Evidence for the involvement of myoendothelial gap junctions in EDHF-mediated relaxation in the rat middle cerebral artery

Elke M. Sokoya, Alan R. Burns, Christopher T. Setiawan, Harold A. Coleman, Helena C. Parkington, Marianne Tare


The mechanisms underlying endothelium-dependent hyperpolarizing factor (EDHF) in the middle cerebral artery (MCA) remain largely unresolved. In particular, very little is known regarding the way in which the signal is transmitted from endothelium to smooth muscle. The present study tested the hypothesis that direct communication via myoendothelial gap junctions contributes to the EDHF response in the male rat MCA. EDHF-mediated dilations were elicited in rat MCAs by luminal application of ATP or UTP in the presence of Nω-nitro-l-arginine methyl ester and indomethacin. Maximum dilation to luminal ATP (10−4 M) was reduced significantly after incubation with a gap peptide cocktail (9 ± 4%, n = 6) compared with a scrambled gap peptide cocktail (99 ± 1%, n = 6, P < 0.05). A gap peptide cocktail had no effect on amplitude of endothelial cell hyperpolarization in response to 3 × 10−5 M UTP (22 ± 3 vs. 22 ± 1 mV, n = 4), whereas smooth muscle cell hyperpolarization was significantly attenuated (17 ± 1 vs. 6 ± 1 mV, n = 4, P = 0.004). Connexin (Cx) 37 was localized to smooth muscle and Cx43 to endothelium, whereas Cx40 was found in endothelium and smooth muscle. Electron microscopy revealed the existence of frequent myoendothelial junctions. The total number of myoendothelial junctions per 5 μm of MCA sectioned was 2.5 ± 0.5. Our results suggest that myoendothelial communication contributes to smooth muscle cell hyperpolarization and EDHF dilation in male rat MCA.

  • connexins
  • endothelium-derived hyperpolarizing factor
  • vascular smooth muscle

endothelium-derived hyperpolarizing factor (EDHF) remains an incompletely described phenomenon, and its identity appears to vary in different vascular beds (2). In rat middle cerebral artery (MCA), we know that it originates in the endothelium and is not nitric oxide (NO), prostacyclin, or a cyclooxygenase metabolite (41). Stimulation of endothelial cells (ECs) evokes an increase in EC intracellular Ca2+ (24), triggering the opening of EC intermediate Ca2+-sensitive K+ (KCa) channels and leading to EC hyperpolarization (25). These events in the ECs are followed by hyperpolarization of the underlying smooth muscle cells (SMCs) (41). There are two possibilities by which hyperpolarization of the EC may be communicated to the SMC: 1) release of a paracrine factor from the EC that diffuses to the SMC or 2) a functional communicating junction between the EC and the SMC.

In some vessels, a paracrine factor has been shown to mediate SMC hyperpolarization. One that has been considered extensively is the family of arachidonic acid metabolites known as the epoxyeicosatrienoic acids (EETs). Elevation in intracellular EC Ca2+ can activate phospholipase A2, releasing arachidonic acid, which is a precursor of EETs. The EETs are thought to diffuse to the SMC and activate the large-conductance, Ca2+-activated K+ (BKCa) channels, leading to hyperpolarization of the smooth muscle (3, 14). However, the EDHF response in rat MCA is resistant to the BKCa inhibitor iberiotoxin (25) and inhibitors of EET formation (39).

In other vessels, K+ has been shown to act as a paracrine factor. The efflux of K+ through EC KCa channels results in accumulation of K+ in the extracellular space, which activates inwardly rectifying K+ (Kir) channels and Na+-K+-ATPase on the SMC, thereby producing hyperpolarization (8). However, events in the MCA are controversial, with evidence for (31) and against (41) involvement of Kir channels and Na+-K+-ATPase. Such discrepancies may reflect differences between agonists or other methodologies.

H2O2 has also been considered to be a paracrine factor because it hyperpolarizes the SMC by activation of KCa channels. There is evidence supporting a role for H2O2 in some vessels (29, 30, 32), but not others (10, 18), including the rat MCA (39).

