Developmental changes in myoendothelial gap junction mediated vasodilator activity in the rat saphenous artery

Abstract

A role for myoendothelial gap junctions (MEGJs) has been proposed in the action of the vasodilator endothelium-derived hyperpolarizing factor (EDHF). EDHF activity varies in disease and during ageing, but little is known of the role of EDHF during development when, in many organ systems, gap junctions are up-regulated. The aims of the present study were therefore to determine whether an up-regulation of heterocellular gap junctional coupling occurs during arterial development and whether this change is reflected functionally through an increased action of EDHF. Results demonstrated that in the saphenous artery of juvenile WKY rats, MEGJs were abundant and application of acetylcholine (ACh) evoked EDHF-mediated hyperpolarization and relaxation in the presence of N^ω-nitro-l-arginine methyl ester (L-NAME) and indomethacin to inhibit nitric oxide and prostaglandins, respectively. Responses were blocked by a combination of charybdotoxin plus apamin, or 1-[(2-chlorophenyl)diphenylmethyl]-1H-pyrazole (TRAM-34) plus apamin, or by blockade of gap junctions with the connexin (Cx)-mimetic peptides, ⁴³Gap26, ⁴⁰Gap27 and ^37,43Gap27. On the other hand, we found no evidence for the involvement of the putative chemical mediators of EDHF, eicosanoids, L-NAME-insensitive nitric oxide, hydrogen peroxide or potassium ions, since 14,15-epoxyeicosa-5(Z)-enoic acid (14,15-EEZE), hydroxocobalamin, catalase or barium and ouabain were without effect. In contrast, in the adult saphenous artery, MEGJs were rare, EDHF-mediated relaxation was absent and hyperpolarizations were small and unstable. The present study demonstrates that MEGJs and EDHF are up-regulated during arterial development. Furthermore, the data show for the first time that this developmentally regulated EDHF is dependent on direct electrotonic coupling via MEGJs.

The endothelium of arteries produces three main vasodilatory factors: nitric oxide (NO; Ignarro et al. 1987) prostaglandins and EDHF (Feletou & Vanhoutte, 1988; Chen et al. 1988; Cohen & Vanhoutte, 1995). The mechanisms of action of the first two are well characterized, whilst that of EDHF is less clear. Indeed, there is evidence for the action of multiple EDHFs and the involvement of multiple transduction pathways (Hill et al. 2001; McGuire et al. 2001; Busse et al. 2002; Ding & Triggle, 2003). Of particular interest in relation to the mechanism of action of EDHF is whether it is dependent on the release of a factor that diffuses across the internal elastic lamina (IEL) or on the existence of myoendothelial gap junctions (MEGJs), which permit electrical and/or chemical coupling between the endothelial cell (EC) and smooth muscle cell (SMC) layers. In spite of the continuing controversy over the mechanism by which EDHF is transferred from the endothelium to the smooth muscle there is consensus that an early target involves the activation of small (S) and intermediate (I) conductance calcium-activated potassium channels (K_Ca) localized to the endothelium (Doughty et al. 1999; Fukuta et al. 1999; Coleman et al. 2001; McGuire et al. 2001; Busse et al. 2002; Ding & Triggle, 2003) or alternatively, the activation of large (B) K_Ca localized to the smooth muscle (McGuire et al. 2001; Campbell & Gauthier, 2002). The former responses are sensitive to the combination of the toxins charybdotoxin plus apamin (McGuire et al. 2001; Busse et al. 2002; Campbell & Gauthier, 2002; Ding & Triggle, 2003).

Involvement of MEGJs in the action of EDHF has been demonstrated in a number of different vascular beds (Beny, 1999; Hill et al. 2001; McGuire et al. 2001). In the mesenteric bed of the rat, for example, MEGJ incidence and EDHF activity increase in a proximo-distal manner (Shimokawa et al. 1996; Sandow & Hill, 2000). Furthermore, in the femoral artery of the adult Wistar rat, an absence of MEGJs accounts for the absence of electrical coupling between ECs and SMCs and a lack of EDHF (Sandow et al. 2002). The relationship between gap junctions and EDHF activity has also been inferred from the blockade of EDHF responses with peptides whose sequences show homology to short segments of the extracellular loops of the Cx proteins; the Cx-mimetic or Gap peptides, ⁴³Gap26, ^37,40Gap26, ⁴⁰Gap27 and ^37,43Gap27 (Chaytor et al. 2001; Taylor et al. 2001; Berman et al. 2002; Griffith et al. 2002; Sandow et al. 2002; Sandow et al. 2003a). Furthermore, several recent electrophysiological experiments have suggested that EDHF is not a factor, but simply the electrotonic spread of hyperpolarization from ECs to SMCs (Yamamoto et al. 1998; Emerson & Segal, 2000; Coleman et al. 2001; Sandow et al. 2002).

