Preconditioning the Diabetic Heart : The Importance of Akt Phosphorylation

Male Wistar rats (300–550 g, n = 47) were obtained from Charles River UK Limited (Margate, U.K.), and male GK rats (300–550 g, n = 71) were obtained from Taconic (Denmark). All animals received humane care in accordance with the Home Office Guidance on the Operation of Animals (Scientific Procedures) Act 1986 (Her Majesty’s Stationery Office, London, U.K.).

Rats were fed a standard chow diet, heparinized with sodium heparin (300 IU), and anesthetized with sodium pentobarbital (50 mg/kg) intraperitoneally. Hearts were rapidly excised, mounted on a Langendorff system, and perfused with a modified Krebs-Henseleit buffer. All hearts were subjected to 35 min of regional ischemia and 120 min of reperfusion as previously described (27). Infarct size (expressed as a percentage of the area at risk) was determined by triphenyltetrazolium chloride staining.

Hearts were randomly assigned to one of the following groups: 1) Wistar vehicle control (n = 8) with 0.02% DMSO given alone during stabilization; 2) Wistar one-cycle IPC (n = 8) comprising 5 min of global ischemia followed by 10 min of reperfusion before the index ischemia; 3) Wistar two-cycle IPC (n = 8) before the index ischemia; 4) Wistar three-cycle IPC (n = 8) before the index ischemia; 5) GK vehicle control (n = 8) with 0.02% DMSO given alone during stabilization; 6) GK one-cycle IPC (n = 9) before the index ischemia; 7) GK two-cycle IPC (n = 9) before the index ischemia; 8) GK three-cycle IPC (n = 8) before the index ischemia; 9) GK three-cycle IPC plus LY294002 (n = 6), in which hearts were given the PI3K inhibitor LY294002 (15 μmol/l) (21) starting 5 min before and throughout the preconditioning protocol with a 5-min washout before the lethal ischemia; or 10) GK control plus LY294002 (n = 6), in which hearts were given LY294002 during stabilization.

Hearts (n = 5 per group) were randomly assigned to the treatment groups 1, 2, 4, 5, 6, 8, 9, and 10 as described above. Myocardial samples were taken at the end of stabilization or 5 min after the last IPC cycle and freeze-clamped in liquid nitrogen before being stored at −80°C. The phosphorylation state of Akt (phospho-Akt, Ser 473) and total level of Akt protein were analyzed by SDS-PAGE immunoelectrophoresis using antibodies obtained from New England BioLabs as described previously (21). Levels of phosphorylated proteins were normalized to their total protein levels, and equal protein loading was confirmed by β-actin probing of membranes (Abcam, Cambridge, U.K). Relative densitometry was determined using the computer software package NIH Image 1.63.

Samples for nonfasting blood glucose (n = 92) and HbA_1c (A1C) (n = 45) were taken immediately after excision of the heart. Blood glucose measurements (millimoles per liter) were determined using an ABL 700 series blood gas analyzer (Radiometer, Copenhagen, Denmark), and A1C measurements (percent) were determined by an antibody-colorimetric assay using a Cobas Mira Plus analyzer (Roche Diagnostic Systems).

All values are expressed as means ± SE. Infarct size and Western blot results were analyzed by one-way ANOVA and Fisher’s protected test of least significant difference. Differences were considered statistically significant when P < 0.05. Infarct size, glucose, and A1C correlations were calculated by linear regression analysis.

In normal Wistar rats, one, two, and three cycles of IPC significantly reduced infarct size represented as a percentage of the area at risk (44.7 ± 3.8% [P < 0.05], 31.4 ± 4.9% [P < 0.01], and 34.3 ± 6.1% [P < 0.01], respectively) compared with control hearts (60.7 ± 4.5%). However, in diabetic GK rats, only three cycles of IPC reduced infarct size significantly compared with GK control hearts (20.8 ± 2.6 vs. 46.6 ± 5.2%; P < 0.01). Both one and two cycles of IPC failed to reduce infarct size significantly compared with GK control hearts (35.8 ± 6.2 and 38.5 ± 4.5 vs. 46.6 ± 5.2%, respectively; NS). However, the infarct reduction afforded by three cycles of IPC in the GK rat (20.8 ± 2.6%) was completely abolished in the presence of the PI3K inhibitor LY294002 administered 5 min before the IPC protocol until 5 min after (44.9 ± 6.4%; P < 0.01), suggesting that PI3K may be important as a trigger in IPC. LY294002 did not influence infarct size in the GK control group (40.6 ± 7.2%) (Fig. 1).

