutive Cxs, is required.
In summary, by selectively loading inhibitory Cx antibodies and peptides into endothelial cells in pressurized rat mesenteric arteries, we have demonstrated directly a major functional role for the MEGJs known to occur in this artery. Furthermore, our data suggest that Cx40 is the critical Cx isoform involved in heterocellular signaling.
Acknowledgments
This work was supported by the Wellcome Trust (UK) and the British Heart Foundation.
References
Duling BR, Berne RM. Propagated vasodilation in the microcirculation of the hamster cheek pouch. Circ Res. 1970; 26: 163eC170.
Gustafsson F, Holstein-Rathlou N. Conducted vasomotor responses in arterioles: characteristics, mechanisms and physiological significance. Acta Physiol Scand. 1999; 167: 11eC21.
Takano H, Dora KA, Spitaler MM, Garland CJ. Spreading dilatation in rat mesenteric arteries associated with calcium-independent endothelial cell hyperpolarization. J Physiol. 2004; 556: 887eC903.
Emerson GG, Segal SS. Electrical coupling between endothelial cells and smooth muscle cells in hamster feed arteries: role in vasomotor control. Circ Res. 2000; 87: 474eC479.
Emerson GG, Segal SS. Electrical activation of endothelium evokes vasodilation and hyperpolarization along hamster feed arteries. Am J Physiol Heart Circ Physiol. 2001; 280: H160eCH167.
Busse R, Edwards G, Feletou M, Fleming I, Vanhoutte PM, Weston AH. EDHF: bringing the concepts together. Trends Pharmacol Sci. 2002; 23: 374eC380.
Griffith TM. Endothelium-dependent smooth muscle hyperpolarization: do gap junctions provide a unifying hypothesis Br J Pharmacol. 2004; 141: 881eC903.
Sandow SL. Factors, fiction and endothelium-derived hyperpolarizing factor. Clin Exp Pharmacol Physiol. 2004; 31: 563eC570.
Edwards G, Dora KA, Gardener MJ, Garland CJ, Weston AH. K+ is an endothelium-derived hyperpolarizing factor in rat arteries. Nature. 1998; 396: 269eC272.
Dora KA, McEvoy L, King M, Ings NT, Garland CJ. The intensity of agonist-stimulation influences the mechanism for relaxation in rat mesenteric arteries. In: Vanhoutte PM, ed. EDHF 2002. London: Taylor & Francis; 2003: 256eC260.
Dora KA, Garland CJ. Properties of smooth muscle hyperpolarization and relaxation to K+ in the rat isolated mesenteric artery. Am J Physiol Heart Circ Physiol. 2001; 280: H2424eCH2429.
Sandow SL, Tare M, Coleman HA, Hill CE, Parkington HC. Involvement of myoendothelial gap junctions in the actions of endothelium-derived hyperpolarizing factor. Circ Res. 2002; 90: 1108eC1113.
Beny J. Endothelial and smooth muscle cells hyperpolarized by bradykinin are not dye coupled. Am J Physiol. 1990; 258: H836eCH841.
Coleman HA, Tare M, Parkington HC. K+ currents underlying the action of endothelium-derived hyperpolarizing factor in guinea-pig, rat and human blood vessels. J Physiol (Lond). 2001; 531: 359eC373.
Tare M, Coleman HA, Parkington HC. Glycyrrhetinic derivatives inhibit hyperpolarization in endothelial cells of guinea pig and rat arteries. Am J Physiol Heart Circ Physiol. 2002; 282: H335eCH341.
Chaytor AT, Bakker LM, Edwards DH, Griffith TM. Connexin-mimetic peptides dissociate electrotonic EDHF-type signalling via myoendothelial and smooth muscle gap junctions in the rabbit iliac artery. Br J Pharmacol. 2005; 144: 108eC114.
