Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Diadenosine polyphosphates such as AP₄A and AP₅A have received considerable attention because of their multiple biological activities. These compounds are present in various types of cells, including platelets (Flodgaard and Klenow, 1982), chromaffin cells (Rodriguez del Castillo et al., 1988) and neural tissue (Pintor et al., 1992), and are implicated primarily in platelet functions (Flodgaard and Klenow, 1982; Luthje and Ogilvie, 1983; Zamecnik et al., 1992) and neurotransmission (Baxi and Vishwanatha, 1995).

In recent years, interest in AP_nA in vasomotor activity has been renewed. Busse et al. (1988) first described AP₄A induced endothelium-independent vasoconstriction in rabbit mesenteric arteries. Subsequent studies also indicated that AP₄A elicits increases in perfusion pressure of rat portal vein (Busshardt et al., 1989) and that AP₄A also induces the vasoconstrictor responses mediated by P_2X-purinergic receptors (Ralevic et al., 1995). AP₅A was recently reported to be a novel vasoconstrictor agent isolated from human platelets (Schluter et al., 1994). These authors reported that AP₅A induced increases in perfusion pressure of the vasculture in isolated perfused rat kidney and aorta. AP₅A-induced vasoconstriction has also been documented in both the rat mesenteric artery (Ralevic et al., 1995) and human umbilical artery (Davies et al., 1995) .

AP₄A also induces endothelium-dependent vasodilation in rabbit mesenteric arteries (Busse et al., 1988), and endothelium-dependent and -independent decreases in rabbit coronary perfusion pressure (Pohl et al., 1991). Intravenous administration of AP₄A to a dog produced a dose-dependent decrease in mean arterial pressure (Kikuta et al., 1994), which thus indicates that AP₄A may induce, either directly or indirectly, relaxation rather than constriction of vascular smooth muscle in vivo. However, the cellular mechanism for AP_nA-induced vasorelaxation has yet to be investigated extensively. In the present study, we determined the mechanism underlying the vasorelaxing effects of AP₄A and AP₅A on porcine coronary smooth muscle cells by use of the simultaneous measurements of [Ca⁺⁺]_i and force. We obtained evidence that AP₄A and AP₅A reduce [Ca⁺⁺]_i by modulating the function of K⁺ channels and decreasing the Ca⁺⁺ sensitivity of the contractile machinery and, as a result, induce vasorelaxation. These effects were thought to be mediated by the increase in the cellular cAMP content.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Tissue preparation. Left circumflex coronary arteries (2-3 cm from the origin) were isolated from fresh porcine hearts at a local slaughterhouse immediately after the animals had been sacrificed. The tissue specimens were placed in ice-cold normal PSS and brought to the laboratory. After the segments were cut open longitudinally, the adventitia was trimmed away. The inner surfaces of the arteries were rubbed with cotton swabs to remove the endothelium. Each specimen was cut into similar sized strips (1 × 5 × 0.1 mm). Complete removal of the endothelium was confirmed by the lack of any relaxing response of the strips to 1 µM bradykinin.

Fura-2 loading. Coronary arterial media strips were loaded with Ca⁺⁺ indicator dye, fura-2, by incubating them in oxygenated (a mixture of 95% O₂ and 5% CO₂) Dulbecco's modified Eagle's medium containing 25 µM fura-2/AM (an acetoxymethyl ester form of fura-2) and 5% fetal bovine serum for 4 hr at 37°C. After loading with fura-2, medial strips were incubated in normal PSS for at least 1 hr at 37°C before starting the measurement, to remove the dye in the extracellular space and for purposes of equilibration. Loading the medial strips with fura-2 per se did not affect the contractility, as described previously (Abe et al., 1990; Hirano et al., 1990).

