Department of Cell Physiology and Pharmacology, Medical Sciences Building, University of Leicester, Leicester LE1 9HN, UK. E-mail: RJE6@le.ac.uk
Diadenosine polyphosphates (ApnAs, n=3–7) and adenosine polyphospho guanosines (ApnGs, n=3–6) are naturally occurring vasoconstrictor substances found in platelets. These vasoconstrictor actions are thought to be mediated through the activation of P2X receptors for ATP. The effects of ApnAs and ApnGs at P2X receptors on rat mesenteric arteries were determined in contraction studies and using the patch clamp technique on acutely dissociated artery smooth muscle cells.
P2X1 receptor immunoreactivity was detected in the smooth muscle layer of artery rings. The sensitivity to α,β-methylene ATP and desensitizing nature of rat mesenteric artery P2X receptors correspond closely to those of recombinant P2X1 receptors.
Ap4A, Ap5A and Ap6A evoked concentration dependent P2X receptor inward currents which desensitized during the application of higher concentrations of agonist. The agonist order of potency was Ap5AAp6AAp4A>>Ap3A. Ap2A and Ap7A were ineffective. Similar results were obtained in contraction studies except for Ap7A which evoked a substantial contraction.
ApnGs (n=2–6)(30 μM) evoked P2X receptor inward currents in mesenteric artery smooth muscle cells. ApnGs (n=4–6) were less effective than the corresponding ApnA.
This study shows that at physiologically relevant concentrations ApnAs and ApnGs can mediate contraction of rat mesenteric arteries through the activation of P2X1-like receptors. However the activity of the longer chain polyphosphates (n=6–7) may be overestimated in whole tissue studies due to metabolic breakdown to yield the P2X receptor agonists ATP and adenosine tetraphosphate.
The vasoconstrictor effects of a range of naturally occurring ApnAs and ApnGs have been determined in a limited number of studies on perfused vascular beds. This work has shown that the activity of these compounds is dependent on the length of the polyphosphate chain (Ralevic et al., 1995; van der Giet et al., 1997; Schlüter et al., 1998). Similar results have been reported for the actions of ApnAs at recombinant P2X1 receptors (Evans et al., 1995; Wildman et al., 1999; Bianchi et al., 1999). However a direct quantitative comparison of the effects of ApnAs at arterial smooth muscle and recombinant P2X1 receptors is difficult to make as drugs were applied in a bolus in the perfused vascular bed experiments and thus the effective agonist concentration at the receptor was not determined. In addition the metabolism of nucleotides by ecto-enzymes can lead to underestimating agonist potency in organ bath studies (Benham & Tsien, 1987; Inoue & Brading, 1990; Evans & Kennedy, 1994; Trezise et al., 1994). For example recent studies have suggested that Ap4A may be subject to ecto-enzyme breakdown (Westfall et al., 1997). Complications associated with agonist breakdown can be overcome in patch clamp studies on acutely dissociated smooth muscle cells where drugs can be applied rapidly under concentration clamp conditions (Evans & Kennedy, 1994). This system has the added advantage that currents recorded in response to drugs give a direct measure of activation of the P2X receptor channels. The primary aim of this study was to characterize the effects of a range of ApnAs, ApnGs and related compounds on rat mesenteric artery smooth muscle P2X receptors to allow comparisons to be made between native smooth muscle and recombinant P2X1 receptors and to investigate further the structure activity relationships of these compounds.
Male wistar rats (250–300 g) were killed by cervical dislocation and femoral exsanguination. The mesentery was removed and second and third order mesenteric arteries were dissected. For contraction experiments second order mesenteric artery rings were mounted in a Mulvany myograph (vessel diameter 226.5±11.4 μm, n=8) (Lagaud et al., 1996). The arteries were superfused at 2 ml min−1 with a warmed physiological saline with the following composition (mM): NaCl 150, KCl 5, HEPES 10, CaCl2 2.5, MgCl2 1, (pH to 7.3 with NaOH). The organ bath temperature was 37°C. Drugs were added to the superfusate. The vasoconstrictor agonists evoked transient contractions of the artery that returned to baseline diameter in the continued presence of purinergic drugs (Figure 2a). Reproducible contractions were obtained when agonists were applied at 30 min intervals.
