INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ISCHEMIC PRECONDITIONING is an endogenous protective mechanism in which one or more brief periods of myocardial ischemia and reperfusion render the myocardium resistant to a subsequent more-sustained ischemic insult. Since preconditioning was first described by Murry et al. (18) in 1986, numerous mechanisms have been proposed to explain this phenomenon. There is a consensus among investigators that infarct size reduction provided by ischemic preconditioning is receptor mediated and initiated by stimulation of several G_i protein-coupled receptors during the brief preconditioning ischemia, including adenosine A₁/A₃, bradykinin B₂, and opioid -receptors (28). In addition, there are other G_i-coupled receptors on myocardial cells that are capable of triggering a preconditioned state, but agonists for them are either simply not released or are released in insufficient quantity during a preconditioning ischemia to contribute to protection. One such receptor is the muscarinic M₂ receptor (4, 29). It is thought that signals from all of these receptors eventually converge on protein kinase C (PKC), making the latter a critical intermediate component in the signal pathways leading to an as-yet-unknown end-effector that produces the protection of preconditioning. Furthermore, it can be shown that only a short pulse of receptor occupancy is needed before ischemia to produce protection. Receptor occupancy therefore triggers the protection. On the other hand, the critical time for PKC to act is during the ischemic insult, making it a mediator (34). It has been assumed that triggers reside upstream of mediators in the signal transduction pathway. Recently, we found that protection from a pulse of bradykinin, phenylephrine, morphine, or acetylcholine (ACh) could be blocked by either mitochondrial ATP-sensitive K⁺ (mitoK_ATP) channel blockers or free radical scavengers given during the trigger phase, whereas that from adenosine was unaffected by either (4). That suggested that the receptors for ACh and adenosine acted through very different signal transduction pathways. Tong et al. (31) recently reported that phosphatidylinositol 3-kinase (PI3-kinase) was also involved in preconditioning because wortmannin could block its protection, and they demonstrated that PI3-kinase was upstream of PKC. The first aim of the present study was to test whether PI3-kinase is involved in the protection induced by either adenosine or ACh, and, if so, whether this kinase acts as a trigger or a mediator.

Tyrosine kinases have long been demonstrated to be involved in the preconditioning signaling pathways (7, 16, 32). Baines et al. (2) found that the tyrosine kinase blocker genistein could prevent protection from ischemic preconditioning but only when introduced during the mediator phase. The second aim of this study was to test whether there is any involvement of a tyrosine kinase during the trigger phase, when protection is triggered by either adenosine or ACh. One tyrosine kinase of interest is the Src nonreceptor tyrosine kinase. It is thought to be an upstream component of the activation mechanism by which G_i proteins activate PI3-kinase through transactivation of receptor tyrosine kinases (5, 14). In a recent study, Mockridge et al. (17) found that the Src inhibitor 4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolol[3,4-d]pyrimidine (PP2) blocked the ability of ischemia to activate PI3-kinase. We therefore tested to see whether a Src nonreceptor tyrosine kinase might also be involved in the trigger phase of either of these two agonists.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Surgical procedures. All experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (20).

Infarct size measurement. At the end of the experiment, the coronary artery was reoccluded, and 1- to 10-µm green fluorescent microspheres (Duke Scientific; Palo Alto, CA) were infused into the perfusate to demarcate the ischemic zone as the area of tissue without fluorescence (region at risk). The heart was weighed, frozen, and then cut into 2-mm-thick slices. The slices were incubated in 1% triphenyltetrazolium chloride (TTC) in sodium phosphate buffer (pH 7.4) at 37°C for 20 min. TTC stains the noninfarcted myocardium brick red, indicating the presence of dehydrogenase enzymes in viable tissue. The slices were then immersed in 10% formalin to enhance the contrast between stained (viable) and unstained (necrotic) tissue. The risk zone was identified by illuminating the slices with ultraviolet light. The areas of infarct and risk zone were determined by planimetry of each slice, and volumes were calculated by multiplying each area by slice thickness and summing them for each heart. Infarct size was expressed as a percentage of the risk zone.

Chemicals. ACh, adenosine, diazoxide, genistein, and wortmannin were obtained from Sigma (St. Louis, MO). PP2 and lavendustin A were ordered from Calbiochem. Adenosine, diazoxide, wortmannin, PP2, lavendustin A, and genistein were dissolved in DMSO. Aliquots of these solutions were diluted in Krebs buffer immediately before the experiments (final DMSO concentration <0.05%). ACh was dissolved directly in the buffer.

