Corresponding author K. A. Dora: Department of Pharmacy and Pharmacology, University of Bath, Claverton Down, Bath BA2 7AY, UK. Email: email@example.com
Levels of ATP achieved within the lumen of vessels suggest a key autacoid role. P2Y receptors on the endothelium may represent the target for ATP, leading to hyperpolarization and associated relaxation of vascular smooth muscle through the endothelium-dependent hyperpolarizing factor (EDHF) pathway. EDHF signals radially from the endothelium to cause dilatation, and appears mechanistically distinct from the axial spread of dilatation, which we showed occurs independently of a change in endothelial cell Ca2+ in rat mesenteric arteries. Here we have investigated the potential of P2Y receptor stimulation to evoke spreading dilatation in rat resistance small arteries under physiological pressure and flow. Triple cannulation of isolated arteries enables focal application of purine and pyrimidine nucleotides to the endothelium, avoiding potential complicating actions of these agents on the smooth muscle. Nucleotides were locally infused through one branch of a bifurcation, causing near maximal local dilatation attributable to EDHF. Dilatation then spread rapidly into the adjacent feed artery and upstream against the direction of luminal flow, sufficient to increase flow into the feed artery. The rate of decay of this spreading dilatation was identical between nucleotides, and matched that to ACh, which acts only on the endothelium. In contrast, focal abluminal application of either ATP or UTP at the downstream end of cannulated arteries evoked constriction, which only in the case of ATP was also associated with modest spread of dilatation. The non-hydrolysable ADP analogue, ADPβS, acting at P2Y1 receptors, caused robust local and spreading dilatation responses whether applied to the luminal or abluminal surface of pressurized arteries. Dilatation to nucleotides was sensitive to inhibition with apamin and TRAM-34, selective blockers of small- and intermediate-conductance Ca2+-activated K+ channels, respectively. These data demonstrate that direct luminal stimulation of P2Y receptor on the endothelium of rat mesenteric arteries leads to marked spreading dilatation and thus suggests that circulating purines and pyrimidines may act as important regulators of blood flow.
Endothelial cells provide a powerful dilator influence on vascular smooth muscle cells, reflecting the release of diffusible factors such as nitric oxide (NO) and endothelium-derived hyperpolarizing factor (EDHF), and the spread of hyperpolarization through gap junctions between the endothelial cells and through myoendothelial gap junctions (MEGJs) to the adjacent muscle (reviewed in McGuire et al. 2001; Busse et al. 2002; Ledoux et al. 2006). Activation of these pathways can be evoked in a variety of ways, including agonist stimulation and changes in luminal blood flow with associated changes in shear stress. The majority of studies involving agonist stimulation have employed acetylcholine as a pharmacological agonist, as it induces robust endothelium-specific effects. However, from a physiological standpoint other agents may be more relevant. For example, ATP is present in the circulating blood and is released locally from cells where it may act as an autacoid. When measured in the femoral vein of humans, plasma concentrations of ATP can reach micromolar levels during exercise (Rosenmeier et al. 2004). The potential sources for circulating ATP include red blood cells responding to low PO2 (Miseta et al. 1993; Ellsworth et al. 1995; Dietrich et al. 2000; Ellsworth, 2004), platelets (Beigi et al. 1999) and the arterial wall itself, including the endothelial cells (Pearson & Gordon, 1979; Bodin et al. 1991; Burnstock, 1999; Yamamoto et al. 2003). In addition to ATP, UTP can also be released from cells, including those in the human heart during ischaemia (Wihlborg et al. 2006), and can evoke vasodilator responses (Lazarowski & Boucher, 2001; Lazarowski et al. 2003; Liu et al. 2004), making both these nucleotides potentially important modulators of arterial tone and blood flow.
Therefore, the present study was designed to characterize the local and spreading dilatation following luminal perfusion of nucleotides in pressurized rat mesenteric arteries to evaluate the potential of nucleotides to act as physiological modulators of blood flow. We demonstrate significant endothelium-dependent spreading dilatation following activation of endothelial cell P2Y receptors by purine and pyrimidine nucleotides in physiologically relevant concentrations.