EDHF activity has been shown to vary along and between vascular beds, as well as between species, and during disease and ageing (Hill et al. 2001; McGuire et al. 2001; Campbell & Gauthier, 2002). Indeed, whilst direct coupling via MEGJs accounts for EDHF in many vascular beds, diffusible factors have also been proposed to act as EDHF. These factors include epoxyeicosatrienoic acids (EETs), K⁺ ions, hydrogen peroxide and C-type natriuretic peptide (CNP) (Campbell et al. 1996; Fisslthaler et al. 1999; Edwards et al. 1998; Matoba et al. 2000; Chauhan et al. 2003a). Additionally, L-NAME-insensitive (non-NO synthase) basal NO has been reported to contribute to EDHF activity in certain vascular beds (Ge et al. 2000; Chauhan et al. 2003b; Stoen et al. 2003).

A general up-regulation of Cxs and gap junctions occurs during many developmental processes, implying that these signalling sites are important for the flow and selective regulation of electrical and chemical signals during changes associated with development (Levin, 2002). Such a role has been demonstrated at the cellular, tissue and organ levels to be integral for the differentiation of embryonic stem cells (Oyamada et al. 2002), the morphogensis of the heart (Levin, 2002) and the development of the central nervous system (Rouach et al. 2002). We have therefore investigated whether MEGJs are up-regulated during development and tested the hypothesis that EDHF should be altered in an analogous fashion. Functional studies were correlated with structural studies of the saphenous artery of juvenile and adult Wistar Kyoto (WKY) rats and the mechanism and nature of EDHF was characterized using selective antagonists of the various candidates for EDHF.

Methods

All experiments were undertaken with the approval of the Australian National University Animal Experimentation Ethics Committee under guidelines published by the National Health and Medical Research Council (NH & MRC) of Australia.

Electron microscopy

Anaesthetized (i.p., 44 and 8 mg kg⁻¹, ketamine and rompun, respectively) 2-week-old juvenile and 16-week-old adult male WKY rats were perfused via the left ventricle with a clearing solution of 0.1% BSA, 0.1% NaNO₃ and 10 U ml⁻¹ heparin, to fully dilate the vessels, and fixed with 1% paraformaldehyde, 3% glutaraldehyde in 0.1 m sodium cacodylate buffer, with 10 mm betaine, pH 7.4, using standard procedures (Sandow et al. 2002). Lateral saphenous branches of femoral arteries, defined as the first branch of the saphenous artery that diverged toward the knee, were removed and processed for electron microscopy as previously described (Sandow et al. 2002). Serial transverse sections (∼100 nm thick) totalling ∼5 μm of vessel length were cut from each of five and four vessels from juvenile and adult rats, respectively. MEGJs and their surrounding EC and SMC regions were counted and photographed at × 10 000 to × 40 000 on a Phillips 7100 transmission electron microscope. Vessel circumference at the IEL and the number of SMC layers were determined (Sandow et al. 2002) from eight juvenile and four adult vessels, each from a different animal.

Immunohistochemistry

To determine EC size, the periphery of ECs was highlighted in intact vessels, cleared, dilated (as above) and perfusion fixed (2% paraformaldehyde in 0.1 m phosphate-buffered saline) using immunohistochemistry and antibodies raised in sheep against rat Cx37 as previously described (Rummery et al. 2002).

Incidence of MEGJs and holes in the IEL

The number of MEGJs per EC was determined using the EC area and vessel circumference at the IEL to determine the surface area occupied by ECs in the 5 μm region examined with serial section electron microscopy. The incidence of holes in the IEL was determined for areas of 10⁴μm² from the immunohistochemical preparations due to the autofluorescence of the IEL. The number of holes per EC was calculated using the measured EC areas.