An alternative mechanism by which events in the EC may be communicated to the SMC is direct contact. Ultrastructural evidence suggests that ECs can extend cellular protrusions through perforations in the internal elastic lamina (IEL) to come into close contact with SMCs [myoendothelial gap junctions (MEGJs)] in rabbit carotid artery (36) and rat mesenteric artery (34, 35). Furthermore, electrical coupling between ECs and SMCs has been demonstrated in hamster retractor muscle feed arteries (11), rat mesenteric arteries (35), and guinea pig submucosal arterioles (7). Further support for a role of gap junctions in mediating EDHF has been provided by the use of connexin inhibitor proteins (1, 4, 9). However, not all blood vessels possess MEGJs (9, 20, 35).

Evidence to date suggests that the existence of a universal EDHF is highly unlikely (2). Vessel size, location, and age may be critical factors determining the underlying mechanism of EDHF. In particular, mechanisms of dilation in cerebral arteries can be different from those in peripheral arteries. For example, activation of KCa channels alone is sufficient to elicit EC hyperpolarization and EDHF-mediated dilation in rat cerebral arteries (25), whereas KCa channels and a small-conductance Ca2+-activated K+ (SKCa) channel are involved in peripheral arteries (8). Since we have no evidence that a paracrine factor mediates EDHF in the rat MCA, we turned to the possibility that there is direct communication between ECs and SMCs. The present study tested the hypothesis that MEGJs exist in MCAs and that they provide a pathway for spread of current between ECs and SMCs, which contributes to EDHF in this vascular bed.


Experiments were carried out in accordance with the National Institutes of Health guidelines for the care and use of laboratory animals and approved by the Animal Protocol Review Committee at Baylor College of Medicine and conducted in accordance with the National Health and Medical Research Council of Australia and approved by the Monash University Animal Ethics Committee.

Male Long-Evans rats (275–325 g body wt) were housed under a 12:12-h light-dark cycle with unrestricted access to food and water.

Harvesting and Mounting Cerebral Vessels

Male Long-Evans rats (n = 12) were placed in an anesthetic chamber, allowed to spontaneously breathe isoflurane, and then decapitated. The brain was removed from the cranium and placed in cold physiological salt solution (PSS). The MCA was excised, cleaned of surrounding connective tissue, and cannulated with micropipettes in a custom-made Teflon-coated vessel chamber (ChuelTech, Houston, TX). The temperature was maintained at 37°C via a microprocessor-controlled heating block. PSS was circulated abluminally (6 ml) and perfused luminally. Continuous monitoring of intraluminal pressure was achieved via in-line transducers, which were connected to two strain gauge panel meters (Omega, Stamford, CT). After the vessels were mounted, the proximal and distal tubings were clamped and intraluminal pressure was monitored to test for leaks. Vessels that did not maintain a steady pressure were discarded. The vessel chamber was mounted on the stage of an inverted microscope. Transmural pressure was set at 85 mmHg with a flow of 150 μl/min through the lumen, and the vessels were allowed to equilibrate for 1 h. During this time, they developed spontaneous tone by constricting from their fully dilated diameter at initial pressurization. After vessel tone was developed, the experiment was initiated (see EDHF-Mediated Dilations).

EDHF-Mediated Dilations

After the vessels developed spontaneous tone, the luminal and abluminal compartments were exposed to a cocktail of connexin mimetic (gap) peptides or their scrambled sequences and incubated for 1 h (each peptide at 300 μM). The gap peptide cocktail (courtesy of Dr. Dale Pelligrino, University of Chicago at Illinois) consisted of a combination of the synthetic peptide homologous to a region of the second extracellular loop of connexins 37 and 43 (Cx37 and Cx43; SRPTEKTIFII; 37,43Gap 27), the peptide homologous to a region of the first extracellular loop of Cx43 (VCYDKSFPISHVR, 43Gap 26), and the peptide homologous to a region of the second extracellular loop of connexin 40 (Cx40; SRPTEKNVFIV, 40Gap 27) (5). A cocktail of scrambled peptides, i.e., negative controls, consisted of the scrambled peptide of 37,43Gap 27 (FKTIRTISIEP), the scrambled peptide of 43Gap 26 (PSDVFRSCVKHYI), and the scrambled peptide of 40Gap 27 (VTNIEVPKSFR). All peptides were made to a purity >95%. S-nitroso-N-acetylpenicillamine (SNAP, 5 × 10−5 M) was used to verify that SMC vasodilator function was intact in the presence of gap peptide inhibition.