Rat mesenteric artery isolation and cannulation
Male Wistar rats (200–250 g) were killed by cervical dislocation and exsanguination according to requirements detailed under Schedule 1 of the Animals (Scientific Procedures) Act 1986 and monitored by the Home Office (UK). The mesentery was removed and placed in cold Mops buffer containing (mm): NaCl 145.0, KCl 4.7, CaCl2·2H2O 2.0, MgSO4·7H2O 1.17, Mops 2.0, NaH2PO4·H2O 1.20, glucose 5.0, pyruvate 2.0, EDTA 0.02 NaOH 2.75, adjusted to pH 7.40 ± 0.02. Third-order branches of the superior mesenteric artery were isolated, dissected and cannulated at each end, as previously described (Takano et al. 2004; Mather et al. 2005). Third order branches with the bifurcation (fourth order braches) were isolated for triple-cannulation experiments. Three pressure myograph chambers were used in this study, each chamber specifically suited to the method of delivering agonists (Series 1–4). The Mops-buffered solution was heated to 36.5 ± 0.2°C in all experiments, and arteries were all pressurized by a gravity-fed inflow and outflow system. Two open 5 ml syringes connected to the perfusion pipettes were attached to either side of a vertical timing belt. This assembly was secured to a block that was moved along a vertical track and held in position at the required height above the artery. All arteries were pressurized to 50 mmHg, and were submaximally contracted with PE (0.3–3 μm) to generate a consistent level of tone between treatments and preparations. Endothelial cell viability was assessed as > 95% relaxation to 1 μm ACh. The responses to PE and ACh were consistent between all the experimental chambers used. Unless otherwise stated (no LN), all experiments were performed in the presence of the selective inhibitor of NO synthase, Nω-nitro-l-arginine methyl ester hydrochloride (l-NAME, 100 μm, Control).
Luminal application of agonists
Series 1: Spreading dilatation to luminally perfused agonists In this series of experiments, three cannulation pipettes were used. These were held in position with three pipette holders attached to ball joints (Narishige) seated on course manipulators (Narishige) attached to the stage of the inverted microscope (IX71, Olympus). The three ends of an isolated artery were cannulated and mounted in a heated, 2 ml chamber (RC-27 chamber, PH-6 platform, Warner Instruments) and continuously superfused at 2 ml min−1 with heated Mops solution. The upstream end of the artery (Feed artery) and one side of the bifurcation (Branch 2) were attached to the gravity-fed pressurizing syringe reservoirs. The upstream and downstream perfusion pressures through the feed artery were adjusted by rotating the timing belt, to generate luminal flow (7–9 cmH2O gradient, 30–50 μl min−1) whilst maintaining a constant average transmural pressure, in order to avoid upstream flow of infused agonists. Phenylephrine was added to the superfusion solution, and each agonist used to study spreading dilatation into the feed artery was infused into one of the sidebranches (Branch 1) for at least 2 min at 50 μl min−1 using a BeeHive® syringe pump system (Bioanalytical systems, USA). In all experiments the movement of perfusion solution was monitored by including 0.1 μm carboxyfluorescein in the agonist solution. Arteries were visualized using a laser scanning confocal microscope (FV500-SU, Olympus, Japan, excitation 488 nm, emission 505 nm) to enable simultaneous fluorescence and brightfield imaging, with a 4×/0.13 NA objective (UplanFl, Olympus, Japan), and images were recorded with Fluoview software (Olympus, USA) at 1 Hz. There was a short delay (∼30–40 s) in observing responses to luminal perfusion of agonists due to voiding the tubing and pipette dead-space volume.
Series 2: Measurement of flow in triple-cannulated arteries A subset of experiments was performed specifically to establish whether spreading dilatation could evoke an increase in flow into the feed artery. Arteries were cannulated as for experiments in Series 1, such that the feed artery was not more than 3 mm long, and Branch 2 as short as possible, the branch tied to the pipette < 250 μm from the bifurcation. To reduce differences in flow due to the resistance generated by the cannulating pipettes, the three pipettes and their configuration remained the same for all experiments. To measure flow, minor modifications were made to the upstream and downstream reservoirs and the effluent solution was weighed. To keep the upstream pressure constant, the solution level of the upstream syringe reservoir was maintained by continuous gravity-fed filling of the syringe from another reservoir. Between experiments, the level of this upstream reservoir ranged from 51 to 54 mmHg, depending on the diameter of the artery, to set the basal (maximum) flow rate near 40–50 μl min−1. The outflow pressure was kept constant at 47 mmHg. To measure flow rate, the outflow from the downstream tubing passed directly to an analytical balance (Sartorius AC210S) with PC interface to acquire weight and time (SartoConnect V3.5.2) at 1 Hz. In order to maintain steady readings, the outflow glass tube was in contact with the liquid in the outflow solution reservoir seated on the balance. Time stamps for simultaneous weight and image acquisitions enabled synchronization of flow and diameter readouts to within 2 s. To avoid the artefact due to infusion of agonists, in these experiments the BeeHive syringe pump flow rate was set to 2 μl min−1 (no observed effect of vehicle infusion on flow). Thus to obtain comparable time courses to that observed in Series 1 experiments, this required voiding of the tubing dead-space prior to commencing an acquisition run. In each experiment, diameter and flow were measured (a) at the beginning of experiments during bath application of phenylephrine and acetylcholine, (b) during the luminal perfusion of agonist to evoke spreading dilatation responses, and (c) during bath application of KCl and papaverine at the end of experiments. Data were only used if changes in flow could be observed to bath application of agents. See Online supplemental material Supplemental Fig. 2 for details of flow rate characterization.