Front-surface fluorometry. Changes in the fluorescence intensity of the fura-2-Ca⁺⁺ complex were monitored simultaneously with the force development, by a front-surface fluorometer specifically designed for fura-2 fluorometry (CAM-OF3, Japan Spectroscopic Co., Tokyo, Japan), as described previously (Ushio-Fukai et al., 1993). The ratio of the fluorescence (500 nm) intensities at alternating 340-nm and 380-nm excitation wavelengths was monitored. The fluorescence ratio was expressed as a percentage, by assigning the values at rest in normal (5.9 mM K⁺) and 118 mM K⁺ PSS to be 0% and 100%, respectively. The absolute values of [Ca⁺⁺]_i of vascular strips were calculated according to the method of Grynkiewicz et al. (1985) with the K_d (apparent dissociation constant) of the fura-2-Ca⁺⁺ complex of 225 nM (at 37°C). The absolute values of [Ca⁺⁺]_i for 0% and 100% levels were determined in separate measurements and were 108 ± 27 nM (n = 10) and 715 ± 103 nM (n = 10), respectively. The obtained [Ca⁺⁺]_i values are considered to be an approximation to the true [Ca⁺⁺]_i value, and the calibration of the absolute levels of [Ca⁺⁺]_i at the end of experiments is likely to be uncertain (Miyagi et al., 1995). Therefore, a statistical analysis of the [Ca⁺⁺]_i signal was performed with use of the percent fluorescence ratio.

Measurement of force development. Coronary arterial media strips were mounted vertically in a quartz organ bath, and the isometric tension was measured, as described previously (Ushio-Fukai et al., 1993). The strips were stimulated with 118 mM K⁺ PSS every 15 min during the fura-2 equilibration period (1 hr), and then the resting tension was increased in a stepwise manner to obtain the maximal tension development. The appropriate resting tension level obtained by this procedure was about 300 mg. At the beginning of each protocol, the responsiveness of each strip to 118 mM K⁺ PSS was recorded. The developed force was expressed in a percentage, by assigning the values at rest in normal PSS (5.9 mM K⁺) to be 0%, and those at steady state of contraction in 118 mM K⁺ PSS to be 100%.

Assay of cAMP and cGMP. The content of cAMP and cGMP in porcine coronary artery were assayed as described previously (Abe et al., 1994). After incubation of vascular strips with oxygenated (95% O₂ and 5% CO₂) normal PSS containing 10 µM AP₄A or AP₅A for 15 min at 37°C, the reaction was stopped by replacing the solution with ice-cold perchloric acid (6%). The strips were then homogenized in perchloric acid. The homogenate was centrifuged at 1500 × g for 15 min. The supernatant was used to measure the cAMP and cGMP content by using radioimmunoassay kits (Yamasa, Tokyo, Japan). cAMP and cGMP levels were expressed as nanomoles or picomoles per milligram of wet weight of tissue (nmol or pmol/mg tissue).

Drugs and solutions. The composition of normal PSS for fura-2 studies was as follows (mM): NaCl, 123; KCl, 4.7; NaHCO₃, 15.5; KH₂PO₄, 1.2; MgCl₂, 1.2; CaCl₂, 1.25; D-glucose, 11.5. The Ca⁺⁺-free solution (Ca⁺⁺-free PSS) contained 2 mM EGTA instead of 1.25 mM CaCl₂. High K⁺ PSS was prepared by replacing NaCl with equimolar KCl. PSS was bubbled with a mixture of 95% O₂ and 5% CO₂, and the resulting pH was 7.4. Fura-2/AM was purchased from Dojindo Laboratories (Kumamoto, Japan). AP₄A, AP₅A and bovine serum albumin were purchased from Sigma Chemical Co. (St. Louis, MO). Bradykinin, ChTX and apamin were purchased from the Peptide Institute, Inc. (Osaka, Japan). U46619 was purchased from Funakoshi (Tokyo, Japan). Caffeine was obtained from Katayama Chemical (Osaka, Japan). Kits for radioimmunoassay of cAMP and cGMP were purchased from Yamasa (Tokyo, Japan). All other chemicals were obtained from Wako (Osaka, Japan).

