INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

FREE PURINE NUCLEOSIDES and nucleotides present in body fluids are responsible for a wide range of pharmacological effects through P1 or P2 purinoceptors expressed on the surface of many cell types (25). P2 receptors are mostly sensitive to ATP and ADP, whereas P1 receptors respond to adenosine. P2 receptors are themselves divided into two main families: a P2X family consisting of ligand-gated ionic channels and a P2Y family consisting, like P1 molecules, of G protein-coupled receptors (25). Seven P2X (1-7) and eleven P2Y (1-11) receptors have been characterized, whereas the P1 family comprises four receptors: A₁, A_2A, A_2B, and A₃. The widespread actions of extracellular purines include effects on the central nervous, cardiovascular, respiratory, and gastrointestinal systems among others (10). It is also well established that they can mediate either vasoconstriction or vasodilatation and contraction or relaxation of visceral smooth muscles via different receptors (5). This has motivated numerous experimental approaches to define the potential therapeutic interest of natural purines in vivo (2, 15), which, today, have a clinical use for the treatment of cardiac arrhythmia, supraventricular tachycardia, pulmonary hypertension (12), and pain (29). However, in all cases, purines are given by intravenous perfusion despite the presence of efficient nucleoside transporters in the gut that salvage nucleosides and nucleotides of dietary origin (7, 14, 33). The possible pharmacological effects of oral administration of free purines has not been investigated, likely because a large fraction of dietary nucleosides is believed to be used for metabolic purpose (6). Yet it remained possible that supplementation with purines may modify the metabolism and turnover of nucleosides. This could modulate the synthesis and release of pharmacologically active purines and subsequently affect some basal physiological parameters through P1 or P2 receptors. Given the pharmacological activity of intravenous adenosine and ATP on cardiovascular and respiratory parameters, experiments were designed to investigate the potential effect of a long-term supplementation with ATP on these important biological functions. ATP was chosen because of its own pharmacological activity and its capacity to generate other active purines by metabolic degradation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals. Male New Zealand White rabbits (3-3.5 kg; Centre d'Elevage et de Selection J. Barrois, Tressin, France) were kept under normal conditions with free access to food and water. Animals fasted for 12 h before surgery and testing of the physiological parameters. All experiments were carried out in accordance with the European Communities Council Directive of 24 November 1986 (86/609/EEC).

Experimental design. On the basis of the dose range used for clinical trials performed by low perfusion (1), for chronic treatment, animals were given 3 mg · kg¹ · day¹ (n = 4) or 20 mg · kg¹ · day¹ (n = 12) ATP mixed with cellulose Avicel PH101 as vehicle (Atepadene, Laboratoires Galéniques Vernin) or 20 mg · kg¹ · day¹ (n = 4) adenosine hemisulfate salt (Sigma Chemical) for 14 days. ATP and adenosine, dissolved in saline, were administered daily via a gastric cannula. Control rabbits received a corresponding amount of saline. For measurement of P1 (theophylline, Sigma Chemical) or P2 (cibacron blue, Sigma Chemical) antagonist effects, acute intravenous injections were performed after establishment of steady-state anesthesia on control (n = 6) and ATP-treated (20 mg/kg) rabbits (n = 6). Theophylline (10 mg/kg) and cibacron blue (5 mg/kg) were dissolved in saline solution and injected into the marginal vein of the ear as described by Konduri et al. (16). Cardiovascular and respiratory parameters were recorded 20 min after administration.

1

1

Surgical preparation. Rabbits were anesthetized with a relaxing intramuscular injection of ketamine (Clorketam, 15 mg/kg) followed by an injection of 30 mg/kg pentobarbital sodium in the marginal vein of the ear. Anesthesia was maintained by a slow perfusion of 10 mg · kg¹ · h¹ pentobarbital sodium throughout the experiment. Lidocaine (2%) was applied topically at sites of incision.

Experimental parameters and calculations. Heart rate (HR, beats/min) was derived from the electrocardiogram. DAP (mmHg) and SAP (mmHg), central venous pressure (mmHg), and IVBF (ml/min) were directly measured on records. The left ventricular work index (LVWI) was calculated as LVWI = SAP × (HR)^1/2 according to the method of Eberlein (11). Venous resistance (VR) was calculated as VR = SAP/IVBF. Both inspiratory and expiratory waveforms of the respiratory cycle were integrated and recorded. Lung resistance (cmH₂O · l¹ · s) and respiratory frequency (breaths/min) were calculated by the on-line respiratory monitoring system (pulmonary monitoring system 800, Mumed, London, UK).