Series 3: Characterization of the EDHF response In this series spreading dilatation responses were not studied, so arteries were mounted in a pressure myograph with 10 ml heated chamber (Danish Myo Technology, 120CP) as previously described (McSherry et al. 2005), and responses were obtained in a static bath. In the presence of PE added to the chamber, the P2Y receptor agonists were perfused through the lumen of arteries at ∼90 μl min−1. This perfusion flow rate equates to a shear stress of ∼10 dyn cm−2 in fully dilated arteries. Non-cumulative concentration response curves were obtained. Arteries were visualized using an inverted microscope (IX70, Olympus, Japan) with a 4×/0.13 NA objective (UplanFl, Olympus, Japan) and CCD camera (Nikon, Japan), and images stored to videotape. Diamtrak software was used to measure outer diameter at one position along the artery online, the results of which were not different to measuring diameter offline with MetaMorph software.
Abluminal application of agonists
Series 4: Spreading dilatation to abluminally applied agonists To avoid upstream diffusion of abluminally applied agonists, arteries were mounted in a custom-made long, narrow 0.5 ml Perspex chamber seated on a 24 mm × 50 mm coverslip, which was heated (PH-6 platform, Warner Instruments) and continuously superfused at 2 ml min−1 with heated Mops solution. Agonists were applied focally from a bevelled borosilicate glass micropipette (5 μm tip) using a pneumatic pico pump (10 psi, PV 820, World Precision Instruments Inc., FL, USA). The micropipette was at least 200 μm from the downstream end of the artery, and was moved to within 10 μm of the wall of the artery during the delivery of agonist, then rapidly moved away. This ensured agonist was only delivered during the short pressure-pulse. The superfusion was carefully controlled and monitored (5 μm diameter microspheres, Molecular Probes) such that flow was from the upstream end of the artery. This experimental setup precludes direct stimulation of cells upstream from the local site (Takano et al. 2004).
For all measurements of spreading dilatation, artery outer diameter was measured offline using motion analysis software (MetaMorph, Universal Imaging, USA). This enabled simultaneous analysis of multiple, calibrated distances along the artery wall, and direct comparisons of local dilatation to spreading dilatation for a single application of agonist, which is not possible with Diamtrak software or higher magnification objectives. The resolution of the system was 5 μm (equivalent to one pixel), ca 1.5% of the maximum diameter of arteries. Fluorescence intensity was also measured offline simultaneously at multiple positions in the lumen of arteries, which was temporally matched to diameter measurements.
All drugs were obtained from Sigma (Poole, UK) with the exception of DEA NONOate (Alexis Biochemicals, Nottingham, UK), apamin (Latoxan, Valence, France), and 1-[(2-chlorophenyl)diphenylmethyl]-1H-pyrazole (TRAM-34) which was a generous gift from Dr H. Wulff (University of California, Irvine, CA, USA). All stock solutions were prepared in distilled water with the exception of TRAM-34, which was dissolved in dimethyl sulfoxide; and DEA NONOate, which was dissolved in chilled 0.01 m NaOH (and aliquots stored at −80°C). Prior to use in experiments all drugs were diluted in physiological buffer, and kept chilled (∼4°C). All inhibitors were incubated in the bath and lumen of the arteries for a minimum of 20 min prior to obtaining responses, except for 2′-deoxy-N6-methyladenosine 3′,5′diphosphate diammonium salt (MRS2179; 5 min), and apamin (1 h).
Results are summarized as means ±s.e.m. of n arteries, one per animal. Statistical comparisons were made using the Kruskal–Wallis test followed by Dunn's multiple-comparison post hoc test, where P < 0.05 was considered statistically significant. The dilatation evoked by each agonist was calculated as the percentage of the maximum dilatation from PE-contracted arteries, 100% being the maximum diameter seen for each experiment (150 μm papaverine added at the end of each experiment). The contraction evoked by phenylephrine was calculated as the percentage of maximum contraction, and percentage contraction to nucleotides that from PE-contracted arteries, 100% being the minimum outer diameter observed in the presence of 45 mm KCl and 10 μm PE. Fluorescence intensity (F) relative to the maximum fluorescence obtained during perfusion periods (Fmax) were expressed as relative changes (F/Fmax) for each luminal region of interest. Concentration–response curves and bar graphs were prepared using Prism v4.0 software (GraphPad Software, USA).
