The Role of Adenosine and ATP-sensitive Potassium Channels in the Protection Afforded by Ischemic Preconditioning Against the Post-ischemic Endothelial Dysfunction in Guinea-pig Hearts

J Mol Cell Cardiol 30, 1735–1747 (1998) Article No. mc980736

The Role of Adenosine and ATP-sensitive Potassium Channels in the Protection Afforded by Ischemic Preconditioning Against the Post-ischemic Endothelial Dysfunction in Guinea-pig Hearts Michał Maω czewski and Andrzej Bereω sewicz Department of Clinical Physiology, Medical Centre of Postgraduate Education, Warsaw, Poland (Received 25 November 1997, accepted in revised form 13 May 1998) M. Mω   A. Bω . The Role of Adenosine and ATP-sensitive Potassium Channels in the Protection Afforded by Ischemic Preconditioning Against the Post-ischemic Endothelial Dysfunction in Guinea-pig Hearts. Journal of Molecular and Cellular Cardiology (1998) 30, 1735–1747. The role of adenosine and ATP-sensitive potassium channels (KATP) in the mechanism of ischemic preconditioning (IPC)-induced protection against the post-ischemic endothelial dysfunction was studied. Langendorff-perfused guinea-pig hearts were subjected either to 40 min of global ischemia and 40 min reperfusion or were preconditioned prior to the ischemia/reperfusion with three cycles of either 5 min ischemia/5 min reperfusion (IPC) or 5 min infusion/5 min wash-out of adenosine, adenosine A1 receptor agonist, N6-cyclohexyladenosine (CHA) or KATP opener, pinacidil. The magnitude of coronary flow reduction caused by NO-synthase inhibitor, Nx-nitro--arginine methyl ester (-NAME), served as an index of a basal endothelium-dependent vasodilator tone. Coronary overflows produced by a bolus of acetylcholine (ACh) and sodium nitroprusside (SNP) were used as measures of agonist-induced endothelium-dependent and endothelium-independent vascular function, respectively. The coronary flow, LVDP, ACh response and -NAME response were reduced by 8, 32, 41 and 54%, respectively, while SNP response was not changed in the hearts subjected to ischemia/reperfusion. ACh response was fully restored, -NAME response was partially restored, and SNP response was not affected in the hearts subjected to IPC. The post-ischemic recoveries of coronary flow and LVDP were not improved by IPC. The protective effect of IPC on the ACh response was mimicked by adenosine, CHA, and pinacidil. The protective effect of IPC, CHA and pinacidil was abolished by KATP antagonist, glibenclamide. The IPC protection was affected neither by a non-specific adenosine antagonist, 8-p-sulfophenyltheophylline, nor by a specific adenosine A1 receptor antagonist, 8-cyclopentyl-1,3-dipropylxanthine (DPCPX). Our data indicate that: (1) IPC affords endothelial protection in the mechanism that involves activation of KATP, but not adenosine A1 receptors; (2) exogenous adenosine and A1 receptor agonist afford the protection, which might be of a potential clinical significance; (3) the endothelial dysfunction is not involved in the mechanism of myocardial stunning in guinea-pig hearts.  1998 Academic Press K W: Guinea-pig heart; Ischemia/reperfusion injury; Endothelial dysfunction; Ischemic preconditioning; Adenosine receptors; ATP-sensitive potassium channels.

Introduction Cardiac ischemia/reperfusion causes damage not only to myocytes but also to vascular endothelial

cells. Indeed, in various experimental models, ischemia/reperfusion has been shown to impair endothelium-dependent, but not endothelium independdent, coronary vasodilation, indicating selective

Please address all correspondence to: Dr Andrzej Bereω sewicz, Department of Clinical Physiology, Medical Centre of Postgraduate Education, Marymoncka 99, 01-813 Warszawa, Poland.

0022–2828/98/091735+13 $30.00/0

 1998 Academic Press

1736

M. Maω czewski and A. Bereω sewicz

endothelial dysfunction (VanBenthuysen et al., 1987; Mehta et al., 1989a; Tsao et al., 1990; Defily and Chilian, 1993; Richard et al., 1994). Adverse consequences of endothelial dysfunction may include reduction in coronary reserve and increased accumulation of polymorphonuclear leukocytes within the reperfused myocardium (Sobey and Woodman, 1993; Lefer and Lefer, 1996). Ischemic preconditioning (IPC) has long been reported to protect the heart from ischemia/reperfusion injury, and to reduce the incidence of reperfusion arrhythmias (Parratt, 1994; Cohen and Downey, 1996). Several reports suggest that the IPC protection extends also to the coronary endothelium (Defily and Chilian, 1993; Richard et al., 1994; Kaeffer et al., 1996; Kolocassides et al., 1996). A large body of evidence suggests the role of adenosine A1 receptors, and/or ATP-sensitive potassium channel (KATP) in the mechanism of IPC. Thus, adenosine receptors antagonists and KATP blockers have been shown to block, while adenosine receptors agonists and KATP openers to mimic IPC protection in cardiomyocytes (Gross, 1995; Cohen and Downey, 1996; Grover, 1997). The role of these mechanisms in the IPC-induced protection of the endothelium has been studied only scarcely (Bouchard and Lamontagne, 1996; Giannella et al., 1997). Therefore, the aim of this study was to assess whether: (1) ischemia/reperfusion produces endothelial dysfunction in isolated guinea-pig heart and IPC prevents this dysfunction; (2) the protective effect of IPC was due to the stimulation of adenosine receptors and/or KATP and (3) adenosine mediates its protective effect via KATP activation, as it has been shown to be the case in a number of other models (Auchampach and Gross, 1993; Toombs et al., 1993).