Data analysis. The values are expressed as the mean ± S.E.M. Student's t-test was used to determine statistical significance between two groups, and analysis of variance was used to determine the dose-dependent effect of AP_nA on [Ca⁺⁺]_i and force. The number of experiments corresponds to the number of the animals used. The IC₅₀ values, the concentrations that decreased the fluorescence ratio and force to 50% of the maximum response, were determined based on the concentration-response curve fitted according to a four-parameter logistic model (DeLean et al., 1978). The significance of the shift of the [Ca⁺⁺]_i-force relation curve was determined by an analysis of covariance. A value of P < .05 was considered to have statistical significance. All data were collected by a computerized data acquisition system (MacLab: Analog Digital Instruments, Castle Hill, Australia; and Macintosh: Apple Computer, Cupertino, CA).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Effects of AP₄A and AP₅A on the [Ca⁺⁺]_i and force induced by U46619. Figure 1, a and b, shows representative recordings of the effects of cumulative application of AP₄A (0.01-10 µM) and AP₅A (0.01-10 µM), respectively, on the [Ca⁺⁺]_i and force induced by 60 nM U46619, a thromboxane A₂ analog. When the bathing medium was changed from normal PSS (5.9 mM K⁺) to 118 mM K⁺ PSS, both [Ca⁺⁺]_i and force rapidly increased and reached the plateau phases within 10 to 15 min. The values at the resting and plateau phases were designated to be 0% and 100% for both [Ca⁺⁺]_i and force. The application of 60 nM U46619 induced a rapid increase in [Ca⁺⁺]_i, which reached a peak level in 5 min and thereafter decreased to a plateau level within 10 min (60.52 ± 0.97%, n = 20). The force also rapidly increased and reached a plateau level within 10 min (95.27 ± 1.01%, n = 20). The applications of AP₄A and AP₅A during the U46619-induced sustained contraction caused concentration-dependent decreases in [Ca⁺⁺]_i and force of the coronary strips. Figure 1, c and d, shows a summary of the results obtained from seven and nine independent experiments performed in a manner similar to that shown in fig. 1, a and b. The cumulative application of AP₄A and AP₅A caused decreases in [Ca⁺⁺]_i and force (P < .05 by an analysis of variance) in a concentration-dependent manner. The IC₅₀ values for [Ca⁺⁺]_i and force were 1.14 ± 0.29 µM and 1.03 ± 0.16 µM for AP₄A (n = 7) and 0.82 ± 0.09 µM and 1.74 ± 0.34 µM for AP₅A (n = 9), respectively. There was no significant difference in these IC₅₀ values between AP₄A and AP₅A.

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Effects of AP4A and AP5A on the changes in [Ca++]i and force induced by U46619 in the porcine coronary artery without endothelium. Panels a and b, representative recordings showing the effects of AP4A (a) and AP5A (b) on the coronary artery precontracted by U46619. Various concentrations of AP4A or AP5A (0.01-10 µM) were cumulatively applied during the steady state contraction induced by 60 nM U46619. Panels c and d, concentration-response curves for changes in [Ca++]i (c) and force (d) obtained by the cumulative addition of AP4A (, n = 7) or AP5A (: n = 9) during 60 nM U46619-induced contraction. Vertical bars represent S.E.M.