Spreading dilatation to luminally perfused agonists
The minimum and maximum outer diameter of bifurcating arteries used in these triple-cannulated arteries is shown in Table 1. The level of phenylephrine tone was adjusted to approximately 50% of maximal diameter in all experiments (Table 1). Luminal infusion of vehicle alone (Mops buffer) had no effect on arterial diameter (n= 5) despite the large increase in shear stress, demonstrating an absence of flow-dependent dilatation in these arteries. However, if nucleotides were included in the solution, near maximal dilatation that was maintained for the period of infusion was observed. The dilatation to infused ATP, ADPβS (each 1 or 3 μm) and UTP (3 or 10 μm) into a side branch evoked dilatation that spread to upstream regions of the artery not directly stimulated by the agonists (Figs 1 and 2). This spreading dilatation was synchronized in all regions of the artery for all agonists, and on some occasions, the artery underwent oscillations in diameter. The spreading dilatation decayed with distance, but was able to pass over 2 mm from the site of stimulation. The response in the infused branch (Branch 1) only slightly reduced as it passed into the feed branch (0 μm) (ATP: Branch 1, 90.0 ± 3.7%, 0 μm, 80.8 ± 3.5%, n= 6; ADPβS: Branch 1, 85.8 ± 5.0%, 0 μm, 80.0 ± 6.6%, n= 6; UTP: Branch 1, 81.5 ± 6.7%, 0 μm, 75.8 ± 10.8%, n= 6). For comparison, ADP was also infused, and stimulated similar responses (Branch 1, 77.6 ± 5.8, 0 μm, 78.1 ± 6.2, n= 6). The near 80% dilatation at 0 μm for each agonist evoked spreading responses that decayed equally with distance upstream (Fig. 2A). In contrast, much higher concentrations of adenosine (100 μm) were required to evoke any spreading response, a response that decayed rapidly with distance, so even at the 0 μm site on the feed artery the dilatation only reached 39.6 ± 7.1% despite a response of 76.7 ± 5.0% in Branch 1 (n= 5, Fig. 2B).
Table 1. Diameter of triple-cannulated arteries used for experiments in Series 1
Min (n= 12)
Max (n= 12)
l-NAME (n= 36)
No LN (n= 14)
EC damaged (n= 8)
Arteries were pressurized to 50 mmHg, and tone generated by adding PE to the superfusion solution. The average values for diameter (μm) and percentage minimum diameter (%) are given for the tone immediately preceding the infusion of agonists either in the presence of l-NAME, the absence of l-NAME (No LN) or when an air bubble was luminally perfused through the feed artery and Branch 2 (EC Damaged). The diameter along the Feed artery did not vary significantly between 0 and 2000 μm, so values at 1000 μm are provided for comparison.
126 ± 20
312 ± 31
235 ± 4
44.9 ± 2.4
202 ± 8
50.1 ± 5.8
226 ± 10
45.2 ± 7.6
146 ± 20
375 ± 32
296 ± 3
32.9 ± 1.5
240 ± 9
57.9 ± 5.1
312 ± 9
29.6 ± 6.3
To assess how the spreading dilatation stimulated by the P2Y receptor agonists compared to that with another agonist known to evoke EDHF and spreading dilatation responses, ACh was also added via the lumen of arteries. Dilatation to 1 μm ACh was greater than any of the P2Y receptor agonists (Branch 1, 95.3 ± 1.2%, 0 μm), and this dilatation was also greater at 0 μm on the feed artery (92.8 ± 4.7%, n= 7), causing more robust spreading dilatation. The early phase of infusion of ACh was further analysed to compare the decay of spreading dilatation with matched levels of dilatation at 0 μm. Thus, when data for 80% dilatation were chosen (paired times at each distance per response), the rate of decay in amplitude of the ∼80% local response was the same as that for the P2Y receptor agonists (Fig. 2B).
In response to both ATP and ACh, the 80% maximal dilatation at the 0 μm position of the feed artery equated to an average increase in diameter of ∼80 μm, while at 2000 μm upstream an average increase of at least 6 μm was observed (Table 2). The dilatation to DEA NONOate in Branch 1 (75.9 ± 13.0 μm, n= 5) was unable to stimulate a similar dilatation in the feed artery (0 μm, 13.4 ± 10.6 μm increase in diameter, n= 5).