Acutely dissociated smooth muscle cells were prepared by a two step papain and collagenase/hyaluronidase enzymatic digestion (see Quayle et al., 1996). Cells were plated onto glass coverslips, stored at 4°C and used within 2–36 h. For electrophysiological recording smooth muscle cells were superfused at 2 ml min−1 with physiological solution and drugs were applied rapidly using a U-tube perfusion system (Evans & Kennedy, 1994). Amphotericin permeabilized patch recordings were made with an Axopatch 200B amplifier and data was collected using pClamp6 software (Axon Instruments U.S.A.). Holding potential was −60 mV. Patch electrodes (2–5 MΩ) were filled with a solution of the following composition (mM): potassium gluconate 140, NaCl 5, HEPES 10, EGTA 9, (pH adjusted to 7.3 with KOH). Reproducible responses to agonists (200–500 ms pulse duration) were obtained when a 5 min interval was given between agonist applications.
Data are reported throughout as mean±s.e.mean, n=number of observations. Peak currents in response to drugs are expressed as percentage of the response to 10 μM α,β-meATP, this is a maximal concentration at rat mesenteric artery P2X receptors (Lewis et al., 1998). It was not always possible to construct a concentration response curve in a single smooth muscle cell therefore pooled concentration response data for agonists were fitted by the least squares method using Origin software (Microsoft U.S.A.) with the equation; response=α[A]H/([A]H+[A50]H) where α and H are the asymptote and Hill coefficient, [A] is the agonist concentration, A50 is the agonist concentration producing 50% of the maximum agonist response (EC50). Differences between means were determined by Student's unpaired t-test (two tailed) and were considered significant when P<0.05.
For immunohistochemical studies second and third order mesenteric arteries were dissected as for electrophysiology experiments and embedded in Tissue-Tek (Miles Inc, Elkhart, In.) and frozen over dry ice and hexane. Frozen transverse arterial 12 μm sections were cut and mounted on pre-subbed slides. The sections were fixed in a 2% paraformaldehyde solution in Sorensons buffer for 10 min. After washing in a phosphate buffered saline (PBS) solution the slides were incubated in a blocking solution (10% donkey serum (Jackson Immunoresearch), 0.5% Triton-X (Sigma) in PBS for 30 min at room temperature to permeabilize the tissue and reduce non-specific binding. Sections were then washed and incubated overnight in primary antisera (Anti-P2X1±control antigen blocking peptide, Alomone Lab. Israel) at 4°C. Anti-P2X1 is a polyclonal antibody raised in rabbit against the C-terminal residues 382–399 of the P2X1 receptor. The anti-P2X1 receptor antibody was used at a dilution of 1 : 200 in 10% donkey serum in PBS. When blocking peptides were used, the primary antibody and its corresponding antigen peptide were incubated together (1 mg of each) for 1 h at room temperature. After washing in PBS, sections were incubated in secondary antisera for visualization, which in each case was fluorescein isothiocyanate (FITC) labelled anti-rabbit IgG raised in donkey (Jackson Immunoresearch) (1 : 100 in 10% donkey serum in PBS) for 2 h at room temperature. As a control to determine the amount of non-specific binding, slides were incubated in 10% donkey serum in PBS only or secondary antisera only. Slides were washed in PBS and mounted in Citifluor (UKC Chem Lab, U.K.). Arterial sections were examined under epifluorescence with neutral density filters. Images were captured using Scionimage software.
Drugs: papain, dithioerythritol, collagenase, hyaluronidase, α,β methylene ATP (α,β-meATP, Ap2A, Ap3A, Ap4A, Ap5A, Ap6A, GTP, adenosine tetraphosphate, tetrapolyphosphate (Hexa-ammonium salt), and trisodium trimetaphosphate (Sigma). HPLC analysis of diadenosine polyphosphates from Sigma has shown that they are of high purity and are not contaminated with ATP (Pintor et al., 1996). Ap7A, Ap2G, Ap3G, Ap4G, Ap5G, and Ap6G (synthesized and HPLC purified by H. Schlüter).