Effects of AP₄A and AP₅A on the [Ca⁺⁺]_i and force and the [Ca⁺⁺]_i-force relation (the sensitivity of the contractile machinery) induced by high K⁺ depolarization. Figure 2 shows the representative recordings of the effects of 10 µM AP₄A (fig. 2, a-e) and AP₅A (fig. 2, f-j), respectively, on the [Ca⁺⁺]_i and force induced by high K⁺ depolarization (20, 30, 40, 60 and 118 mM). The extent of the AP₄A- and AP₅A-induced decrease in force during the high K⁺ depolarization was much greater than that expected from the extent of reduction of [Ca⁺⁺]_i. This effect can be typically seen in figure 2, c, d, e, h, i and j, where AP₄A (panels c, d and e) or AP₅A (panels h, i and j) induced relaxation with little or no reduction of [Ca⁺⁺]_i. These observations indicated that AP₄A and AP₅A might decrease the Ca⁺⁺ sensitivity of the contractile machinery. To further analyze the AP₄A- and AP₅A-induced change in the Ca⁺⁺ sensitivity, the [Ca⁺⁺]_i (abscissa)-force (ordinate) relation curves were constructed with the data points obtained from several experiments done in a manner similar to shown in figure 2. As shown in figure 3, AP₄A (panel a) and AP₅A (panel b) shifted the [Ca⁺⁺]_i-force relation curves to the right (P < .05 by an analysis of covariance), which indicated that the reduction of force was much greater than that expected from the given reduction in [Ca⁺⁺]_i levels (the decrease in Ca⁺⁺ sensitivity). AP₄A and AP₅A increased the ratio (%) values at which force was reduced by 50%, from 62.4 ± 6.2% to 76.3 ± 1.7% (AP₄A); from 59.9 ± 8.7% to 77.1 ± 1.2% (AP₅A).

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Representative recordings showing the effects of 10 µM AP4A or AP5A on the elevations of [Ca++]i and force induced by various concentrations of extracellular K+. AP4A (a-e) or AP5A (f-j) was applied 15 min after the application of various concentration of K+ PSS as shown in each panel. Response of ratio and force to 118 mM K+ was recorded before each experiment, as a control, and was designated to be 100% (indicated by C).

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Effects of AP4A or AP5A on the [Ca++]i-force relationship of the steady state contractions induced by various concentrations of extracellular K+. Each point was obtained by the same protocol as that shown in figure 2. The [Ca++]i (abscissa)-force (ordinate) relationships of high K+-induced contractions (5.9, 20, 30, 40, 60 and 118 mM, as indicated by the number shown in the figure) were obtained in the absence (, n = 4-8 for AP4A and n = 4-6 for AP5A) or presence (, n = 4-8 for AP4A and n = 4-6 for AP5A) of 10 µM AP4A (a) or 10 µM AP5A (b). The arrow tails and heads indicate the levels just before and 15 min after the application of AP4A or AP5A, respectively. The data are expressed as the means ± S.E.M.

Differential effects of AP₄A and AP₅A on the contractions induced by high K⁺ depolarization to those by U46619. As shown in figure 1, 10 µM AP₄A or AP₅A almost completely inhibited the increases in [Ca⁺⁺]_i and force induced by U46619, whereas 10 µM AP₄A or AP₅A had little effect on the increases in [Ca⁺⁺]_i and force induced by high K⁺ depolarization (fig. 2, a-j) These observations indicated that the increase in [Ca⁺⁺]_i induced by high K⁺ depolarization is more resistant to inhibition by AP₄A or AP₅A. To further characterize this, we made a new figure (fig. 4) using parts of the data points shown in figure 1, c and d, and figure 3, a and b. At comparable [Ca⁺⁺]_i (30 mM K⁺ depolarization vs. 60 nM U46619) and force (60 mM K⁺ depolarization vs. 60 nM U46619) levels, both AP₄A and AP₅A induced smaller decreases in the [Ca⁺⁺]_i and force in the contractions induced by K⁺ depolarization than those induced by U46619 (fig. 4). These results suggest that the AP₄A- and AP₅A-induced decrease in [Ca⁺⁺]_i is caused, at least in part, by membrane hyperpolarization.

View larger version (29K): [in this window] [in a new window]   Fig. 4.   Comparison of AP4A- or AP5A-induced reductions of [Ca++]i and force during the activation by U46619 to those during high K+ depolarization. A comparison of the [Ca++]i level (a) was performed at the contractions between 60 nM U46619 and 30 mM K+ depolarization, whereas a comparison of the force level (b) was performed at the contractions between 60 nM U46619 and 60 mM K+-depolarization. This is because the levels of increases in [Ca++]i and force induced by 60 nM U-46619 were comparable to the increase in [Ca++]i level induced by 30 mM K+-depolarization and to the increase in force level induced by 60 mM K+-depolarization, respectively. The corresponding data points were taken from fig. 1, c and d, and figure 3. Each column represents the levels of [Ca++]i or force just before () and after the application of either AP4A () or AP5A (). The data are expressed as the means ± S.E.M. ***; P < .001; n.s., difference not significant.