Table 2. Summary of dilatation responses to luminal perfusion of agonists in triple-cannulated arteries
ATP (n= 6)
ADPβS (n= 6)
UTP (n= 6)
ADP (n= 6)
ACh (n= 7)
ATP (n= 5)
ACh (n= 4)
ATP (n= 4)
ACh (n= 4)
Series 1 experiments. Values are the average change in diameter from phenylephrine-stimulated tone that were used to calculate percentage dilatation in Fig. 2 (l-NAME) and Fig. 5 (no l-NAME, EC Damaged). See the corresponding Figure Legends for details and concentrations of agonists used. Values for acetylcholine (ACh) relate to those used for 80% maximum dilatation at the local site in the feed artery (0 μm).
96.1 ± 25.8
82.5 ± 35.0
72.8 ± 31.8
62.1 ± 27.2
83.1 ± 17.4
48.0 ± 9.0
58.4 ± 19.4
51.6 ± 19.8
72.7 ± 24.6
83.2 ± 39.8
81.0 ± 32.2
77.0 ± 43.6
59.4 ± 22.1
77.2 ± 11.8
78.8 ± 13.6
85.5 ± 17.9
−12.7 ± 6.0
16.4 ± 6.7
55.3 ± 22.7
47.1 ± 12.1
55.4 ± 30.5
40.4 ± 10.4
51.1 ± 6.6
61.4 ± 11.8
71.8 ± 9.9
−1.4 ± 3.9
1.6 ± 1.2
26.6 ± 13.6
23.2 ± 9.1
30.7 ± 18.5
22.2 ± 5.7
28.2 ± 5.1
37.6 ± 9.4
51.6 ± 9.6
−1.0 ± 4.3
1.1 ± 1.5
13.6 ± 7.6
13.3 ± 6.0
14.4 ± 6.4
10.4 ± 4.6
13.5 ± 3.8
15.9 ± 3.1
24.7 ± 3.4
−0.5 ± 3.0
0.1 ± 1.0
6.1 ± 3.7
4.3 ± 5.4
6.1 ± 1.8
3.7 ± 1.8
7.1 ± 3.2
7.7 ± 1.5
11.1 ± 5.1
1.4 ± 3.0
1.0 ± 0.9
Increase in flow during spreading dilatation
In Series 2 experiments, arteries contracted with phenylephrine had an average diameter of 270 ± 37 μm and a luminal flow rate of 28 ± 4.8 μl min−1 (n= 7). Since the spreading dilatation response to ACh was similar to the nucleotides, but not dependent on rapid infusion (to avoid hydrolysis), only this agonist was used for this series of experiments. Infusion of 1 μm ACh into Branch 1 evoked a spreading dilatation response and increase in flow into the feed artery (Fig. 3). In these arteries, the average maximum outer diameter and luminal flow rates were 405 ± 55 μm and 49 ± 1 μl min−1 (see Supplemental Fig. 2 for further details).
Activation of KCa by luminally perfused agonists
The mean maximum arterial outer diameter in experiments where unbranched arteries were cannulated was 345 ± 5 μm (n= 94; range 280–460 μm). As with the branched, triple-cannulated arteries, the concentration of phenylephrine was adjusted to provide comparable tone between experiments and treatments, averaging near 50% maximal constriction.
Perfusion with ATP (10 nm to 3 μm) evoked concentration- and perfusion flow-dependent dilatation (1 μm ATP: 79.5 ± 6.0%, n= 14), which was unaffected by l-NAME (1 μm ATP: 77.4 ± 4.2%, n= 9, Fig. 4A). ADPβS (1 μm) and UTP (3 μm) evoked similar responses to 1 μm ATP (Fig. 4B). The P2Y1 receptor antagonist MRS2179 (1 μm) almost abolished ATP (1 μm) and ADPβS (1 μm)-evoked dilatation, but had no effect on dilatation to UTP (3 μm) (Fig. 4C).