Nω-nitro-l-arginine methyl ester (l-NAME, 3 × 10−5 M) and indomethacin (10−5 M) were added to the luminal and abluminal solutions for 30 min to remove the NO synthase and cyclooxygenase contributions, respectively. EDHF-mediated dilations to luminal application of 10−7–10−4 M ATP were assessed by concentration-response curves. Experiments were terminated by replacement of PSS with Ca2+-free PSS containing 1 mM EGTA to determine the maximum diameter of the vessel.

Electrophysiological Experiments

Long-Evans male rats (n = 30) were anesthetized with chloroform and decapitated. The MCA was isolated as described above. One end was cut open longitudinally and pinned to the base of a vessel chamber with the ECs facing up, allowing for EC impalement. The other end remained intact, with the outer SMCs facing up, allowing for SMC impalement. Preparations were superfused with PSS and allowed to equilibrate for 30 min. The vessels were exposed to l-NAME and indomethacin in all experiments.

EC and SMC membrane potentials were recorded independently using intracellular glass micropipettes; the tips of these micropipettes were filled with 2% Lucifer yellow and backfilled with 1 M KCl; tip resistance was ∼100 MΩ (35). Lucifer yellow diffused into the cell during impalement, permitting identification of each impaled cell. Lucifer yellow was the dye of choice to unequivocally identify the impaled cell because it does not spread to the other cell type (23). Criteria for successful impalement of a cell included 1) abrupt change in voltage on cell entry, 2) return to baseline voltage on exiting the cell, and 3) maintenance of preentry electrode resistance.

The EC was stimulated using a 1-min application of 3 × 10−5 M UTP. We previously showed that ATP and UTP produce comparable EDHF-mediated dilations (15, 41) and SMC hyperpolarization in rat MCAs (unpublished data). MCAs were incubated for 1 h with the gap peptide cocktail or vehicle (PSS). In some experiments, membrane potential responses to UTP were assessed in the presence of charybdotoxin (5 × 10−8 M) and apamin (5 × 10−7 M; both from Auspep), inhibitors of KCa and SKCa channels, respectively.

Connexin Immunofluorescence

Preparation of whole mounts.

Rats were anesthetized with pentobarbital sodium (60 mg/kg ip), and the vascular system was perfused by a transcardial approach with heparinized saline solution (20 U/ml). Once cleared of blood, the animals were perfusion fixed with 2% paraformaldehyde in 0.1 M phosphate buffer (PBS). After the animals were decapitated, MCAs were removed and immersion fixed in the same solution. The MCA was cut open longitudinally, with the intimal side facing up, and pinned flat on silicone elastomer. These preparations were stored in PBS at 4°C until they were used in the immunofluorescence protocol.

Preparation of cross sections.

Rats were prepared as described above. MCAs were dissected from the brain, placed in optimum cutting temperature compound, and then snap frozen in dry ice-chilled 2-methylbutane. Tissues were cut at 10-μm-thick cross sections using a cryostat (−20°C), mounted on glass slides, and allowed to air dry. The sections were stored at −20°C until they were used in the immunofluorescence protocol.

Immunofluorescence light microscopy.

MCA whole mounts or cross sections were washed in cold PBS, permeabilized (0.1% Triton X-100), and blocked (0.1% Tween 20 and 10% goat or donkey serum in PBS). Primary antibody was added for 2 h at room temperature (Cx43), overnight at room temperature (Cx40), or overnight at 4°C (Cx37). Cx37 and Cx43 were localized using rabbit polyclonal antipeptide antibodies (6 μg/ml; Alpha Diagnostics and Sigma, respectively). Cx40 was localized using a guinea pig polyclonal antibody (courtesy of Dr. Robert Gourdie, University of South Carolina).