Experimental protocols. In all experiments, infarcts were induced by 30 min of regional ischemia. Subsequently, all hearts experienced 120 min of reperfusion. Ischemia was confirmed by a decrease in left ventricular developed pressure and a reduction in coronary flow. Animals were divided into 17 groups (Fig. 1). After 20 min of equilibration, the control group experienced only ischemia and reperfusion as noted. The PC group was preconditioned with 5 min of global ischemia and 10 min of reperfusion before the 30-min regional ischemia. In the ACh group, ACh was included in the perfusate (0.55 mmol/l) for 5 min, starting 15 min before the 30-min regional ischemia. In the Aden group, adenosine (100 µmol/l) was perfused in lieu of ACh. In the ACh+Gen group, the tyrosine kinase inhibitor genistein (50 µmol/l) was present in the perfusate for 15 min, starting 20 min before the 30-min ischemia to bracket the ACh infusion. In the ACh+W(E₁) group, the PI3-kinase inhibitor wortmannin (100 nmol/l) was infused in lieu of genistein. In the Aden+Gen group, adenosine and genistein were coinfused as described above. In the ACh+W(E₂) group, wortmannin (100 nmol/l) infusion was prolonged to the start of the 30-min period of ischemia. This prolonged infusion of wortmannin (100 nmol/l) was combined with adenosine in the Aden+W(E₂) group. In the ACh+W(L) group, ACh treatment as outlined above was combined with wortmannin (100 nmol/l) infused for 35 min, starting 5 min before the 30-min ischemia. In the ACh+PP2(E) and Aden+PP2(E) groups, PP2 (1 µmol/l) was combined with either ACh or adenosine and infused from 5 min before to 5 min after the agonist. In the ACh+PP2(L) group, PP2 was added from 5 min before until the end of the index ischemia. In the Diaz+PP2(E) group, 10 µM diazoxide was infused for 5 min, starting 15 min before the 30-min regional ischemia. This latter infusion was bracketed by a PP2 infusion beginning 5 min before and extending until 5 min after diazoxide. In the ACh+Lav group, ACh was combined with 2 µmol/l lavendustin A infused from 5 min before to 5 min after the agonist. To determine whether the inhibitors had any direct effect on infarction, we tested both lavendustin A and PP2 at the above concentrations and schedules in otherwise untreated hearts (the Lav and PP2 groups). Both genistein and wortmannin have been tested previously in a similar animal model (3) and, therefore, were not tested again.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Experimental protocols. Timing of interventions is shown in relation to the 30-min period of regional ischemia. PC, ischemic preconditioning; ACh, acetylcholine (0.55 mM); Aden, adenosine (100 µM); Diaz, diazoxide (10 µM); Gen, genistein (50 µM); PP2, 4-amino-5(4-chlorophenyl)-7-(t-butyl)pyrazolol[3,4-d]pyrimidine (1 µM); Lav, lavendustin A (2 µM); W(E1), wortmannin (100 nM) infusion followed by a washout period of 5 min before the onset of ischemia; W(E2), preischemic infusion of wortmannin (100 nM) prolonged to the onset of ischemia; W(L) and PP2(L), wortmannin (100 nM) or PP2 (1 µM) infusion from 5 min before ischemia until the onset of reperfusion.

Statistics. All data are presented as means ± SE. One-way ANOVA with Tukey's post hoc test was performed on baseline hemodynamics and infarct measurements. ANOVA for repeated measures was used to test for differences in hemodynamics within any given group. A value of P < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Hemodynamic data. Table 1 summarizes the hemodynamic data for the isolated rabbit hearts. There were no significant differences in baseline values for heart rate, developed pressure, or coronary flow for any parameter among the experimental groups. Wortmannin caused a decrease in developed pressure with no significant change in heart rate or coronary flow.

