Effects of diabetes on myocardial infarct size and cardioprotection by preconditioning and postconditioning

In spite of the current optimal therapy, the mortality of patients with ischemic heart disease (IHD) remains high, particularly in cases with diabetes mellitus (DM) as a co-morbidity. Myocardial infarct size is a major determinant of prognosis in IHD patients, and development of a novel strategy to limit infarction is of great clinical importance. Ischemic preconditioning (PC), postconditioning (PostC) and their mimetic agents have been shown to reduce infarct size in experiments using healthy animals. However, a variety of pharmacological agents have failed to demonstrate infarct size limitation in clinical trials. One of the possible reasons for the discrepancy between the results of animal experiments and clinical trials is that co-morbidities, including DM, modified myocardial responses to ischemia/reperfusion and to cardioprotective agents. Here we summarize observations of the effects of DM on myocardial infarct size and ischemic PC and PostC and discuss perspectives for protection of DM hearts.


Introduction
Cardiovascular diseases are the leading cause of death, accounting for approximately 30% of all deaths worldwide [1]. Among the various cardiovascular diseases, acute myocardial infarction (AMI) has a high rate of mortality, and infarct size is a primary determinant of prognosis in these patients. The only established and clinically approved method to limit infarct size is restoration of coronary blood flow by percutaneous coronary interventions (PCI), thrombolytic agents or coronary bypass surgery. However, our recent review of clinical infarct size data indicated that infarct size after current reperfusion therapy was larger than 75% of the area at risk and larger than 20% of the left ventricle in one fourth of AMI patients [2]. Lack of substantial myocardial salvage in those patients cannot be simply attributable to long symptom-to-reperfusion time, and involvement of reperfusion injury [3] is also possible. Nevertheless, significant percentages of AMI patients suffer from insufficient cardiac function and have a poor prognosis even after successful restoration of coronary flow [2]. Therefore, the establishment of a novel strategy to limit the extent of infarction during ischemia/reperfusion is of great clinical importance.
As a cardioprotective strategy, ischemic preconditioning (PC) has received much attention for its potent infarct size-limiting effect since its first report by Murry et al. in 1986 [4]. Zhao et al. found that "conditioning" with repetitive ischemia/reperfusion at the time of reperfusion also affords protection and named the phenomenon ischemic postconditioning (PostC) [5]. Laboratory investigations to date have clarified outlines of the intracellular mechanisms of PC and PostC [6][7][8][9]. In addition, several agents targeting signaling relevant to PC and PostC have been shown to act as PostC mimetics [6][7][8]. However, clinical benefits of PC, PostC and their mimetics were not always demonstrated in clinical trials [8,10]. The discrepancy between results of animal studies and clinical trials may be explained by notable differences between animal models of AMI and AMI in humans [2,10]. First, there are species differences in regulatory mechanisms of cardiomyocytes such as contribution of sarcolemmal Ca 2+ ATPase to overall Ca 2+ handling. Second, the timing of administration of the test agent was not exactly the same in the positive animal studies and negative clinical trials. Third, most of the patients with AMI were pretreated with pharmacological agents for their co-morbidities at the onset of AMI, whereas animals were untreated, except for anesthetic agents, before AMI in most of the previous studies. Pharmacological agents used for treatment of hypertension, dyslipidemia and/or diabetes mellitus (DM) potentially modify intracellular signaling relevant to cytoprotection and thus myocardial responses to PC, Post and PC mimetics. Fourth, AMI patients recruited to clinical studies were generally old and had concomitant diseases, whereas relatively young and healthy animals were used in most of the previous animal studies. Aging and co-morbidities have been shown to modify or even abrogate the infarct size-limiting effects of interventions in animal experiments [6,11] and thus possibly underlie the negative results in clinical trials.
Of morbidities that potentially compromise the protective mechanisms of the heart, DM appears primarily important to study today. The number of patients with DM has been increasing worldwide in the past two decades, and these patients are predisposed to serious cardiovascular morbidity and mortality [12]. Despite recent progress in coronary intervention strategies, DM is associated with higher mortality after AMI due to more extensive atherosclerotic lesions and also hypertrophied and dysfunctional left ventricle [13][14][15]. Interestingly, Lamblin et al. recently showed that rates of cardiovascular death and heart failure after the first AMI were higher in DM patients than in non-DM patients, though ventricular function and remodeling were comparable in the two groups [16]. The increased mortality after AMI in DM patients is unlikely to be prevented simply by improvement of glycemic control, since intensive glycemic controls failed to reduce cardiovascular events as well as mortality rate in DM patients in recent large-scale clinical trials [17][18][19]. Hence, we need novel strategies for protecting DM patients with coronary artery diseases from excessive myocardial ischemiareperfusion injury. To gain an insight into a novel strategy, we thoroughly review reported effects of DM on myocardial infarct size and on protective mechanisms against infarction and discuss future perspectives in this article. Our discussion on pathophysiology and clinical approach to diabetic cardiomyopathy was recently published elsewhere [20].