EDHF-mediated hyperpolarization

Juvenile and adult male WKY rats were anaesthetized with ether and decapitated. Intact segments of artery were pinned flat to the base of a recording chamber using the outer connective tissue layer. While the juvenile and the adult arteries used for the electrophysiological recordings were not standardized for initial resting tension, the degree of stretch applied was sufficient to result in resting membrane potentials of around −60 mV, which is within the range found when arteries are subjected to normal levels of intravascular pressure (Nelson & Quayle, 1995). One end of the artery was cut open to expose the ECs to the superfused (3 ml min⁻¹) Krebs solution (mm): NaCl 120; KCl 5; NaHCO₃ 25; NaH₂PO₄ 1; CaCl₂ 2.5; MgCl₂ 2; glucose 11; at 34–35°C, gassed with 95% O₂: 5% CO₂. L-NAME (100 μm) and indomethacin (10 μm) were present to inhibit NO and prostanoid production, respectively. Each preparation was taken from a different animal.

After equilibration for 40 min, SMC membrane potentials were recorded (Hill et al. 1999). Briefly, SMCs were impaled from the adventitial side using glass microelectrodes (tip resistance, 120–240 MΩ). Criteria for successful impalement included an abrupt drop in voltage when the microelectrode impaled the vascular SMC, a stable membrane potential for at least 2 min, and a sharp return to zero potential on withdrawal of the electrode. Electrical signals were low-pass filtered (cut-off frequency 1 kHz), amplified with an Axoclamp 2B (Axon Instruments) and stored on computer disk for analysis. The electrodes contained 0.2% propidium iodide in 0.5 m KCl for identification of impaled cells.

ACh (0.1–30 μm) was applied in a dose-dependent manner under conditions of depolarization with phenylephrine (PE; 0.5 μm for juvenile, and 1 μm for adult). The concentrations of PE used were determined as those that produced a 60% maximal constriction, in order to standardize conditions of depolarization, irrespective of the origin of the vessel. Each dose of ACh was applied separately after an appropriate washout period. In some experiments, ACh was applied in the absence of PE and of L-NAME and indomethacin.

In a separate set of experiments, drug effects on ACh (1 μm)-induced hyperpolarization in juvenile saphenous arteries were investigated. After recording the control ACh response (followed by 15 min washout), the following drugs were applied prior to a second application of ACh. The drugs and preincubation times were: charybdotoxin (60 nm, a BK_Ca and IK_Ca inhibitor, 10 min), apamin (0.5 μm, an SK_Ca inhibitor, 10 min), TRAM-34 (50 nm, a specific IK_Ca inhibitor, 10 min; kindly supplied by Dr Heike Wulff, University of California (Wulff et al. 2000), Cx-mimetic peptides, ⁴³Gap26, ⁴⁰Gap27 and ^37,43Gap27 (100 μm each; gap junction inhibitors synthesized by the Biomolecular Resource Facility, John Curtin School of Medical Research; purity > 97%, 60 min), 14,15-EEZE (10 μm, a specific EET antagonist, 20 min; kindly supplied by Dr J. R. Falck, University of Texas South-western Medical School; Gauthier et al. 2002), hydroxocobalamin (100 μm, a NO scavenger, 60 min), catalase (2000 U ml⁻¹, a catalyst of H₂O₂ degredation, 20 min), barium (30 μm, an inward rectifier K⁺ channel inhibitor, 10 min), and ouabain (500 μm, a Na⁺–K⁺ pump inhibitor, 10 min). All experiments were performed under conditions of depolarization due to PE (0.5 μm), except for the combination of barium plus ouabain which was applied in the absence of this depolarization. In the presence of PE, application of barium plus ouabain generated spontaneous rhythmical activity, which prevented the stable impalement of SMCs. Except where stated otherwise, chemicals were from Sigma (St Louis, MO, USA).

EDHF-mediated relaxation

ACh-induced relaxation of adult and juvenile arteries was measured simultaneously with membrane potential using video microscopy and a computer program to monitor vessel diameter (DIAMTRAK; Neild, 1989). Preparations were preconstricted with PE (0.5 μm for juvenile, 1 μm for adult) to 60% of maximal constriction.