In the presence of l-NAME, incubation with the selective IKCa channel blocker, TRAM-34 (1 μm), had no effect on the vasodilatation responses to ATP, ADPβS or UTP (Fig. 4A and B). In contrast, incubation with the selective SKCa channel blocker, apamin (50 nm), inhibited the response to ATP, causing a rightward shift of the concentration–response curve (Fig. 4A). This inhibition of the vasodilatation to ATP was most apparent when comparing responses evoked by submaximal concentrations of ATP (1 μm) (Apamin: 33.6 ± 10.3%, n= 12; Fig. 4B). Apamin had no significant effect on the dilatation to ADPβS or UTP, although in 2 out of 11 arteries, vasodilatation to ADPβS was almost completely abolished by apamin alone. Blockade of both IKCa and SKCa channels with a combination of TRAM-34 and apamin significantly inhibited the vasodilatation evoked by all agonists (ATP: 6.4 ± 4.5, n= 8; ADPβS: 12.9 ± 3.5%, n= 9; UTP: 12.7 ± 6.1%, n= 9, Fig. 2B). The response to ACh (1 μm) was similarly affected (Control: 89.6 ± 1.0%, n= 38; TRAM-34 + Apamin: 21.1 ± 6.4%, n= 20). Raising extracellular K+ (35–40 mm) in the presence of l-NAME (100 μm) in both the bath and the lumen of the vessel completely abolished the relaxation response to each of the luminally perfused P2Y receptor agonists (n= 3) and ACh (1 μm, n= 11).
Contribution of NO to spreading dilatation
Despite the observation that the direct agonist-evoked dilatation to either luminal infusion of ATP (Fig. 4A) or bath application of ACh (Mather et al. 2005) are not affected by an inhibitor of NO synthase, the release of NO could potentially affect the spreading dilatation response to either of these agonists. Therefore a set of experiments was performed in the absence of l-NAME, and it was found that the spreading dilatation was effectively identical, suggesting a minor role, if any, by released NO (Fig. 5A). To assess whether released NO itself could evoke spreading dilatation, the NO donor DEA NONOate was luminally infused and found to evoke dilatation in Branch 1 (61.8 ± 10.6%, n= 5) but did not spread into the feed artery beyond the modest dilatation observed at 0 μm (8.5 ± 7.2%, n= 5).
We have previously shown an important role for the endothelium in spreading dilatation responses, but were limited to the use of the smooth muscle-dependent hyperpolarizing agent, the KATP channel opener levcromakalim (Takano et al. 2004). The triple-cannulation arrangement has enabled us to provide the first clear demonstration of the role of the endothelium in a pressurized mesenteric artery in response to endothelium-dependent agonists. By selectively damaging the feed artery and Branch 2, near maximal endothelium-dependent dilatation was stimulated in Branch 1 by ATP (67.6 ± 15.2%, n= 4) and ACh (85.3 ± 8.4%, n= 4). Interestingly, the dilatation to ATP could not spread into the feed artery, but instead a constriction was observed at 0 μm (−19.6 ± 8.2% dilatation, n= 4), whereas in the same arteries ACh dilatation spread as far as 0 μm (17.5 ± 5.0%, n= 4) but did not reach 500 μm (Fig. 5B). Note that the feed artery maximally dilated to levcromakalim or papaverine, when added to the bath.
Spreading dilatation to abluminal application of drugs
When applied to the outside of arteries, the P2 receptor agonists were applied as short pulses of 1 mm solutions (Fig. 6A), likely transiently reaching higher concentrations than when applied through the lumen. In this case, is it likely that ATP would stimulate P2X as well as P2Y receptors located on smooth muscle cells. Following the application of ATP at a local site, arteries rapidly and transiently contracted (peak 50.1 ± 6.6%, n= 4), and then immediately dilated by 17.5 ± 7.5% (n= 4). This local response to ATP coincided with synchronous contraction and dilatation along the entire artery length, with a slight reduction in the extent of response with distance (Figs 6B and 7). The local dilatation response to ADPβS was greater in amplitude (90.3 ± 5.9%, n= 4) and mimicked acetylcholine more than ATP, although a small, non-spreading constriction was observed at the local site (Figs 6C and 7). Dilatation to ATP and ADPβS was significantly inhibited by MRS2179 (not shown, n= 3), and fully inhibited by a combination of TRAM-34 and apamin (Fig. 8A and B). UTP was unable to evoke local or spreading dilatation when applied to the outside of arteries (Figs 6D and 7).
The results presented in this study show that ATP and UTP each evoke local and spreading dilatation when infused into a sidebranch of isolated and pressurized rat mesenteric arteries. The spreading dilatation response was rapid and decayed with distance, and was evident 2 mm upstream from the site of initiation of dilatation, an effect sufficient to increase flow into the feed artery. Further, the response was fully blocked by inhibiting SKCa and IKCa channels in the endothelium. Thus, we show for the first time that the presence of either ATP or UTP in the lumen of mesenteric arteries will stimulate marked dilatation both locally and away from the local site of release and action, suggesting an effective and significant ability to reduce vascular resistance and increase blood flow.