Effects of K⁺ channel blockers on AP₄A- and AP₅A-induced decreases in the [Ca⁺⁺]_i and force. To determine whether or not K⁺ channels are involved in the relaxation induced by AP_nA, we examined the effect of various K⁺ channel blockers on AP₄A- and AP₅A-induced decreases in [Ca⁺⁺]_i and force. The following K⁺ channel blockers were examined: 1 mM TBA (nonspecific K⁺ channel blocker), 100 nM ChTX (large conductance Ca⁺⁺-activated K⁺ channel blocker), 30 µM 4-AP (voltage-dependent K⁺ channel blocker), 3 µM glibenclamide (ATP-sensitive K⁺ channel blocker) and 1 µM apamin (small conductance Ca⁺⁺-activated K⁺ channel blocker). These K⁺ channel blockers were applied 10 min before and during the application of 60 nM U46619. We chose the highest concentrations of K⁺ channel blockers that do not directly affect U46619-induced increases in [Ca⁺⁺]_i and force (Kawasaki et al., 1997). As shown in figure 5, a and b, TBA partially inhibited the AP₄A-induced decrease in [Ca⁺⁺]_i (P < .05 by an analysis of covariance) and TBA, ChTX and 4-AP partially inhibited the AP₄A-induced decrease in force (P < .05 by an analysis of covariance). As shown in figure 5, c and d, TBA, ChTX and 4-AP partially inhibited the AP₅A-induced decrease in [Ca⁺⁺]_i and force (P < .05 by an analysis of covariance). Glibenclamide and apamin had no significant blocking effect on the decreases in [Ca⁺⁺]_i and force induced by AP₄A or AP₅A (data not shown).

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Effects of various types of K+ channel blockers on the decreases in [Ca++]i and force induced by various concentrations of AP4A or AP5A. The concentration-response curves for AP4A (a and b) or AP5A (c and d) were constructed in the same manner as in figure 1, c and d, in the absence (; n = 7 for AP4A, n = 9 for AP5A) or presence of 1 mM TBA (; n = 4), 100 nM ChTX (; n = 7) or 30 µM 4-AP, (; n = 7 for AP4A, n = 5 for AP5A). Vertical bars represent S.E.M.

Effects of AP₄A and AP₅A on the elevation of [Ca⁺⁺]_i and force induced by histamine and caffeine in the absence of extracellular Ca⁺⁺. Figure 6a shows representative time courses of the changes in [Ca⁺⁺]_i and force induced by histamine in Ca⁺⁺-free PSS containing 2 mM EGTA. When vascular strips were exposed to Ca⁺⁺-free PSS, [Ca⁺⁺]_i gradually declined to reach a steady state (21.60 ± 2.18%, n = 8), whereas the force remained unchanged. The application of 10 µM histamine after a 10-min incubation in Ca⁺⁺-free PSS caused transient elevations of [Ca⁺⁺]_i (22.51 ± 2.76%, n = 8) and force (48.04 ± 1.75%, n = 8) with peaks at 15 s and 30 s, respectively. As shown in figure 6, b and c, pretreatment with AP₄A or AP₅A for 10 min did not affect the transient elevations of [Ca⁺⁺]_i induced by histamine. However, AP₄A and AP₅A did inhibit the transient elevations of force induced by histamine. Figure 6d summarizes the results obtained from eight independent experiments.