After the sections were washed twice, labeling was assessed by secondary detection with Alexa Fluor 488 fluorochrome conjugated to goat anti-rabbit IgG (Cx43), donkey anti-rabbit FITC (Cx37), or goat anti-guinea pig biotin (Sigma; 1:250 dilution) followed by streptomycin-fluorescein (Cx40; Amersham; 1:250 dilution).

The specimens were treated with 4′,6-diamidino-2-phenylindole (1 μM; Molecular Probes) for nuclear detection and mounted in Airvol. Aortic sections were used as positive control for Cx43 to demonstrate the specificity of the antibody. For negative controls, sections were treated with nonimmune rabbit or guinea pig IgG, and the primary antibody was omitted. Deconvolution microscopy (DeltaVision, Applied Precision, Issaquah, WA) was used to evaluate connexin immunolabeling.

Electron Microscopy

Anesthetized rats (n = 3) were perfusion fixed as described above. Fixation was achieved using Sorenson's PBS containing 3% glutaraldehyde, and the brain was removed and immersed in the fixative overnight at 4°C. On the following day, the MCA was dissected from the brain and placed in PBS at 4°C. Tissue samples were postfixed in 1% tannic acid (5 min) followed by 1% osmium tetroxide (1 h) and then aqueous uranyl acetate (1 h). Samples were subsequently dehydrated in a graded ethanol series and embedded in Araldite resin, and ultrathin (100-nm) sections were obtained using an ultramicrotome (model RMC 7000, RMC) equipped with a diamond knife. The sections were stained with uranyl acetate and lead citrate before they were viewed with a JEOL 200CX electron microscope. Gap junctions are defined by the pentalaminar appearance of the membranes at points of cell-cell contact, where the distance between opposing membranes is ≤3.5 nm (33). In the present study, we use this criterion to define an MEGJ. In this case, the EC and SMC basal laminae are breached, and the two cell types are in close contact (≤3.5 nm). However, where the distance between the opposing membranes exceeded 3.5 nm, this type of cell-cell association was defined as a myoendothelial junction (MEJ).

Reagents and Buffers

All chemicals used to assess EDHF-mediated dilations were purchased from Sigma. The composition of PSS was as follows (in mM): 119 NaCl, 26 NaHCO3, 4.7 KCl, 1.18 KH2PO4, 1.17 MgSO4, 1.6 CaCl2, 5.5 glucose, and 0.026 EDTA. Stock solutions of ATP (10−2 M), UTP (10−2 M), and l-NAME (3 × 10−2 or 10−1 M) were prepared in distilled water, divided into aliquots, and frozen. Indomethacin stock solution (10−2 M) was prepared in Na2CO3-distilled water (1:1 by weight). Gap peptides and scrambled peptides were dissolved directly in PSS.

Data Analysis and Calculations

Values are means ± SE; n indicates the number of animals tested. Diameter measurements were averaged over 2 min immediately after luminal exposure to ATP. Changes in vascular diameter are presented as percentage of the maximum diameter of the MCAs, as described previously (15).

Statistical comparisons of the concentration-response curves to ATP were made using a two-way analysis of variance with repeated measures, and multiple comparisons were made using a Student-Newman-Keuls test. For the electrophysiological data, statistical significance was tested using a one-way analysis of variance with repeated measures. Post hoc comparisons were made using a Student-Newman-Keuls test. Differences were considered significant at error probabilities <0.05 (P < 0.05).


Effect of Gap Peptides on EDHF-Mediated Dilations

After equilibration, resting MCA diameter was similar between groups: 220 ± 15 and 229 ± 21 μm for scrambled and gap peptides, respectively. To assess the participation of gap junctions in the EDHF-mediated response, MCAs were tested randomly after incubation with a gap peptide cocktail (n = 6) or a scrambled peptide cocktail as the control (n = 6). Figure 1 illustrates the diameter changes in response to increasing concentrations of luminal ATP. The maximum dilation to ATP was significantly reduced in the presence of the gap peptide cocktail (9 ± 4%, n = 6) compared with scrambled gap peptides (99 ± 1%, n = 6, P < 0.05). Gap peptides selectively attenuated EDHF-mediated dilations without affecting dilation to luminal delivery of the NO donor SNAP (5 × 10−5 M; 95 ± 2%, n = 6, data not shown).