The only other report to demonstrate upstream spread of dilatation to luminally applied ATP relied on impalement through the wall of an arteriole with a small-tipped sharp pipette, and the subsequent delivery of agonist into the lumen with pressure-pulse ejection in exteriorized hamster retractor muscle (McCullough et al. 1997). This study reported dilatation responses only ∼150 μm upstream from the delivery pipette (< 2 endothelial cell lengths), which were sensitive to l-NAME, suggesting that nitric oxide was the sole factor responsible. In contrast, there are a few reports that the abluminal delivery of ATP can evoke spreading dilatation, but these were all complicated by local vasoconstriction due to direct smooth muscle stimulation (Dietrich et al. 1996; McCullough et al. 1997).
In the present study, local dilatation to luminally infused ATP was continuous for 2 min periods and was able to evoke maintained levels of both local and spreading dilatation in the presence of l-NAME. Near maximal dilatation to ATP in a sidebranch spread rapidly into the feed artery and then at least 2 mm upstream. Any possibility of the agonist reaching upstream sites was eliminated by continuous fluid flow through the feed artery, and confirmed by monitoring the fluorescence of the infused solutions. Importantly, when applied directly to the lumen of arteries, ATP stimulated near maximal dilatation without constriction, but when focally applied to the outside of arteries, this robust dilatation to ATP was not evident. A transient vasoconstriction preceded dilatation, presumably due to activation of the smooth muscle P2X receptors which are present in this tissue (Gitterman & Evans, 2001). The modest dilatation following abluminal application suggests the ATP was hydrolysed to vaso-inactive products before reaching the endothelium in any significant amounts. As luminal perfusion of ATP evoked a significant EDHF dilatation, which was fully blocked by the P2Y1 receptor antagonist MRS2179, we explored the response further using ADPβS. This agonist was also sensitive to block with MRS2179, but was not hydrolysed and is inactive at P2X receptors. Luminal and abluminal delivery of ADPβS evoked robust local and spreading dilatation responses, very similar to those stimulated by acetylcholine. So these data demonstrate that stimulation of P2Y1 receptors by endogenous purine nucleotides provides a very effective mechanism to evoke dilatation in rat resistance arteries, provided the intraluminal levels are maintained in the low micromolar range.
The possibility that stimulation of P2Y2 receptors could also stimulate spreading dilatation responses was confirmed using UTP, which infused through the sidebranch of arteries at slightly higher concentrations (∼3 fold higher than ATP) mimicked the response to ATP, ADP and ADPβS. However, UTP only stimulated vasoconstriction following focal abluminal application. This may reflect an action of UTP at smooth muscle cell P2Y receptors (Gitterman & Evans, 2000; Buvinic et al. 2002), and as with ATP, rapid degradation to inactive products before reaching the endothelium, but in contrast to ADP, UDP is unable to evoke dilatation in this artery (Malmsjo et al. 2000a, 2002). The actions of adenosine, a vasodilator that can also stimulate spreading dilatation (Delashaw & Duling, 1991; Rivers & Frame, 1999; Duza & Sarelius, 2003), were also investigated, and despite stimulating local dilatation, the ability of adenosine to stimulate spreading dilatation was clearly less than evoked by the nucleotides or acetylcholine. Thus although adenosine is clearly involved in hypoxia-mediated increases in blood flow to skeletal muscle (Edmunds & Marshall, 2001), it appears that the process we describe here is independent of the rises in endothelial cell Ca2+ levels and NO release described in vivo and in the rat aorta (Ray & Marshall, 2005, 2006), and is likely to act as a parallel pathway.
