## Abstract

Objectives: A hypothesis was tested that a reaction product between superoxide (O2) and nitric oxide (NO) mediates post-ischemic coronary endothelial dysfunction that ischemic preconditioning (IPC) protects the endothelium by preventing post-ischemic cardiac O2 and/or NO formation, and that the opening of the mitochondrial ATP-dependent potassium channel (mKATP) plays a role in the mechanism of IPC. Methods: Langendorff-perfused guinea-pig hearts were subjected either to 30 min global ischemia/30 min reperfusion (IR) or were preconditioned prior to IR with three cycles of either 5 min ischemia/5 min reperfusion or 5 min infusion/5 min wash-out of mKATP opener, diazoxide (0.5 μM). Coronary flow responses to acetylcholine (ACh) and nitroprusside were used as measures of endothelium-dependent and -independent vascular function, respectively. Myocardial outflow of O2 and NO, and functional recoveries were followed during reperfusion. Results: IR impaired the ACh response by approximately 60% and augmented cardiac O2 and NO outflow. Superoxide dismutase (150 U/ml) and NO synthase inhibitor, l-NMMA (100 μM) inhibited the burst of O2 and NO, respectively, and afforded partial preservation of the ACh response in IR hearts. NO scavenger, oxyhemoglobin (25 μM), afforded similar endothelial protection. IPC and diazoxide preconditioning attenuated post-ischemic burst of O2, but not of NO, and afforded a complete endothelial protection. Diazoxide given after 30-min ischemia increased the O2 burst and was not protective. The effects of IPC and diazoxide preconditioning were not affected by HMR-1098 (25 μM), a selective blocker of plasmalemmal KATP, and were abolished by glibenclamide (0.6 μM) and 5-hydroxydecanoate (100 μM), a nonselective and selective mKATP blocker, respectively. 5-Hydroxydecanoate produced similar effects, whether it was given as a continuous treatment or was washed out prior to IR. Conclusion: The results suggest that in guinea-pig heart: (i) a reaction product between O2 and NO mediates the post-ischemic endothelial dysfunction; (ii) the mKATP opening serves as a trigger of the IPC and diazoxide protection; and (iii) the mKATP opening protects the endothelium in the mechanism that involves the attenuation of the O2 burst at reperfusion.

## 1. Introduction

Cardiac ischemia/reperfusion (IR) causes damage to cardiomiocytes and coronary endothelium. Indeed, IR impairs endothelium-dependent, but not endothelium-independent, coronary vasodilation, indicating selective endothelial dysfunction. Oxygen free radicals have been implicated in the mechanism of the post-ischemic endothelial dysfunction [1–4]. The dysfunction can be prevented by superoxide dismutase (SOD), but not by catalase and hydroxyl radical scavengers [2,4–6], implicating superoxide (O2) as the oxidant mediating the dysfunction. It is not known if O2 per se, or rather a product of its reaction with nitric oxide (NO) mediates the injury [7].

Ischemic preconditioning (IPC) is a protective mechanism whereby brief nonlethal episode(s) of ischemia protect the heart against a subsequent lethal ischemia [8]. The hallmark of IPC is infarct size reduction [9,10]. Other beneficial effects of IPC, including the protection against the endothelial dysfunction, have been also identified in IR heart of experimental animals and humans [4,11–14]. It is uncertain if the mechanism of all these effects is the same. For instance, the early endothelial protection by IPC occurs through pathways that only partially parallel those established in cardiomyocytes (for review, see Ref. [15]). The signaling pathway of the cardiomyocyte early IPC protection includes the opening of the ATP-sensitive potassium channel (KATP). Evidence indicates that mitochondrial (mKATP), rather than plasmalemmal KATP (pKATP), is critical for the protection, and that mKATP serves as a trigger rather than an end-effector of IPC [10,16–18]. A nonselective KATP opener, pinacidil, was reported to mimic [13] and a nonselective KATP inhibitor, glibenclamide, to block the early endothelial protection by IPC [13,19], implying that KATP is critical also for this protection. The role of mKATP in the endothelial protection has not been established, and its end-effector remains elusive.

The aims of this study were to ascertain: (i) whether O2 or a product of its reaction with NO mediates the post-ischemic endothelial dysfunction in the isolated guinea-pig heart model [6,13]; (ii) whether classic IPC protects endothelium by preventing post-ischemic O2 and/or NO formation. While it has been independently reported that preconditioning prevents reperfusion-induced formation of free radicals [20–22] and NO [23], these effects have not been observed together in a single experimental model; (iii) a role of mKATP vs. pKATP in the endothelial protection by IPC. Accordingly we examined if mKATP opener, diazoxide [24,25], would mimic the effects of IPC, and if the effects of IPC and diazoxide would be blocked by a selective blocker of either mKATP (5-hydroxydecanoate) [26] or pKATP (HMR-1098) [27], and a nonselective KATP blocker, glibenclamide [26] and (iv) whether mKATP serves as a trigger or an effector of IPC.

## 2. Methods

### 2.1. Chemicals

Acetylcholine chloride (ACh), catalase, diazoxide, ferricytochrome c (type VI, from horse heart), rabbit hemoglobin, NG-methyl-l-arginine acetate salt (l-NMMA), sodium nitroprusside (SNP), and superoxide dismutase were purchased from Sigma, and glibenclamide, 5-hydroxydecanoic acid sodium (5HD) from RBI (Natick, MA, USA). HMR-1098 was a gift from Aventis Pharma (Frankfurt, Germany). Oxyhemoglobin (oxyHb) was prepared as described before [28,29]. HMR-1098, l-NMMA and SOD were dissolved in the perfusate. The other agents were made up as stock solutions in: distilled water (ACh, SNP), DMSO (glibenclamide) or 96% ethanol (diazoxide). SOD, oxyHb, ACh and SNP were infused via a sidearm of the aortic cannula with a digital infusion pump (Kwapisz, Poland). The glassware and tubing containing SNP were protected from light.

The concentration of SOD (150 U/ml) selected for the study prevented endothelial dysfunction in our model [6]. Similarly, glibenclamide, 0.6 μM, prevented IPC-induced endothelial protection [13]. l-NMMA, 100 μM, produced maximum coronary vasoconstriction, and thus afforded maximum NO synthase (NOS) blockade. The concentration of oxyHb (25 μM) is equal to that reported to inhibit bradykinin-induced coronary vasodilation in guinea-pig heart [28]. 5HD, 100 μM, was reported to selectively block mKATP in rabbit cardiomyocytes [26]. HMR-1098, 25 μM, was selected to attenuate the ACh response similar to 0.6 μM glibenclamide.

### 2.2. Isolated heart preparation

The investigation conformed to the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, revised 1996). The preparation used in this study has been described elsewhere [13]. In brief, guinea pig (260–320 g) hearts were perfused by the Langendorff method, at perfusion pressure of 70 mm Hg, with Krebs–Henseleit buffer (KHB) containing, in mmol/l: 118 NaCl; 23.8 NaHCO3; 4.7 KCl; 1.2 KH2PO4; 2.5 CaCl2; 1.2 MgSO4, 11 glucose and gassed with 95% O2+5% CO2 gas mixture. In some experiments KHB was supplemented with cytochrome c (see later). A fluid-filled latex balloon, connected to a pressure transducer (P23 Pressure Transducer, Gould Statham Instruments), was inserted into the LV for measurement of its pressure. The hearts were not paced. Global ischemia was induced by clamping the aortic cannula and simultaneous immersing the heart in a small volume of the venous effluent.

### 2.3. Experimental protocols

To probe for the ability of diazoxide to open pKATP in coronaries, a cumulative concentration–response curve for diazoxide-induced increase in coronary flow in isolated guinea-pig hearts was constructed and the study was repeated in the presence of 5HD, glibenclamide or HMR-1098.