Relaxation was also assessed in adult arteries using a Mulvany-Halpern style myograph (Sandow et al. 2003a). Unfortunately, we were unable to assess relaxation of juvenile arteries due to their small size and delicate endothelium. Segments of vessel 1 mm long were mounted between two 40 μm wires and equilibrated for 30 min prior to stretching. Preparations were constricted with PE (1 μm) to 60% of maximum constriction, equivalent to 0.85 g of tension. Endothelial integrity was initially assessed in each experiment, with preparations exhibiting an ACh (3 μm) relaxation of <40% being discarded. After the constriction reached a steady level, the effects of EDHF were assessed by adding ACh cumulatively in the presence of L-NAME (100 μm) and indomethacin (10 μm), which were both present for 40 min prior to ACh. The extent of the relaxation was expressed as the percentage reversal of the constriction evoked by PE.

Statistical analysis

Results are expressed as means ±s.e.m. Concentration–response curves were analysed by two-way ANOVA followed by Scheffé's test for multiple comparisons. Agonist concentrations causing half-maximal responses (EC₅₀ value) were calculated using non-linear regression analysis and expressed as the negative logarithm of the molar concentration (pD₂ values). Other data were analysed using one-way ANOVA followed by Student's pairwise t tests with Bonferroni correction for multiple group comparisons, or with paired or unpaired t tests for groups of two. A level of P < 0.05 was considered significant.

Results

Anatomy of juvenile and adult rat saphenous arteries

EC area, circumference and number of SMC layers in juvenile arteries were significantly less than those in adults (P < 0.05; Table 1; Figs 1A and C, 2A and 3A). Projections from ECs and SMCs, which came to within ≤ 20 nm of each other, but did not make gap junctions with the adjacent cell layer (Fig. 2F), were common in juvenile vessels, but were absent in adults. Holes in the IEL were common in both juvenile and adult rats, although significantly more numerous in juvenile vessels (P < 0.05; Table 1; Figs 1 and 3D).

Table 1.

Developmental characteristics of saphenous arteries from juvenile and adult WKY rats

	Circumference (μm)	Number SMC layers	EC area (μm2)	Number IEL holes/104μm2 IEL	Number IEL holes/EC	Number MEGJs/ 5 μm vessel length	Number MEGJs/ EC
Juvenile	390 ± 28*	3.6 ± 0.4*	346 ± 16*	129 ± 15*	4.5 ± 0.5*	11 ± 0.6*	2.1 ± 0.28*
	(8)*	(8)	(4; 41 cells)	(3; 6 areas)		(5)	(5)
Adult	745 ± 36	7.2 ± 0.1	392 ± 9	86 ± 10	3.4 ± 0.4	1 ± 0.5	0.1 ± 0.08
	(4)	(4)	(5; 150 cells)	(5; 10 areas)		(4)	(4)

^*P < 0.05. Number in parenthesis, number of preparations, each from a different animal.

Figure 1. Morphology of ECs and IEL in juvenile (A and B) and adult (C and D) saphenous arteries.

EC borders are highlighted by Cx staining (A and C). Holes in the IEL (arrowheads) appear as dark spots (B and D). The longitudinal axis of the vessel is left to right. Bar 50 μm.

Figure 2. Typical myoendothelial relationships in juvenile saphenous arteries.

Fewer SMC layers were seen in juvenile (A) compared to the adult (Fig. 3A) vessels. MEGJs (B and E arrows; insets) and close associations (F) were common in the juvenile arteries whilst they were rare in the adult (Fig. 3B, arrow; inset). Holes in the IEL (Fig. 3D, asterisk) were present in both juvenile and adult vessels. Large gap junctions between ECs (D, Fig. 3C and E arrows) and smaller gap junctions between SMCs (C, arrows) were present at both ages. Bar 5 μm, A; 1 μm, B, E and F; 50 nm, C and D; and 25 nm B and E, insets.

Figure 3. Typical myoendothelial relationships in adult saphenous arteries.

More SMC layers were seen in adult (A), compared to the juvenile (Fig. 2A) vessels. MEGJs (B arrow; inset) were rare in the adult arteries, although holes in the IEL (D, asterisk) were present in both juvenile and adult vessels. Large gap junctions between ECs (C and E, arrows) and smaller gap junctions between SMCs (Fig. 2C, arrows) were present at both ages. Bar 5 μm, A; 1 μm, B and D; 50 nm, C, E and B inset.