Materials and Methods Isolated heart preparation The investigation conforms to the Guide for the care and use of laboratory animals (US National Institutes of Health publication No 85-23, revised 1985). Guinea-pigs (300–380 g) of either sex were injected with 500 units of heparin sulphate, i.p., 20 min before being killed by a blow on the head. Hearts were removed and placed in an ice-cold perfusion solution. The aorta was cannulated and the hearts were perfused aerobically in the Langendorff mode, at perfusion pressure of 70 mmHg, with

prefiltered (5.0 l Millipore filter) perfusion fluid containing, in mmol/l: 118 NaCl; 23.8 NaHCO3; 4.7 KCl; 1.2 KH2PO4; 2.5 CaCl2; 1.2 MgSO4 and 11 glucose and gassed with 95% O2+5% CO2 gas mixture giving pH 7.4 and P2 580–640 mmHg at 37°C. A fluid-filled latex balloon, connected to a pressure transducer (P23 Pressure Transducer, Gould Statham Instruments Inc.), was inserted into the left ventricle via the left atrium and inflated to set an end-diastolic pressure of 4–7 mmHg. The hearts were enclosed in a small, water-jacketed chamber. The temperature of the perfusate and the atmosphere surrounding the heart was thermostatically controlled to ensure 37°C. The hearts were not stimulated, if not otherwise stated. Global ischemia and reperfusion were induced by clamping and unclamping the aortic inflow line just above the aortic cannula. In some experiments, an epicardial ECG recording was made using silver wire electrodes attached to the ventricular apex, the appendage of the right atrium, and the connective tissue around the base of the aorta. ECG, leftventricular developed pressure (LVDP), and its firsttime derivative (dP/dt) were continuously recorded with an Elema Scho¨nander Mingograph-81 polygraph (Stockholm, Sweden). Coronary flow was quantified by a timed collection and weighing of perfusate exiting the right heart. Experimental protocols The hearts were divided into 21 groups which were assigned to one of six protocols summarized in Figure 1. All the hearts had an initial 20-min equilibration perfusion followed by infusion of either vehicle (0.0025% DMSO, 0.034% ethanol, or no additive), or a non-specific adenosine antagonist, SPT (50 l), a specific adenosine A1 receptor antagonist, DPCPX (0.5 l), or a KATP blocker, glibenclamide (0.6 l). This was then followed by: Protocol (a) (Sham)—a further 130-min aerobic perfusion. In six untreated hearts, a NO-synthase inhibitor, -NAME, was included into the perfusate at 120 min of the protocol. Protocol (b) (Ischemia/reperfusion, IR)—a further 50-min aerobic perfusion followed by a 40-min global ischemia and 40-min reperfusion. In six untreated hearts, -NAME was included into the perfusate at 30 min of the reperfusion period. Protocol (c) (IPC+IR)—three cycles of preconditioning ischemia (3×5 min global ischemia, each followed by 5 min reperfusion), immediately prior to the standard 40-min ischemia/40-min reperfusion. In six untreated hearts, -NAME was

Endothelial Dysfunction and Ischemic Preconditioning Time (min) 20 30 40 50 60 70 80 90 100 110 120 130 140 150

(a) *

*

* L–NAME

Vehicle, SPT, DPCPX or glibenclamide (b)

*

* L–NAME

Vehicle, SPT, DPCPX or glibenclamide

(c)

*

* L–NAME

1737

agonist, CHA [CHA(PC)+IR] or KATP opener, pinacidil [Pin(PC)+IR] infusion and 5-min drug washout prior to the standard ischemia/reperfusion. To prevent changes in the heart rate, the hearts were stimulated at 300 beats/min during the cycles of adenosine and CHA infusion. Protocol (f) [Drug(PC)]—three cycles of adenosine [Ado(PC)], CHA [CHA(PC)] or pinacidil [Pin(PC)] infusion and 5 min of the drug wash-out as in (e), and a further 80 min of the aerobic perfusion. Protocols (d) and (f) were used as time-matched controls for (c) and (e), respectively.

Vehicle, SPT, DPCPX or glibenclamide

(d)

*

*

Evaluation of coronary vascular function

Vehicle

(e)

*

*

Vehicle or glibenclamide (f)

*

*

Vehicle

Figure 1 Experimental protocols. Each experiment started with a 20-min stabilization period, followed by infusion of either vehicle, SPT, DPCPX or glibenclamide, as indicated by the arrows. (a) Sham protocol; (b) ischemia/reperfusion (IR)—hearts underwent a standard IR challenge involving 40 min of global ischemia (solid box) and 40 min of reperfusion; (c) ischemic preconditioning (IPC+IR)—hearts were exposed to three cycles of 5 min global ischemia/5 min reperfusion, before the standard IR; (d) IPC without standard IR—hearts were exposed to three cycles of 5 min global ischemia/5 min reperfusion as in (c), followed by 80 min aerobic perfusion; (e) druginduced preconditioning [Drug(PC)]—hearts were subjected to three cycles of 5 min adenosine, CHA or pinacidil infusion (cross-hatched box)/5 min wash-out, before the standard IR; (f) Drug(PC) without standard IR. Asterisks indicate time-points at which the vasodilator response to SNP and ACh [Protocols (a), (b), (c)] or only to ACh [Protocols (d), (e), (f)] was evaluated. In some untreated hearts, -NAME was administered at the end of the perfusion period [Protocols (a), (b), (c)].

added to the perfusate at 30 min of the reperfusion period. Protocol (d) (IPC)—three cycles of preconditioning ischemia, as in (c) and further 80 min of the aerobic perfusion. Protocol (e) [Drug(PC)+IR]—three cycles of 5 min adenosine [Ado(PC)+IR], adenosine A1 receptor

The magnitude of coronary flow reduction caused by NO-synthase inhibitor, -NAME (10 l), served as an index of basal endothelium- and NO-dependent vasodilator tone. Vasodilator responses to acetylcholine (ACh) and sodium nitroprusside (SNP) served as a measure of an agonist-induced endothelium-dependent and endothelium-independent vascular function, respectively. To minimize a potential preconditioning effect of ACh and NO (Richard et al., 1995; Bilin´ska et al., 1996), either the ACh or SNP response was evaluated in a single heart and the test was performed only once before the ischemia and compared with that performed during the reperfusion (see Fig. 1). In sham-perfused hearts only, vasodilator response to SNP or ACh was evaluated three times in each heart (20 min, 30 min and 150 min of the perfusion protocol). Each ACh and SNP test began with a steady-state coronary flow assessment. Then, bolus of ACh (5 n in 50 ll) or SNP (20 n in 50 ll) was injected into the aortic cannula, while 5-s samples of the effluent were collected and weighed over the next 60 s. During the consecutive tests, the volume of ACh or SNP bolus was adjusted in proportion to the actual coronary flow. Data from these measurements were used to calculate a 1min coronary overflow produced by the drug and then a normalized vasodilator response to the drug (a drug-induced coronary overflow at 150 min/the overflow at 30 min×100%). To quantify post-ischemic coronary flow hyperemic response, coronary effluent collected over the initial 5 min of the reperfusion period following the 40 min ischemia (coronary reflow) was compared with the effluent collected between 35 and 40 min of the experimental protocol (pre-ischemic flow). The formula used was: hyperemic response (%)=(coronary reflow/pre-ischemic flow)×100%.