The main study involved three series of experiments (Fig. 1). Series 1 was aimed at testing if O2 and/or NO mediate the post-ischemic endothelial dysfunction. All the hearts had a 25-min stabilization perfusion followed by:

1. (1a)

Sham—a further 110-min aerobic perfusion. Between 65 and 115 min of the protocol, the hearts were perfused either with no additive or with SOD, l-NMMA or oxyHb, and the agent was washed out for the remaining 20 min of the protocol.

2. (1b)

Ischemia/reperfusion (IR)—after 50-min aerobic perfusion, the hearts underwent a test IR challenge involving 30-min global ischemia and 30-min reperfusion. SOD, l-NMMA or oxyHb was infused as in (1a).

Series 2 was aimed at studying: (i) the effect of IPC, and of diazoxide, given either prior or after the test ischemia, on the endothelial function, and O2 and NO generation in IR hearts and (ii) the effect of a continuous treatment with 5HD, glibenclamide and HMR-1098 on the IPC and diazoxide protection. Previously, we verified [13] that the vehicles used in this study (0.0025% DMSO and 0.034% ethanol) affected neither endothelial nor contractile function. After the stabilization perfusion the hearts were perfused either with no additive or with 5HD, glibenclamide or HMR-1098. This was followed by:

1. (2a)

Sham—a further 110 min aerobic perfusion.

2. (2b)

IR—the test IR as in (1b).

3. (2c)

IPC—three cycles of preconditioning ischemia (3×5 min global ischemia, the first two incidents followed by 5 min reperfusion, and the third one by 10 min reperfusion), prior to the test IR.

4. (2d)

Diazoxide preconditioning (Dx-PC)—three cycles of diazoxide infusion (3×5 min infusion, the first two infusions followed by 5 min washout, and the third one by 10 min washout) prior to the test IR.

5. (2e)

Diazoxide at the reperfusion (Dx-R)—the hearts were subjected to the test IR, as in 2b, and diazoxide was infused during the initial 10 min of the reperfusion.

Fig. 1

Experimental protocols. Three series of experiments were performed, all starting with 25-min stabilization and lasting 135 min (see Methods for the details). Asterisks indicate time-points at which ACh or SNP response was evaluated. Solid boxes, global ischemia. Open boxes, diazoxide pulses of the preconditioning protocols.

Fig. 1

Experimental protocols. Three series of experiments were performed, all starting with 25-min stabilization and lasting 135 min (see Methods for the details). Asterisks indicate time-points at which ACh or SNP response was evaluated. Solid boxes, global ischemia. Open boxes, diazoxide pulses of the preconditioning protocols.

Series 3 was aimed at studying whether the IPC and diazoxide-PC protection can be aborted by 5HD given to bracket the brief ischemic periods and diazoxide pulses of the preconditioning protocols. The hearts were subjected either to sham perfusion, IR, IPC or diazoxide-PC as in (2a–d), respectively, and 5HD was applied between 25 and 70 min of the protocol (early treatment).

Each series consisted of two separate sets of hearts. In the experiments aimed at studying endothelium and NO generation the hearts were perfused with KHB. To study O2, KHB was supplemented with cytochrome c, starting from 70 min of the protocol (Fig. 1).

### 2.4. Endothelial function

Vasodilator responses to ACh and SNP served as measures of an agonist-induced endothelium-dependent and endothelium-independent vascular function, respectively [13]. To minimize a preconditioning effect of ACh and/or NO [30,31], either ACh or SNP response was evaluated in a single heart and the test was performed only once before the test ischemia and compared with that performed during the reperfusion. The bolus of ACh (5 nM in 50 μl) or SNP (20 nM in 50 μl) was applied while 10-s samples of the effluent were measured over the next 60 s. These boluses were established to produce a submaximum increase in coronary flow in our model [13]. During the consecutive tests, the volume of ACh or SNP bolus was adjusted in proportion to the actual coronary flow to assure that the heart was exposed to the same drug concentration as during the initial test. A 1-min coronary overflow produced by the drug and a normalized drug's response (a drug induced overflow at the end of the protocol/the overflow during the initial test x 100%) was calculated.

### O2− measurement

O2 formation was determined as described by Southworth et al. [32]. The hearts were perfused with KHB containing 10 μM succinylated ferricytochrome c and 600 U/ml catalase (cytochrome c). The optical density of the coronary effluent was measured at 550 nm (Beckman DU-65 UV spectrophotometer). The O2 formation was calculated using the molar absorbance coefficient for cytochrome c of 21 mM−1 cm−1[32].

### 2.6. NO measurement

NO formation was determined by measuring cardiac release of nitrite and nitrate (nitrite+nitrate=NOx). The method involved two independent procedures. One was the conversion of nitrite to NO gas (in the presence of potassium iodide in acetic acid) and the chemiluminescent detection of NO using Nitric Oxide Analyzer (Sievers, USA). The other was the conversion of NOx to NO gas in the presence of vanadium (III) chloride in hydrochloric acid and further detection of NO. The effluent nitrate was calculated from the formula: [nitrate]=[NOx}−[nitrite].

### 2.7. Statistics

The data are expressed per gram of wet weight, and are means±S.E.M. Significant differences among groups were usually calculated by one-way analysis of variance followed by Dunnet's procedure. To test for the differences in the normalized responses to ACh and SNP (Fig. 3), Kruskal–Wallis test followed by Mann–Whitney test was performed. Values of p<0.05 were considered significant.

Fig. 3

Normalized post-ischemic ACh response (a) and SNP response (b) in the hearts (from left to right): untreated, given 100 μM 5HD as the continuous (5HD-C) or the early treatment (5HD-E) or continuously treated with 0.6 μM glibenclamide or 25 μM HMR-1098. Open, filled, cross-lined, hatched-rising to left, hatched-rising to right, crosshatched, horizontally lined, and vertically lined columns represent: sham, 30-min IR, IR+SOD, IR+l-NMMA, IR+oxyHb, ischemic preconditioning (IPC), diazoxide preconditioning (Dx-PC), and diazoxide given at the reperfusion (Dx-R), respectively. Data from the untreated sham-perfused group and sham groups perfused with SOD, l-NMMA, and oxyHb were pooled (untreated sham). Values are mean±S.E.M. See Table 2 for N in each group. *p<0.05 vs. respective sham; #p<0.05 vs. respective IR.

Fig. 3

Normalized post-ischemic ACh response (a) and SNP response (b) in the hearts (from left to right): untreated, given 100 μM 5HD as the continuous (5HD-C) or the early treatment (5HD-E) or continuously treated with 0.6 μM glibenclamide or 25 μM HMR-1098. Open, filled, cross-lined, hatched-rising to left, hatched-rising to right, crosshatched, horizontally lined, and vertically lined columns represent: sham, 30-min IR, IR+SOD, IR+l-NMMA, IR+oxyHb, ischemic preconditioning (IPC), diazoxide preconditioning (Dx-PC), and diazoxide given at the reperfusion (Dx-R), respectively. Data from the untreated sham-perfused group and sham groups perfused with SOD, l-NMMA, and oxyHb were pooled (untreated sham). Values are mean±S.E.M. See Table 2 for N in each group. *p<0.05 vs. respective sham; #p<0.05 vs. respective IR.

## 3. Results

### 3.1. Diazoxide concentration–response studies

Diazoxide caused a concentration-dependent rise in coronary flow (Fig. 2). While this effect was not affected by 5HD, HMR-1098 and glibenclamide caused a 5.2-fold and 40-fold rightward displacement of the diazoxide curve, respectively (p<0.05), suggesting that diazoxide activated pKATP to produce coronary vasodilation. A subvasodilator 0.5 μM diazoxide was selected for further studies.

Fig. 2

Concentration–response relationship of diazoxide for increase in coronary flow as affected by 100 μM 5HD, 25 μM HMR-1098 or 0.6 μM glibenclamide. As variance analysis revealed no inter-group difference between the control curves, they were pooled and fitted to a single curve (EC50 4.1±0.5 μM). Values are mean±S.E.M. of percent changes from the basal flow. n=5 for each inhibitor.