View larger version (37K): [in this window] [in a new window]   Fig. 6.   Effects of AP4A or AP5A on elevations of [Ca++]i and force induced by histamine or caffeine in Ca++-free PSS containing 2 mM EGTA. Panels a, b and c: The representative recordings of changes in [Ca++]i and force induced by 10 µM histamine in Ca++-free PSS, in the absence (a) or presence of 10 µM AP4A (b) or 10 µM AP5A (c). Panel d is a summary of the results obtained from experiments performed in the same manner as in panels a, b and c. The [Ca++]i and force development induced by the application of 10 µM histamine or 20 mM caffeine in the absence (; n = 8 for histamine and n = 5 for caffeine) or presence of AP4A (; n = 8 for histamine and n = 5 for caffeine) or AP5A (; n = 8 for histamine and n = 5 for caffeine). The bottom and top of each column indicate the [Ca++]i and force just before and at the peak levels obtained by the application of histamine or caffeine, respectively. The data are expressed as the means ± S.E.M. shown as vertical bars. **; P < .01; ***; P < .001; n.s., difference not significant.

Effects of AP₄A and AP₅A on cellular cAMP and cGMP contents. Figure 7 shows effects of AP₄A and AP₅A on the levels of cellular cAMP and cGMP contents of the porcine coronary artery in normal PSS. The cellular cAMP and cGMP contents were 0.21 ± 0.03 nmol/mg tissue (n = 14) and 2.04 ± 0.51 pmol/mg tissue (n = 14) at resting level (5.9 mM K⁺ PSS), respectively. When the vascular strips were exposed to 10 µM AP₄A and AP₅A for 15 min in normal PSS, the intracellular cAMP levels significantly increased to 0.31 ± 0.03 nmol/mg tissue (n = 14) and 0.32 ± 0.04 nmol/mg tissue (n = 14; vs. control, P < .05 for both, by student's t test), respectively (fig. 7a). Neither AP₄A nor AP₅A induced any change in the intracellular cGMP levels (1.99 ± 0.34 and 2.38 ± 0.44 pmol/mg tissue) (fig. 7b).

View larger version (24K): [in this window] [in a new window]   Fig. 7.   Effects of AP4A or AP5A on the tissue contents of cAMP and cGMP in the porcine coronary arterial smooth muscle. The contents of cAMP (a) and cGMP (b) were determined by radioimmunoassay. Each column represents the levels of cAMP or cGMP in the absence (; n = 14 for both) or presence of 10 µM AP4A (; n = 14 for both) or 10 µM AP5A (; n = 14 for both). The data are expressed as the means ± S.E.M. shown as vertical bars. *; P < .05, n.s., difference not significant.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The present study investigated the mechanisms underlying the AP₄A- and AP₅A-induced direct relaxation of the porcine coronary artery without endothelium. The results obtained indicate that AP₄A and AP₅A induce relaxation of the porcine coronary artery by decreasing [Ca⁺⁺]_i (fig. 1), which may be partially caused by the opening of K⁺ channels (fig. 5), and also by decreasing the Ca⁺⁺ sensitivity of the contractile machinery (fig. 3). These effects were thought to be mediated by cAMP, because AP₄A and AP₅A increased the cellular cAMP content (fig. 7). In addition, AP₄A and AP₅A had no effect on the intracellular Ca⁺⁺ release mechanism (fig. 6).

The relaxing effect of AP₄A on the coronary artery has already been described previously. Pohl et al. (1991) first reported that AP₄A induces endothelium-dependent and -independent decreases in rabbit coronary perfusion pressure. Nakae et al. (1996) also reported AP₄A-induced coronary vasodilation in the porcine. However, the AP₅A-induced vasorelaxation apparently has not been described previously. The results obtained in the present study indicate that AP₄A- and AP₅A-induced relaxation of the coronary artery is accompanied by a reduction of [Ca⁺⁺]_i. This mechanism could be clearly seen when precontraction was induced by U46619 (fig. 1). However, when precontraction was induced by 60 mM or 118 mM K⁺ depolarization, the reduction of [Ca⁺⁺]_i was less obvious, although the reduction of force could still be detected (fig. 2). This result indicates that reduction of [Ca⁺⁺]_i is not the sole mechanism for the AP₄A- or AP₅A-induced coronary vasorelaxation.

