Endothelial Insulin Resistance Protects the Heart Against Prolonged Ischemia–Reperfusion Injury But Does Not Prevent Insulin Transport Across the Endothelium in a Mouse Langendorff Model

Abstract

Aim:

The endothelium plays an important role in the maintenance of cardiovascular homeostasis in healthy individuals. Insulin resistance can lead to the development of endothelial dysfunction, which is an important step in the pathogenesis of atherosclerosis. We investigated specifically whether the presence of vascular insulin resistance and endothelial dysfunction has any influence on the myocardial tolerance to ischemia–reperfusion (IR) injury, using Endothelial Specific Mutant Insulin Receptor Over-expressing (ESMIRO) mice, which exhibit vascular insulin resistance and vascular dysfunction.

Methods:

ESMIRO or wild-type (WT) littermate mouse hearts were isolated and perfused on a Langendorff apparatus. These were subjected to either 35-minute or 45-minute ischemia followed by reperfusion, after which infarct size was determined. The ability of insulin to activate its target kinase pathway, that is, phosphoinositide 3 (PI3) kinase/protein kinase B (AKT) in ESMIRO hearts was also assessed by Western blot analysis.

Results:

Compared to 35-minute ischemia, the extended 45-minute ischemic protocol significantly exacerbated myocardial infarction in WT mice, (56% ± 4%, n = 6 vs 32% ± 4%, n = 9; P < .01) but not in ESMIRO littermates (34% ± 7%, n = 6 vs 32% ± 3%, n = 9; not significant), suggesting some form of protective phenotype. Insulin treatment was associated with a significant increase in AKT phosphorylation in the myocardium in both the ESMIRO mice and WT littermates, and this was attenuated in both by inhibition of PI3 kinase using LY294002. Thus, insulin was able to directly activate PI3 kinase/AKT in the myocardium despite the absence of functional endothelial insulin receptors in the ESMIRO mice.

Conclusion:

(1) Insulin at pharmacologic concentrations can be transported across the endothelium independent of vascular insulin receptors and (2) vascular insulin resistance and/or endothelial dysfunction are protective against prolonged IR injury in the Langendorff model.

Introduction

The endothelium plays an important role in the maintenance of cardiovascular homeostasis. ¹ The presence of insulin resistance in humans and in various animal models has been shown to adversely affect the homeostatic function of the endothelium by causing endothelial dysfunction. Endothelium dysfunction further leads to atherosclerosis and is a risk factor for the development of cardiovascular disease. ^{1 –3} Also, although the endothelium is believed to directly influence the myocardium’s susceptibility to cell death secondary to lethal ischemia–reperfusion (IR), ^4,5 the effect of endothelial insulin resistance and vascular dysfunction on the myocardial vulnerability to IR injury has not been investigated previously.

Since normal endothelial function is necessary for the maintenance of cardiovascular homeostasis, we hypothesized that the presence of vascular dysfunction due to insulin resistance may adversely influence the susceptibility of the myocardium to infarction and investigated this hypothesis using transgenic Endothelium Specific Mutant Insulin Receptor (“ESMIRO”) mice, which overexpress insulin receptors with a mutation (Ala-Thr¹¹³⁴) exclusively in the endothelium. Vascular insulin signaling is completely disrupted in these mice, resulting in vascular insulin resistance. As a consequence, vascular dysfunction develops. ⁶ Importantly, however, the ESMIRO mice are identical to their wild-type (WT) littermates with regard to their blood pressure as well as fasting blood glucose and insulin levels. ⁶ Thus, they are a useful model to investigate the specific contribution of endothelial insulin resistance and endothelial dysfunction to the sensitivity of the myocardium to IR injury.

In addition to examining the effects of vascular insulin resistance and endothelial dysfunction on the myocardial susceptibility to IR injury, we aimed to investigate the mechanism of transport of insulin across the endothelium using the ESMIRO mice. Insulin crosses the continuous endothelial layer of the cardiac microvasculature and binds to insulin receptors on the myocytes, influencing cardiac glucose uptake. However, currently, it is unclear whether insulin is transported across the endothelium in an insulin receptor-dependent process or whether it is in fact a receptor-independent process related to simple diffusion between endothelial cells. ^{7 –14} Consequently, it is not evident whether the vascular insulin-resistance component of diabetes could affect the transport of insulin across the vascular endothelium to the underlying myocardium.

A Langendorff-isolated mouse heart model was utilized for this study. The extent of myocardial infarct in response to two different lengths of IR was assessed in WT and ESMIRO mice. We also assessed the ability of insulin to reach and act on the myocardium in the absence of functional endothelial insulin receptors by comparing the extent of protein kinase B (AKT) phosphorylation in response to exogenous insulin administration.