Fig. 2

Concentration–response relationship of diazoxide for increase in coronary flow as affected by 100 μM 5HD, 25 μM HMR-1098 or 0.6 μM glibenclamide. As variance analysis revealed no inter-group difference between the control curves, they were pooled and fitted to a single curve (EC50 4.1±0.5 μM). Values are mean±S.E.M. of percent changes from the basal flow. n=5 for each inhibitor.

### 3.2. Sham experiments

There were no differences in the baseline characteristics between the study groups (Tables 1 and 2). The characteristics were not affected by SOD. l-NMMA reduced coronary flow and LVDP by 30% and ACh response by 85%, while SNP response remained unaffected. In the presence of oxyHb, coronary flow was increased by 15%, LVDP remained unaffected, ACh caused a coronary vasoconstriction, and SNP response was reduced by 75%. All the effects of l-NMMA and oxyHb were reversible during 20-min washout (Tables 1 and 2). The values of coronary flow, LVDP, ACh response and SNP response were reduced by glibenclamide (by 22–38%) and HMR-1098 (12–22%), and were not affected by 5HD. There were no significant differences between the characteristics obtained at 25 min (in SOD-, l-NMMA-, and oxyHb-treated hearts) or at 40 min (untreated, 5HD-, glibenclamide-or HMR-1098-treated) and at the conclusion of the perfusion protocol, in any of the sham groups. Consequently, the normalized ACh and SNP responses were approximately 100%, in all sham groups (Fig. 3).

Table 1

Effects of various treatments on coronary flow, left ventricular developed pressure and total post-ischemic coronary flow in guinea-pig hearts subjected to 30 min ischemia and 30 min reperfusion

Type of protocol Treatment N Coronary flow (ml/min/g wet wt)

LVDP (mm Hg)

Total post-ischemic coronary flow (ml/5min/g wet wt)
25 min 40 or 75 min 135 min 25 min 40 or 75 min 135 min
Untreated
1a/2a Sham 11 12.0±0.7 12.0±0.8 11.7±0.7 85.4±6.8 84.0±5.3 83.1±4.3
1b/2b IR 19 12.4±0.8 12.1±0.7 10.3±0.4a,b 86.2±4.1 85.1±3.2 50.2±5.5a,b 78±8.9
1a Sham+SOD 12.5±0.7 13.3±0.8 12.1 ±0.6 78.3±5.2 80.2±5.5 75.4±3.8
1b IR+SOD 13.5±0.8 13.2±0.9 12.1±0.4 85.3±6.7 82.3±4.5 52.5±5.6a,b 116±12.0c
1a Sham+l-NMMA 13.0±0.9 9.4±0.5d 11.7±0.8a 85.2±4.1 65.0±3.8d 76.5±4.3a
1b IR+l-NMMA 12.8±1.2 9.7±1.0d 11.2±1.3a 86.4±4.0 65.6±4.2d 51.4±4.6b 92±11.4c
1a Sham+oxyHb 12.5±1.5 14.7±2.6d 12.0±1.6a 84.7±4.6 75.9±4.9 80.3±3.8
1b IR+oxyHb 12.9±1.6 14.6±3.0d 11.6±1.6a 81.2±6.1 77.7±5.6 52.7±3.6a,b 124±13.2c
2c IPC 15 12.7±0.9 12.5±0.6 11.0±0.6a,b 81.3±4.5 81.2±4.6 49.5±4.1a,b 109±12.0c
2d Dx-PC 13.0±1.0 12.4±0.7 11.1±0.8b 80.0±4.0 78.3±3.8 45.4±4.8a,b 109±11.8c
2e Dx-R 10 12.8±1.1 12.5±1.0 11.8±0.9 78.9±4.8 77.6±5.0 50.3±5.2a,b 105±10.6c

5HD-Continuous
2a Sham 10 11.3±0.6 11.6±0.6 11.0±0.6 83.1±5.1 80.0±5.5 77.1±5.9
2b IR 12 11.0±0.8 10.7±0.7 9.6±0.4a,b 85.4±5.1 84.6±4.6 50.1±5.1a,b 80±12.1
2c IPC 10 11.5±0.8 12.0±0.9 10.3±0.5a,b 80.2±6.7 80.1±4.8 46.5±5.5a,b 92±10.7
2d Dx-PC 11.4±0.9 11.5±0.6 10.4±0.5a,b 86.3±5.5 84.8±6.1 49.9±5.7a,b 96±10.7

5HD-Early
3a Sham 13.2±1.1 13.4±1.1 12.1±1.1 81.2±5.6 79.4±4.4 78.0±5.5
3b IR 12.4±0.6 12.5±0.7 10.9±0.3a,b 81.2±5.8 74.5±4.2 57.4±5.0a,b 84±11.2
3c IPC 11.3±0.3 11.2±1.0 8.8±0.8a,b 85.4±5.4 82.3±4.5 52.5±5.6a,b 95±10.2
3d Dx-PC 11.4±0.5 11.0±0.6 10.0±0.6b 86.3±4.8 84.3±6.5 49.4±4.5a,b 99±9.8

Glibenclamide
2a Sham 10 12.8±0.8 8.3±0.4d 7.9±0.4b 82.0±6.8 54.0±4.9d 52.1±4.6b
2b IR 10 13.1±0.8 8.1±0.5d 7.0±0.5a,b 85.1±7.9 55.6±3.3d 34.2±3.1a,b 56±9.1
2c IPC 12.3±0.7 7.7±0.5d 6.9±0.4b 82.3±8.0 51.2±3.8d 32.5±4.1a,b 59±8.1
2d Dx-PC 10 13.8±0.8 8.5±0.6d 7.6±0.6a,b 88.4±9.1 52.2±7.4d 31.2±5.6a,b 64±7.4

HMR-1098
2a Sham 12.0±1.0 9.7±0.7d 9.6±0.7b 81.2±5.8 71.5±4.2d 69.4±5.0b
2b IR 12.9±0.7 9.9±0.3d 8.6±0.3a,b 80.2±5.6 72.5±5.3d 42.5±4.6a,b 70±10.0
2c IPC 13.0±0.8 10.1±0.4d 8.9±0.4a,b 82.3±5.5 69.6±5.1d 39.8±5.4a,b 104±9.0c
2d Dx-PC 12.7±0.6 10.2±0.6d 8.7±0.6a,b 82.3±6.6 72.3±5.4d 41.7±5.2a,b 100±10.0c
Type of protocol Treatment N Coronary flow (ml/min/g wet wt)

LVDP (mm Hg)

Total post-ischemic coronary flow (ml/5min/g wet wt)
25 min 40 or 75 min 135 min 25 min 40 or 75 min 135 min
Untreated
1a/2a Sham 11 12.0±0.7 12.0±0.8 11.7±0.7 85.4±6.8 84.0±5.3 83.1±4.3
1b/2b IR 19 12.4±0.8 12.1±0.7 10.3±0.4a,b 86.2±4.1 85.1±3.2 50.2±5.5a,b 78±8.9
1a Sham+SOD 12.5±0.7 13.3±0.8 12.1 ±0.6 78.3±5.2 80.2±5.5 75.4±3.8
1b IR+SOD 13.5±0.8 13.2±0.9 12.1±0.4 85.3±6.7 82.3±4.5 52.5±5.6a,b 116±12.0c
1a Sham+l-NMMA 13.0±0.9 9.4±0.5d 11.7±0.8a 85.2±4.1 65.0±3.8d 76.5±4.3a
1b IR+l-NMMA 12.8±1.2 9.7±1.0d 11.2±1.3a 86.4±4.0 65.6±4.2d 51.4±4.6b 92±11.4c
1a Sham+oxyHb 12.5±1.5 14.7±2.6d 12.0±1.6a 84.7±4.6 75.9±4.9 80.3±3.8
1b IR+oxyHb 12.9±1.6 14.6±3.0d 11.6±1.6a 81.2±6.1 77.7±5.6 52.7±3.6a,b 124±13.2c
2c IPC 15 12.7±0.9 12.5±0.6 11.0±0.6a,b 81.3±4.5 81.2±4.6 49.5±4.1a,b 109±12.0c
2d Dx-PC 13.0±1.0 12.4±0.7 11.1±0.8b 80.0±4.0 78.3±3.8 45.4±4.8a,b 109±11.8c
2e Dx-R 10 12.8±1.1 12.5±1.0 11.8±0.9 78.9±4.8 77.6±5.0 50.3±5.2a,b 105±10.6c