MEGJs were present in both juvenile and adult arteries (Table 1; Figs 2B and E and 3B), although there were significantly more MEGJs in juvenile vessels (P < 0.05; Table 1). When expressed per EC there were substantially more MEGJs in juvenile compared to adult vessels (P < 0.05; Table 1). Large gap junctions between ECs were commonly found at both ages (Figs 2D and 3C and E). Small gap junctions between SMCs were also found in both juvenile and adult rats (Fig. 2C).

EDHF activity in juvenile and adult rat saphenous arteries

SMCs of juvenile arteries were slightly depolarized compared with those of adults (−61.1 ± 0.6 versus−63.8 ± 0.7 mV, for juvenile, n= 55, versus adult, n= 17, respectively, P < 0.05). L-NAME and indomethacin had no effect on resting membrane potential. When the vessels were constricted to 60% of maximum with PE (0.5 μm, juvenile; 1 μm, adult), membrane potentials were significantly more depolarized in adult compared with those in juveniles (−46.7 ± 0.6 versus−38.9 ± 0.4 mV, for juvenile, n= 37, versus adult, n= 8, respectively, P < 0.05). Identity of SMCs was confirmed by dye labelling (Fig. 4A, inset).

Figure 4. EDHF-mediated hyperpolarization and relaxation to ACh in juvenile and adult saphenous arteries.

A, representative tracings showing hyperpolarization to 3 μm ACh under conditions of depolarization with PE in the saphenous artery of juvenile (upper) and adult (lower) WKY rats. Insets show examples of propidium iodide-filled SMCs (arrows) from which recordings were made. Longitudinal vessel axis runs left to right. B, concentration–response curves showing hyperpolarization to ACh under conditions of depolarization with PE in the saphenous artery of juvenile (open circles, n= 6–9) and adult (filled circles, n= 7–8) WKY rats. C, concentration–response curves showing relaxation to ACh in juvenile (open circles) and adult (filled circles for DIAMTRAK, filled triangles for myograph) saphenous arteries of WKY rat, preconstricted with PE (n= 5, each). L-NAME and indomethacin were present throughout. *P < 0.05.

In the presence of L-NAME and indomethacin, ACh-induced hyperpolarization was observed in both age groups, but ACh-induced hyperpolarization was considerably greater in juvenile than in adult arteries (Fig. 4A and B; maximal hyperpolarization: −13.7 ± 1.1 versus−5.4 ± 1.5 mV for juvenile, n= 9, versus adult, n= 8, respectively, P < 0.05). The pD₂ value for the juvenile arteries was 6.5 (n= 9), but could not be determined in adult vessels due to the variability in responses observed. With 3 μm ACh, the majority of responses were less than −5 mV (−2.7 ± 0.9, n= 6), but on two occasions were more than −10 mV. Repeat application of this dose, however, produced a significantly smaller response. This was not the case for repeated application of ACh in the juvenile arteries. This variability precluded further investigation. The absence of significant ACh-induced hyperpolarization at high concentrations was not due to a competing depolarization since ACh produced an insignificant change in membrane potential (0.7 ± 0.4 mV, n= 3) in the presence of TRAM-34 and apamin.

In the juvenile arteries ACh produced a dose-dependent relaxation. Under the same recording conditions, relaxation was virtually absent in the adult (Fig. 4C; pD₂ value of 6.3 for juvenile, open circles; not determined for the adult, filled circles, n= 5, respectively; maximal relaxation: 51.3 ± 3.5 versus 5.9 ± 2.1%, n= 5, for juvenile versus adult, P < 0.05). In order to verify that the absence of a relaxation in the adult did not arise due to the method in which the arteries were prepared, relaxation studies were also performed using tension myography. Data showed that there was no difference in the response to ACh irrespective of whether relaxation was measured under stretched or isometric conditions (Fig. 4C; filled triangles, maximal relaxation: 9.9 ± 2.8%, n= 5).

Characterization of EDHF in juvenile rat saphenous arteries

Under resting conditions, ACh (3 μm) produced a hyperpolarization which was unaffected by treatment with L-NAME and indomethacin (−9.0 ± 1.4 versus−8.1 ± 0.8 mV in the absence, n= 4, and presence, n= 19, of L-NAME and indomethacin, respectively, P > 0.05).