Purine extraction and characterization. Immediately after blood collection, plasma and erythrocytes were separated by a 30-s centrifugation at 2,000 g. Erythrocytes were washed in Hanks' balanced salt solution. Purines were extracted from plasma (100 µl) and washed erythrocytes (25 µl) with 1 ml trichloroacetic acid (7%, wt/vol) for 30 min at 4°C. Samples were then centrifuged for 30 s at 2,000 g, and supernatants were neutralized with a tri-n-octylamine-trichlorotrifluoroethane mixture (0.5 M) before analysis, according to Rapaport (27). ATP concentration was determined by luminescence through use of a luciferin-luciferase kit assay (Boehringer Mannheim). Luminescence was counted on a Top-Count luminometer (Packard Instruments). Adenosine concentration was measured by fluorescence coupled to HPLC fractionation. Samples were first treated with chloroacetaldehyde to obtain the N⁶-ethene derivatives (20). HPLC was realized on a C18-E column (125 × 4 mm, Purospher, Merck, Nogent sur Marne, France) with a pump L600 200 A (Merck). Fractionation was followed by using a L-7480 fluorimeter (excitation 240 nm, emission 410 nm, Merck). The mobile phase consisted of a mixture of solution A (12.5 mM Na₂HPO₄, 12.5 mM NaH₂PO₄, pH 6.9) and methanol (B). For the first 5 min, solution A was passed at a flow rate of 0.5 ml/min. A linear gradient was then applied to achieve 80% solution A-20% solution B after 5 min.

Statistical analysis. Results are expressed as means ± SE. Statistical comparisons of hemodynamic, respiratory, and blood gas data were made by ANOVA followed by multiple comparisons of means by using a one-way ANOVA. Post hoc tests, parametric or nonparametric as appropriate, were used to compare mean values. The statistical significance for plasma and erythrocyte ATP and adenosine concentrations was assessed by an unpaired Student's t-test after an ANOVA. Statistical significance was assumed for P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Acute treatment with ATP. In a first series of experiments, we compared the effect of a single oral administration vs. intravenous administration of ATP on cardiovascular parameters. After surgery and determination of the baseline of parameters, two groups of five rabbits were given 3 or 20 mg/kg ATP via a gastric cannula. DAP, HR, central venous pressure, IVBF, lung resistance, and arterial partial pressure of oxygen (Pa_O2) remained constant over a 2-h period of observation after ATP administration, indicating that a single oral dose of ATP had no effect on these physiological parameters (data not shown).

1

1

1

1

View larger version (24K): [in this window] [in a new window]   Fig. 1.   Cardiovascular response to intravenous administration of 1 mg · kg1 · min1 ATP for 2 min to normal rabbits. Each value represents mean of diastolic arterial pressure (DAP, A) and heart rate (HR, B) in 2 individual rabbits. Left ventricular work index (C) and peripheral resistance index (D) were calculated as described in METHODS.

Physiological response to chronic oral administration of ATP. In rabbits chronically treated with ATP, no modification of electrocardiogram morphology or HR was detected compared with controls, irrespective of the dose administered (Table 1). Central venous and arterial pressures were comparable in all groups of rabbits. In contrast, increases of 31 and 50% in the IVBF were observed in ATP-treated rabbits after treatment with 3 and 20 mg · kg¹ · day¹ ATP, respectively (P < 0.001 for 20 mg · kg¹ · day¹, Fig. 2A). A correlated diminution of the vascular resistance of 18 and 22% was observed in ATP-treated rabbits (P < 0.01 for 20 mg · kg¹ · day¹, Fig. 2B). Consequently the LVWI significantly increased by 10% in animals given 20 mg · kg¹ · day¹ ATP (P < 0.05, Table 1). All together, these results demonstrated that chronic oral treatment with ATP had no effect on central blood pressure and cardiac activity but induced peripheral vasodilatation.

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Peripheral venous blood flow (A) and venous resistance (B) in rabbits given 3 mg · kg1 · day1 ATP (n = 4), 20 mg · kg1 · day1 ATP (n = 11), or vehicle (n = 17) for 14 days. Values are means ± SE. ** P < 0.01 and *** P < 0.001 vs. control rabbits.