5HD-Continuous
2a Sham 10 11.3±0.6 11.6±0.6 11.0±0.6 83.1±5.1 80.0±5.5 77.1±5.9
2b IR 12 11.0±0.8 10.7±0.7 9.6±0.4a,b 85.4±5.1 84.6±4.6 50.1±5.1a,b 80±12.1
2c IPC 10 11.5±0.8 12.0±0.9 10.3±0.5a,b 80.2±6.7 80.1±4.8 46.5±5.5a,b 92±10.7
2d Dx-PC 11.4±0.9 11.5±0.6 10.4±0.5a,b 86.3±5.5 84.8±6.1 49.9±5.7a,b 96±10.7

5HD-Early
3a Sham 13.2±1.1 13.4±1.1 12.1±1.1 81.2±5.6 79.4±4.4 78.0±5.5
3b IR 12.4±0.6 12.5±0.7 10.9±0.3a,b 81.2±5.8 74.5±4.2 57.4±5.0a,b 84±11.2
3c IPC 11.3±0.3 11.2±1.0 8.8±0.8a,b 85.4±5.4 82.3±4.5 52.5±5.6a,b 95±10.2
3d Dx-PC 11.4±0.5 11.0±0.6 10.0±0.6b 86.3±4.8 84.3±6.5 49.4±4.5a,b 99±9.8

Glibenclamide
2a Sham 10 12.8±0.8 8.3±0.4d 7.9±0.4b 82.0±6.8 54.0±4.9d 52.1±4.6b
2b IR 10 13.1±0.8 8.1±0.5d 7.0±0.5a,b 85.1±7.9 55.6±3.3d 34.2±3.1a,b 56±9.1
2c IPC 12.3±0.7 7.7±0.5d 6.9±0.4b 82.3±8.0 51.2±3.8d 32.5±4.1a,b 59±8.1
2d Dx-PC 10 13.8±0.8 8.5±0.6d 7.6±0.6a,b 88.4±9.1 52.2±7.4d 31.2±5.6a,b 64±7.4

HMR-1098
2a Sham 12.0±1.0 9.7±0.7d 9.6±0.7b 81.2±5.8 71.5±4.2d 69.4±5.0b
2b IR 12.9±0.7 9.9±0.3d 8.6±0.3a,b 80.2±5.6 72.5±5.3d 42.5±4.6a,b 70±10.0
2c IPC 13.0±0.8 10.1±0.4d 8.9±0.4a,b 82.3±5.5 69.6±5.1d 39.8±5.4a,b 104±9.0c
2d Dx-PC 12.7±0.6 10.2±0.6d 8.7±0.6a,b 82.3±6.6 72.3±5.4d 41.7±5.2a,b 100±10.0c

Values are means±S.E.M. N, number of hearts. LVDP, left ventricular developed pressure; IR, ischemia/reperfusion; IPC, ischemic preconditioning; Dx-PC, diazoxide preconditioning; Dx-R, diazoxide given at the reperfusion.

Baseline measurements were taken at 25 min of each study protocol (the number of study protocol is given in the first column, compare Fig. 1). The effect of a studied drug was measured either at 40 min (5HD, glibenclamide or HMR-1098) or at 75 min (SOD, l-NMMA, oxyHb). Then the remaining elements of the protocol followed and the measurements were repeated at 135 min.

a

p<0.05, 135 min vs. 40 or 75 min.

b

p<0.05, 135 min vs. 25 min.

c

p<0.05 vs. the respective IR group.

d

p<0.05, 40 or 75 min vs. 25 min.

Table 2

Effect of various interventions on coronary flow response to acetylcholine and sodium nitroprusside in guinea-pig hearts subjected to 30 min ischemia and 30 min reperfusion

Type of protocol Treatment N/n Acetylcholine response (ml/g wet wt)

Sodium nitroprusside response (ml/g wet wt)

25 min 40 or 75 min 135 min 25 min 40 or 75 min 135 min
Untreated
1a/2a Sham 6/5 4.1±0.5 3.9±0.5 3.8±0.5 4.1±0.5 4.2±0.4 4.0±0.4
1b/2b IR 13/6  4.9±0.4 1.8±0.1a  4.6±0.2 4.1±0.2
1a Sham+SOD 3/2 3.9±0.5 4.0±0.6 3.8±0.5 3.8±0.8 4.0±0.7 4.0±0.8
1b IR+SOD 6/3 4.1±0.4  3.0±0.3b 4.2±0.5  4.1±0.4
1a Sham+l-NMMA 5/2 4.6±0.5 0.7±0.3c 3.8±0.6a 3.9±0.5 4.3±0.4 4.0±0.5
1b IR+l-NMMA 6/3 4.3±0.4  2.9±0.3b 4.2±0.2  4.1±0.5
1a Sham+oxyHb 5/2 5.2±0.2 -1.1±0.3c 5.2±0.3a 4.1±0.3 1.1±0.3c 4.0±0.4a
1b IR+oxyHb 6/3 5.2±0.3  4.2±0.3b 4.6±0.4  4.5±0.3
2c IPC 11/4  4.3±0.3 3.7±0.2  4.4±0.5 3.9±0.4
2d Dx-PC 6/3  5.1±0.3 4.7±0.2  4.5±0.5 4.4±0.4
2e Dx-R 6/4  4.9±0.3 1.3±0.3a  4.2±0.4 4.0±0.3

5HD-continuous
2a Sham 6/4 4.6±0.3 4.6±0.4 4.4±0.4 3.5±0.5 3.3±0.4 3.2±0.3
2b IR 7/5  4.3±0.3 1.0±0.2a  4.1±0.2 3.9±0.3
2c IPC 6/4  4.4±0.4 1.1±0.3a  4.0±0.3 3.8±0.1
2d Dx-PC 6/3  4.2±0.4 1.2±0.4a  3.9±0.4 3.7±0.4

5HD-early
3a Sham 4.1±0.4 4.0±0.5 3.9±0.4
3b IR  3.9±0.3 1.0±0.3a
3c IPC  4.1±0.4 1.2±0.1a
3d Dx-PC  3.7±0.3 1.1±0.2a

Glibenclamide
2a Sham 6/4 4.1±0.40 2.9±0.4c 2.7±0.3 4.7±0.3 3.2±0.4c 3.2±0.3
2b IR 6/4  2.5±0.3 0.9±0.2a  3.5±0.5 3.2±0.5
2c IPC 6/3  2.9±0.3 1.0±0.1a  3.4±0.2 3.2±0.3
2d Dx-PC 6/4  3.0±0.3 1.1±0.2a  3.3±0.2 3.1±0.3

HMR-1098
2a Sham 6/3 4.7±0.4 3.4±0.4c 3.3±0.4 4.5±0.6 3.7±0.5c 3.6±0.4
2b IR 6/3  3.1±0.4 1.4±0.2a  3.2±0.4 3.0±0.4
2c IPC 6/3  3.8±0.5 3.5±0.3a  3.3±0.4 3.2±0.3
2d Dx-PC 6/3  4.0±0.3 3.6±0.3a  3.2±0.4 3.0±0.4
Type of protocol Treatment N/n Acetylcholine response (ml/g wet wt)

Sodium nitroprusside response (ml/g wet wt)