1738

M. Maω czewski and A. Bereω sewicz

Agents used and selection of their concentrations Acetylcholine chloride (ACh), adenosine, sodium nitroprusside (SNP), Nx-nitro--arginine methyl ester (-NAME) were purchased from Sigma and N6-cyclohexyladenosine (CHA), 8-Cyclopentyl-1, 3-dipropylxanthine (DPCPX), 8-p-Sulfophenyltheophylline (SPT), pinacidil and glibenclamide, from Research Biochemicals International (Natick, MA, USA). Desired amounts of adenosine, -NAME and SPT were dissolved in the perfusing fluid immediately before use. The other agents were made up as concentrated stock solutions in: perfusing solution (ACh and SNP), DMSO (glibenclamide, 24.3 m), the mixture of 2% ethanol and 5 m HCl (pinacidil, 1 m), 96% ethanol (CHA, 46.7 m and DPCPX, 8.2 m). The stock solutions were then diluted directly into the perfusate immediately prior to use. The glassware and tubing containing SNP were protected from light. In order to select a proper concentration of an agent, preliminary concentration–response studies were performed. The boluses of ACh (5 n in 50 ll) and SNP (20 n in 50 ll) were found to produce c. 75% of maximum increase in coronary flow and 10 l -NAME to produce maximal reduction in coronary flow. Adenosine, 10 l, and CHA, 0.25 l, produced a steady maximum increase in coronary flow (increase by 75 and 66%, respectively, Table 2) and in atrio–ventricular conduction time, suggesting full activation of vascular as well as myocardial adenosine receptors. Infusion of pinacidil produces coronary flow increase which peaks after c. 2 min and afterwards slowly decays. To obtain coronary flow changes comparable to those produced by adenosine and CHA, 170 n pinacidil was selected. With this concentration, the flow increased by 67 and 45% at 2 and 5 min of the pinacidil infusion, respectively. Glibenclamide, 0.6 l, caused a 8.6-fold parallel rightward shift of the pinacidil concentration–response curve for the coronary flow increase. The concentrations of SPT and DPCPX were chosen from the known Ki value for A1 adenosine receptors (Armstrong and Ganote, 1994), with c. 20- and 700-fold increase, respectively, to ensure full activity.

Statistics All data are expressed as mean±... In most cases, differences among groups were calculated by one-way analysis of variance followed by Dunnet’s procedure. Paired or non-paired Student’s t-test was also used when appropriate. To test for the

differences in the normalized responses to ACh and SNP (Figs 4, 5, 6) and in the hyperemic responses, Wilcoxon rank test was performed. Differences between groups were considered significant if P value was <0.05.

Results Post-ischemic endothelial dysfunction and the effect of IPC In the sham-perfused hearts, -NAME caused a 36.7±2.4 and 31.6±3% reduction in coronary flow and LVDP, respectively. The -NAME induced drop in coronary flow amounted to 19.2±1.8 and 27.9±2.1% in the untreated hearts subjected to ischemia/reperfusion and to IPC prior to ischemia/ reperfusion, respectively (n=6/group). Thus, ischemia/reperfusion reduced the -NAME response by 54% (P<0.05) and IPC partially prevented this effect (P<0.05 v IR and v sham). The vasodilator response to ACh was reduced by 71% while SNP response was increased by 21% in -NAME perfused hearts (Fig. 2), indicating that mainly NO-release accounts for the ACh-induced vasodilation in guinea-pig hearts. Neither of the vehicles used in this study affected ACh and SNP responses, and these responses did not change significantly during the sham perfusion [Protocol (a)]. Therefore, for further comparisons, data from all sham groups were pooled and treated as one group (Table 1, untreated sham). In the untreated hearts, ischemia/reperfusion resulted in a 47% reduction in ACh response, IPC fully prevented this effect, and SNP response was affected neither by ischemia/reperfusion nor by IPC (Fig. 3, Table 1). Of importance, ACh response was not affected by IPC applied to the non-ischemic hearts [Protocol (d), Table 1, untreated IPC].

Role of adenosine in the IPC-induced protection Administration of SPT (50 l) and DPCPX (0.5 l) resulted in a 30 and 12% reduction in ACh response, respectively, while SNP response was not affected (Table 1, SPT sham, DPCPX sham). In the SPT- and DPCPX-treated hearts, ischemia/reperfusion resulted in a further 43 and 46% reduction in ACh response, respectively, IPC fully prevented these effects (Fig. 4, Table 1), and the vasodilation to SNP was comparable in all SPTand DPCPX-treated groups (Table 1).

Endothelial Dysfunction and Ischemic Preconditioning (a)

Role of KATP in the IPC-induced protection Administration of glibenclamide (0.6 l) resulted in a 36 and 41% reduction in the vasodilator responses to ACh and SNP, respectively (Table 1, glibenclamide sham). In glibenclamide-treated hearts, ischemia/reperfusion resulted in a further 41% reduction in ACh response, the effect not prevented by IPC (Fig. 5, Table 1). Neither ischemia/ reperfusion nor IPC produced significant change in SNP response over that produced by glibenclamide itself (Table 1). Pinacidil (170 n), applied to mimic IPC, completely prevented the impairment of ACh response caused by ischemia/reperfusion and did not affect ACh response in the non-ischemic hearts (Table 2). Both the protective (Fig. 5) and the vasodilator effects of pinacidil and CHA (Table 2) were prevented by glibenclamide.

1.0

0.8 Ach response (ml/g)

1739

0.6

0.4 * * 0.2

* * *

0.0 (b) 1.4 * 1.2

Hemodynamic functions in guinea-pig hearts

* *

SNP response (ml/g)

1.0 0.8 * 0.6 0.4 0.2 0.0 0

10

20 30 Time (s)

40

50

Figure 2 Effect of NO-synthase inhibition with -NAME (10 l) on coronary vasodilator response to acetylcholine (a) and sodium nitroprusside (b) in isolated guinea-pig hearts. Values are means±... of five experiments. ∗P<0.05, -NAME v control. Closed circles, control; open circles, -NAME.

When applied to mimic IPC, adenosine (10 l) partially and CHA (0.25 l) completely prevented the deleterious effect of ischemia/reperfusion on ACh response (Fig. 4, Table 2). Neither adenosine nor CHA affected ACh response in the non-ischemic hearts [Table 2, Ado(PC), CHA(PC)].