The mechanism other than the reduction of [Ca⁺⁺]_i for the AP₄A- or AP₅A-induced coronary vasorelaxation was explored in the experiments shown in figures 2 and 3. As shown in figure 2, c, d, e, h, i and j, where AP₄A and AP₅A induced relaxation with little or no reduction of [Ca⁺⁺]_i. AP₄A and AP₅A shifted the [Ca⁺⁺]_i-force curve to the right during stimulation with high-K⁺ depolarization (fig. 3), which indicates that AP₄A and AP₅A decrease the Ca⁺⁺ sensitivity of the contractile machinery. This mechanism also explains the observation that AP₄A and AP₅A inhibited the development of force without affecting the transient increase in [Ca⁺⁺]_i induced by histamine in Ca⁺⁺-free medium (fig. 6). However, this is not case in the contractions induced by 20 mM caffeine in Ca⁺⁺-free medium (fig. 6d). Neither AP₄A nor AP₅A were able to inhibit the caffeine-induced transient force development. In Ca⁺⁺-free medium without AP_nA, 20 mM caffeine induced a much smaller force than 10 µM histamine, although the former induced a greater [Ca⁺⁺]_i elevation than the latter (fig. 6). We reported previously that caffeine increases cAMP, which markedly decreases Ca⁺⁺ sensitivity of the contractile machinery (Watanabe et al., 1992). The inability for AP₄A and AP₅A to decrease the caffeine-induced contraction could be explained by the saturation of the effect of cAMP on the reduction of the Ca⁺⁺ sensitivity, because the relaxing effect of AP₄A and AP₅A was also mediated by cAMP (fig. 7). We did not use permeabilized preparations, which could directly demonstrate the decrease in the Ca⁺⁺ sensitivity, because the solution used for permeabilized cell preparations contains millimolar concentrations of ATP. These conditions may desensitize the purinergic and possibly adenosine receptors, that are proposed to be involved in AP₄A- and AP₅A-induced vasorelaxation (discussed later).

It is well known that contraction of vascular smooth muscle is regulated by agonist-mediated modulation of the [Ca⁺⁺]_i-force relation (= the Ca⁺⁺ sensitivity of the contractile machinery) as well as the changes in [Ca⁺⁺]_i (Somlyo and Somlyo, 1994). As to the candidate mediator for the reduction of Ca⁺⁺ sensitivity, cAMP and cGMP have been shown to decrease the Ca⁺⁺ sensitivity of the smooth muscle contractile machinery (Nishimura and van Breemen, 1989). We have also shown in a previous study that nitroglycerine (Abe et al., 1990) and isoprenaline (Ushio-Fukai et al., 1993) shift the [Ca⁺⁺]_i-force relation to the right in the porcine coronary artery. It thus seems likely that the decreases in Ca⁺⁺ sensitivity of the contractile machinery induced by AP₄A or AP₅A may be mediated by cAMP or cGMP. In the present study we determined the second messenger for AP₄A- or AP₅A-induced vasorelaxation. As shown in figure 7, both AP₄A and AP₅A increased the cellular contents of cAMP. These results also support the hypothesis that AP₄A and AP₅A decrease the Ca⁺⁺ sensitivity of the contractile machinery.