25 min 40 or 75 min 135 min 25 min 40 or 75 min 135 min
Untreated
1a/2a Sham 6/5 4.1±0.5 3.9±0.5 3.8±0.5 4.1±0.5 4.2±0.4 4.0±0.4
1b/2b IR 13/6  4.9±0.4 1.8±0.1a  4.6±0.2 4.1±0.2
1a Sham+SOD 3/2 3.9±0.5 4.0±0.6 3.8±0.5 3.8±0.8 4.0±0.7 4.0±0.8
1b IR+SOD 6/3 4.1±0.4  3.0±0.3b 4.2±0.5  4.1±0.4
1a Sham+l-NMMA 5/2 4.6±0.5 0.7±0.3c 3.8±0.6a 3.9±0.5 4.3±0.4 4.0±0.5
1b IR+l-NMMA 6/3 4.3±0.4  2.9±0.3b 4.2±0.2  4.1±0.5
1a Sham+oxyHb 5/2 5.2±0.2 -1.1±0.3c 5.2±0.3a 4.1±0.3 1.1±0.3c 4.0±0.4a
1b IR+oxyHb 6/3 5.2±0.3  4.2±0.3b 4.6±0.4  4.5±0.3
2c IPC 11/4  4.3±0.3 3.7±0.2  4.4±0.5 3.9±0.4
2d Dx-PC 6/3  5.1±0.3 4.7±0.2  4.5±0.5 4.4±0.4
2e Dx-R 6/4  4.9±0.3 1.3±0.3a  4.2±0.4 4.0±0.3

5HD-continuous
2a Sham 6/4 4.6±0.3 4.6±0.4 4.4±0.4 3.5±0.5 3.3±0.4 3.2±0.3
2b IR 7/5  4.3±0.3 1.0±0.2a  4.1±0.2 3.9±0.3
2c IPC 6/4  4.4±0.4 1.1±0.3a  4.0±0.3 3.8±0.1
2d Dx-PC 6/3  4.2±0.4 1.2±0.4a  3.9±0.4 3.7±0.4

5HD-early
3a Sham 4.1±0.4 4.0±0.5 3.9±0.4
3b IR  3.9±0.3 1.0±0.3a
3c IPC  4.1±0.4 1.2±0.1a
3d Dx-PC  3.7±0.3 1.1±0.2a

Glibenclamide
2a Sham 6/4 4.1±0.40 2.9±0.4c 2.7±0.3 4.7±0.3 3.2±0.4c 3.2±0.3
2b IR 6/4  2.5±0.3 0.9±0.2a  3.5±0.5 3.2±0.5
2c IPC 6/3  2.9±0.3 1.0±0.1a  3.4±0.2 3.2±0.3
2d Dx-PC 6/4  3.0±0.3 1.1±0.2a  3.3±0.2 3.1±0.3

HMR-1098
2a Sham 6/3 4.7±0.4 3.4±0.4c 3.3±0.4 4.5±0.6 3.7±0.5c 3.6±0.4
2b IR 6/3  3.1±0.4 1.4±0.2a  3.2±0.4 3.0±0.4
2c IPC 6/3  3.8±0.5 3.5±0.3a  3.3±0.4 3.2±0.3
2d Dx-PC 6/3  4.0±0.3 3.6±0.3a  3.2±0.4 3.0±0.4

Values are means±S.E.M. N/n, number of hearts in which vasodilator response to acetylcholine and sodium nitroprusside were evaluated, respectively. IR, ischemia/reperfusion; IPC, ischemic preconditioning; Dx-PC, diazoxide preconditioning; Dx-R, diazoxide given at the reperfusion.

Baseline measurements were taken at 25 min of each study protocol (the number of study protocol is given in the first column, compare Fig. 1). The effect of the studied drug was measured either at 40 min (5HD, glibenclamide or HMR-1098) or at 75 min (SOD, l-NMMA or OxyHb). Then the remaining elements of the protocol followed and the measurements were repeated at 135 min (see asterisks in Fig. 1).

a

p<0.05, 135 min vs. 40 or 75 min.

b

p<0.05; 135 min vs. 25 min.

c

p<0.05, 40 or 75 min vs. 25 min.

### 3.3. Endothelial dysfunction

The normalized ACh response was reduced by approximately 60% in the untreated IR group and in IR hearts perfused with 5HD, glibenclamide and HMR-1098 (Fig. 3a). Actually, the ACh response tended to be reduced more in 5HD-treated then in the untreated IR group (by 77% vs. 62%, p>0.05). The impairment of ACh response was partially and similarly prevented by SOD, l-NMMA, and oxyHb, was completely prevented by IPC and diazoxide-PC, and tended to be aggravated by diazoxide given at the reperfusion (p>0.05). The protection by IPC and diazoxide-PC was not affected by HMR-1098, and it was blocked by glibenclamide and 5HD, no matter whether 5HD was given as the continuous or early treatment. However, hyperaemic coronary flow responses associated with IPC were not affected by 5HD and were almost completely abolished by HMR-1098 and glibenclamide. Preconditioning pulses of diazoxide caused no change in coronary flow (Fig. 4).

Fig. 4

Coronary flow changes accompanying three 5-min episodes of the preconditioning ischemia and reperfusion (a) and three 5-min preconditioning pulses of diazoxide (b) in the untreated hearts and in those given 100 μM 5HD, 25 μM HMR-1098 or 0.6 μM glibenclamide prior to the preconditioning (arrow). Filled bars on the top mark time-intervals of the preconditioning ischemias and diazoxide pulses. Values are mean±S.E.M. See Table 1 for N in each group. *p<0.05 vs. untreated group.

Fig. 4

Coronary flow changes accompanying three 5-min episodes of the preconditioning ischemia and reperfusion (a) and three 5-min preconditioning pulses of diazoxide (b) in the untreated hearts and in those given 100 μM 5HD, 25 μM HMR-1098 or 0.6 μM glibenclamide prior to the preconditioning (arrow). Filled bars on the top mark time-intervals of the preconditioning ischemias and diazoxide pulses. Values are mean±S.E.M. See Table 1 for N in each group. *p<0.05 vs. untreated group.

SNP responses were comparable in all sham and IR groups (Fig. 3b).

### O2− and NO production

A burst of reduced cytochrome c and NOx outflow occurred in all groups upon the reperfusion following 30-min ischemia, suggesting increased rate of O2 and NO production with IR. As exemplified in Fig. 5, the outflows and post-ischemic coronary reflow peaked at 40–60 s, and returned nearly to the pre-ischemic values within the following 5 min. For statistical comparisons, the total outflows, and the total coronary flow during the initial 5 min of the reperfusion were estimated.

Fig. 5

Time course of post-ischemic outflow of reduced cytochrome c (a), NOx outflow (b), and coronary flow (c) in the untreated IR hearts. Note differences in coronary reflow between the hearts perfused with KHB (○, NOx measurements) and KHB supplemented with cytochrome c (□, O2 measurements). Values are mean±S.E.M. of 6 (O2 measurement) to 12 experiments (NOx measurement).

Fig. 5

Time course of post-ischemic outflow of reduced cytochrome c (a), NOx outflow (b), and coronary flow (c) in the untreated IR hearts. Note differences in coronary reflow between the hearts perfused with KHB (○, NOx measurements) and KHB supplemented with cytochrome c (□, O2 measurements). Values are mean±S.E.M. of 6 (O2 measurement) to 12 experiments (NOx measurement).

SOD, IPC, and diazoxide-PC attenuated post-ischemic O2 outflow by 80–34%, and did not affect NOx outflow. l-NMMA reduced the NOx outflow by 80%, and did not affect the O2 outflow (Fig. 6). Diazoxide, given at the reperfusion, increased the O2 outflow by 66%, while the NO outflow remained unaffected. OxyHb actually increased NOx outflow by 40%, and in its presence the NOx outflow comprised more than 98% of nitrate (as compared to 51% in untreated IR group, 16–18% in IPC, diazoxide-PC and SOD groups, and 5% in sham-perfused groups, Fig. 6b), the results compatible with the notion that oxyHb rapidly binds to and inactivates NO to nitrate [28,29]. OxyHb interfered with the cytochrome c assay precluding O2 measurement.