In the presence of L-NAME, indomethacin and PE, ACh (1 μm) produced a hyperpolarization and relaxation corresponding to 60% of the maximal hyperpolarization and of the constriction induced by PE (Fig. 4). The combination of charybdotoxin plus apamin significantly depolarized the membrane and abolished the hyperpolarization (Fig. 5A, Table 2). In a similar manner, the combination of TRAM-34 plus apamin significantly depolarized the membrane, although TRAM-34 alone did not produce a significant depolarization (Fig. 5B, Table 2). Tram-34 induced a significant inhibition of the hyperpolarization, with abolition of the residual hyperpolarization by addition of apamin (Fig. 5B, Table 2).

Figure 5. Effects of K_Ca antagonists on ACh-induced hyperpolarization in juvenile saphenous arteries.

Representative tracings of the effects of charybdotoxin (60 nm) plus apamin (0.5 μm) (A), TRAM-34 (50 nm) alone and TRAM-34 plus apamin (B) on ACh (1 μm)-induced hyperpolarization under conditions of depolarization with PE (0.5 μm) in the saphenous artery of juvenile WKY rat. Recordings in each of A and B were obtained from the same preparation. L-NAME and indomethacin were present throughout.

Table 2.

Effects of drugs on ACh-induced hyperpolarization and membrane potential

	ACh-induced hyperpolarization (mV)		Membrane potential (mV)
Drugs	Control	+drug	Control	+drug	n
Charybdotoxin (60 nm) + apamin (0.5 μm)	−8.6 ± 1.4	0*	−45.8 ± 2.6	−39.4 ± 1.8*	5
Tram-34 (50 nm)	−7.0 ± 0.7	−3.2 ± 0.9*	−45.3 ± 2.4	−42.3 ± 2.3	4
Tram-34 (50 nm) + apamin (0.5 μm)	−7.0 ± 0.7	0*	−45.3 ± 2.4	−39.3 ± 3.3*	4
43Gap26, 40Gap27 and 37,43Gap27 (100 μm each)	−8.4 ± 0.3	−1.8 ± 0.4*	−45.8 ± 1.5	−45.5 ± 1.6	4
14,15-EEZE (10 μm)	−8.1 ± 1.2	−8.1 ± 1.1	−46.8 ± 0.7	−46.8 ± 0.7	6
Hydroxocobalamin (100 μm)	−9.5 ± 1.9	−8.7 ± 1.9	−46.2 ± 2.0	−46.0 ± 1.9	5
Catalase (2000 U ml−1)	−9.8 ± 1.0	−10.4 ± 0.9	−47.5 ± 0.6	−46.8 ± 0.3	4
Barium (30 μm) + ouabain (500 μm)	−6.6 ± 1.2	−7.5 ± 1.0	−61.5 ± 1.8	−47.3 ± 1.4*	4

^*P < 0.05 (versus control). n, number of preparations, each from a different animal.

The addition of the Gap-mimetic peptide combination (⁴³Gap26, ⁴⁰Gap27 and ^37,43Gap27), significantly reduced the ACh-induced hyperpolarization by 79%, but had no effect on membrane potential (Fig. 6A, Table 2). In contrast, the hyperpolarization was unaffected by 14,15-EEZE, hydroxocobalamin or catalase (Figs 6B, C and D, Table 2). The combination of barium plus ouabain which was applied in the resting state due to the increased excitability when combined with PE, significantly depolarized the membrane, but was also without effect on ACh-induced hyperpolarization (Fig. 6E, Table 2).

Figure 6. Effects of drugs on ACh-induced hyperpolarization in juvenile saphenous arteries.

Representative tracings of the effects of, the Gap-mimetic peptide combination (⁴³Gap26, ⁴⁰Gap27 and ^37,43Gap27; 100 μm each) (A), 14,15-EEZE (10 μm) (B), hydroxocobalamin (100 μm) (C), catalase (2000 U ml⁻¹) (D) and barium (30 μm) plus ouabain (500 μm) (E) on ACh (1 μm)-induced hyperpolarization in the saphenous artery of juvenile WKY rat in the presence (A, B, C and D) or absence (E) of PE (0.5 μm). Recordings in each pair of traces were obtained from the same preparation. L-NAME and indomethacin were present throughout.