Infarct size. Table 2 summarizes infarct size data. Five groups did have slightly smaller body weights and heart weights relative to the control group, but there were no significant differences in the risk zone size among the groups. In control hearts, infarct size was 34.3 ± 4.1% of the area at risk. A brief infusion of ACh (Fig. 2) or adenosine (Fig. 3) before ischemia significantly reduced the level of infarction to 9.3 ± 3.5% and 13.9 ± 3.3%, respectively (P < 0.002). Early administration of genistein during the trigger phase blocked the protective effect of ACh [36.2 ± 5.0% infarction, P = not significant (NS) vs. control]. Neither 15 min of wortmannin during the trigger phase [ACh+W(E₁)] nor 35 min of drug during the mediator phase [ACh+W(L)] affected the protective effect of ACh (8.1 ± 1.4% and 7.4 ± 2.1% infarction, respectively, P < 0.001 vs. control). However, when the early wortmannin schedule was extended to include drug up to the onset of ischemia [ACh+W(E₂)], protection from ACh was blocked (30.3 ± 5.4% infarction; Fig. 2). In contrast to the above results, neither genistein nor the extended early infusion of wortmannin blocked the protection of adenosine (P < 0.001 vs. control; Fig. 3). Thus adenosine, unlike ACh, triggers protection by a signal transduction pathway that uses neither PI3-kinase nor a tyrosine kinase.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Effect of ACh and genistein or wortmannin on infarct size expressed as a percentage of the risk zone. , Individual experiments; , group means ± SE. Brief infusion of ACh was protective. Genistein or prolonged preischemic infusion of wortmannin blocked the protection of ACh. However, bracketing ACh infusion with wortmannin followed by a washout period of 5 min before the onset of ischemia or confining wortmannin to the immediate preischemic period and continuing until the onset of reperfusion did not abolish the protection of ACh. See Fig. 1 for abbreviations. *P < 0.05 vs. control.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Effect of adenosine and genistein or wortmannin on infarct size expressed as a percentage of the risk zone. , Individual experiments; , group means ± SE. Brief infusion of adenosine was protective. Neither genistein nor prolonged preischemic infusion of wortmannin blocked the protection of adenosine. *P < 0.05 vs. control.

View larger version (10K): [in this window] [in a new window]   Fig. 4.   Effect of lavendustin A on infarct size expressed as a percentage of the risk zone in rabbit hearts protected by ACh. , Individual experiments; , group means ± SE. Brief infusion of ACh was protective. Bracketing the ACh infusion with lavendustin A followed by a washout period of 5 min before the onset of ischemia blocked the protection of ACh. Lavendustin A in the same protocol did not modify the infarct size in hearts not receiving ACh.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Effect of PP2 on infarct size expressed as a percentage of the risk zone in rabbit hearts protected by either ACh or adenosine. , Individual experiments; , group means ± SE. Brief infusion of ACh or adenosine was protective. Bracketing ACh infusion with PP2 followed by a washout period of 5 min before the onset of ischemia blocked the protection of ACh. However, bracketing adenosine or diazoxide infusion with PP2 or confining PP2 to the immediate preischemic or ischemic periods in ACh-treated hearts did not abolish protection. *P < 0.05 vs. control. PP2 alone did not modify the infarct size in control ischemic hearts.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study shows that, although a brief infusion of ACh or adenosine in the isolated rabbit heart offers similar protection against an ischemic insult, these two agonists trigger that protection through very different signal transduction pathways. ACh-induced protection was dependent on PI3-kinase, at least one genistein-sensitive tyrosine kinase, and a Src family tyrosine kinase, whereas that from adenosine was dependent on none of these. Although both agents activate what are thought to be G_i-coupled receptors, curiously they use very different signal transduction pathways to trigger the preconditioning phenomenon.

It is useful to divide preconditioning into two time periods: the trigger phase and the mediator phase. Preconditioning has a memory such that once triggered, the heart stays preconditioned for an hour or two. In our classification system, factors that turn the memory on are triggers, whereas those that mediate the protection once ischemia begins are mediators. Thus triggers such as receptor ligands act before the onset of the index ischemia, whereas mediators such as PKC (34) must act during ischemia. In a previous study, we (4) found that a group of receptor agonists typified by ACh triggered a preconditioned state by a mechanism that required both opening of mitoK_ATP channels and the generation of ROS before the lethal ischemic insult. Protection from adenosine, however, could not be blocked by either mitoK_ATP channel closure or free radical scavenging, suggesting very different signal transduction pathways for these two agonists. Whereas opening of mitoK_ATP channels clearly serves as a trigger, it may also mediate protection because a mK_ATP channel blocker administered during the mediator phase can prevent protection, although the concentration required to block the mediator phase is fourfold higher than that required to block triggering (33).