There were no significant differences in baseline values for coronary flow and LVDP between any of the experimental groups (Table 1, 20 min). Neither coronary flow nor LVDP was affected by SPT and DPCPX. Glibenclamide reduced these variables by 40 and 36%, respectively (Table 1, 30 min). Adenosine, CHA and pinacidil increased coronary flow by 75, 66 and 45%, respectively (Table 2, 45 min), the effects partially reversible upon 5 min wash-out of the drugs (not shown). Thus, as a consequence of these different treatments, there were significant differences in the immediately pre-ischemic values for the hemodynamic variables between the experimental groups. Nevertheless, the percent postischemic recoveries of coronary flow and LVDP did not differ between any experimental groups and amounted approximately to 90 and 65% of the preischemic values (obtained at 30 min of the protocol), respectively (Tables 1 and 2). As exemplified in Figure 6, the only difference between the ischemia/ reperfusion and IPC+IR groups was that the hyperemic response upon the reperfusion following the 40 min ischemia was significantly augmented in IPC+IR group. Such augmentation was typical for all these groups (except that treated with pinacidil) in which the preconditioning measures afforded to protect ACh response (Tables 1 and 2).

Discussion The major finding of this study is that IPC protects guinea-pig hearts against the post-ischemic endo-

10.8±0.6 10.1±0.6 10.4±0.7

11.3±0.6 11.3±0.8 11.6±0.6

DPCPX (0.5 l/) Sham 5/5 IR 6/5 IPC+IR 6/5

Glib (0.6 l/) Sham IR IPC+IR

6.8±0.7† 7.4±0.3† 7.4±0.3†

10.9±0.6 10.0±0.7 10.1±0.6

11.2±0.6 10.2±0.6 10.0±0.6

10.5±0.7 10.9±0.6 11.0±0.6 10.5±0.6

30 min

6.7±0.7† 6.3±0.4∗ 6.1±0.4∗

10.6±0.5 8.6±0.4∗ 8.6±0.4∗

11.0±0.6 9.3±0.6∗ 9.7±0.3∗

10.1±0.7 10.0±0.6∗ 9.7±0.3∗ 10.1±0.5

150 min

30 min

76.5±5.5 45.5±3.8∗ 43.4±3.4∗

73.6±3.8 49.4±3.7∗ 47.9±4.6∗

81.1±3.5 49.6±3.1∗ 47.9±4.6∗ 87.8±3.9

150 min

84.8±4.3 54.3±4.0† 54.4±4.2† 96.3±3.2 61.8±3.6† 35.2±1.8∗ 93.2±2.9 57.3±3.9† 37.3±3.2∗

75.5±6.1 79.9±5.5 78.3±4.6 78.1±5.1 74.3±4.2 76.4±4.2

79.5±4.1 75.3±3.8 80.2±3.7 77.0±3.5 85.2±3.2 81.0±3.0

83.2±3.8 83.0±3.7 79.2±3.2 78.0±3.3 81.5±3.5 81.3±3.2 89.6±4.1 89.4±4.1

20 min

LVDP (mmHg)

3.3±0.4 .. ..

4.1±0.5 .. ..

4.3±0.5 .. ..

3.5±0.4 .. .. ..

20 min

20 min

3.5±0.4 3.3±0.2 2.0±0.3∗ .. 3.9±0.4‡ .. 3.7±0.4 ..

150 min

2.1±0.3† 2.0±0.2† 3.4±0.2 2.9±0.2 1.7±0.2∗ .. 2.9±0.3 1.5±0.3∗ ..

3.6±0.4† 3.7±0.4 4.3±0.5 3.5±0.4 1.9±0.3∗ .. 3.2±0.3 3.7±0.5‡ ..

3.7±0.4 3.5±0.4 3.8±0.3

4.0±0.2 3.9±0.3 3.8±0.3

3.3±0.2 3.1±0.2 3.3±0.2 ..

150 min

142±10 170±10‡

136±11 163±11‡

143±9 189±11‡

Hyperemic response (%)

2.0±0.2† 2.3±0.2† 2.2±0.2 2.1±0.2 140±15 2.3±0.2 2.1±0.3 138±11

3.9±0.5 3.8±0.4 4.0±0.3

4.2±0.3 4.2±0.2 3.9±0.3

3.2±0.2 3.2±0.2 3.4±0.2 ..

30 min

SNP response (ml)

3.0±0.4† 3.1±0.4† 3.9±0.3 2.8±0.2 1.6±0.2∗ .. 2.7±0.2 2.5±0.3‡ ..

3.6±0.5 3.8±0.4 3.8±0.6 4.0±0.4

30 min

ACh response (ml)

Values are means±...; N and n, number of hearts in which vasodilator responses to acetylcholine (ACh) and sodium nitroprusside (SNP) were evaluated, respectively; SPT, 8-psulfophenyltheophylline; DPCPX, 8-Cyclopentyl-1,3-dipropylxanthine; Glib, glibenclamide; IR, 40 min ischemia+40 min reperfusion; IPC, ischemic preconditioning; LVDP, left ventricular developed pressure; .., not evaluated. After baseline measurements (20 min), the vehicle, SPT, DPCPX or glibenclamide was administered and the measurements were taken 10 min later (30 min). Then, the remaining elements of the respective study protocols followed and the measurements were repeated at 150 min (see Fig. 1). ACh and SNP responses were calculated as the drug-induced coronary overflow at the indicated time points. Hyperemic response, coronary flow collected over the initial 5 min of the reperfusion period following the standard 40 min ischemia expressed as a percent of a 5-min fraction of the pre-ischemic coronary flow. ∗P<0.05, v corresponding ‘‘30 min’’; †P<0.05, v corresponding ‘‘20 min’’; ‡P<0.05, IPC+IR v corresponding IR.

5/5 7/6 7/6

11.5±0.6 10.3±0.5 10.3±0.6

5/5 7/6 7/6

SPT (50 l) Sham IR IPC+IR

10.7±0.7 10.9±0.6 11.1±0.7 10.5±0.6

12/12 7/6 7/6 5/0

20 min

Coronary flow (ml/min/g)

Untreated Sham IR IPC+IR IPC

N/n

Table 1 Effect of 8-p-sulfophenyltheophylline, DPCPX, and glibenclamide on IPC-induced changes in hemodynamics and coronary vasodilator response to acetylcholine and sodium nitroprusside in isolated guinea-pig hearts

1740 M. Maω czewski and A. Bereω sewicz

Endothelial Dysfunction and Ischemic Preconditioning (a)

Normalized ACh response (%)

120 †

100

80

60

*

40

20

0 (b)

Normalized SNP response (%)

120

100

80

60

40

20

0 Sham

IR

IPC + IR

Figure 3 Normalized vasodilator response to acetylcholine (a) and sodium nitroprusside (b) in the untreated isolated guinea-pig hearts subjected to the aerobic sham perfusion (closed boxes, sham), standard ischemia/reperfusion (open boxes, IR) or ischemic preconditioning before the standard ischemia/reperfusion (hatched boxes, IPC+IR). ∗P<0.05, v sham; †P<0.05, v IR; n=6–12 hearts per column (see Table 1).

thelial dysfunction and that this protective effect involves activation of KATP, but not adenosine A1 receptors.