1

1

1

1

1

1

1

1

1

1

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Respiratory frequency (A), lung resistance (B), and arterial PO2 (PaO2) (C) in rabbits given 3 mg · kg1 · day1 ATP (n = 4) or 20 mg · kg1 · day1 ATP (n = 11) or vehicle (n = 17) for 14 days. Values are means ± SE. * P < 0.05, ** P < 0.01, and *** P < 0.001 vs. control rabbits.

Physiological parameters in animals chronically treated with adenosine. To further explore the mechanisms underlying the response to chronic administration of ATP, we tested whether replacement of ATP by adenosine could lead to similar effects. No modification of HR or electrocardiogram was detected in rabbits treated with 20 mg · kg¹ · day¹ adenosine per os for 14 days (P > 0.05, data not shown). The DAP was not significantly changed by treatment, but an increase in the LVWI was noticed from 2,004.6 ± 49.08 in control rabbits to 2,168.5 ± 80.79 in adenosine-treated animals. A highly significant increase (63%, P < 0.001, Fig. 4A) of the IVBF was also induced by adenosine. This increase, associated with a 32% diminution of the VR (P < 0.05, Fig. 4B), indicated that chronic oral administration of adenosine, like that of ATP, induced peripheral vasodilatation.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Peripheral venous blood flow (A) and venous resistance (B) in rabbits given 20 mg · kg1 · day1 adenosine (Ado; n = 4) or vehicle (n = 17) for 14 days. Values are means ± SE. * P < 0.05 and *** P < 0.001 vs. control rabbits.

1

1

Purine level in plasma and erythrocytes from chronically ATP-treated animals. Adenosine and ATP concentrations were determined in arterial plasma and erythrocytes collected after postsurgical stabilization from control and ATP-treated rabbits (Table 2). In control individuals, adenosine concentration represented 1.0 ± 0.29 µM in plasma and 0.16 ± 0.020 µM in erythrocytes. ATP concentration was 0.2 ± 0.08 µM in plasma and 1.6 ± 0.62 mM in erythrocytes (Table 2). No significant variation of these concentrations was observed in plasma and erythrocytes from ATP-treated counterparts (P > 0.05).

Effect of P1 and P2 antagonists. One possible interpretation of these results is that changing the body pool of purine nucleosides or nucleotides would lead to the activation or desensitization of P2 or P1 receptors. To address this issue, we tested the effect of cibacron blue, a specific P2 antagonist, and theophylline, a P1 antagonist, on cardiovascular and respiratory parameters in normal and ATP-treated rabbits.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Results presented herein indicate that acute oral administration of ATP has no effect on cardiac activity, vascular resistance, or respiratory parameters. This contrasts with the effect of intravenous infusion, which leads to a rapid cardiac response similar to that obtained by administration of adenosine. It reflects the well-documented negative chronotropic, dromotropic, and inotropic cardiac effects mediated by A₁ purine receptors (26) and likely results from the rapid degradation of ATP administered into adenosine by ectonucleotidases (34).

However, repeated oral administration of ATP and adenosine induces, after 14 days, a reduction of vascular resistance with no effect on cardiac activity. The augmentation of the LVWI observed is likely the consequence of the reduced peripheral vascular resistance. In many respects these effects resemble those reported by slow perfusion with free purines (31) and should be mediated by activation of purine receptors. A₂ adenosine receptors have been shown to be present in the aorta of rat (18) and in the vasculature of many other species (19). A₂-receptor activation is also known to induce vasorelaxation (28). It is thus possible that chronic oral treatment with ATP, which is rapidly degraded into adenosine in the gut, induces vasorelaxation and improves cardiac output by stimulating A_2A receptors as observed with selective A_2A-receptor agonists (19). Alternatively, ATP is also known to exert different and even opposite effects on blood vessels by activating different P2-nucleotide receptors (4). ATP induces vasoconstriction through P2X-receptor subtypes present on the surface of vascular smooth muscle cells (21). ATP also mediates vasodilatation involving both endothelium-dependent and endothelium-independent mechanisms (23) by activating different P2Y-purinoceptor subtypes present on endothelial or vascular smooth muscle cells (24). Whether the vascular effects observed in rabbits after oral treatment with ATP are mediated by A₂ or by P2Y receptors could not be ascertained. In an attempt to inhibit the vascular response to ATP supplementation, we tested the effect of theophylline and cibacron blue, which are P1- and P2-receptor antagonists respectively (16). Cardiac modifications induced by these drugs in control animals did not allow conclusions about the implication of purine receptors on the peripheral vasodilatation observed in ATP-treated rabbits. Yet this vascular effect can result from a direct activation of purine receptors on endothelial and smooth muscle cells by local release of purines. It can also reflect the intervention of the central nervous system. Indeed, cardiovascular parameters are regulated by the central and peripheral neurons, which are sensitive to extracellular purines, like those of the nucleus tractus solitarius (22), and use adenosine and ATP as neurotransmitter.