Fig. 1.

Concentration-response curve for endothelium-derived hyperpolarizing factor (EDHF)-mediated vasodilation evoked by luminal ATP in the presence of Nω-nitro-l-arginine methyl ester (l-NAME) and indomethacin. Dilation to luminal ATP was significantly attenuated after incubation with gap peptide cocktail (37,43Gap 27, 43Gap 26, and 40Gap 27 at 300 μM each, n = 6) compared with scrambled peptide cocktail (300 μM each, n = 6). *P < 0.05 vs. scrambled peptides (2-way repeated-measures ANOVA).

Effect of Gap Peptides on EDHF-Mediated Hyperpolarizations

Resting membrane potentials of dye-identified ECs and SMCs (Fig. 2A) were −34 ± 0.4 mV (n = 21) and −37 ± 1 mV (n = 22), respectively. UTP (3 × 10−5 M) produced hyperpolarizations of similar amplitudes in ECs (22 ± 2 mV, n = 12) and SMCs (16 ± 4 mV, n = 12). Exposure to gap peptides had no effect on the resting membrane potential in ECs (−35 ± 1 vs. −35 ± 4 mV, n = 4) or SMCs (−39 ± 2 vs. −37 ± 2 mV, n = 4).

Fig. 2.

A, top: membrane potential responses in endothelium and smooth muscle with and without gap peptides. UTP (3 × 10−5 M) was applied to the vessel for 1 min. Bottom: Lucifer yellow-filled endothelial and smooth muscle cells. Dashed lines show orientation of blood vessel; arrow indicates cell that was directly impaled with microelectrode. Scale bar, 50 μm. B: in the presence of gap peptide cocktail, endothelial cell hyperpolarization was sustained (n = 4) and smooth muscle cell hyperpolarization was significantly attenuated (n = 4) in the same arteries. *P < 0.05 vs. control (1-way ANOVA).

In the presence of gap peptides, UTP-mediated hyperpolarizations in SMCs were significantly attenuated (34 ± 8% of initial response, n = 4; Fig. 2B). In contrast, UTP-mediated hyperpolarizations in ECs were maintained (103 ± 16% of initial response, n = 4; Fig. 2B). Hyperpolarization to UTP in ECs and SMCs was abolished by charybdotoxin and apamin. EC hyperpolarization to UTP was 21 ± 5 mV (n = 4) and 0 ± 0 mV (n = 4) before and after charybdotoxin and apamin, and in SMCs the values were 17 ± 5 mV (n = 3) and 0 ± 0 mV (n = 3).

Localization of Connexin Immunofluorescence

Connexin immunofluorescence was detected using whole mount sections and frozen cross sections of rat MCAs. In whole mount sections, EC nuclei appeared in focus first (arranged along the longitudinal axis of the vessel wall), and then SMC nuclei appeared (arranged circumferentially). This allowed easy distinction between the two cell types. In pilot studies, the endothelial layer was confirmed with anti-von Willebrand factor (data not shown). In frozen cross sections, the SMCs were positively identified with anti-smooth muscle α-actin (data not shown). The IEL, separating the endothelial layer from the SMC, could be visualized by autofluorescence through the FITC filter set (green autofluorescence). The immunocytochemistry in cross sections revealed good evidence of punctate staining, which suggests connexin labeling at gap junctions.

When viewed en face, Cx40 and Cx43 fluorescence decorated the borders of each EC with punctate staining (Fig. 3, A and B). Electron microscopy subsequently confirmed the presence of gap junction plaques at EC-EC contacts (see Fig. 6). Cx40 was also found in SMCs (Fig. 4B), whereas Cx43 was also found in the adventitia and parenchyma (Fig. 4C). Note the intense Cx43 staining in the parenchyma, which has been reported previously in tissue directly beneath the pial surface (38). Positive staining for connexin proteins was confirmed when compared with sections incubated without the primary antibody, the antigen + the primary antibody, or the appropriate nonimmune IgG (data not shown). Neither Cx37 (data not shown) nor the IgG control showed EC staining (Fig. 3C). However, Cx37 was apparent in SMCs (Fig. 4A).