Fig. 6

Total post-ischemic outflow of reduced cytochrome c (a), and NOx (b) in the hearts (from left to right): untreated, given 100 μM 5HD as the continuous (5HD-C) or the early treatment (5HD-E) or continuously treated with 0.6 μM glibenclamide or 25 μM HMR-1098. Open, filled, cross-lined, hatched-rising to left, hatched-rising to right, crosshatched, horizontally lined, and vertically lined columns represent: sham, 30-min IR, IR+SOD, IR+l-NMMA, IR+oxyHb, ischemic preconditioning (IPC), diazoxide preconditioning (Dx-PC), and diazoxide given at the reperfusion (Dx-R), respectively. Open and filled parts of the columns in (c) represent nitrite and nitrate outflow, respectively. Values are mean±S.E.M. of 6 (O2 measurement) to 12 experiments (NOx measurement). *p<0.05 vs. respective sham; #p<0.05 vs. respective IR; $p<0.05 vs. untreated IR. Fig. 6 Total post-ischemic outflow of reduced cytochrome c (a), and NOx (b) in the hearts (from left to right): untreated, given 100 μM 5HD as the continuous (5HD-C) or the early treatment (5HD-E) or continuously treated with 0.6 μM glibenclamide or 25 μM HMR-1098. Open, filled, cross-lined, hatched-rising to left, hatched-rising to right, crosshatched, horizontally lined, and vertically lined columns represent: sham, 30-min IR, IR+SOD, IR+l-NMMA, IR+oxyHb, ischemic preconditioning (IPC), diazoxide preconditioning (Dx-PC), and diazoxide given at the reperfusion (Dx-R), respectively. Open and filled parts of the columns in (c) represent nitrite and nitrate outflow, respectively. Values are mean±S.E.M. of 6 (O2 measurement) to 12 experiments (NOx measurement). *p<0.05 vs. respective sham; #p<0.05 vs. respective IR;$p<0.05 vs. untreated IR.

5HD and glibenclamide did not affect pre-ischemic O2 outflow and they exaggerated post-ischemic O2 burst by 78% and 58%, respectively (Fig. 6a). In 5HD- and glibenclamide-treated IR groups, IPC and diazoxide-PC attenuated O2 outflows to a level similar to that noted in the untreated IR group. 5HD produced virtually the same effects, was it applied as the continuous or early treatment. 5HD did not affect the NOx outflow (Fig. 6b). HMR-1098 affected the O2 burst neither in IR nor IPC group.

### 3.5. Recoveries of hemodynamic functions

Most of the hearts subjected to IR developed short-lasting (ca. 10–20 s) ventricular fibrillation upon the reperfusion. In five hearts the fibrillation was sustained (one in the untreated, two in 5HD-, one in glibenclamide-, and one in diazoxide-after-ischemia-treated heart) and they were excluded from further evaluation. As a consequence of different treatments, there were significant differences in the pre-ischemic values for coronary flow and LVDP between the experimental groups (Table 1). Nevertheless, the percent post-ischemic recoveries of coronary flow and LVDP did not differ between any experimental groups and amounted to approximately 90% and 60% of the pre-ischemic values, respectively (Table 1).

The only difference between IR groups was that the coronary reflow during the initial 5 min of the reperfusion following the test ischemia was significantly greater in all groups in which the tested intervention afforded endothelial protection (Table 1). This was, however, true only for the groups perfused with the standard KHB. As exemplified in Fig. 5c, the time-course of the coronary reflow was prolonged and its amplitude was increased in the hearts perfused with cytochrome c vs. KHB. There were no significant differences in the coronary reflow between any experimental groups perfused with cytochrome c (not shown).

## 4. Discussion

This study demonstrates that: (i) anti-O2 and anti-NO interventions afforded equally potent endothelial protection; (ii) IPC attenuated the O2, but not NO production, and protected the endothelium; (iii) diazoxide given before, but not after, the test ischemia mimicked the effects of IPC; (iv) the effects of IPC and diazoxide-PC were abolished by the early treatment with 5HD, and not affected by HMR-1098. These results implicate that a product of the reaction between O2 and NO mediated the endothelial dysfunction, that IPC and diazoxide-PC afforded the protection by attenuating the O2 generation, and that the opening of mKATP served as a trigger of the IPC-and diazoxide-induced protection.

### 4.1. Post-ischemic endothelial dysfunction

The test with ACh performed here served as an index of agonist-induced endothelium dependent vascular function. Previously, we have verified that the impairment of the ACh induced vasodilation correlated with the impairment of a basal endothelium-dependent vasodilator tone, the disruption of the endothelium glycocalyx, and the impairment of ACh induced NO production [4,13], implicating that the test with ACh is a reliable measure of the endothelial injury. In this study, IR impaired ACh response whilst coronary smooth muscle function, as probed with SNP, remained intact, indicating a selective endothelial dysfunction. SOD, l-NMMA, oxyHb, IPC and diazoxide-PC prevented this dysfunction. As discussed before [4,13], an important feature of our model is that it allows dissociation of the endothelial and cardiomyocyte injury. One evidence would be that none of the interventions found to protect the endothelium in this and our earlier studies [6,13], affected the post-ischemic hemodynamic recoveries. This implies that the endothelial protection reported here involved a mechanism intrinsic to the vasculature.

### 4.2. End-effector of the protection

Consistent with previous reports, IR caused a burst of O2[32,33] and NO generation [23,34,35] also in our model. This was evidenced by the increased cardiac outflow of the reduced cytochrome c and NOx, respectively. SOD and l-NMMA inhibited these outflows, indicating that they were specific assays for O2 and NO, respectively.

The O2 scavenger SOD, the NOS inhibitor l-NMMA, and the specific NO vs. O2 scavenger oxyHb (the reported rate constants for NO-oxyHb vs. O2 oxyHb reaction are 3.7×107 and 4×103 M−1 s−1, respectively [29,36]), all afforded similar protection, suggesting that neither O2 nor NO alone but rather a product of their reaction (e.g. peroxynitrite) mediated the endothelial dysfunction [7]. Actually, the blockade of either O2 or NOS has been demonstrated to attenuate peroxynitrite formation and injury in IR rat heart [34,37].

Consistent with previous reports [20–22], IPC and diazoxide-PC attenuated post-ischemic O2 formation also in our system. Our results, however, did not confirm those of Csonka et al. [23] who reported classic IPC to protect IR rat heart by the blockade of myocardial NO accumulation. Although the post-ischemic NO burst appeared to be harmful to the endothelium, as evidenced by our l-NMMA and oxyHb studies, it was not affected by IPC and diazoxide-PC. It is uncertain whether it is differences between the experimental models (rat vs. guinea-pig heart) or between NO assays (in the myocardium by EPR vs. in the effluent by chemiluminescence), which explains these discrepant results.

Mitochondria, xanthine oxidase, a nonphagocytic NADPH oxidase, and/or NOS may be considered as likely sources of the O2 mediating the endothelial dysfunction in our crystalloid-perfused heart preparation. The involvement of NOS can be excluded, as the cardiac O2 outflow was not affected by l-NMMA. Previously, we have shown that it is O2 generated in some endothelin- and protein kinase C-activated process that mediated the endothelial dysfunction in our model [6]. Interestingly, the NADPH oxidase, rather than NOS, xanthine oxidase or the mitochondrial O2 production is activated by endothelin and protein kinase C [38,39]. Furthermore, IPC and diazoxide were reported to attenuate O2 generation by mitochondria isolated from IR heart [40], and by mitochondria subjected to anoxia/reoxygenation [41], respectively, suggesting mitochondrial O2 as a possible end-effector of the protection.