In diabetic GK hearts, one cycle of IPC induced significant phosphorylation (in arbitrary units [AU]) of Akt compared with GK control hearts (24.7 ± 5.0 AU vs. 6.2 ± 0.9 AU, respectively; P < 0.05), although this was not commensurate with a reduction in infarct size. However, three cycles of IPC induced significant phosphorylation of Akt compared with both GK control and GK one-cycle IPC hearts (42.4 ± 3.2 vs. 6.2 ± 0.9 AU [P < 0.01] and 24.7 ± 5.0 AU [P < 0.05], respectively), and this was associated with a significant reduction in infarct size. Inhibiting PI3K with LY294002 during the IPC protocol partially but significantly abrogated the phosphorylation of Akt in GK three-cycle IPC to levels similar to those seen with one-cycle IPC (42.4 ± 3.2 vs. 20.8 ± 2.1 AU, respectively; P < 0.01). In normal Wistar rats, one and three cycles of IPC led to significant phosphorylation of Akt compared with Wistar control hearts (72.4 ± 7.5 and 74.2 ± 9.1 vs. 37.1 ± 5.7 AU, respectively; P < 0.01 for both) (Fig. 2).

Diabetic GK rats were characterized by significantly higher levels of blood glucose (16.1 ± 0.6 mmol/l in GK rats vs. 9.0 ± 0.3 mmol/l in Wistar rats; P < 0.01) and A1C (4.4 ± 0.1% in GK rats vs. 3.0 ± 0.04% in Wistar rats; P < 0.01) when compared with normal Wistar rats. A1C values >3.45% were considered diabetic using our assay. Regression analysis, however, did not demonstrate any correlation between infarct size and glucose (R² = 0.003) or A1C (R² = 0.163).

We report two findings for the first time. First, the type 2 diabetic myocardium can benefit from the cardioprotective effects of ischemic preconditioning, provided that the preconditioning stimulus is increased to reach the threshold necessary to achieve myocardial protection. In our study, this threshold required a preconditioning stimulus of three cycles of 5 min of ischemia and 10 min of reperfusion before the prolonged ischemic insult. Second, it appears that impairment of prosurvival signaling cascades may be responsible for this elevated threshold. We demonstrated that although one cycle of IPC induced a significant phosphorylation of Akt, this did not result in a reduction of infarct size. However, three cycles of IPC in the diabetic heart resulted in a significantly greater phosphorylation of Akt than one cycle, and this was commensurate with a significant reduction in infarct size, suggesting that a certain level of activated Akt is critical to mediate the protective effects of preconditioning. Furthermore, Western blot analysis did not demonstrate any significant differences in the total levels of Akt protein in diabetic GK hearts in all groups, indicating that the impairment in phosphorylation of Akt was not a result of lower levels of total Akt in diabetic hearts but rather in the signaling process leading to Akt activation. Whether this impairment in Akt activation is due to abnormalities in upstream mediators of the Akt pathway (i.e., PI3K or phosphatidylinositol-3-phosphate–dependent kinase [PDK]) remains to be elucidated. However, it should be noted that our data does not prove a direct causal relationship between impaired Akt phosphorylation and the elevated threshold for protection. The use of specific Akt inhibitors would help to clarify the role of Akt in protection in this setting.

Interestingly, our current study supports the findings of Kondo and Kahn (28) who demonstrated that different defects in components of cell survival kinase cascades in diabetic models are not species specific but organ specific within the same species. In both type 1 and 2 diabetic mice, they showed that in the retina, the reduction in the phosphorylation of PDK1 and Akt was due to reduced total levels of PDK1 and Akt when compared with controls after insulin stimulation. However, within the same mice, the total levels of PDK1 and Akt were the same in the liver, and the reduction in PDK1 and Akt phosphorylation was due to impaired activation of these proteins.

Evidence suggests that other components of cellular prosurvival pathways are defective in diabetic tissues as well as the heart. Hyperglycemia has been shown to inhibit the prosurvival effect of vascular endothelial growth factor, leading to retinal cell apoptosis via tyrosine nitration of PI3K that results in Akt inactivation and increased p38 mitogen-activated protein kinase activation (29). Similar results have been demonstrated in rat hearts exposed to hyperglycemic conditions, which led to tyrosine nitration and apoptosis through the action of inducible nitric oxide (NO) synthase and NO release (30). To our knowledge, our data demonstrate for the first time that the impairment in cellular signaling cascades is responsible for the inconsistent results reported in previous diabetic preconditioning experiments.