Discussion

The present study demonstrates that MEGJs and the endothelium-dependent vasodilator EDHF change during development in the saphenous artery of the WKY rat. The key to this increased activity during development is the dependence of EDHF on the presence of MEGJs. Our results demonstrate for the first time that the incidence of MEGJs is not constant throughout development and show that the increase in the incidence of MEGJs during development is responsible for significant functional changes in EDHF activity.

The ACh-induced hyperpolarization in the juvenile arteries was abolished by charybdotoxin plus apamin, indicating that it was attributable to EDHF. However, the non-selective action of charybdotoxin on BK_Ca and IK_Ca channels and some voltage-dependent K⁺ channels (Garcia et al. 1995; Nelson & Quayle, 1995; Ding & Triggle, 2003) precludes the definitive identification of the target K⁺ channels associated with EDHF. Our present findings that both TRAM-34, a recently synthesized clotrimazole derivative that is a specific inhibitor of IK_Ca channels (Wulff et al. 2000), and apamin were required to abolish the EDHF-mediated hyperpolarization in the saphenous artery of the juvenile rat provide strong evidence that both IK_Ca and SK_Ca are the target channels for this developmentally regulated EDHF. Preliminary data from control experiments have shown that TRAM-34 and apamin do not affect the relaxation to cromakalim, an endothelium-independent smooth muscle cell response. The involvement of IK_Ca and SK_Ca in EDHF action in adult rats was similarly identified in the carotid artery, with the combination of TRAM-34 plus apamin (Eichler et al. 2003), and in the mesenteric artery using TRAM-34 and the slightly less potent IK_Ca inhibitor, TRAM-39, with apamin (Crane et al. 2003; Hinton & Langton, 2003). The significant depolarization caused by these potassium channel antagonists is unlikely to contribute to the abolition of the hyperpolarization since the EDHF-induced outward current shows no rectification (Yamamoto et al. 1999). Indeed, the more depolarized potential in the juvenile arteries in the presence of potassium channel antagonists, or in the adult arteries after preconstriction with PE, would be expected to increase the size of the EDHF-induced hyperpolarization rather than to decrease it as we found.

The lack of effect of 14,15-EEZE, barium plus ouabain, catalase and hydroxocobalamin on the EDHF-mediated hyperpolarization demonstrates that the putative chemical candidates for EDHF – EETs, K⁺ ions, hydrogen peroxide and L-NAME-insensitive NO – are not involved in the developing rat saphenous artery. Furthermore, our data do not support the proposal that CNP is an EDHF, since the mechanism involved is sensitive to barium and ouabain (Chauhan et al. 2003a), and the EDHF in the present study is not. Unfortunately, since there is no selective antagonist for the CNP receptor-C subtype (Chauhan et al. 2003a), it was not possible to further examine a role for CNP.

A role for MEGJs in the juvenile EDHF response was supported by inhibition with Cx mimetic peptides. We have shown recently that these peptides do not affect the relaxation to endogenously released nitric oxide (Sandow et al. 2003a). An additional action of the peptides at homocellular gap junctions cannot be excluded, but such an action for the spread of an electrotonic signal would be dependent in the first instance on conduction of the current through the MEGJs. Indeed, there was a strong correlation between the incidence of MEGJs and the magnitude of the EDHF response. Thus, in the juvenile vessels MEGJs were common and EDHF-mediated hyperpolarization and relaxation present, while in the adult, MEGJs were rare and EDHF ineffectual. Together with the lack of effect of the other blockers of EDHF-mediated hyperpolarization, these data support the conclusion that EDHF is an electrical signal transmitted from ECs to SMCs through MEGJs in the saphenous artery of the rat. These data further imply that there is a limit to the amount of heterocellular coupling required to establish a biophysically relevant arterial response since some MEGJs were present in the adult vessels. The results thus support the contention that MEGJs confer a significant degree of plasticity to arterial function (Christ et al. 1996).