Altogether, our data support the hypothesis that IPC and diazoxide-PC protected the endothelium by attenuating the post-ischemic O2 generation and thus the generation of a harmful product of the reaction between O2 and NO. An incomplete inhibition of this generation by SOD, l-NMMA and oxyHb might explain why these agents afforded the partial protection, in contrast to the full protection by IPC and diazoxide-PC. Alternatively, some mechanism beyond the radicals accounted for IPC and diazoxide-PC protection.

### KATP and the protection

In the heart, KATP are expressed in plasmalemma of cardiomyocytes, coronary smooth muscle cells [42], and endothelial cells [43], and in mitochondria of all these types of cells. As already discussed, the endothelial protection probably involves a mechanism intrinsic to the vasculature. The problem with the non-cardiomyocyte cell types, such as studied here, is that the specificity of available blockers and inhibitors for mKATP vs. pKATP has not been determined.

The following evidence suggests that the activation of mKATP rather than pKATP accounted for the protection afforded by 0.5 μM diazoxide used in this study. The hyperpolarization of coronary smooth muscle by opening of pKATP results in vasodilation, and endothelial pKATP might contribute to the vasodilation via NO release [42,44]. Diazoxide caused coronary vasodilation, which was inhibited by the putative pKATP blocker, HMR-1098, and the nonselective KATP blocker, glibenclamide, and was not affected by the putative mKATP blocker, 5HD, implicating that diazoxide opens vascular pKATP in our model. Consequently, a subvasodilator 0.5 μM diazoxide was selected for this study. Although this concentration was similar to the reported K1/2 of 0.4–0.8 μM for the diazoxide-induced activation of mKATP[24], its selectivity for vascular mKATP vs. pKATP appeared quite low in our preparation (Fig. 2). Its selectivity for cardiomiocyte mKATP was probably high, as diazoxide opens cardiomiocyte pKATP with K1/2 of 850 μM [24].

The effects of 0.5 μM diazoxide-PC on post-ischemic endothelium and O2 production were blocked by 5HD and glibenclamide, and were not affected by HMR-1098. This was true although, in contrast to 5HD, glibenclamide and HMR-1098 blocked vascular pKATP, as evidenced by their inhibition of diazoxide-induced vasodilation (Fig. 2) and hyperemic coronary flow responses associated with IPC (Fig. 4). These results support the notion that 5HD and glibenclamide blocked diazoxide-PC protection because they blocked mKATP rather than pKATP, and that the opening of mKATP (in an undefined type of cells) rather than pKATP in the coronaries and/or cardiomiocytes accounted for the diazoxide-PC.

Unexpectedly, 5HD and glibenclamide appeared to increase O2 generation, and tended to aggravate endothelial dysfunction in the nonpreconditioned IR hearts over those seen in the untreated IR group. The explanation most compatible with the present data is that mKATP opening is beneficial to IR endothelium and that mKATP opens naturally at basal conditions. Actually, glibenclamide has been reported to aggravate myocardial IR injury [45,46]. This tonic protection by mKATP would be blocked by 5HD, potentiated by IPC and diazoxide-PC, and these opposing effects would cancel each other in the 5HD- and glibenclamide-treated and preconditioned hearts. In fact, the O2 production and endothelial function did not differ between untreated IR group and the groups perfused with 5HD or glibenclamide and preconditioned.

Another possibility is that 5HD and glibenclamide aggravated the effects of IR in some mKATP-independent manner and, in this way, have simply negated the beneficial effects of preconditioning, giving the impression that these effects were blocked. This interpretation seems unlikely because 5HD and glibenclamide exerted similar effects although they differ in their chemical structure, KATP selectivity, and the effect on basal coronary flow and hyperemic coronary flow responses associated with IPC. Furthermore, although HMR-1098 and glibenclamide similarly attenuated diazoxide-induced coronary vasodilation, the basal coronary flow, and the hyperemic coronary flow response associated with IPC, it was glibenclamide, but not HMR-1098, that aggravated the O2 generation, and prevented the protection by IPC and diazoxide-PC, suggesting the later effects are related to glibenclamide mKATP rather than pKATP inhibition. Finally, while the O2 generation was increased by 5HD and glibenclamide it was attenuated by diazoxide-PC, implicating that all three agents affected the O2 generation by acting on the same target, i.e., the mKATP.

Altogether, several arguments support the notion that the effects of diazoxide and 5HD reported here were related to their action on mKATP. Diazoxide-PC, but not diazoxide at the reperfusion, fully mimicked the effects of IPC, and the early treatment with 5HD was equally effective in preventing the effects of IPC and diazoxide-PC as the continuous 5HD treatment. Given the fact that the effects of diazoxide and 5HD are readily reversible upon washout [25], these data suggest that the mKATP opening in response to IPC plays a vital role to protect coronary endothelium against IR injury, and that mKATP serves as a trigger of the protection. Our results support the concept that mKATP opening prior to IR protects the endothelium by attenuating the post-ischemic O2 burst. Neither the mechanistic link between these two processes nor the source of O2 mediating the endothelial dysfunction has been addressed experimentally in this study.

## Acknowledgments

This study was supported by the KBN Grant 4 PO5A 015 15 and the CMKP Grant 501-2-1-05-65/01.

## References

[1]
Mehta
J.L
Nichols
W.W
Donnelly
W.H
Lawson
D.L
Thompson
L
ter Riet
M
et al
Protection by superoxide dismutase from myocardial dysfunction and attenuation of vasodilator reserve after coronary occlusion and reperfusion in dog
Circ. Res.

1989
65
1283
1295
[2]
Tsao
P.S
Aoki
N
Lefer
D.J
Johnson
G.I
Lefer
A.M
Time-course of endothelial dysfunction and myocardial injury during ischemia and reperfusion in the cat
Circulation

1990
82
1402
1412
[3]
Gross
G.J
O'Rourke
S.T
Pelc
L.R
Warltier
D.C
Myocardial and endothelial dysfunction after multiple, brief coronary occlusions—role of oxygen radicals
Am. J. Physiol.

1992
263
H1703
H1709
[4]
Beresewicz
A
Czarnowska
E
Maczewski
M
Ischemic preconditioning and superoxide dismutase protect against endothelial dysfunction and endothelium glycocalyx disruption in the postischemic guinea-pig hearts
Mol. Cell. Biochem.

1998
186
87
92
[5]
Tsao
P.S
Lefer
A.M
Time-course and mechanism of endothelial dysfunction in isolated ischemic-and hypoxic-perfused rat hearts
Am. J. Physiol.

1990
259
H1660
H1666
[6]
Maczewski
M
Beresewicz
A
The role of endothelin, protein kinase C, and free radicals in the mechanism of the post-ischemic endothelial dysfunction in guinea-pig hearts
J. Mol. Cell. Cardiol.

2000
32
297
310
[7]
Beckman
J.S
Koppenol
W.H
Nitric oxide, superoxide, and peroxynitrite: the good, the bad, and the ugly
Am. J. Physiol.

1996
40
C1424
C1437
[8]
Murry
C.E
Jennings
R.B
Reimer
K.A
Preconditioning with ischemia: a delay of lethal cell injury in ischemic myocardium
Circulation

1986
74
1124
1136
[9]
Przyklenk
K
Kloner
R.A
Prog. Cardiovasc. Dis.

1998
40
517
547
[10]
Cohen
M.V
Baines
C.P
Downey
J.M
Ischemic preconditioning: from adenosine receptor to K-ATP channel
Annu. Rev. Physiol.

2000
62
79
109
[11]
DeFily
D.V
Chilian
W.M
Preconditioning protects coronary arteriolar endothelium from ischemia-reperfusion injury
Am. J. Physiol.

1993
265
H700
H706
[12]
Richard
V
Kaeffer
N
Tron
C
Thuillez
C
Ischemic preconditioning protects against coronary endothelial dysfunction induced by ischemia and reperfusion
Circulation

1994
89
1254
1261
[13]
Maczewski
M
Beresewicz
A
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.