Fig. 3.

En face immunostaining in rat middle cerebral artery (MCA) of connexin (Cx) 40 (A), Cx43 (B), and IgG (C). Cx40 and Cx43 decorated borders of endothelial cells; IgG control showed no staining.

Fig. 4.

Immunofluorescence staining (green) of rat MCA cryosections with rabbit anti-rat Cx37 (A), guinea pig anti-rat Cx40 (B), and rabbit anti-rat Cx43 (C). Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (blue). Images represent a maximum projection composite taken from 60 optical sections spaced at 0.15-μm intervals. Arrowheads show punctate staining of Cx37 and Cx40 in smooth muscle and Cx43 staining in adventitia and parenchyma.

Identification of MEGJs

Using electron microscopy, we examined 100-nm-thick MCA serial cross sections and systematically evaluated perforations (holes) in the IEL. Sets of serial sections were analyzed for each perforation in the IEL. Of the total number of perforations assessed, 15% (15 of 101) were not filled with any cellular projections, and 83% (84 of 101) contained an EC projection. Of these 84 perforations, 35 had serial sections encompassing the entire perforation. The average dimension of EC projections (measured in the long axis of the vessel) as they passed through the IEL was 434 ± 41 nm. In the remaining 49 perforations, the serial sections were incomplete. The average dimension of these projections was 492 ± 32 nm. Analysis of these 84 perforations showed that in 24% (20 of 84) the EC contacted, but did not breach, the basal lamina of the SMC.

MEJs require cellular penetration of the basal lamina of the EC and SMC, thereby allowing the two cell types to come into close proximity; this occurred on 14 of 84 occasions (17%). Our data suggest that 6 of these 14 appeared to be touching, and, in some cases, a pentalaminar membrane structure characteristic of gap junctions was discernible (Fig. 5). On the basis of these observations, the total number of MEJs per 5 μm of MCA sectioned was 2.5 ± 0.5 (n = 3 rats). In five of six cases, the MEJ consisted of an EC projecting through the IEL and perforating the SMC basal lamina. In one case, the MEJ appeared at a site where the EC and SMC perforated their respective basal laminae and extended toward one another (Fig. 5). In an additional eight cases, a myoendothelial association was observed where the distance between the EC and SMC membranes was 20–250 nm.

Fig. 5.

Transmission electron micrographs of a rat MCA taken from three 100-nm serial cross sections showing a myoendothelial gap junction. A: low-magnification image showing a small endothelial cell (EC) protrusion extending toward a larger smooth muscle cell (SMC) protrusion through a gap in the internal elastic lamina (*in D), where they meet. C: ×4 magnification of area in A enclosed in box. Note point-to-point contact between EC and SMC (arrow). Width of gap is ∼3.5 nm. Serial sections preceding (B) and succeeding (D) the section containing myoendothelial gap junction show distance between the 2 cell membranes to be greater (20 nm), demonstrating point-to-point contact nature of myoendothelial gap junction. Scale bar, 500 nm.

Very few (2%) perforations showed an SMC projecting through the IEL toward the EC, and none of these perforated the EC basal lamina. During inspection of the MCA cross sections, gap junctions between ECs were routinely observed (Fig. 6).

Fig. 6.

Transmission electron micrographs of a rat MCA taken from three 100-nm cross sections showing EC-EC contact. A: low-magnification image showing a region of close contact between EC 1 and EC 2. *, Internal elastic lamina. B and C: ×7 magnification of area enclosed in box in A. Note multilaminar nature of gap junction. Width of gap is ∼3.5 nm. In contrast to myoendothelial gap junctions being visible in a single cross section, a single gap junction plaque between ECs was visible in multiple serial sections spanning the 600-nm distance between B and C. Arrows indicate depth of gap junction within endothelial cleft. Scale bar, 200 nm.