Changes in myocardial susceptibility to infarction by DM
Clinical studies showed that DM increased the susceptibility of the myocardium to ischemia/reperfusion injury [21,22]. However, results of infarct size studies in diabetic animal models were contradictory (Tables 1 and  2), and the reason for the inconsistency remains unclear.
Multiple mechanisms are involved in cardiomyocyte necrosis induced by ischemia/reperfusion. Depletion of intracellular ATP, intracellular overloading with Na + and Ca 2+ , and increased fragility of the cell membrane are major changes during ischemia relevant to cell necrosis. Accumulating evidence supports the notion that the mitochondrial permeability transition pore (mPTP) is primed during ischemia and is opened by Ca 2+ overload and burst of reactive oxygen species at the time of reperfusion, leading to irreversible loss of mitochondrial functions and cell necrosis [23]. Theoretically, modification of ischemia-induced changes or reperfusion-induced opening of the mPTP should underlie the change in myocardial susceptibility to infarction in DM hearts. However, such modification has not been understood well.

Experimental studies
Infarct size in animal models of DM was larger or smaller than or similar to, depending on reports, that in nondiabetic controls ( Table 1). As shown in Table 1, there were numerous differences in experimental preparations and protocols, and a single factor cannot entirely explain the discrepancy in the effects of DM on infarct size. Duration of the diabetic state and plasma level of insulin (i.e., type 1 DM vs. type 2 DM) appear to influence myocardial tolerance against infarction. In a study by Ravingerová et al. [24], infarct size after 30-min ischemia was smaller in diabetic rat hearts at 1 week after streptozotocin (STZ) injection than in controls, but this infarct size limitation was not detected 8 weeks later. Two other studies have also shown that resistance of diabetic hearts to ischemia/reperfusion injury was detectable at the early phase of DM and disappeared later [25,26]. However, enlargement of infarct size was observed as early as 8 days after STZ injection [27][28][29], indicating that duration of DM is not the only factor responsible for inconsistent effects of DM on infarct size.
Difference in metabolic profiles associated with DM (i.e., presence or absence of hyperinsulinemia and types of dyslipidemia) may also be responsible for different changes in infarct size in DM animals. With a few exceptions [30,31], DM models with obesity and hyperinsulinemia showed increased myocardial susceptibility to ischemia/reperfusion-induced necrosis [32][33][34][35][36][37][38][39][40][41]. In contrast, high glucose level induced simply by glucose or dextrose infusion did not affect infarct size in the majority of studies, including two recent studies [42,43] ( Table 2).

Clinical studies
Increased myocardial susceptibility to infarction in DM patients was indicated by two clinical studies: infarct size determined by SPECT imaging after reperfusion therapy was larger by 30~70% in DM patients than in non-DM patients [21,22]. The change in myocardial tolerance to infarction is consistent with worse short-and long-term prognosis after AMI in DM patients [44,45]. Of particular relevance, Haffner et al. [46] showed that cardiovascular event risk in DM patients without prior myocardial infarction was comparable to the risk in non-diabetic patients with prior myocardial infarction.
However, the untoward effects of DM on prognosis after AMI cannot be explained solely by increased myocardial vulnerability to ischemia/reperfusion injury, since the risk of heart failure and death was significantly higher in DM patients than in non-DM patients with comparable infarct sizes and left ventricular ejection fractions [47]. Myopathic changes ("diabetic cardiomyopathy"), impaired collateral recruitment, microcirculatory abnormalities, and underuse of evidence-based therapies are  possible explanations, in addition to increased susceptibility to infarction, for detrimental prognosis after AMI in DM patients.

Defects in intracellular protective signaling in DM hearts
DM is one of the co-morbidities that potentially modify myocardial response to protective interventions, including PC and PostC. A number of studies have been carried out to determine the efficacy of ischemic PC and PostC for protecting the myocardium in animal models of DM (Tables 3 and 4) and diabetic patients ( Table 5). The majority of those studies showed that DM interferes with protective mechanisms of cardioprotective interventions. Myocardial protection by PC and PostC is achieved by activation of multiple protective signaling pathways that appear to converge to inhibit mPTP opening upon reperfusion via phosphorylation of glycogen synthase kinase-3β (GSK-3β) at Ser9 [6,8,9], and DM-induced defects in the protective signaling may be different depending on the model and/or phase of DM.