The present study reveals that the critical time for ACh to activate PI3-kinase is also before ischemia, thus making PI3-kinase a trigger as well. It is interesting that it was necessary to continue the wortmannin infusion until the onset of the index ischemia to effectively block the protection of ACh. It is unclear why the additional 5-min exposure to wortmannin was necessary. However, blockade of protection was probably not related to the presence of wortmannin during the ischemic period, because the ACh+W(L) protocol was ineffective at aborting protection. Because wortmannin administered during the ischemic period had no effect against protection, PI3-kinase does not have a mediator role.

We previously found that a 5-min infusion of ACh, as used here, caused a prolonged activation of PI3-kinase as assessed by Akt phosphorylation (10). That could explain why the wortmannin needed to be present up to the onset of ischemia to block protection. Tong et al. (31) recently reported that PI3-kinase was also involved in preconditioning because wortmannin could block protection in hearts preconditioned with ischemia. In that study, they concluded that PI3-kinase was upstream of PKC, which is compatible with the present observation because a trigger should be upstream from a mediator like PKC.

Wortmannin has been shown to have nonspecific effects apart from PI3-kinase inhibition including suppression of myosin light chain kinase (19). Unfortunately, the other available PI3-kinase inhibitor, LY-294002, is too expensive to apply in our whole heart model. It should be pointed out that ACh does induce PI3-kinase activation in the rabbit heart (10), which further supports our hypothesis that PI3-kinase is involved in the protection of ACh. Nevertheless, we cannot totally rule out the possibility that wortmannin may have exerted its effect at a site other than PI3-kinase.

Oldenburg et al. (22) recently reported that, in a vascular smooth muscle cell model, ACh could trigger ROS generation by mitochondria, and the ROS production could be blocked by either 5-hydroxydecanoate, a K_ATP channel blocker, or wortmannin. Because wortmannin could not block ROS generation from the direct K_ATP channel opener diazoxide, they concluded that PI3-kinase must reside between the muscarinic receptor and the K_ATP channel. If PI3-kinase is in the pathway leading to opening of K_ATP channels, then that also would make it a trigger. Indeed, our results are compatible with these findings.

Oldenburg et al. (22) also noted that genistein could block ROS production triggered by ACh but not diazoxide. That would put at least one tyrosine kinase in the pathway between the muscarinic receptor and the K_ATP channel. The present results further support the involvement of such a tyrosine kinase in the rabbit heart because early administration of genistein or lavendustin A during the trigger phase did block protection from ACh. As expected, they had no effect against the protection of adenosine. Because the Src nonreceptor tyrosine kinase inhibitor PP2 also blocked the protection of ACh, a Src kinase could well be the site of the action of lavendustin A and genistein.

In this study, we showed that neither the tyrosine kinase inhibitors nor the PI3-kinase blocker abolished the protection of adenosine when these inhibitors were present in the trigger phase. Adenosine activates both A₁ and A₃ receptors to trigger cardioprotection (1, 12, 13, 30). A recent report (24) demonstrated that A₁ and A₃ receptors exerted their cardioprotection by differential coupling to phospholipases C and D, respectively. As a result, A₃ receptors induced a longer lasting PKC activation than A₁ receptors (24). Because the inhibitors in our protocols were given before the administration of adenosine, it is difficult to explain our results on the basis that adenosine could have overcome the inhibitors simply because it produces a more persistent activation of the pathway. That would be especially true of the wortmannin protocol, where drug was present right up to the onset of ischemia, which would have trapped the drug in the tissue during the ischemic period as well. At no time before or during ischemia would the pathway have been unblocked, yet adenosine continued to protect. These results suggest that adenosine protects the heart via a signal transduction pathway different from that of ACh, which seems odd because both activate G_i-coupled receptors. Biochemical measurements (10) reveal that adenosine does indeed activate PI3-kinase in the rabbit heart. The only explanation seems to be that, in addition, adenosine activates a second parallel pathway involving PKC that continues to protect even when PI3-kinase is blocked. Our working hypothesis is that adenosine receptors couple directly to phospholipases C and/or D, whereas the other G_i-coupled receptors including the muscarinic receptor can only activate downstream kinases through mitoK_ATP channel-dependent ROS production.