The results of the present study suggest that, in the rat MCA, 1) there is strong electrical communication between the EC and the SMC, 2) myoendothelial junctions are present, and 3) gap junctions mediate the EDHF response, suggesting an important role for myoendothelial communication in mediating EDHF dilations in this vessel.

Effect of Gap Peptides on EDHF-Mediated Dilations

Historically, inhibition of gap junctions has been achieved using a variety of compounds, many of which have indirect effects. Aliphatic alcohols (octanol and heptanol) are believed to limit gap junction communication by dissolving in the lipid membrane, whereas lipophilic compounds (18α-glycyrrhetinic acid and oleamide) may produce their effect through activation of protein kinases or G proteins (12). In contrast, gap peptides are short synthetic peptides that possess sequence homology with conserved domains of the extracellular connexin loops. Although these peptides do not discriminate between homocellular or heterocellular gap junctions, 37,43Gap 27 has been shown to inhibit myoendothelial dye transfer in an EC-SMC coculture system (27). Although modulation of dye transfer does not necessarily reflect modulation of electrical transfer, the data suggest that the gap peptides may inhibit communication between ECs and SMCs. We elected to use a gap peptide cocktail directed toward Cx37, Cx40, and Cx43, because we observed positive immunofluorescence staining in the MCA for these proteins.

EDHF-mediated dilations were assessed in perfused and pressurized MCAs by application of ATP specifically to the lumen of the vessel in the presence of l-NAME and indomethacin. We showed previously that, during this process, 1) a viable endothelium is required, 2) dilation persists when NO and prostanoid components are removed, 3) the smooth muscle is hyperpolarized, and 4) KCa channels are involved (40, 41). Therefore, this response can be attributed to EDHF. Incubation with scrambled gap peptides had no effect on concentration-dependent dilation in the MCA (Fig. 1). After incubation with the gap peptide cocktail, maximum dilation to ATP was reduced significantly. The effect of gap peptides was not a general inhibition of dilation, because the NO donor SNAP produced near-maximal dilations.

Gap junctions have been implicated previously in the EDHF response of rabbit carotid artery (4), guinea pig carotid artery (9), rabbit mesenteric artery (21), rat mesenteric artery (28, 35), and rabbit MCA (37). To our knowledge, this is the first study implicating a role for gap junctions in the rat MCA and demonstrating this role in perfused and pressurized vessels. This is an important consideration, because this preparation closely resembles the physiological situation of luminal pressure and shear stress.

Effect of Gap Peptides on EDHF-Mediated Hyperpolarizations

In the present study, hyperpolarization was sustained in the endothelium in the presence of gap peptides, whereas it was significantly reduced in the smooth muscle, consistent with a role for myoendothelial communication in mediating the EDHF response. The gap peptides may impair EC-EC, SMC-SMC, and EC-SMC gap junction communication. In rat mesenteric artery, where MEGJs are involved in the EDHF response, the hyperpolarization was also selectively impaired in the smooth muscle by gap peptides (35). In the same artery, selective loading of a connexin antibody into the EC depressed EDHF-mediated relaxation (28). However, in the study of Mather and colleagues (28), extracellular application of gap peptides was without effect on EDHF-mediated responses. In a wider range of arteries, gap peptides markedly reduce EDHF-mediated hyperpolarization (9, 17). Our data are consistent with those in other vessels (9, 17) and, together, infer a role for MEGJ in the EDHF response in a number of arteries, including the MCA.

Our finding that EC and SMC resting membrane potentials were similar is consistent with functional electrical connectivity between the two cell types (see below). Sandow and colleagues (35) reported that, in the rat femoral artery, the resting membrane potentials in ECs and SMCs were significantly different, and they detected no MEJs. Therefore, the existence of MEJs may be commensurate with comparable resting membrane potentials in endothelium and smooth muscle. In the present study, some preparations exhibited spontaneous action potential activity in ECs (data not shown). This is further support for electrical coupling between ECs and SMCs, inasmuch as ECs cannot generate action potentials. Spread of action potentials from SMCs to ECs has been reported in highly coupled vessels, such as guinea pig arterioles (6).