Ischemic preconditioning in ischemic/reperfused guineapig heart The tests with -NAME and ACh performed here provide indices of basal and receptor-stimulated endothelial release of biologically active NO, respectively. We demonstrate that in guinea-pig

1741

hearts, ischemia/reperfusion resulted in the selective endothelial dysfunction in which the basal and receptor-stimulated endothelium-dependent vasodilation was impaired whilst coronary smooth muscle function, as probed with SNP, remained intact. The endothelial dysfunction, as defined above, was prevented by IPC. NO production has not been directly measured in this study. Therefore, it remains unknown whether the IPC effect involves an increased NO production, a decreased NO destruction or both. The protection of the endothelium by IPC has been described before in coronary microcirculation in dogs (Defily and Chilian, 1993), in rat isolated coronary arteries in vitro (Richard et al., 1994; Kaeffer et al., 1996), and in blood (Kolocassides et al., 1996) and crystalloid (Bouchard and Lamontange, 1996) perfused rat hearts. In most of these studies, IPC protected also cardiomyocytes (Richard et al., 1994; Kaeffer et al., 1996; Kolocassides et al., 1996). In this study, the discernible effects of IPC included only endothelial protection, and, perhaps related, augmentation of the post-ischemic hyperemic response. The myocardium, as such, did not seem to be protected in our model, as the post-ischemic recoveries of LVDP (Fig. 6) and the degree of the ultrastructural damage to the cardiomyocytes (Bereω sewicz et al., 1998) did not differ between ischemia/reperfusion and IPC+IR groups. IPC has not been studied thoroughly in guineapig hearts. Recently, Miyamae et al. (1997) demonstrated IPC to attenuate the post-ischemic deterioration of LVDP and creatine kinase release in isolated guinea-pig hearts, the results suggesting that the myocardium is protectable by IPC also in this species. The post-ischemic recovery of LVDP amounted to 30% of the pre-ischemic values in the study of Miyamae et al. (1997), comparing to c. 65% in our study. In addition, in our model of ischaemia/ reperfusion, the ultrastructure of the cardiomyocytes was relatively well preserved, with only 16 and 4.5% of the myocytes presenting the mild and severe injury, respectively (Bereω sewicz et al., 1998). These data suggest much milder heart injury in our experiments compared to those in the study of Miyamae et al. (1997) (stunning v irreversible injury?). In some species (dog, pig), IPC was found to be not protective against the myocardial stunning (Ovize et al., 1992; Miyamae et al., 1993). We speculate that similar is true for guinea-pig hearts. The studies in in vivo dog hearts suggested that the post-ischemic endothelial dysfunction does not occur without irreversible myocardial damage (Bauer et al, 1993; Ehring et al., 1995). However, this and our earlier study (Bereω sewicz et al., 1998)

M. Maω czewski and A. Bereω sewicz

1742

*

*

N.S.

*

N.S.

*

*

N.S.

*

*

*

140

Normalized ACh response (%)

120 100 80 60 40

CHA(PC) + IR

ADO(PC) + IR

DPCPX + IPC + IR

DPCPX + IR

DPCPX – Sham

SPT + IPC + IR

SPT + IR

SPT – Sham

IPC + IR

IR

0

Sham

20

Figure 4 Normalized post-ischemic ACh response in the hearts (from left to right): untreated, treated with 50 l SPT, treated with 0.5 l DPCPX, and preconditioned with 10 l adenosine or 0.25 l CHA. Filled, open, and hatched columns represent sham, standard ischemia/reperfusion, and IPC+standard ischemia/reperfusion hearts, respectively. Cross-hatched columns represent hearts preconditioned with adenosine and CHA before the standard ischemia/ reperfusion. ∗P<0.05; .., non-significant; see Tables 1 and 2 for the number of experiments in each group.

suggest a stronger vulnerability of endothelium to ischemia/reperfusion damage as compared with myocytes and smooth muscle cells in guinea-pig hearts. Similar results have been reported also for isolated rat hearts (Mankad et al., 1997). Furthermore, we demonstrate that the IPC protection of the endothelium is not translated into the improved mechanical function of the hearts. From this, we conclude that, at least in the crystalloid perfused guinea-pig hearts, the endothelial dysfunction does not participate in the mechanism of the myocardial stunning, and vice versa. Role of adenosine and KATP in the mechanism of IPCinduced protection Both adenosine and KATP have been implicated as mediators of the protection afforded by IPC in the

post-ischemic myocardium (Gross, 1995; Cohen and Downey, 1996). We demonstrate that the protective effect of IPC was fully prevented by glibenclamide and mimicked by pinacidil. In addition, the protection from pinacidil was prevented by glibenclamide, confirming that the effect of pinacidil was related to KATP activation. These data indicate that the activation of KATP fully accounts for the IPC-induced endothelial protection. However the triggers resulting in such an activation are not apparent from this study. The activation of KATP via adenosine A1 receptors has been found to mediate IPC in several experimental models (Auchampach and Gross, 1993; Toombs et al., 1993). In this study, the protective effect of IPC was mimicked by adenosine and adenosine A1 receptor agonist, CHA, and the effect of CHA was prevented by glibenclamide. Adenosine and CHA were used in the concentrations producing

7 5

7 5

7 7

CHA(PC)+IR CHA(PC)

Pin(PC)+IR Pin(PC)

Glib+Pin(PC)+IR Glib+CHA(PC)+IR

8.0±0.4 7.9±0.3

11.0±0.5 10.9±0.6

10.7±0.6 10.9±0.5

10.6±0.5 11.1±0.5

10.5±0.7 10.9±0.6 11.0±0.6

30 min

8.1±0.4 8.2±0.4

15.9±1.3∗ 15.7±0.6∗

18.1±0.8∗ 17.9±0.4∗

18.9±0.9∗ 19.4±1.2∗

45 min

6.9±0.2∗ 6.7±0.3∗

10.6±0.4 11.4±0.4

10.2±0.4 10.5±0.5

10.2±0.4 11.2±0.5

10.1±0.7 10.0±0.6∗ 9.7±0.3∗

150 min

Coronary flow (ml/min/g)