Effects of ApnAs on acutely dissociated rat mesenteric smooth muscle cells
The diadenosine polyphosphates Ap4A, Ap5A and Ap6A evoked concentration dependent inward currents which declined during the continued presence of high concentrations of agonist (Figure 1a). Ap5A evoked responses had an EC50 of 6.7 μM (pA50 5.23±0.12) and a slope of 1.1±0.23, the maximal response to Ap5A was 80.8±7.2% of the response to 10 μM α,β-meATP (n=5). The mean concentration response relationship for Ap6A gave an EC50 of ∼13 μM and a maximum of 43±7.5% of the response to 10 μM α,β-meATP (n=5–7) (Figure 1a,b). No clear maximum response to Ap4A was recorded at concentrations up to 300 μM (n=4) (∼50 μM Ap4A evoked a response equivalent to 50% of the maximal α,β-meATP response). Ap3A evoked a small current at 100 μM which was 21.3±5.7% of the maximal α,β-meATP evoked response. In cross desensitization studies following α,β-meATP (10 μM) treatment no response to ApnAs (all 30 μM, n=3–6) was recorded indicating that all these agonists acted at the same receptor. No response was recorded to applications of Ap2A or Ap7A (30 and 100 μM) (n=4–6).
Comparison of ApnA evoked contractions in intact arteries and currents recorded from isolated smooth muscle cells
The majority of studies on arterial smooth muscle P2X receptors have recorded contractile responses. These whole tissue pharmacological studies may be complicated by problems associated with agonist metabolism by ecto-enzymes and the contractile response only gives an indirect measure of P2X receptor activation. It was therefore of interest to compare ApnA mediated P2X receptor inward currents and contractions in the rat mesenteric artery. Ap5A evoked rapid, concentration dependent, transient constrictions of rat mesenteric artery wings which returned to baseline during the continued presence of the agonist (Figure 2a). The EC50 values for evoking contraction and inward currents were essentially the same (6.5 and 8.5 μM respectively, Figure 2b). The maximal constrictor response was 113±11.6% of the response to 10 μM α,β-meATP compared to 80.8±7.2% for inward currents. Following desensitization of smooth muscle P2X receptors in the continued presence of 10 μM α,β-meATP the response to 100 μM Ap5A was abolished (n=4) (Figure 2c). These cross desensitization experiments suggest that Ap5A and α,β-meATP are acting through the same P2X receptor to mediate contraction.
The effects of 10 μM ApnAs (n=2–6) in contractile and patch clamp studies are shown in Figure 2d. Ap2A (upto 100 μM) and Ap3A (upto 30 μM) did not evoke contractions or inward currents in the mesenteric artery. Ap2A (10 μM) had no effect on contractions evoked by an EC50 concentration of α,β-meATP (1 μM) (responses 103±2.6% of control, n=4). In contrast Ap3A (10 μM) potentiated by 59±15% (n=6) contractions evoked by α,β-meATP (1 μM). Ap4A appeared less effective in mediating vasoconstriction than inward currents (8.8±1.3% and 19.1±5.2% respectively of the response to 10 μM α,β-meATP, n=6). Although this effect is not significant, given that Ap5A and Ap6A were more effective in evoking contractions than inward currents, it may indicate that Ap4A is subject to ecto-enzyme breakdown in whole tissue studies as has been reported previously (Westfall et al., 1997). Ap6A was significantly more effective in mediating vasoconstriction than inward currents (50.3±4.3% and 21.2±4% of the response to 10 μM α,β-meATP n=6 and 5 respectively P<0.005). Ap7A was an effective agonist in mediating transient contractions (response to 10 μM Ap7A was 50.1±11.2% of response to 10 μM α,β-meATP, n=4) (Figure 2d). In cross desensitization studies following pre-treatment with α,β-meATP (10 μM) contractile responses to Ap7A were abolished (n=3). These results indicate there is a major difference in the activity of Ap7A in contractile and patch clamp studies and suggest that ApnAs may be subject to metabolism in whole tissue studies.
Effects of ApnGs on acutely dissociated rat mesenteric smooth muscle cells
The ApnGs mediate vasoconstriction in the rat isolated perfused kidney preparation and have been suggested to mediate this response through the activation of P2X receptors (Schlüter et al., 1998). In the present study ApnGs (n=2–6, 30 μM) evoked P2X receptor mediated inward currents in isolated smooth muscle cells (Figure 3). Ap3G was significantly more effective than Ap3A in evoking inward currents (P<0.001). In contrast Ap4G, Ap5G and Ap6G were significantly less effective at evoking inward currents than the corresponding diadenosine polyphosphate compound (P<0.05). In cross desensitization studies following α,β-meATP treatment (10 μM) ApnAs (n=4–6, 30 μM) had no effect on holding current indicating that they were acting through the same P2X receptor (n=3–6). In whole tissue experiments ApnGs (n=4–6) were effective in evoking contractions, the response to 30 μM ApnG (n=4–6) were 7.3±1.6, 44.1±10.3 and 55.4±7.9% respectively of the contractile response to 10 μM α,β-meATP (n=4). Thirty μM Ap3G failed to evoke a contractile response (n=4).