Experimental studies
With a few exceptions, previous studies demonstrated that cardioprotection achieved by ischemic PC or PostC is impaired in DM: the infarct size-limiting effect was lost or required extra cycles of ischemia/reperfusion in experimental diabetes (Table 3). Tsang et al. [48] reported that one, two and three cycles of ischemic PC significantly reduced infarct size in normal rats; however, three cycles of ischemic PC were required to limit infarct size in diabetic Goto-Kakizaki rats. Their findings suggested that the threshold for PC protection is increased in the diabetic myocardium. Liu et al. [49] were the first to examine the effects of DM on PC and they reported preservation of PC protection in STZinduced DM. However, since they used a PC protocol with multiple cycles of ischemia/reperfusion only, the change in the threshold for inducing PC protection might have been missed in their study. In fact, impairment of PC by DM has been observed in multiple species (rats, rabbits and dogs) and different models of DM (STZ-induced type 1 DM and genetic models of type 2 DM) [30,37,[50][51][52][53][54]. Mimetics of ischemic PC and PostC (diazoxide, erythropoietin, [D-Ala 2 , D-Leu 5 ]-enkephalin acetate [DADLE] and isoflurane) were also ineffective for limitation of infarct size in DM hearts [33,34,37,[55][56][57][58][59], confirming that DM impairs intracellular signaling mechanisms relevant to myocardial protection. Several steps in cytoprotective signaling have been found to be disrupted by DM in the heart. Blocked or attenuated signaling via phosphoinsitide-3 kinase (PI3K)-Akt, extracellular signal-regulated kinase (ERK) and/or signal transducer and activator of transcription-3 (STAT3) pathways was detected in association with loss or reduced efficiency of PC in protection against infarction [31,33,34,36,48,[57][58][59]. Our recent studies have shown that phosphorylation of Jak2, being upstream of PI3K-Akt signaling, is inhibited by enhanced calcineurin activity and that phosphorylation of GSK-3β by ERK is lost by an endoplasmic reticulum stress-dependent mechanism in OLETF, a rat model of obese type 2 DM [33,34]. Furthermore, protein level of active GSK-3β, a pro-necrotic and pro-apoptotic kinase, was increased in mitochondria, leading to reduction in the threshold for mPTP opening in response to Ca 2+ overload [34]. On the other hand, a protective mechanism downstream of GSK-3β phosphorylation appears to be intact in diabetic hearts, since direct pharmacological inhibitors of GSK-3β limited infarct size in diabetic hearts as in nondiabetic animals [33,34,54,57,58].
An important and clinically relevant issue is whether hyperglycemia per se is responsible for disruption of protective signaling in PC and PostC. Acute hyperglycemia induced by dextrose infusion attenuated infarct size limitation by ischemic PC, a mitochondrial ATPsensitive K + channel (K ATP channel) opener and anesthetic agents (Table 4). These results indicate a primary role of hyperglycemia in impairment of protective signaling. Interestingly, Przyklenk et al. [31] reported that the cardioprotective effect of ischemic PostC was re-established in STZ-induced diabetic mice by pancreas islet cell transplantation. Transplantation of islet cells in diabetic mice normalized both blood glucose level and free fatty acid level and cellular signaling, such as ERK phosphorylation, activated by ischemic PostC. Since dyslipidemia was also shown to attenuate the infarct sizelimiting effect of ischemic PC [60,61], restoration of myocardial response to PostC in the diabetic heart by islet cell transplantation might be achieved by normalization of both plasma glucose level and lipid profile. Nevertheless, circumstantial evidence in animal experiments supports the notion that normalization of the metabolic profile restores protective signaling mechanisms in the DM heart.