61.4±3.1 62.0±2.6

78.9±3.8 82.3±4.5

77.7±3.0 77.2±4.5

77.1±3.8 76.7±4.3

83.0±3.7 78.0±3.3 81.3±3.2

30 min

60.5±2.8 62.8±2.6

82.4±3.1 86.0±5.5

81.0±3.0 81.8±4.6

81.5±4.0 82.0±4.5

45 min

LVDP (mmHg)

37.5±2.5∗ 38.3±2.1∗

46.7±3.5∗ 84.2±3.6

46.1±3.2∗ 81.7±4.3

45.7±3.1∗ 80.7±4.6

81.1±3.5 49.6±3.1∗ 47.9±4.6∗

150 min

2.6±0.3 2.4±0.3

4.1±0.4 4.5±0.5

3.9±0.3 3.5±0.5

3.8±0.4 4.2±0.5

3.6±0.5 3.8±0.4 3.8±0.6

30 min

1.6±0.2∗ 1.6±0.2∗

3.7±0.3† 4.4±0.5

4.2±0.5† 3.5±0.5

2.9±0.3† 4.2±0.5

3.5±0.4 2.0±0.3∗ 3.9±0.4†

150 min

ACh response (ml)

140±11 145±11

161±9

169±9†

167±10†

143±9 189±11†

Hyperemic response (%)

Values are means±...; n, number of hearts; LVDP, left-ventricular developed pressure; ACh, acetylcholine; IR, 40 min ischemia+40 min reperfusion; IPC, Ado(PC), CHA(PC), Pin(PC), preconditioning with ischemia, adenosine, N6-cyclohexyladenosine and pinacidil, respectively (3×5 min ischemia/reperfusion or 3×5 min drug infusion/wash-out, respectively); Glib, glibenclamide. The measurements obtained at 30 min of the experimental protocols are compared with those obtained during the first 5 min episode of the respective preconditioning drug infusion (45 min) and those obtained at the end of the experiment (150 min, cf. Fig. 1). ∗P<0.05, v corresponding ‘‘30 min’’; †P<0.05, v untreated-IR.

8 5

24 13 13

Ado(PC)+IR Ado(PC)

Untreatd-sham Untreated-IR Untreated-IPC+IR

n

Table 2 Effect of preconditioning with adenosine, N6-cyclohexyladenosine and pinacidil on hemodynamics and coronary vasodilator response to acetylcholine in isolated guinea-pig hearts subjected to ischemia/reperfusion

Endothelial Dysfunction and Ischemic Preconditioning

1743

M. Maω czewski and A. Bereω sewicz

1744

*

*

*

N.S.

120 *

*

*

*

N.S.

*

Normalized ACh response (%)

100

80

60

40

Gli + CHA(PC) + IR

CHA(PC) + IR

Gli + Pin(PC) + IR

Pin (PC) + IR

Gli + IPC + IR

Gli + IR

Gli + Sham

IPC + IR

IR

0

Sham

20

Figure 5 Normalized post-ischemic ACh response in the hearts (from left to right): untreated, treated with 0.6 l glibenclamide, preconditioned with 170 n pinacidil, and preconditioned with 0.25 l CHA. Filled, open, and hatched columns represent sham, standard ischemia/reperfusion, and IPC+standard ischemia/reperfusion hearts, respectively. Diagonally crosshatched columns represent hearts preconditioned with pinacidil and CHA before the standard ischemia/ reperfusion. Vertically-lined columns represent hearts preconditioned with pinacidil and CHA which were pretreated with 0.6 l glibenclamide. ∗P<0.05; .., non-significant; see Tables 1 and 2 for the number of experiments in each group.

full activation of vascular as well as myocardial adenosine receptors (see Materials and Methods). These data demonstrate that adenosine A1 receptors, if sufficiently activated, have a potential to open KATP to trigger the protection. Nevertheless, the IPC-induced protection could be abolished neither by SPT, nor by A1 receptor antagonist, DPCPX. SPT was used in the concentration (50 l) which was approximately 20- and 5-times higher than the respective adenosine A1 and A3 receptor binding affinities for SPT, respectively. The concentration of DPCPX (0.5 l) was 725-times higher than A1 receptor Ki for the agent (Armstrong and Ganote, 1994). Thus, insufficient adenosine A1 receptor blockade is not a likely explanation for SPT and DPCPX inability to prevent IPC-induced protection.

We rather believe that in guinea-pig hearts the activation of A1 receptors during IPC episodes does not reach a threshold necessary to induce the protection. This might be due to a limited access of adenosine to receptors critical to the protection (e.g. on endothelium?). Adenosine acting via A3 receptors has been proposed to mediate the protection afforded by IPC (Armstrong and Ganote, 1994; Liu et al., 1994). The observation that the protection was not abolished by 50 l SPT makes the protective role of A3 receptors in our experimental model doubtful. The advantage of the isolated heart model used here is that it allows to study coronary microcirculation. The limitation of the model is, however, the presence of confounding factors (e.g. hemo-

Endothelial Dysfunction and Ischemic Preconditioning (a)

LVDP (mmHg)

80

60

40

20

0 (b) 30

CF (ml/g wet weight)

25

* ** * *

20 15 10 5 0 40 50 60 70 110

120 130 Time (min)

140

150

Figure 6 Post-ischemic recovery of coronary flow (a) and LVDP (b) in the untreated isolated guinea-pig hearts subjected to the standard ischemia/reperfusion (closed circles, IR, n=13) and to ischemic preconditioning before the standard ischemia/reperfusion (open circles, IPC+IR, n=13). ∗P<0.05, v IR.

dynamic depressive effect of glibenclamide) that are difficult to eliminate. Another complication of this study is that treatment with either glibenclamide or adenosine antagonists reduced the basal ACh response, and that glibenclamide reduced the basal SNP response as well, the effects observed also by others in isolated rat hearts (Bouchard and Lamontagne, 1997). In this study, we demonstrate -NAME and glibenclamide to produce comparable depression of coronary flow and LVDP. Nevertheless, SNP response was enhanced in -NAME-perfused hearts and it was attenuated in glibenclamideperfused hearts. This suggests that the depressed vascular NO responsiveness after glibenclamide is independent from the alterations in the hemodynamics. We believe rather that the changes in