1998
30
1735
1747
[14]
Kharbanda
R.K
Peters
M
Walton
B
Kattenhorn
M
Mullen
M
Klein
N
et al
Ischemic preconditioning prevents endothelial injury and systemic neutrophil activation during ischemia-reperfusion in humans in vivo
Circulation

2001
103
1624
1630
[15]
Rubino
A
Yellon
D.M
Ischemic preconditioning of the vasculature: an overlooked phenomenon for protecting the heart?
TIPS

2000
21
225
230
[16]
Gross
G.J
Fryer
R.M
Sarcolemmal versus mitochondrial ATP-sensitive K+channels and myocardial preconditioning
Circ. Res.

1999
84
973
979
[17]
O'Rourke
B
Myocardial KATP channels in preconditioning
Circ. Res.

2000
87
845
855
[18]
Oldenburg
O
Cohen
M.V
Yellon
D.M
Downey
J.M
Mitochondrial KATP channels: role in cardioprotection
Cardiovasc. Res.

2002
55
429
437
[19]
Bouchard
J.F
Lamontagne
D
Mechanisms of protection afforded by preconditioning to endothelial function against ischemic injury
Am. J. Physiol.

1996
40
H1801
H1806
[20]
Tosaki
A
Cordis
G.A
Szerdahelyi
P
Engelman
R.M
Das
D.K
Effects of preconditioning on reperfusion arrhythmias, myocardial functions, formation of free radicals, and ion shifts in isolated ischemic/reperfused rat hearts
J. Cardiovasc. Pharmacol.

1994
23
365
373
[21]
Vanden Hoek
T
Becker
L.B
Shao
Z.H
Li
C.Q
Schumacker
P.T
Preconditioning in cardiomyocytes protects by attenuating oxidant stress at reperfusion
Circ. Res.

2000
86
541
548
[22]
Narayan
P
Mentzer
R.M
Jr.
Lasley
R.D
Adenosine A1 receptor activation reduces reactive oxygen species and attenuates stunning in ventricular myocytes
J. Mol. Cell. Cardiol.

2001
33
121
129
[23]
Csonka
C
Szilvassy
Z
Fulop
F
Pali
T
Blasig
I.E
Tosaki
A
et al
Classic preconditioning decreases the harmful accumulation of nitric oxide during ischemia and reperfusion in rat hearts
Circulation

1999
100
2260
2266
[24]
Garlid
K.D
Paucek
P
Yarovyarovoy
V
Sun
X.C
Schindler
P.A
The mitochondrial K-ATP channel as a receptor for potassium channel openers
J. Biol. Chem.

1996
271
8796
8799
[25]
Liu
Y.G
Sato
T
Orourke
B
Marban
E
Mitochondrial ATP-dependent potassium channels: novel effectors of cardioprotection?
Circulation

1998
97
2463
2469
[26]
Sato
T
O'Rourke
B
Marban
E
Modulation of mitochondrial ATP-dependent K+ channels by protein kinase C
Circ. Res.

1998
83
110
114
[27]
Gogelein
H
Hartung
J
Englert
H.C
Schoelkens
B.A
HMR 1883, a novel cardioselective inhibitor of the ATP-sensitive potassium channel: Part I. Effects on cardiomyocytes, coronary flow and pancreatic beta-cells
J. Pharmacol. Exp. Ther.

1998
286
1453
1464
[28]
Kelm
M
J
Control of coronary vascular tone by nitric oxide
Circ. Res.

1990
66
1561
1575
[29]
Feelisch
M
Kubitzek
D
Werringloer
J
The Oxyhemoglobin Assay
Feelisch
M
Stamler
J.S
Methods in nitric oxide research
1996
Chichester
Wiley
455
478
[30]
Richard
V
Blanc
T
Kaeffer
N
Tron
C
Thuillez
C
Myocardial and coronary endothelial protective effects of acetylcholine after myocardial ischaemia and reperfusion in rats: role of nitric oxide
Br. J. Pharmacol.

1995
115
1532
1538
[31]
Bilinska
M
Maczewski
M
Beresewicz
A
Donors of nitric oxide mimic effects of ischaemic preconditioning of reperfusion induced arrhythmias in isolated rat heart
Mol. Cell. Biochem.

1996
160/161
265
271
[32]
Southworth
R
Shattock
M.J
Hearse
D.J
Kelly
F.J
Developmental differences in superoxide production in isolated guinea-pig hearts during reperfusion
J. Mol. Cell. Cardiol.

1998
30
1391
1399
[33]
Zweier
J.L
Measurement of superoxide-derived free radicals in the reperfused heart. Evidence for a free radical mechanism of reperfusion injury
J. Biol. Chem.

1988
263
1353
1357
[34]
Wang
P.H
Zweier
J.L
Measurement of nitric oxide and peroxynitrite generation in the postischemic heart—evidence for peroxynitrite-mediated reperfusion injury
J. Biol. Chem.

1996
271
29223
29230
[35]
Maczewski
M
Beresewicz
A
Role of nitric oxide and free radicals in cardioprotection by blocking Na+–H+ and Na+–Ca2+ exchange in rat heart
Eur. J. Pharmacol.

2003
46
139
147
[36]
Sutton
H.C
Roberts
P.B
Winterbourn
C.C
The rate of reaction of superoxide radical ion with oxyhaemoblobin and methaemoglobin
Biochem. J.

1976
155
503
510
[37]
Yasmin
W
K.D
Schulz
R
Generation of peroxynitrite contributes to ischemia-reperfusion injury in isolated rat hearts
Cardiovasc. Res.

1997
33
422
432
[38]
Heitzer
T
Wenzel
U
Hink
U
Krollner
D
Skatchkov
M
Stahl
R.A
et al
Increased NAD(P)H oxidase-mediated superoxide production in renovascular hypertension: evidence for an involvement of protein kinase C
Kidney Int.

1999
55
252
260
[39]
Li
L
Fink
G.D
Watts
S.W
Northcott
C.A
Galligan
J.J
Pagano
P.J
et al
Endothelin-1 increases vascular superoxide via endothelin(A)-NADPH oxidase pathway in low-renin hypertension
Circulation

2003
107
1053
1058
[40]
Park
J.W
Chun
Y.S
Kim
Y.H
Kim
C.H
Kim
M.S
Ischemic preconditioning reduces O2 generation and prevents respiratory impairment in the mitochondria of post-ischemic reperfused heart of rat
Life Sci.

1997
60
2207
2219
[41]
Ozcan
C
Bienengraeber
M
Dzeja
P.P
Terzic
A
Potassium channel openers protect cardiac mitochondria by attenuating oxidant stress at reoxygenation
Am. J. Physiol.

2002
282
H531
H539
[42]
Nelson
M.T
Quayle
J.M
Physiological roles and properties of potassium channels in arterial smooth muscle
Am. J. Physiol.

1995
37
C799
C822
[43]
Katnik
C
D.J
An ATP-sensitive potassium conductance in rabbit arterial endothelial cells
J. Physiol.

1995
485
595
606
[44]
Herrera
G.M
Resta
T.C
Candelaria
J.J
Walker
B.R
Maintained vasodilatory response to cromakalim after inhibition of nitric oxide synthesis
J. Cardiovasc. Pharmacol.

1998
31
921
929
[45]
Thornton
J.D
Thornton
C.S
Sterling
D.L
Downey
J.M
Blockade of ATP-sensitive potassium channels increases infarct size but does not prevent preconditioning in rabbit hearts
Circ. Res.

1993
72
44
49
[46]
Munch-Ellingsen
J
Bugge
E
Ytrehus
K
Blockade of the K-ATP-channel by glibenclamide aggravates ischemic injury, and counteracts ischemic preconditioning
Basic Res. Cardiol.

1996
91
382
388

## Author notes

Time for primary review 30 days