The inhibitory effects of AP₄A or AP₅A on [Ca⁺⁺]_i and force were attenuated when the strips were precontracted by high K⁺ depolarization, as compared with those during U46619 stimulation (fig. 4). This observation evoked speculation that the reduction of [Ca⁺⁺]_i might involve membrane hyperpolarization mediated by the opening of K⁺ channels, because this mechanism is eliminated during high K⁺ depolarization. Recent evidence has suggested that vascular tone and membrane potential are regulated by several types of K⁺ channels, including Ca⁺⁺-activated K⁺ channels, voltage-dependent K⁺ channels and ATP-sensitive K⁺ channels (Nelson and Quayle, 1995). Therefore, we investigated the effects of selective blockers for these K⁺ channels on the AP₄A- and AP₅A-induced decrease in [Ca⁺⁺]_i and force during the stimulation with U46619. Comparative studies with these selective K⁺ channel blockers, as shown in figure 5, suggest that the activation of large conductance Ca⁺⁺-activated K⁺ channels and voltage-dependent K⁺ channels may contribute to the decrease in [Ca⁺⁺]_i and force. It has been hypothesized that cAMP may decrease [Ca⁺⁺]_i by a hyperpolarization (Somlyo et al., 1970) probably by stimulation of Ca⁺⁺-activated K⁺ channel openings (Sadoshima et al., 1988). Thus, the present results that AP₄A and AP₅A significantly increased the intracellular cAMP levels (fig. 7) support the idea that AP₄A and AP₅A may induce membrane hyperpolarization mediated by the opening of K⁺ channels. Our results do not support the notion that ATP-sensitive K⁺ channels and small conductance Ca⁺⁺-activated K⁺ channels play a major role in the AP₄A- and AP₅A-induced vasorelaxation and decrease in [Ca⁺⁺]_i.

Another potential mechanism for the reduction of [Ca⁺⁺]_i, namely the inhibition of Ca⁺⁺ release from intracellular stores, was examined in the experiments shown in figure 6. AP₄A and AP₅A did not inhibit the histamine- and caffeine-induced release of intracellular Ca⁺⁺ in the absence of extracellular Ca⁺⁺. These observations suggested that AP₄A and AP₅A do not affect intracellular Ca⁺⁺ release through inhibition of a receptor-coupled signal transduction pathway or through a direct effect on intracellular storage sites.

The receptors responsible for the AP₄A- and AP₅A-induced coronary vasorelaxation were not explored in the present study. It has been reported that mammalian cells typically contain a specific AP₄A hydrolase that hydrolyzes AP₄A in an asymmetric fashion to yield AMP and ATP, which is converted to AMP and inorganic pyrophosphate; and dephosphorylation of AMP yields adenosine (Hankin et al., 1995; Luthje and Ogilvie, 1985, 1988; Ogilvie et al., 1989; Thorne et al., 1995). This enzyme also hydrolyzes higher homologs (e.g., AP₅A in an asymmetric fashion to yield ADP and ATP) and is presumed to be involved in the regulation of the intracellular level of such nucleotides (Guranowski and Sillero, 1992). Thus, it is possible that the application of AP₄A and AP₅A may produce adenosine, AMP, ADP and ATP, and stimulate such multiple receptors as nucleotide receptors and adenosine receptors. This may also partly explain why AP₄A and AP₅A have a different functional effect (constriction in some smooth muscle and relaxation in the other) at different potencies. For example, AP₅A was approximately 10 to 1000 times more potent than AP₄A for the contraction of guinea pig urinary bladder and vas deferens (Bo et al., 1994; Hoyle et al., 1995; Ralevic et al., 1995; Stone and Paton, 1989). For the vasoconstriction of rat and rabbit mesenteric vessels, AP₅A was more potent than AP₄A (Busse et al., 1988; Ralevic et al. 1995). On the other hand, there was no significant difference in the potency of AP₄A and AP₅A to induce relaxation of the guinea pig left atrium (Hoyle et al., 1996). We also observed almost equal potency for AP₄A- and AP₅A-induced coronary vasorelaxation. However, it is still possible that AP₄A and AP₅A might stimulate receptors as their own form without degradation because it has been reported that AP_nA can be considered as the long-lived substances (Busse et al., 1988; Hoyle et al., 1996; Pohl et al., 1991). In agreement with this speculation, the specific and saturable membrane receptors for AP₄A have been reported to be present in brain, cardiac, liver, kidney, spleen and adipose tissue (Hilderman et al., 1991; Walker and Hilderman, 1993).

In summary, we obtained evidence that AP₄A and AP₅A induce relaxation of the porcine coronary artery by decreasing [Ca⁺⁺]_i, which may be partially caused by the opening of K⁺ channels, and decreasing the Ca⁺⁺ sensitivity of the contractile machinery. These effects were thought to be mediated by cAMP. These substances may be naturally occurring coronary dilators.