1745

ACh and SNP responses produced by glibenclamide, and also by SPT and DPCPX, could be explained in terms of a possible functional interplay between NO and KATP (Kubo et al., 1994; Ma¸czewski and Bereω sewicz, 1997) and between adenosine and NO (Abebe et al., 1995). Whatever the mechanism of the drug-induced alterations in ACh and SNP responses might be, it is important that for each treatment, the pre-ischemic hemodynamic variables and responses to ACh and SNP were similar in ischemia/reperfusion and remaining groups. This, in turn, makes it unlikely that the depressive effect of glibenclamide on coronary flow and LVDP contributes to its antagonistic action on the protective effects of IPC, CHA and pinacidil. Altogether, our data demonstrate that the activation of KATP accounts for the IPC protection against the post-ischemic endothelial dysfunction while the activation of A1, and perhaps of A3 receptors is not critical to the protection. These data, however, seem to contrast with two recent reports implicating that adenosine is protective. Thus, the IPC-induced protection of coronary response to the endothelium-dependent vasodilator serotonin has been found to be blocked by adenosine antagonist 8-phenyltheophylline (and only partially by glibenclamide) in coronary arteries isolated from the post-ischemic rat hearts (Bouchard and Lamontange, 1996). Furthermore, the role of adenosine A1 and A3 receptors in the mechanism in which IPC prevents the post-ischemic impairment of hypoxic coronary vasodilation (NO-dependent phenomenon) has been demonstrated in isolated guinea-pig hearts (Giannella et al., 1997). It remains to be established whether this is only the differences in the experimental models which explains these discrepant results.

Mechanism of IPC protection mediated by KATP KATP are expressed in plasma membrane of cardiomyocytes and coronary smooth muscle cells (Nelson and Quayle, 1995), and in endothelial cells (Katnik and Adams, 1995), including those from guinea-pig coronaries (Langheinrich and Daut, 1997). The mechanisms by which IPC-induced activation of KATP protects myocardium from the ischemia/reperfusion injury are not completely understood, but seem to be mediated by energy conservation (Grover, 1997). As already discussed, IPC-induced protection of the endothelium seems to involve a mechanism intrinsic to the vascular tissue. This points to KATP within the coronary microcirculation as mediators

1746

M. Maω czewski and A. Bereω sewicz

of the protection. Coronary endothelial cells and the underlying smooth muscle cells are coupled through myoendothelial gap junctions (Beny, 1997). Therefore, the hyperpolarization related to KATP will be transmitted from smooth muscle cells to the endothelial cells, and vice versa. In endothelial cells, the hyperpolarization, either transmitted or related to KATP, will increase the electrochemical gradient and facilitate Ca2+ influx (Nilius et al., 1997). This, in turn, will enhance endothelial NO production (Luckhoff and Busse, 1990). However, a possible role of NO- in IPC-induced protection of the endothelium remains unknown. There is growing evidence that the sarcolemmal KATP may not be involved with the cardioprotection mechanism of KATP openers (Grover, 1997). Recent evidence supports the role for mitochondrial KATP as mediators of myocardial protection induced by these agents (Garlid et al., 1997). However, the role of mitochondrial KATP in IPC-induced protection, and particularly in the protection of the endothelium, remains to be elucidated. All preconditioning stimuli tested here failed to affect ACh response in the non-ischemic hearts. Therefore, the protection they afforded is not due to a simple stimulation of the endothelial function, but rather involves the induction of a memory which makes the endothelium resistant specifically to the ischemia/reperfusion injury. The role of free radicals in the development of the post-ischemic endothelial dysfunction has been suggested because in several species (Mehta et al., 1989b; Tsao et al., 1990; Gross et al., 1992), including guinea-pig (Bereω sewicz et al., 1998), superoxide dismutase and catalase were able to prevent this injury. Thus, one likely explanation for the endothelial protection afforded by IPC is that it attenuates vascular production of free radicals. In fact, IPC has been reported to attenuate the postischemic radical production in isolated rat hearts (Tosaki et al., 1994) and superoxide anion generation by the mitochondria of the post-ischemic rat hearts (Park et al., 1997). The possible relation, if any, between the activity of KATP and free radical production remains unknown. In conclusion, we found that in the post-ischemic guinea-pig heart: (1) the basal and the receptorstimulated endothelium-dependent vasodilation is impaired while the vascular responsiveness to exogenous NO is not affected; (2) this endothelial dysfunction is prevented by IPC; (3) the protective effect of IPC involves activation of KATP, but not adenosine A1 receptors; (4) exogenous adenosine and the adenosine A1 receptor agonist afford the protection, which might be of potential clinical significance; (5) the endothelial dysfunction is not

involved in the mechanism of myocardial stunning in guinea-pig hearts.

Acknowledgements We thank Alicja Protasowicka and Marek Woz´niak for technical assistance. The study was supported by the CMKP 501-1-1-05-16/97 grant.

References A W, H T, O H, M SJ, 1995. Role of nitric oxide in adenosine receptor-mediated relaxation of porcine coronary artery. Am J Physiol 38: H1672–H1678. A S, G CE, 1994. Adenosine receptor specificity in preconditioning of isolated rabbit cardiomyocytes: evidence of A3 receptor involvement. Cardiovasc Res 28: 1049–1056. A JA, G GJ, 1993. Adenosine-A1 receptors, KATP channels, and ischemic preconditioning in dogs. Am J Physiol 264: H1327–H1336. B B, S BZ, K RA, P K, 1993. Does preconditioning protect the coronary vasculature from subsequent ischemia/reperfusion injury? Circulation 88: 659–672. B JL, 1997. Electrical coupling between smooth muscle cells and endothelial cells in pig coronary arteries. Pflu¨gers Arch-Eur J Physiol 433: 364–367. Bω  A, C E, Mω  M, 1998. Ischemic preconditioning and superoxide dismutase protect against endothelial dysfunction and endothelium glycocalyx disruption in the postischemic guinea-pig hearts. Mol Cell Biochem (in press). B´  M, Mω  M, Bω  A, 1996. Donors of nitric oxide mimic effects of ischaemic preconditioning of reperfusion induced arrhythmias in isolated rat heart. Mol Cell Biochem 160/161: 265– 271. B JF, L D, 1996. Mechanisms of protection afforded by preconditioning to endothelial function against ischemic injury. Am J Physiol 40: H1801–H1806. C MV, D JM, 1996. Myocardial preconditioning promises to be a novel approach to the treatment of ischemic heart disease. Annu Rev Med 47: 21–29. D DV, C WM, 1993. Preconditioning protects coronary arteriolar endothelium from ischemia-reperfusion injury. Am J Physiol 265: H700–H706. E T, K M, B D, K S, H M, H G, 1995. Cholinergic and a-adrenergic coronary constriction with increasing ischemia– reperfusion injury. Am J Physiol 268: H886–H894. G KD, P P, Y V, M HN, D RB, D AJ, L NJ, S MA, G GJ, 1997. Cardioprotective effect of diazoxide and its interaction with mitochondrial ATP-sensitive K+ channels: possible mechanism of cardioprotection. Circ Res 81: 1072–1082. G E, M HC, L R, 1997. Ischemic