Methods and Materials

All procedures were performed at The Hatter Cardiovascular Institute, UCL in strict accordance with the Home Office (United Kingdom) Guidance on Research and Testing using animals and the Animals (Scientific Procedures) Act of 1986.

Langendorff-Isolated Mouse Heart Preparation

Mice were given terminal anesthesia with a single 0.01 mL/g intraperitoneal dose of a mixture of ketamine (10 mg/mL), xylazine (2 mg/mL), and atropine (0.06 mg/mL). Additionally, an intraperitoneal injection of heparin (500 IU) was administered 15 minutes prior to excising the hearts. Hearts were extracted, immediately submerged in ice-cold modified Krebs-Henseleit buffer and then swiftly mounted via the aorta on a perfusion canula attached to a Langendorff perfusion system supplied by ADInstruments (ADInstruments Ltd, Unit B, Bishop Mews, Transport Way, Oxford). The time between cessation of extracorporeal circulation and the start of perfusion in the Langendorff mode was less than 5 minutes. Temperature was constantly monitored and maintained at 37.5 ± 0.5°C. Heart rate was monitored via a saline filled balloon inserted through the left atrium into the left ventricular cavity and connected at its other end to a pressure transducer. Hearts with a heart rate of less than 300 beats per minute were excluded. Hearts were perfused with a modified Krebs-Henseleit buffer (composed of 118 mmol/L NaCl, 25 mmol/L NaHCO₃, 11 mmol/L glucose, 4.7 mmol/L KCl, 1.22 mmol/L MgSO₄.7H₂0, 1.21 mmol/L KH₂PO₄, and 1.84 mmol/L CaCl₂.2H₂0) aerated with a mixture of O₂ (95%) and CO₂ (5%). pH was maintained at 7.42 ± 0.2. Isolated hearts with a buffer flow rate of less than 1 mL/min or more than 6.5 mL/min on the Langendorff preparation were excluded. After observing for exclusion criteria in an initial stabilization period, the hearts were subjected to the experimental protocols.

Ischemia–Reperfusion Protocol

Hearts were subjected to global IR by cessation of perfusion for the required duration followed by reperfusion. Two protocols were used consisting of 35-minute or 45-minute ischemia followed by 30-minute or 40-minute reperfusion, respectively. Hearts were stained with 1% triphenyltetrazolium chloride at the end of the IR protocol, frozen at −20°C and then sliced into 1 mm slices which were submerged overnight in 10% formaldehyde solution. Planimetry analysis was carried out to accurately quantify the percentage infarct size in each heart as a proportion of the total heart volume.

Insulin Treatment and Western Blot Analysis Protocol

Hearts were treated with 100 mU/mL of insulin (insulin solution from bovine pancreas 10 mg/mL in 25 mmol/L 2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid from Sigma-Aldrich [Sigma-Aldrich Company Ltd, The Old Brickyard, New Road, Gillingham, Dorset]) for 15 minutes followed by a 10-minute washout period. A second group of mice were given 100 mU/mL of insulin for the same period but in the presence of LY294002 (Cell Signalling [New England Biolabs (UK) Ltd, Hitchin, Hertfordshire]), a specific phosphoinositide 3 (PI3) kinase inhibitor, added to the modified Krebs-Henseleit buffer in a 15 µmol/L concentration (made in dimethyl sulfoxide) 15 minutes prior to insulin administration. LY294002 was used to demonstrate that the AKT phosphorylation was specific to the action of insulin on PI3 kinase in the myocardium. At the end of each protocol, whole hearts were freeze-clamped with liquid nitrogen and stored at −80°C for further analysis. Western blot analysis was carried out to assess total AKT (using AKT [pan] C67E7 antibody, Cell Signalling) and phosphorylated AKT (using phospho-AKT [Ser 473] antibody, Cell Signalling] in the homogenized whole hearts.

Animals Used

The ESMIRO mice were gifted by Professor Kearney’s group from Leeds University. Heterozygous ESMIRO mice were bred with WT mice in the University College London Biological Services Unit. The ESMIRO mice (n = 9; 3 males and 6 females, age 7-16 weeks, weight [wt] 22.8 ± 1.5 g) and wild-type littermates (n = 9, 3 males and 6 females, age 7-15 weeks, wt 21.81 ± 1.1 g) identical in age and weight were used to determine infarct size in the heart following 35 minutes ischemia and 30 minutes reperfusion. Similarly, matching ESMIRO mice (n = 6, all males, age 8-15 weeks, wt 28.2 ± 1.2 g) and WT littermates (n = 6, all males, age 9-15 weeks, wt 29.12 ± 1.4 g) were used to assess the infarct size in the hearts in response to 45-minute ischemia and 40-minute reperfusion. Three ESMIRO mice each were used for Western blot analysis in the control (n = 3, 1 male and 2 females, age 16 weeks, wt 25.8 ± 4.7 g), insulin-treated (n = 3, 1 male and 2 females, age 14-16 weeks, wt 24.2 ± 2.0 g), and insulin + LY294002 (n = 3, 1 male and 2 females, age 15-16 weeks, wt 24.6 ± 0.2) treatment groups. Likewise, 3 WT mice each were used in the respective control (n = 3, 1 male and 2 females, age 16 weeks, wt 23.2 ± 2.4 g), insulin-treated (n = 3, 1 male and 2 females, age 16 weeks, wt 22.87 ± 2.7 g), and insulin + LY294002 (n = 2; 1 male and 1 female; age 16 weeks; wt 23.7 ± 0.6 g) groups.