Clinical studies on PC
PC protection in human hearts has been demonstrated by results of in vitro experiments using human ventricular myocytes [62] and atrial trabeculae [63] and by detailed analyses of myocardial responses in patients with naturally occurring ischemic syndromes [64,65]. Since ischemic PC and PC-like phenomena in human hearts have been recently reviewed elsewhere [7,65], we briefly summarize the results that are relevant to the effects of DM on ischemic PC. As for PC-like effects of preinfarct angina, different groups reported that pre- infarct angina was associated with smaller infarct size determined by creatine kinase (CK) release, improved left ventricular function and reduced mortality [64,66]. Furthermore, in overall analysis of AMI patients, angina before AMI was associated with better long-term prognosis [67,68]. However, in patients with DM, beneficial effects of pre-infarct angina were not detected [69]. Ishihara et al. [69] showed that CK release and recovery of cardiac function and in-hospital mortality after AMI were similar in DM patients with and without pre-infarct angina. DM-induced impairment of ischemic PC protection in human hearts has also been indicated by studies in which myocardial injury was assessed during angioplasty [70] and during a treadmill exercise test [71] by use of electrocardiography. Direct evidence for diabetes-induced loss of ischemic PC protection in the human myocardium was provided by an in vitro experiment using atrial trabeculae obtained at open heart surgery. CK release and contractile dysfunction after hypoxia/reoxygenation in vitro was   significantly suppressed by PC in trabeculae from nondiabetic patients but not in trabeculae from diabetic patients [63,72].

Clinical studies on PostC
In contrast to PC, PostC or administration of its mimetic at the time of reperfusion is clinically feasible. In 2005, Staat et al. [73] translated for the first time the concept of ischemic PostC into the clinical setting. PostC with four cycles of 1-min inflation and 1-min deflation before full reperfusion successfully reduced CK release in a small and selected cohort of patients with AMI. Efficacy of ischemic PostC in patients with AMI was subsequently confirmed in a number of small-scale studies, using as endpoints cardiac biomarkers, ECG STsegment resolution, and anatomic infarct size (Table 5). However, three recent studies failed to show protection by ischemic PostC in terms of CK and troponin release, LV function or infarct size that was determined by delayed enhancement magnetic resonance imaging (MRI) [74][75][76]. The reasons for the discrepant results remain uncertain, but there are some possibilities: differences in the protocols of ischemic PostC, presence of co-morbidities and use of different pharmacological agents for the co-morbidities.
In the PostC protocol, duration from reperfusion to the first re-occlusion and number of re-occlusions have been shown to influence the infarct size-limiting effects of PostC in animal models of AMI [77][78][79]. Retrospective analyses by Darling et al. [80] and Wang et al. [81] suggest that larger numbers of balloon inflationsdeflations were associated with greater protection in patients with AMI. However, an appropriate algorithm of PostC stimulus to maximally protect the human myocardium has not been established. In addition, protection by PostC in animal models of AMI is attenuated with aging and presence of co-morbidities, such as DM, hypertension and hypercholesterolemia [6,11]. Although characteristics of patients, such as age and prevalence of DM, seem to be similar in the human studies (Table 5), difference in durations of the comorbidities might have been involved in the different outcomes. Nevertheless, the effect of DM on protection afforded by PostC has not been specifically examined in clinical studies.
The effects of pharmacological PostC (i.e., administration of pharmacological agents at reperfusion) on myocardial injury have been examined in a number of clinical trials. Results of the trials are a mixture of positive and negative results, and the reason for the discrepancy remains unknown [82][83][84][85][86][87][88][89][90][91][92]. Adenosine infusion limited infarct size, but neither the death rate during 6-month follow-up nor the rate of rehospitalization for heart failure were reduced by the adenosine treatment [82,84]. A post hoc analysis suggested that only patients with early reperfusion therapy received clinical benefits of adenosine, such as a reduction in death or heart failure [85]. In the J-WIND trial [86], the effects of nicorandil and ANP on infarct size in different study arms were examined. Administration of nicorandil failed to significantly limit infarct size, though recombinant ANP infusion was shown to induce a small but significant reduction in infarct size. The results of studies on the effect of erythropoietin administered at reperfusion were discrepant [87][88][89][90][91][92]. Taken together, there are a few promising agent for infarct size limitation in AMI patients, but none of their effects has been unequivocally demonstrated to be sufficient for improving prognosis.