In two out of eight preparations from adult rats, 3 μm ACh evoked more than 10 mV of hyperpolarization despite the low incidence of MEGJs in these arteries. These occasional large hyperpolarizations cannot be explained by the impalement of ECs, because all of the impaled cells were identified as SMCs by dye labelling. In the porcine ciliary artery Beny et al. (1997) found that EDHF-mediated hyperpolarization could only occasionally be recorded in the SMC layers (adjacent to the endothelium), and this was attributed to attenuation of the signal within the multiple SMC layers. Since our data show that MEGJs are rare in the saphenous artery of the adult rat, we propose that occasional large hyperpolarizations occurred when the recording electrode was in the vicinity of these junctions (position 1, Fig. 7), while the small responses occurred when the recording electrode was distant from an MEGJ (position 2, Fig. 7). Responses would be expected to be rapidly attenuated along and around the vessel and incapable of eliciting any significant relaxations, as was the case in the adult vessels. These data provide evidence for the first time of the minimal limit of heterocellular coupling required to effect a functional EDHF response.

Figure 7. Schematic diagram showing the relationship between EDHF-mediated vasodilatation and the incidence of MEGJs in juvenile and adult saphenous arteries.

The relative density of MEGJs per EC and the number of SMC layers are shown for juvenile and adult animals, with MEGJs being 20-fold more prevalent in juvenile than in adult arteries. The incidence and proximity of MEGJs to the site of the recording electrode in the adult vessels is proposed to account for the presence of a hyperpolarization. Thus, in the adult vessels, hyperpolarization was present when the recording electrode was near an MEGJ (position 1), but absent when at a remote site (position 2). Vasodilatation and hyperpolarization was robust in the juvenile, irrespective of the location of the recording electrode (positions 1 and 2).

The absence of EDHF activity in the adult vessels cannot be simply attributed to the increase in the number of SMC layers found in the mature vessels since the wall of the vessel, being only 20 μm thick, would approximate an electrically short cable. In support of this view, there was no significant attenuation of electrically transmitted neural responses across the medial thickness in the caudal artery of the adult rat, a vessel of comparative diameter and medial composition (Sandow et al. 2002). In addition, some 3-fold greater number of MEGJs were present in the adult caudal artery than in the adult saphenous artery and an appreciable gap junctional-dependent EDHF activity could be measured. The absence of significant ACh-induced hyperpolarization at high concentrations could also not be attributable to the existence of an opposing depolarization since ACh produced only a millivolt change in the adult vessels when EDHF was abolished with TRAM-34 and apamin.

Holes in the IEL were abundant in both juvenile and adult vessels, even though MEGJs were rare in the adult. This suggests that such holes in the adult may indicate a history of the presence of a more abundant MEGJ population, or perhaps the potential for a dynamic state, whereby projections could readily form under conditions where EDHF-mediated vasodilator activity is altered. Indeed, the prevalence of close non-gap junctional projections between the adjacent cell layer in the juvenile, but not in adult, arteries further supports the likelihood of a dynamic MEGJ population. Similar projections have recently been described in the microcirculation of the hamster, where physiological studies (Welsh & Segal, 1998; Emerson & Segal, 2000) suggest that MEGJ function is modulated (Sandow et al. 2003b).

Little is known about alterations in EDHF-mediated responses during development. Thomsen et al. (2002) have shown that in the microcirculation of the sciatic nerve, ACh induced a non-NO-, non-prostanoid-dependent vasodilatation which was smaller in 18- to 20-week-old rats than in 1- to 2-week-old rats. However, hyperpolarization was not measured, nor was data provided to relate this change to EDHF activity. In the mesenteric artery of the rat, EDHF-mediated hyperpolarization and relaxation have been shown to be impaired with ageing, although this decline was not evident until around 12 months of age (Fujii et al. 1993; Goto et al. 2000a). This finding could be reversed with angiotensin II inhibition. Our present finding demonstrates for the first time that significant anatomical remodelling of MEGJs occurs during development and that this anatomical remodelling accounts for changes in EDHF-mediated responses in saphenous arteries of juvenile and adult rats. Similar remodelling may also underlie changes occurring during ageing and in pathophysiological conditions, where EDHF responses are altered (Fujii et al. 1992; Van de Voorde et al. 1992; Fukao et al. 1997; Goto et al. 2000b; Sandow et al. 2003a). This remodelling represents a potential therapeutic target for controlling vasodilator function and thus changes in vascular tone associated with vascular disease.