Statistical Analysis

All values are presented at mean ± standard error of the mean. One way analysis of variance was used to assess statistical significance followed by Tukey test to compare treatment groups. Unpaired t-test with Welch correction was used to compare the baseline buffer flow rates between the ESMIRO mice and the WT littermates. A P value of <.05 was considered statistically significant. Graphpad PRISM 6.0 software was used for statistical analysis.

Results

There was no significant difference in the infarct size between the ESMIRO mice and the WT littermates when subjected to 35-minute ischemia followed by 30-minute reperfusion (mean infarct size 32%± 4% vs 32% ± 4%, n = 9, not significant (ns; Figure 1). A more severe ischemic insult with 45-minute ischemia led to a significant increase in the infarct size in the WT littermates (55% ± 4%, n = 6 vs 32% ± 4%, n = 9; P < .01). However, the same increase in ischemia time from 35 to 45 minutes, surprisingly, did not significantly influence infarct size in ESMIRO mice (mean infarct size 40% ± 7%, n = 6 vs 32% ± 4%, n = 9; ns). Consequently, after 45-minute ischemia and 40-minute reperfusion, the extent of infarction was significantly higher in the hearts of WT mice than that in the ESMIRO mice (56% ± 4% vs 40%± 7%, n = 6; P < .01). Baseline buffer flow rates in the WT littermates and the ESMIRO mice on the Langendorff isolated heart preparations were not significantly different (3.0 ± 0.1 mL/min vs 3.3 ± 0.2 mL/min; P = .16).

Figure 1.

Effect of ischemic time on infarct size (ESMIRO mice vs WT littermates). There was a significant increase in the infarct size in the hearts from WT littermates with an increase in the duration of ischemia–reperfusion (IR). However, ESMIRO mice were resistant to the effects of a similar increase in the duration of IR and displayed no significant increase in the infarct size despite the increased intensity of IR (IR 45/40, n = 6; IR 35/30, n = 9; P = .005). ESMIRO indicates Endothelial Specific Mutant Insulin Receptor Over-expressing; WT, wild type.

Binding of insulin to its receptor on the myocytes leads to the activation of PI3 kinase and AKT through phosphorylation of AKT without affecting the total AKT levels in the myocytes. Hence, as expected, the total myocardial content of AKT was the same in the WT and ESMIRO mice before and after insulin (Figure 2B and D). The mean ratio of phosphorylated AKT to total AKT in the WT littermate control group was 0.030 ± 0.001 AU. With insulin treatment, there was significant activation of AKT indicated by a significant rise in the amount of phosphorylated AKT (normalized to total AKT) in the WT littermate hearts (mean 0.130 ± 0.001 AU, P < .001) compared with the control group. In the presence of PI3 kinase inhibitor LY294002, there was a significant reduction in AKT phosphorylation in response to insulin treatment (mean 0.080 ± .008 AU, P = .02) compared with insulin treatment in the absence of this inhibitor (Figure 2A and C).

Figure 2.

Western blot analysis demonstrated that phosphorylated AKT levels increase after insulin treatment in WT or ESMIRO hearts and this phosphorylation is blocked by LY294002. A, Phosphorylated AKT and tubulin levels. B, Total AKT and tubulin levels. Lanes (left to right) 1 to 3, WT control hearts; 4 to 6, WT hearts treated with insulin; 7 and 8, WT hearts treated with insulin + LY294002; 9, 50 KDa marker; 10 to 12, ESMIRO control hearts; 13 to 15, ESMIRO hearts treated with insulin; 16 to 18, ESMIRO hearts treated with insulin + LY294002. C, Densitometry analysis of Western blot (A) showing a significant increase in phosphorylation of AKT with insulin treatment compared with the control group in both ESMIRO and WT mouse hearts isolated and perfused in a Langendorff mode. Treatment with LY294002 in addition to insulin significantly reduced AKT phosphorylation in response to insulin in both cases. D, Densitometry analysis of Western blot (B) showing no significant change in the total AKT levels in both the ESMIRO and the WT littermate hearts after treatment with either insulin or insulin + LY294002 compared with untreated control hearts. ESMIRO indicates Endothelial Specific Mutant Insulin Receptor Over-expressing; WT, wild type; AKT, protein kinase B.