Effects of anti-diabetic medications on AMI in diabetic patients
Possible adverse effects of sulfonylureas on cardiac events in DM patients have been debated for decades, and recent studies have indicated that cardiovascular risk was significantly higher in DM patients treated with sulfonylureas than in those who received metformin monotherapy [93,94]. Sulphonylureas have been shown to inhibit protection of PC and PostC by blocking the ATP-sensitive K + channel (K ATP channel) in cardiomyocytes, raising concern that such actions of sulphonylureas may be detrimental at the time of AMI in DM patients. However, the effects on PC and PostC are actually not uniform across sulfonylureas. Glibenclamide, a non-specific blocker of the K ATP channel, abrogated the infarct size-limiting effect of PC and PostC in nondiabetic animals [6,7,95]. In contrast, glimepiride, a second-generation sulfonylurea, did not block the protective effect of PC in the animal myocardium or human myocardium [96,97], presumably because this agent has few cardiac actions [98]. In addition, glimepiride was shown to have a PostC-mimetic action by activation of PI3K/Akt in the rabbit myocardium [99]. However, the effects of sulfonylureas on PC in diabetes have been tested in only a few studies [51,96], since PC and PostC failed to limit infarction in most of the DM models. Interestingly, Nieszner et al. [51] reported that infarctsparing effect of PC in alloxan-treated diabetic rabbits was restored by pretreatment with a low dose, but not a high dose, of glibenclamide. The mechanism by which glibenclamide restored myocardial response to PC in the DM rabbit is unclear. However, beneficial effects of glycemic control, possibly on PI3K/Akt signaling, might have overwhelmed the effect of partial inhibition of cardiac K ATP channels when an appropriate dose of glibenclamide was selected.
Metformin is an insulin-sensitizing agent that reduces hepatic glucose output and increases the uptake of glucose in peripheral tissues, including the skeletal muscle. In non-diabetic animals, metformin reduced infarct size by activation of AMP-activated protein kinase and/or Akt [95]. However, whether metformin is additive to ischemic PC or PostC has not been examined. Infarct size-limiting effect of metformin in diabetic animals was controversial: there are two positive reports [40,100] and one negative report [101].
Thiazolidinediones (pioglitazone, rosiglitazone) administered before ischemia reduced infarct size in both normal and diabetic animals [39,41,95]. However, the impact of the cardioprotective effect of thiazolidinediones on clinical outcome is unclear, since the effects of pioglitazone and rosiglitazone on risk of cardiovascular events appear discrepant [102].
In the past few years, growing evidence has demonstrated cardiovascular effects of incretin-based therapy using glucagon-like peptide-1 (GLP-1) receptor agonists and inhibitors of dipeptidyl peptidase-4 (DPP-4) [95,[103][104][105]. However, the effects of GLP-1 receptor agonists or DPP-4 inhibitors on ischemia/reperfusion injury have not been examined in diabetic animals except for a study by Huisamen et al. [106] They reported that enlargement of infarct size in obese pre-diabetic rats was reduced by treatment with a DPP-4 inhibitor for 4 weeks [106]. Further investigation is necessary for clarifying the effects of GLP-1 analogues and DPP-4 inhibitors on intracellular protective signaling and its modification by DM, though there are a number of on-going large-scale clinical trials to test the effects of incretin-based therapy on cardiovascular outcomes [107].

Conclusions and perspectives
DM potentially increases myocardial susceptibility to ischemia/reperfusion injury, though it cannot be explained by chronic hyperglycemia alone. DM also modifies myocardial responses to ischemic and pharmacological PC and PostC by disruption of intracellular signaling responsible for enhancement of resistance to cell death. These alterations in the diabetic heart appear to underlie the poor prognosis of DM patients after AMI. In addition, a defect of the protective mechanisms caused by aging, DM and possibly other co-morbidities is one of possible reasons why infarct size-limiting effects of agents in animal studies could not have been translated into AMI patients.
Analyses of modifications of cytoprotective signaling by DM, heart failure, dyslipidemia and hypertension generally showed that defects in the signaling by the pathological factors are upstream of a step of GSK-3β phosphorylation and mPTP regulation [6,9]. Hence, a reasonable approach to protect the heart in DM patients is manipulation of protective mechanisms at the level of GSK-3β or mPTP. Blockade of mPTP opening by cyclosporine A at the time of reperfusion is one such approach, and its effect was first examined in 58 AMI patients [108]. Infarct size determined by MRI at 5 days after AMI was significantly limited by cyclosporine A and the beneficial effects persisted until 6 month after AMI [109]. The results are very promising and support the feasibility of a strategy for cardioprotection targeting the mPTP. However, how DM modifies the mPTP regulation and how we can suppress opening of the mPTP at the time of reperfusion need to be further investigated for devising a novel therapy for cardioprotection in DM patients.

Competing interests
TaM has received research funding from Daiichi-Sankyo, Japan, Eisai, Japan, and Pfizer, Japan. TI and DS have no conflicts of interest or financial ties to disclose. TeM has received research funding from Takeda Pharmaceuticals, Japan, Novartis, Japan, Chugai Pharmaceutical, Japan, and Astellas Pharma, Japan.

Authors' contributions
TaM collected and arranged the data in sequence, drafted and correspond the manuscript. TI and DS contributed to collection and analysis of the data. TeM contributed to analysis and interpretation of data and revision. All authors have read and approved the final manuscript.