One interesting effect of oral administration of ATP is a reduction of lung resistance and a Pa_O2 increase. Both are compatible with bronchodilatation subsequent to treatment. However, no regulation of bronchial tone by ATP or adenosine has been documented, although purine receptors are present on airway epithelial cells (17). Pa_O2 increase observed in animals treated with ATP deserves further comments. It could result from both bronchodilatation and pulmonary vascular relaxation. However, such vascular relaxation has not been directly evaluated in our experiments but can be anticipated from the peripheral relaxation attested by increased iliac blood flow. Adenosine is also known to increase O₂ supply and angiogenesis via A₂ signaling, and A₁-receptor activation reduces O₂ consumption (26). Hence, increased O₂ supply or reduced O₂ consumption could alternatively account for the Pa_O2 increase observed in ATP-treated rabbits.

The modification of vascular parameters seen in rabbits treated with ATP for 14 days contrasts with the absence of significant change observed after administration of a single oral dose. This suggests an adaptive response that may result from an increase of purine precursors provided by treatment. Although the amount of ATP given to animals is small, compared with the intracellular purine concentration, it might be sufficient to modify their cellular turnover and to stimulate the ATP salvage pathway and ATP exportation toward the blood stream. The absence of variation of ATP and adenosine basal concentrations in the plasma of treated rabbits does not conflict with this conclusion. The average concentration of plasma purines seems rather constant and regulated (32). Moreover, their half-life in extracellular body fluids ranges from 0.1 ms to 1 s (34). Mild and local variations of extracellular purines can induce pharmacological responses but are not detectable at the average plasma level (4). Hence, limited variations may explain the peripheral vascular response and the absence of direct cardiac effect contrasting with that observed after intravenous infusion. A_2A-receptor reserve, for instance, is high, whereas that of A₁, mainly responsible of the cardiac effects, is low (30). A small local delivery of extracellular adenosine could thus stimulate A_2A receptors and promote vascular relaxation without inducing an A₁-mediated cardiac response. This interpretation might also explain why rather similar results were obtained in rabbits treated with 3 or 20 mg · kg¹ · day¹ ATP. The adaptive metabolic response, resulting from ATP administration, would stimulate enzymatic activities involved in nucleoside salvage and ATP synthesis. Once induced, these activities would not depend on substrate availability. Thus repeated oral ATP supplementation would lead to physiological changes that should not be strictly dependent on the dose above a threshold that was not determined in our experiments.

Interestingly, similar vascular effects were achieved with ATP and adenosine oral administration. However, when adenosine was used, no modification of pulmonary resistance and Pa_O2 was observed. One possible interpretation would be that ATP and adenosine, when orally administered, have different pharmacological effects. For example, motor activity of the gut is differently modulated by P1 or P2 receptors (13). However, ATP is rapidly degraded into adenosine in the gut by ectonucleotidases present on enterocyte membranes and then captured by nucleoside transporters (33). It seems thus alternatively possible that differences observed after chronic administration of 20 mg · kg¹ · day¹ ATP and adenosine are dependent on quantitative aspects even if twice as many micromoles of adenosine than of ATP were provided by this dose. Adenosine is directly transported by nucleoside transporters present on the brush border and on the membrane of commensal intestinal microorganisms. However, these bacteria do not possess ectonucleotidases (3). As proposed regarding other tissues, a coupling or a proximity between ectonucleotidases and nucleoside transporters might exist on enterocytes (8). Consequently, ATP administration would privilege the absorption of newly generated adenosine by enterocytes, whereas in adenosine-treated animals a competition between enterocytes and commensal microorganisms would limit the availability of adenosine for intestinal cells.

In summary, our results demonstrate that chronic oral administration of ATP leads to pharmacological effects that are different from those achieved after intravenous administration. These effects might be the consequence of a profound modification of the metabolic nucleoside salvage pathway and of subsequent ATP exportation from various tissues. Experiments in progress in different mammalian species indicate that this adaptive process is associated with a modulation of nucleoside transport.