The lack of correlation between glucose, A1C levels, and infarct size in our study suggests that the severity of diabetes does not predict the extent of myocardial infarction. Rather, once a minimum level of diabetes is reached, the threshold for preconditioning is elevated.

Until recently, most animal studies have focused on ischemic preconditioning in the diabetic heart using chemically induced type 1 diabetic animal models (10–14). However, the conflicting data obtained from these studies should be interpreted with caution. Many of the animals died as a result of the chemical induction of diabetes, and others displayed characteristics of stress, probably as a consequence of the toxicity of the drugs (alloxan or streptozotocin) used. The nonspecific effects of these drugs on the myocardium are not well known. In addition, these animal models of diabetes simulate type 1 diabetes (31), which in the clinical scenario is the least common form in the human population. In this regard, chemically induced models of diabetes do not represent the most appropriate model in which to study the effects of myocardial protection. Animal models of type 2 diabetes should be used, as it is this form of diabetes that is prevalent worldwide and is associated with increased cardiovascular risk.

The GK rat is a selectively inbred model of type 2 diabetes developed from the Wistar rat. Type 2 diabetes in this rat has many similarities to the human form of the disease (25,26), and it has been used extensively as a type 2 diabetic research model (32). This appears to be a more appropriate diabetic research model because worldwide the most common form of diabetes is type 2 diabetes, and its prevalence is steadily increasing (33). To date, only one study has addressed the issue of preconditioning in a model of type 2 diabetes using an obese and lean animal. In that study, Kristiansen et al. (34) used the GK rat to address the issue of ischemic preconditioning and demonstrated that, in their model, the diabetic heart failed to show any reduction in infarct size when subjected to an IPC protocol of four cycles of 2 min of ischemia followed by 3 min of reperfusion, a preconditioning stimulus that we would argue was insufficient to reach the threshold necessary to activate cardioprotective mechanisms in the type 2 diabetic heart. Furthermore, this study did not provide any mechanistic data to explain the lack of protection afforded by IPC in the model. In contrast to their study, the diabetic hearts in our study were able to be protected by IPC, although the IPC stimulus required was elevated. This would suggest that, provided the IPC stimulus is sufficient, the diabetic heart can be protected from ischemic-reperfusion injury.

Interestingly, our data support the observations made by Kristiansen et al. (34) with respect to the smaller infarct size they observed in their diabetic animal hearts compared with nondiabetic hearts after an episode of ischemia-reperfusion. This appears to be a common finding in studies comparing the myocardial infarct size in diabetic animal models with “normal” hearts (10,34,35) and suggests that, in diabetic hearts, a myocardial adaptation occurs after a prolonged ischemic insult that attempts to limit the damage sustained from the ischemic injury. However, the exact mechanisms responsible for this adaptation have yet to be elucidated. In our study, the Wistar control animals also demonstrated larger infarct sizes compared with the diabetic GK animals, raising an important question as to whether a “normal” Wistar heart is the appropriate control with which to compare the type 2 diabetic GK hearts. GK rats have been developed by selective inbreeding of glucose-intolerant Wistar rats since 1975, resulting in a species with many biochemical and metabolic similarities to human type 2 diabetes. Therefore, as in clinical studies, it could be argued that a more appropriate control to compare a treatment in diabetic populations would be a nontreated diabetic rather than a “healthy” species. Nevertheless, we still thought it appropriate to add this control specifically to demonstrate that the concept of preconditioning is observed in nondiabetic hearts. Importantly, this should not detract from the fact that the most important findings from our study, namely the requirement for a sufficient preconditioning stimulus to induce protection, was observed within the diabetic GK group.

In conclusion, we report for the first time that the type 2 diabetic myocardium can be protected by IPC, but that the threshold required to achieve this protection is elevated compared with that in nondiabetic hearts. We find that this elevation in threshold may be required to achieve sufficient phosphorylation of Akt to execute the IPC protective signal. The findings from this study suggest that the human diabetic population may be more resistant to the protective effects of IPC but that, provided the preconditioning stimulus is sufficient, the diabetic myocardium can be protected. One intriguing aspect would be to investigate whether this resistance to IPC-induced protection is reversed in diabetic species treated with either insulin or oral hypoglycemic drugs.

A.T. is supported by a project grant from the British Heart Foundation. The authors thank Novo Nordisk for additional funds for this study.