Baines et al. (2) observed that ischemic preconditioning could be blocked by genistein but only when administered during the mediator phase. This observation is not really at odds with the present results because the earlier study evaluated ischemic preconditioning rather than individual agonists. Ischemic preconditioning in the rabbit uses a mixture of adenosine, bradykinin, and opioid receptors to trigger protection. It is likely that the adenosine pathway predominated in that study, thus preventing early genistein from blocking protection. It is only when we trigger protection through a single receptor that the individual signal transduction pathways can be discerned. We did not test genistein with the late protocol because Baines et al. (2) have already shown that a tyrosine kinase is involved in the mediator phase.

In the rat heart, adenosine seems to be a minor player in ischemic preconditioning, and opioid receptors predominate (11, 26, 27). Because we found that opioid receptors fall into the ACh class of receptors (4), the rat may well respond very differently to genistein. Indeed, Fryer et al. (6) found that genistein blocked protection from ischemic preconditioning in the rat during the trigger phase.

Even though genistein is a potent tyrosine kinase inhibitor, it has some nonspecific effects. For example, genistein was demonstrated to antagonize adenosine receptors (25). To rule out the possibility that genistein may have acted in our system by blocking adenosine receptors, we repeated the experiment with a structurally different tyrosine kinase blocker, lavendustin A, and obtained an identical result. At a low dose, lavendustin A is selective for receptor tyrosine kinases such as the EGF receptor (IC₅₀ = 0.011 µmol/l) (23). However, at the 2 µmol/l concentration we used, lavendustin A will also antagonize nonreceptor tyrosine kinases such as pp60^src (IC₅₀ = 0.5 µmol/l) (21).

In the present study, PP2 clearly blocked protection from ACh only when it was present during the trigger phase. That would suggest that a Src tyrosine kinase is also required for the protection of ACh. We would propose that a Src kinase is at least one of the tyrosine kinases targeted by genistein and that it is in the same pathway as PI3-kinase. It would be logical for PP2 to block the protection of ACh because several members of the Src family of tyrosine kinases have been shown to be necessary components of the G_i protein-induced transactivation of certain receptor tyrosine kinases such as the EGF receptor (15). In this scheme, G_i -subunits formed by binding of an agonist to its G_i-coupled receptor cause ligand-independent transactivation of receptor tyrosine kinases, which in turn activate PI3-kinase with subsequent production of 3-phosphoinositides (14).

Our observations are consistent with those of Mockridge et al. (17), who found that PP2 could block activation of PI3-kinase in cardiac myocytes subjected to simulated ischemia-reperfusion. Hattori et al. (9) were able to show that a similar compound, PP1, could block protection from ischemic preconditioning in isolated rat hearts but again only when PP1 was present during the trigger phase. On the other hand, Fryer et al. (8) could not block protection in in situ rat hearts with PP2 when protection was triggered with intravenous opioid agonists. These latter results obviously contradict our hypothesis, because we would have predicted that opioid agonists would behave similarly to ACh (4). It should be noted, however, that we did not actually test opioid agonists in these experiments. The differences could be related to receptor type, species, or even their in situ protocol, in which opioids were present continuously throughout the experiment rather than transiently as in the present study. Also, the actual plasma concentration of PP2 was not determined in their study, nor was blockade of Src kinase confirmed. The studies of Oldenburg et al. (22) suggest that Src kinase is activated somewhere between the muscarinic receptor and the mitoK_ATP channel. Thus we would not expect PP2 to block protection triggered by a pulse of diazoxide, a direct mitoK_ATP channel opener. That indeed turned out to be the case.

Several animal models are currently used for preconditioning studies. The rabbit heart is a useful experimental model for infarct studies because of its low collateral blood flow. Also, because rabbit myocardium contains very little xanthine oxidase, it would be expected to be a better biochemical model of the human heart than that of the rat or dog. Because the pig heart is anatomically and biochemically similar to the human heart, it also should be a suitable model for preconditioning. However, because of its size, isolated heart studies would be impractical. Although the rabbit heart seems to be an appropriate model, we cannot rule out the possibility that these results may not apply to humans.

In conclusion, the ACh-triggered anti-infarct effect was abolished by genistein, PP2, lavendustin A, and wortmannin when each was present to bracket the ACh infusion, i.e., the trigger phase. This suggests that both tyrosine and PI3-kinases are located in the trigger phase of ACh-induced preconditioning. In contrast, early treatment with genistein, PP2, or wortmannin failed to block adenosine-induced cardioprotection. It is concluded that ACh uses a signaling pathway that is distinctly different from that used by adenosine to trigger protection in the rabbit heart.