Unexpectedly, in the ESMIRO hearts, treatment with insulin similarly led to a significant increase in AKT phosphorylation compared with the control group (mean ratios of phosphorylated AKT normalized to total AKT 0.118 ± 0.008 vs 0.037 ± .001 AU respectively, P < .001). Again, treatment with LY294002 significantly reduced AKT phosphorylation in response to insulin compared with insulin treatment in the absence of the inhibitor (mean ratio of phosphorylated AKT against total AKT normalized to tubulin 0.070 ± 0.001 vs 0.118 ± 0.008, respectively, P = .04; Figure 2A and C).

Summary and Discussion

The results demonstrate that an increase in the duration of ischemia in the ESMIRO mice did not lead to a significant rise in the infarct size, in contrast to the WT hearts in which a significant increase in the infarct size was observed after prolongation of the ischemic period. Additionally, insulin was able to phosphorylate AKT in both the ESMIRO and the WT mice despite the absence of functional vascular insulin receptors in the ESMIRO mice.

After the shorter duration of IR ESMIRO and WT hearts had similar sizes of myocardial infarction, suggesting the hearts are not inherently more resistant to all forms and durations of ischemia and reperfusion. However, our results suggest that the ESMIRO mice are resistant to the worsening effect of prolonged IR injury. The ESMIRO mice possess the characteristics of vascular insulin resistance and endothelial dysfunction, in the absence of any of the other defining features of type 2 diabetes mellitus (ie, hyperglycemia, hyperinsulinemia, and global insulin resistance). The results imply that contrary to our hypothesis, vascular insulin resistance and/or the resulting endothelial dysfunction, in isolation, decreases the myocardial susceptibility to prolonged ischemia.

The mechanisms by which endothelial cells may impart protection to the myocardium during IR are poorly understood. A limited number of studies have attempted to address this subject. ^15,16 These have identified proteins such as neuregulin and platelet-derived growth factor-BB as potential prosurvival proteins released from the healthy endothelium. The vascular endothelium can also be an important source of reactive oxygen species (ROS), particularly from the nicotinamide adenine dinucleotide phosphate oxidase (Nox). The ROS play a dual role in IR injury—while excessive generation of ROS plays an important role in mediating IR injury, ROS themselves also play an essential role in activating cardioprotective signaling pathways. ^{17 –20} Two important members of the family of NADPH oxidases responsible for the generation of ROS by the endothelium are Nox2 and Nox4. The effect of perturbations in these Nox isoforms on myocardial susceptibility to IR injury has previously been investigated and may have some relevance to this study. A decrease in Nox2 activity in Nox2 knockout mice has previously been shown to increase myocardial susceptibility to lethal IR injury. ¹⁸ In contrast, however, a more recent study has shown that globally reduced activity of Nox2 (but not Nox4), might decrease myocardial susceptibility to IR injury. ²¹ The ESMIRO mice express higher levels of both Nox2 and Nox4 isoforms of NADPH oxidase and have an increased generation of reactive oxygen species in their endothelium. ⁶ Whether alterations in the Nox2/Nox 4 isoforms of NADPH oxidase restricted to the endothelium in the ESMIRO mice might have influenced the myocardial susceptibility to IR injury remains to be established.

In this study, insulin was able to activate its target enzyme (AKT) in the myocardium in the ESMIRO mice despite the absence of functional insulin receptors in the endothelium which forms a continuous, nonfenestrated barrier between the blood and myocardium. This supports the existence of a receptor-independent mode of transport of insulin across the endothelium. As a pharmacological dose of insulin was used in this study, it is possible that a different mode of transport may be used by insulin in the normal physiological state and at nonpharmacologic concentrations. However, the phenotype of the ESMIRO mice itself suggests that vascular insulin receptors may not be required for insulin to cross the endothelium and act on the target tissues as these mice had normal glucose metabolism even in the absence of functional insulin receptors on the endothelium and in fact demonstrated a trend toward improved insulin sensitivity. ⁶ Additionally, it is possible that at least a fraction of the insulin transport across the endothelium in this case may have been mediated through insulin-like growth factor 1 receptors which are structurally similar to the insulin receptors and can bind with insulin though with much lesser affinity than the insulin receptors. ^22,23

These findings provide some insights into the questions related to ischemia tolerance of the heart in the presence of vascular insulin resistance and vascular dysfunction, which also occurs in diabetes and regarding the method of transport of insulin across the endothelium. Despite the damage to the endothelium incurred during the development of vascular insulin resistance, adaptive mechanisms may paradoxically reduce the susceptibility of the myocardium to IR injury. This occurs in the absence of hyperglycemia, hyperinsulinemia, or blood pressure changes.