Effects of other ‘polyphosphate’ compounds on P2X receptors
The action of ApnAs and ApnGs at P2X receptors raises questions about ligand recognition by P2X receptors and possible metabolism of these compounds in whole tissue studies. In order to investigate this further we have determined the effects of a number of related ‘polyphosphate’ compounds. GTP, trimetaphosphate and tetrapolyphosphate (100 μM) were ineffective as agonists at mesenteric artery P2X receptors (n=4–6) and ineffective as antagonists (Cook & Evans, 1999 unpublished observations) (Figure 4a). These results show that an adenosine moiety and a phosphate chain are required for ligand recognition by mesenteric artery P2X receptors. Adenosine tetraphosphate evoked similar currents to those seen with α,β-meATP (Figure 4b,c) with an EC50 of 0.51 μM. Responses to adenosine tetraphosphate (10 μM) were abolished following desensitization with 10 μM α,β-meATP (n=3). At adenosine tetraphosphate concentrations 1 μM recovery from receptor inactivation was prolonged, for example following an application of 1 μM adenosine tetraphosphate a subsequent response to 10 μM α,β-meATP required 15 min to return to pre adenosine tetraphosphate control levels, and at concentrations of 3 and 10 μM adenosine tetraphosphate responses to α,β-meATP (10 μM) were only ∼70% of control values after 30 min (n=4). The maximal adenosine tetraphosphate response was 159.5% of the response to 10 μM α,β-meATP.
Immunohistochemical localization of P2X1 receptors in rat mesenteric artery smooth muscle cells
A polyclonal antibody raised against the carboxy terminus of the P2X1 receptor produced selective staining of the smooth muscle cell layer of mesenteric artery rings that was abolished by pre-incubation of the antibody with a sequence specific blocking peptide, similar detection of P2X1 receptor immunoreactivity in smooth muscle has been detected previously (Vulchanova et al., 1996) (Figure 5). A high level of background autofluorescence was associated with the endothelial cell layer (Figure 5b).
This study has shown that a range of ApnAs evoked inward currents in rat mesenteric artery smooth muscle cells through the activation of P2X receptor cation channels and that activity of these compounds is dependent on the length of the phosphate chain. The study extends on previous work and documents the effective concentration range of agonists and presents new data on the structure activity relationships of P2X receptors. These data show that at concentrations likely to be present physiologically ApnAs and ApnGs may contribute to the control of blood pressure through the direct activation of P2X1-like receptor cation channels.
The use of patch clamp methods in this study has allowed a direct comparison of the effects of ApnAs between P2X receptor mediated currents in rat mesenteric smooth artery cells and at recombinant P2X1 receptors (Evans et al., 1995; Wildman et al., 1999; Bianchi et al., 1999) to be made. Previous comparisons of native smooth muscle and recombinant P2X1 receptors have been based on the timecourse of response and sensitivity to the agonists α,β-meATP, 1-β,γ-methylene ATP and the antagonist trinitrophenyl ATP. We extended this to compare further native smooth muscle P2X receptors and recombinant P2X1 receptors. Our results are similar to those reported for human P2X1 receptors expressed in mammalian HEK293 or 1321N1 astrocytoma cells (Evans et al., 1995; Bianchi et al., 1999). There are however marked differences in the actions of Ap4A and Ap6A in rat mesenteric artery smooth muscle cells and at recombinant rat P2X1 receptors expressed in Xenopus oocytes (Wildman et al., 1999). A degree of caution however must be used in drawing conclusions based on these findings as discrepancies between the properties of recombinant P2X receptors expressed in mammalian and amphibian systems have been reported (e.g. comparison of the studies of Brake et al., 1994; Evans et al., 1995; Pintor et al., 1996). Thus it remains to be confirmed whether the native rat mesenteric artery smooth muscle P2X receptor is different from recombinant rat P2X1 receptors or if the differences can be accounted for by the difference in the expression system. However the general properties of rat mesenteric artery P2X receptors and the presence of P2X1 receptor immunoreactivity on these vessels suggest strongly that the P2X1 receptor characteristics dominate this native phenotype (Evans & Surprenant, 1996; Lewis et al., 1998).