Spreading dilatation to nucleotides or acetylcholine decayed at the same rate, presumably reflecting similar pathways for vasodilatation. Dilatation was abolished by raised extracellular KCl, confirming a role for smooth muscle cell hyperpolarization (Malmsjo et al. 1999; Mistry et al. 2003). Since P2Y1 and P2Y2 receptors are all coupled to PLC, endothelial cell Ca2+ would be expected to increase in response to all the nucleotides, as observed with ATP (1 μm) (Liu et al. 2006) and acetylcholine (Mather et al. 2005; McSherry et al. 2005) in this artery. So as with acetylcholine, a rise in endothelial cell Ca2+ in response to nucleotides stimulates KCa channels (Malmsjo et al. 1999; Mistry et al. 2003) which are present in endothelial, but not present in smooth muscle cells in these arteries (Walker et al. 2001; McSherry et al. 2005; Sandow et al. 2006). In mesenteric arteries (Takano et al. 2004), spreading dilatation to acetylcholine is associated with hyperpolarization along the arterial length, but rises in endothelial cell Ca2+ only occur at the site of direct agonist stimulation, suggesting EDHF is not released at the upstream sites. Here in the bifurcated arteries, the directly stimulated cells can only reach a maximum of one cell length (< 100 μm) upstream into the feed artery. Thus for the signal responsible for spreading dilatation to travel a distance 500 μm upstream, at least five non-directly stimulated endothelial cells must be traversed (see Supplemental Fig. 3). Although we have not recorded membrane potential in this study, at least the local dilatation response to nucleotides is dependent on opening KCa channels. Both in the microcirculation and in resistance arteries, hyperpolarization due to agonist-mediated opening of endothelial cell KCa channels spreads longitudinally between coupled endothelial cells, and then to the surrounding smooth muscle cells along the artery length (reviewed in Domeier & Segal, 2007). In the mesenteric artery, the smooth muscle cells cannot sustain spreading dilatation over long distances (Takano et al. 2004), highlighting the importance of the endothelium as the conduit for increases in potential (Haas & Duling, 1997; Yamamoto et al. 1999; Emerson & Segal, 2000). Indeed here we show that we can evoke endothelium-dependent dilatation to ATP and ACh in one (undamaged) branch of the artery, but the spread of dilatation does not pass to any great extent into regions of damaged endothelium, further supporting the importance of the endothelium in transferring signals upstream. Importantly, hyperpolarization following opening of other K+ channels, such as ATP-sensitive K+ channels, also leads to spreading hyperpolarization and dilatation in this artery, but still relies on the endothelium for passage of current and dilatation upstream (Takano et al. 2004). In the mesenteric artery, it appears that hyperpolarization per se does not increase endothelial cell Ca2+ levels, and that a change in endothelial cell Ca2+ levels is not essential for spreading dilatation (Takano et al. 2004). So the mechanisms leading to hyperpolarization are not important in determining cell-cell coupling. However, this mechanism for spreading dilatation is clearly not universal, as in some tissues other pathways clearly play a role, where agonists may stimulate additional signalling pathways that can either augment (e.g. by cAMP) (Popp et al. 2002) or reduce (e.g. by α1-adrenoceptor stimulation; PKC) (Haug et al. 2003; Bao et al. 2004) the ability of the hyperpolarization (and/or other conducted signals) to spread between cells. The endothelium-dependent spreading dilatation responses were not altered significantly in the presence of an NO synthase inhibitor, consistent with other vascular beds (Domeier & Segal, 2007), but our study does not preclude any complex underlying or parallel pathways related to slow waves of Ca2+ and the release of NO which can also occur.
The requirement for continuous delivery of the purine and pyrimidine triphosphates to maintain dilatation (Liu et al. 2004) suggests that when they are released into the extracellular environment, for example from circulating red blood cells or platelets (Miseta et al. 1993; Ellsworth et al. 1995; Beigi et al. 1999; Dietrich et al. 2000; Ellsworth, 2004), the ensuing dilatation responses will only be maintained if the triphosphate supply continues. For any released ATP to effectively improve tissue blood flow in vivo, spreading dilatation would provide a more effective means to reduce vascular resistance. In the microcirculation in vivo, if dilatation only occurs focally, as has been shown with NO, the increase in blood flow into that artery is negligible, whereas if an agonist such as acetylcholine was used, which also stimulates spreading dilatation, blood flow was markedly improved (Kurjiaka & Segal, 1995; Dora et al. 2000). Although this parallel has not been explored in vivo in the rat mesenteric artery arcade, it is interesting to speculate that the observed graded dilatation to approximately 80 μm at the downstream end of the feed artery can manifest as a reduction in vascular resistance, and improve blood flow into that artery, as was observed in the isolated and triple-cannulated arteries. This ability of ATP and UTP to stimulate spreading dilatation would more effectively improve tissue blood flow in response to their release from blood bourn or surrounding cells than the actions of NO, which did not evoke significant spreading dilatation.
In conclusion, we show that the presence of low micromolar concentrations of nucleotides stimulate dilatation attributable to EDHF that spreads between cells of the arterial wall to evoke responses that, like acetylcholine, can be observed 2 mm upstream from the site of agonist delivery. The local responses are associated with opening KCa channels, and the local and spreading dilatation responses to both ATP and UTP are more pronounced when infused into the lumen of arteries. The presence of P2Y receptors on the endothelium therefore provides an important mechanism for regulating tissue blood flow.
We are grateful for helpful discussions with Dr Richard Rivers, Dr Amanda MacKenzie and Prof Chris Garland. This work was funded by the Wellcome Trust (UK), and a British Heart Foundation PhD studentship.