To our knowledge, this study is the first to implicate a role for gap junctions in mediating EDHF dilations in a rat cerebral artery. Gap junctions have been implicated in the EDHF response in a variety of vessels (see above). It is always possible that a diffusible factor may also contribute to EDHF-mediated responses in the MCA, which could account for the 7-mV hyperpolarization that persists in the presence of gap peptides. This response was consistently, but variably, delayed in onset and appeared to consist of a mixture of opposing depolarization and hyperpolarization (Fig. 2A). Although we previously found no effect of ouabain or Ba2+ on EDHF-mediated responses evoked by ATP in the MCA of Long-Evans rats (41), the protease-activated receptor 2 agonist SLIGRL evokes release of an ouabain- and Ba2+-sensitive paracrine dilator in MCAs of Wistar rats (31).

Localization of Connexin Immunofluorescence

Immunocytochemistry of the rat MCA revealed the presence of Cx40 in endothelium and smooth muscle. In agreement with our studies, Little and colleagues (22) observed Cx40 in ECs and SMCs of rat brain pial arterioles. Interestingly, Hong and Hill (19) reported that Cx43 staining was virtually absent from ECs and SMCs of the rat MCA. It is possible that the disparity in findings may be attributed to the difference in rat strain (Wistar-Kyoto rats vs. our Long-Evans rats) or age (5- to 7-wk-old vs. our 8- to 10-wk-old animals).

Although our data suggest that at least three connexin proteins are expressed in the rat MCA, it is not known which combination is located at MEGJs. Positive immunogold labeling would be a prerequisite to confirm the presence of connexin proteins at these junctions. However, this is a challenging prospect in the rat MCA. Since an MEGJ can be contained within a single 100-nm section and appeared to involve a single point-point contact between ECs and SMCs, the number of connexin proteins available for gold labeling is most likely to be very small, perhaps even below the limit of detection.

Identification of MEGJs

We have provided a detailed ultrastructural evaluation of the perforations in the IEL of the rat MCA. Although we observed many instances of EC projections passing through the IEL, the prevalence of MEGJs was significantly less. We employed rigorous morphological criteria to permit accurate identification of gap junctions. The membranes of potential MEGJs could not be resolved on five occasions, even when goniometer tilting was utilized. Hence, we may be underestimating the presence of MEGJs by 5. MEGJs are more technically challenging to resolve than gap junctions between ECs perhaps because of the nature of the cell-cell contact in MEGJs. EC gap junctions (defined by pentalaminar membrane apposition) are typically larger and remain visible in multiple serial sections, making extensive lateral contacts at their cell-cell margins (Fig. 6). In contrast, ECs and SMCs make small point-point contacts that are contained within a single 100-nm serial section (Fig. 5) and appear to be smaller than those reported previously in the rat mesenteric artery (34).

As discussed in a recent review (13), many published electron micrographs of putative MEGJs show membrane separation greater than that defined for a gap junction. In our preparation, we observed projections between ECs and SMCs through holes in the IEL, where the distance between the two membranes was 20–250 nm. This could ensure rapid paracrine communication, contain a small number of hemichannels or MEGJs so small in area as not to be distinguishable as a pentalaminar MEGJ, or represent a dynamic structure capable of forming closer communications within a limited time frame.

In conclusion, the results of the present study demonstrate that MEGJ communication is involved in EDHF-mediated SMC hyperpolarization and dilation of the male rat MCA. Our findings expand the potential mechanisms associated with disrupted EDHF responses after pathological conditions such as stroke (26) and traumatic brain injury (16) and emphasize the critical role of cell-cell coupling in mediating cerebrovascular responses.


This work was supported by American Heart Association Scientist Development Grant 0130250N (E. M. Sokoya), National Institutes of Health Grants HL-72954 (E. M. Sokoya) and AI-46773, HL-42550, and HL-070537 (A. R. Burns), and National Health and Medical Research Council of Australia Grant 194425 (H. C. Parkington).


We are grateful to Evelyn Brown and Sharon Phillips for technical assistance.


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


View Abstract