Edwards G, Thollon C, Gardener MJ, Feletou M, Vilaine J, Vanhoutte PM, Weston AH. Role of gap junctions and EETs in endothelium-dependent hyperpolarization of porcine coronary artery. Br J Pharmacol. 2000; 129: 1145eC1154.
Okada CY, Rechsteiner M. Introduction of macromolecules into cultured mammalian cells by osmotic lysis of pinocytic vesicles. Cell. 1982; 29: 33eC41.
Vequaud P, Thorin E. Endothelial G protein -subunits trigger nitric oxide- but not endothelium-derived hyperpolarizing factor-dependent dilation in rabbit resistance arteries. Circ Res. 2001; 89: 716eC722.
Dora KA, Ings NT, Garland CJ. KCa channel blockers reveal hyperpolarization and relaxation to K+ in the rat isolated mesenteric artery. Am J Physiol Heart Circ Physiol. 2002; 283: H606eCH614.
Coats P, Hillier C. Determination of an optimal axial-length tension for the study of isolated resistance arteries on a pressure myograph. Exp Physiol. 1999; 84: 1085eC1094.
McSherry IN, Spitaler MM, Takano H, Dora KA. Endothelial cell Ca2+ increases are independent of membrane potential in pressurized rat mesenteric arteries. Cell Calcium. 2005; 38: 23eC33.
Gustafsson F, Mikkelsen HB, Arensbak B, Thuneberg L, Neve S, Jensen LJ, Holstein-Rathlou NH. Expression of connexin 37, 40 and 43 in rat mesenteric arterioles and resistance arteries. Histochem Cell Biol. 2003; 119: 139eC148.
Sandow SL, Looft-Wilson R, Doran B, Grayson TH, Segal SS, Hill CE. Expression of homocellular and heterocellular gap junctions in hamster arterioles and feed arteries. Cardiovasc Res. 2003; 60: 643eC653.
Chaytor AT, Evans WH, Griffith TM. Peptides homologous to extracellular loop motifs of connexin 43 reversibly abolish rhythmic contractile activity in rabbit arteries. J Physiol. 1997; 503: 99eC110.
Earley S, Resta TC, Walker BR. Disruption of smooth muscle gap junctions attenuates myogenic vasoconstriction of mesenteric resistance arteries. Am J Physiol Heart Circ Physiol. 2004; 287: H2677eCH2686.
Conner SD, Schmid SL. Regulated portals of entry into the cell. Nature. 2003; 422: 37eC44.
Sandow SL, Hill CE. Incidence of myoendothelial gap junctions in the proximal and distal mesenteric arteries of the rat is suggestive of a role in endothelium-derived hyperpolarizing factor-mediated responses. Circ Res. 2000; 86: 341eC346.
Evans WH, Martin PE. Gap junctions: structure and function. Mol Membr Biol. 2002; 19: 121eC136.
Unger VM, Kumar NM, Gilula NB, Yeager M. Three-dimensional structure of a recombinant gap junction membrane channel. Science. 1999; 283: 1176eC1180.
Rychlik B, Balcerczyk A, Klimczak A, Bartosz G. The role of multidrug resistance protein 1 (MRP1) in transport of fluorescent anions across the human erythrocyte membrane. J Membr Biol. 2003; 193: 79eC90.
Bachmeier CJ, Trickler WJ, Miller DW. Drug efflux transport properties of 2',7'-bis(2-carboxyethyl)-5(6)-carboxyfluorescein acetoxymethyl ester (BCECF-AM) and its fluorescent free acid, BCECF. J Pharm Sci. 2004; 93: 932eC942.
Dora KA, Hinton JM, Walker SD, Garland CJ. An indirect influence of phenylephrine on the release of endothelium-derived vasodilators in rat small mesenteric artery. Br J Pharmacol. 2000; 129: 381eC387.
Hinton JM, Langton PD. Inhibition of EDHF by two new combinations of K+-channel inhibitors in rat isolated mesenteric arteries. Br J Pharmacol. 2003; 138: 1031eC1035.