Ap7A has been shown recently to be released by platelets and has a vasoconstrictor action in the isolated perfused rat kidney (Jankowski et al., 1999). In the present study we show that Ap7A evokes α,β-meATP sensitive contractions of the rat mesenteric artery. However Ap7A failed to evoke inward currents in patch clamp studies. Similarly Ap6A was more effective in evoking contractions than P2X receptor currents. These results suggest that in whole tissue studies Ap6A and Ap7A may be subject to metabolism, as has been shown for Ap4A (Westfall et al., 1997), and that when added under concentration clamp conditions in patch clamp studies the true potency of the agonist is revealed. Cleavage of the phosphate chain could give rise to ATP and adenosine tetraphosphate (Mateo et al., 1997) which then act locally to activate P2X receptor channels (Evans et al., 1995 and this study). As Ap6A is an agonist of P2X receptor currents it is likely that in contraction studies the response results from activation of the P2X receptor by both Ap6A and metabolic products. Thus the activity of Ap6A and particularly Ap7A directly at P2X receptors may be overestimated in whole tissue studies due to metabolism.
The ApnGs evoked inward currents and contractions through mesenteric artery P2X receptors. Previously ApnGs have been shown to mediate vasoconstriction in the perfused kidney preparation (Schlüter et al., 1998) however this is the first time a direct role for P2X receptors in this process has been demonstrated. Ap2G and Ap3G were more effective than the corresponding ApnA at evoking inward currents. The increased efficacy of Ap3G relative to Ap3A may reflect that binding of the γ-phosphate to the P2X receptor is influenced by the chemical group substituted at the ester oxygen. When this group is a phosphate i.e. adenosine tetraphosphate the potency is equivalent to α,β-meATP. The potency order of the chemical groups at this position is phosphate>>guanosine>adenosine. In contrast Ap4G, Ap5G and Ap6G were less effective than the corresponding ApnA. The reduced potency of ApnG (n=4–6) compared to the corresponding ApnA may simply reflect the reduced probability of the adenosine moiety binding with the P2X receptor.
A number of ‘polyphosphate’ compounds including Ap2A, Ap3A, GTP, trimetaphosphate and tetrapolyphosphate were ineffective as agonists at the P2X receptor. When these compounds were co-applied with α,β-meATP (1 μM, an EC50 concentration) they either had no effect on the amplitude of the P2X receptor mediated responses or slightly potentiated them. These results demonstrate that binding of an adenine group in the case of Ap2A or a triphosphate group in the case of GTP is ineffective in activating the receptor. The fact that these compounds are ineffective as antagonists suggests that for high affinity binding at the P2X receptor both an adenine group and a triphosphate moiety are required.
The activation of platelets leads to the release of a variety of vasoactive nucleotides. For example ApnAs and ApnGs have been estimated to be present at 0.5–3 μM in the supernatant following platelet aggregation (Schlüter et al., 1998). Recent studies on platelets have shown that it is likely that the nucleotide concentration at the local microenvironment of the cell surface may be at least an order of magnitude higher than that measured in the supernatant (Beigi et al., 1999). ApnAs and ApnGs may therefore be present locally in the 5–30 μM range, concentrations that were effective in activating P2X receptors in this study. When the endothelium is damaged there may be a local rise in ApnAs and ApnGs following platelet activation resulting in the activation of P2X receptors and subsequent vasoconstriction.
In summary these results have shown that diadenosine polyphosphates and adenosine polyphospho guanosines at physiologically relevant concentrations can act directly at P2X1-like receptor ion channels to mediate vasoconstriction in rat mesenteric artery. The activity of these compounds is dependent on the length of the polyphosphate chain. In addition it is possible that ApnAs and ApnGs (n=6–7) may be metabolised in whole tissue studies to yield the P2X receptor agonists ATP and adenosine tetraphosphate that may mediate some of their vasoconstrictor actions.
We would like to thank the Wellcome Trust, the MRC and DFG Schl 406/1–2 for support.