Endothelial Dysfunction and Ischemic Preconditioning preconditioning prevents the impairment of hypoxic coronary vasodilation caused by ischemia/reperfusion. Role of adenosine A1/A3 and bradykinin B2 receptor activation. Circ Res 81: 415–422. G GJ, 1995. ATP-sensitive potassium channels and myocardial preconditioning. Basic Res Cardiol 90: 85– 88. G GJ, O’R ST, P LR, W DC, 1992. Myocardial and endothelial dysfunction after multiple, brief coronary occlusions—role of oxygen radicals. Am J Physiol 263: H1703–H1709. G GJ, 1997. Pharmacology of ATP-sensitive potassium channel (KATP) openers in models of myocardial ischemia and reperfusion. Can J Physiol Pharmacol 75: 309–315. K N, R V, F A, L F, H JP, T C, 1996. Preconditioning prevents chronic reperfusion-induced coronary endothelial dysfunction in rats. Am J Physiol 40: H842–H849. K C, A DJ, 1995. An ATP-sensitive potassium conductance in rabbit arterial endothelial cells. J Physiol (London) 485: 595–606. K KG, G M, H DJ, 1996. Ischemic preconditioning, cardioplegia or both? Differing approaches to myocardial and vascular protection. J Mol Cell Cardiol 28: 623–634. K M, N Y, M S, S K, K Y, 1994. Atrial natriuretic factor and isosorbide dinitrate modulate the gating of ATP-sensitive K+ channels in cultured vascular smooth muscle cells. Circ Res 74: 471–476. L U, D J, 1997. Hyperpolarization of isolated capillaries from guinea-pig heart induced by K+ channel openers and glucose deprivation. J Physiol (London) 502: 397–408. L AM, L DJ, 1996. The role of nitric oxide and cell adhesion molecules on the microcirculation in ischaemia-reperfusion. Cardiovasc Res 32: 743–751. L GS, R SC, O RA, M KH, W RS, D JM, 1994. Evidence that the adenosine A3 receptor may mediate the protection afforded by preconditioning in the isolated rabbit heart. Cardiovasc Res 28: 1057–1061. L A, B R, 1990. Calcium influx into endothelial cells and formation of endothelium-derived relaxing factor is controlled by the membrane potential. Pflu¨gers Arch 416: 305–311. Mω  M, Bω  A, 1997. Inhibitors of nitric oxide and ischemia/reperfusion attenuate coronary vasodilator response to pinacidil in isolated rat heart. J Physiol Pharmacol 48: 737–749. M PS, A M, R S, S NJ, Y MH, 1997. Relative susceptibilty of endothelium and myocardial cells to ischaemia–reperfusion injury. Acta Physiol Scand 161: 103–112. M JL, N WW, D WH, L DL, S TGP, 1989a. Impaired canine coronary vasodilator response to acetylcholine and bradykinin after occlusion–reperfusion. Circ Res 64: 43–54. M JL, N WW, D WH, L DL, T L, T R M, S TGP, 1989b. Protection by superoxide dismutase from myocardial dysfunction and attenuation of vasodilator reserve after

1747

coronary occlusion and reperfusion in dog. Circ Res 65: 1283–1295. M M, F H, K M, Y R, T M, K M, H K, O M, K K, Y Y, S S, 1993. Preconditioning improves energy metabolism during reperfusion but does not attenuate myocardial stunning in porcine hearts. Circulation 88: 223–234. M M, D I, W MW, C SA, F VM, 1997. Regular alcohol consumption mimics cardiac preconditioning by protecting against ischaemia–reperfusion injury. Proc Natl Acad Sci USA 94: 3235–3239. N MT, Q JM, 1995. Physiological roles and properties of potassium channels in arterial smooth muscle. Am J Physiol 37: C799–C822. N B, V F, D G, 1997. Ion channels in vascular endothelium. Annu Rev Physiol 59: 145–170. O M, P K, H SL, K RA, 1992. Preconditioning does not attenuate myocardial stunning. Circulation 85: 2247–2254. P JW, C YS, K YH, K CH, K MS, 1997. Ischemic preconditioning reduces O2− generation and prevents respiratory impairment in the mitochondria of post-ischemic reperfused heart of rat. Life Sci 60: 2207–2219. P JR, 1994. Protection of the heart by ischaemic preconditioning—mechanisms and possibilities for pharmacological exploitation. Trends Pharmacol Sci 15: 19–25. R V, K N, T C, T C, 1994. Ischemic preconditioning protects against coronary endothelial dysfunction induced by ischemia and reperfusion. Circulation 89: 1254–1261. R V, B T, K N, T C, T C, 1995. Myocardial and coronary endothelial protective effects of acetylcholine after myocardial ischaemia and reperfusion in rats: role of nitric oxide. Br J Pharmacol 115: 1532–1538. S CG, W OL, 1993. Myocardial ischaemia: what happens to the coronary arteris? Trends Pharmacol Sci 14: 448–453. T CF, MG DS, J WE, V J, 1993. Protection from ischaemic–reperfusion injury with adenosine pretreatment is reversed by inhibition of ATP sensitive potassium channels. Cardiovasc Res 27: 623–629. T A, C GA, S P, E RM, D DK, 1994. Effects of preconditioning on reperfusion arrhythmias, myocardial functions, formation of free radicals, and ion shifts in isolated ischemic/reperfused rat hearts. J Cardiovasc Pharmacol 23: 365–373. T PS, A N, L DJ, J G III, L AM, 1990. Time-course of endothelial dysfunction and myocardial injury during ischemia and reperfusion in the cat. Circulation 82: 1402–1412. V KM, MM IF, H LD, 1987. Reperfusion after coronary occlusion in dogs impairs endothelium-dependent relaxation to acetylcholine and augments contractile reactivity in vitro. J Clin Invest 79: 265–274.