White R, Hiley CR. Hyperpolarisation of rat mesenteric endothelial cells by ATP-sensitive K+ channel openers. Eur J Pharmacol. 2000; 397: 279eC290.
Garland CJ, McPherson GA. Evidence that nitric oxide does not mediate the hyperpolarization and relaxation to acetylcholine in the rat small mesenteric artery. Br J Pharmacol. 1992; 105: 429eC435.
Crane GJ, Gallagher NT, Dora KA, Garland CJ. Small and intermediate calcium-dependent K+ channels provide different facets of endothelium-dependent hyperpolarization in rat mesenteric artery. J Physiol. 2003; 553: 183eC189.
Doughty JM, Boyle JP, Langton PD. Potassium does not mimic EDHF in rat mesenteric arteries. Br J Pharmacol. 2000; 130: 1174eC1182.
Kansui Y, Fujii K, Nakamura K, Goto K, Oniki H, Abe I, Shibata Y, Iida M. Angiotensin II receptor blockade corrects altered expression of gap junctions in endothelial cells from hypertensive rats. Am J Physiol Heart Circ Physiol. 2004; 287: H216eCH224.
Doughty JM, Boyle JP, Langton PD. Blockade of chloride channels reveals relaxations of rat small mesenteric arteries to raised potassium. Br J Pharmacol. 2001; 132: 293eC301.
Crane GJ, Walker SD, Dora KA, Garland CJ. Evidence for a differential cellular distribution of inward rectifier K channels in the rat isolated mesenteric artery. J Vasc Res. 2003; 40: 159eC168.
Richards GR, Weston AH, Burnham MP, Feletou M, Vanhoutte PM, Edwards G. Suppression of K+-induced hyperpolarization by phenylephrine in rat mesenteric artery: relevance to studies of endothelium-derived hyperpolarizing factor. Br J Pharmacol. 2001; 134: 1eC5.
Walker SD, Dora KA, Ings NT, Crane GJ, Garland CJ. Activation of endothelial cell IKCa and 1-ethyl-2-benzimidazolinone evokes smooth muscle hyperpolarization in rat isolated mesenteric artery. Br J Pharmacol. 2001; 134: 1548eC1554.
Haefliger JA, Demotz S, Braissant O, Suter E, Waeber B, Nicod P, Meda P. Connexins 40 and 43 are differentially regulated within the kidneys of rats with renovascular hypertension. Kidney Int. 2001; 60: 190eC201.
Seki A, Duffy HS, Coombs W, Spray DC, Taffet SM, Delmar M. Modifications in the biophysical properties of connexin43 channels by a peptide of the cytoplasmic loop region. Circ Res. 2004; 95: e22eCe28.
Edwards G, Feletou M, Gardener MJ, Thollon C, Vanhoutte PM, Weston AH. Role of gap junctions in the responses to EDHF in rat and guinea-pig small arteries. Br J Pharmacol. 1999; 128: 1788eC1794.
Dora KA, Sandow SL, Gallagher NT, Takano H, Rummery NM, Hill CE, Garland CJ. Myoendothelial gap junctions may provide the pathway for EDHF in mouse mesenteric artery. J Vasc Res. 2003; 40: 480eC490.
de Wit C, Roos F, Bolz SS, Kirchhoff S, Kruger O, Willecke K, Pohl U. Impaired conduction of vasodilation along arterioles in connexin40-deficient mice. Circ Res. 2000; 86: 649eC655.
Figueroa XF, Paul DL, Simon AM, Goodenough DA, Day KH, Damon DN, Duling BR. Central role of connexin40 in the propagation of electrically activated vasodilation in mouse cremasteric arterioles in vivo. Circ Res. 2003; 92: 793eC800.
Emerson GG, Segal SS. Endothelial cell pathway for conduction of hyperpolarization and vasodilation along hamster feed artery. Circ Res. 2000; 86: 94eC100.(Simon Mather, Kim A. Dora)