Superior diastolic function with K_ATP channel opener diazoxide in a novel mouse Langendorff model

Abstract

Background:

Adenosine triphosphate–sensitive potassium (K_ATP) channel openers have been found to be cardioprotective in multiple animal models via an unknown mechanism. Mouse models allow genetic manipulation of K_ATP channel components for the investigation of this mechanism. Mouse Langendorff models using 30 min of global ischemia are known to induce measurable myocardial infarction and injury. Prolongation of global ischemia in a mouse Langendorff model could allow the determination of the mechanisms involved in K_ATP channel opener cardioprotection.

Methods:

Mouse hearts (C57BL/6) underwent baseline perfusion with Krebs-Henseleit buffer (30 min), assessment of function using a left ventricular balloon, delivery of test solution, and prolonged global ischemia (90 min). Hearts underwent reperfusion (30 min) and functional assessment. Coronary flow was measured using an inline probe. Test solutions included were as follows: hyperkalemic cardioplegia alone (CPG, n = 11) or with diazoxide (CPG + DZX, n = 12).

Results:

Although the CPG + DZX group had greater percent recovery of developed pressure and coronary flow, this was not statistically significant. Following a mean of 74 min (CPG) and 77 min (CPG + DZX), an additional increase in end-diastolic pressure was noted (plateau), which was significantly higher in the CPG group. Similarly, the end-diastolic pressure (at reperfusion and at the end of experiment) was significantly higher in the CPG group.

Conclusions:

Prolongation of global ischemia demonstrated added benefit when DZX was added to traditional hyperkalemic CPG. This model will allow the investigation of DZX mechanism of cardioprotection following manipulation of targeted K_ATP channel components. This model will also allow translation to prolonged ischemic episodes associated with cardiac surgery.

Introduction

The adenosine triphosphate–sensitive potassium (K_ATP) channel exists in metabolically active tissues throughout the body and provides endogenous myocardial protection via coupling of cell membrane potential to myocardial metabolism (inhibited by ATP and open during times of metabolic stress).^1–5 Pharmacologic opening of K_ATP channels with diazoxide (DZX) prevents detrimental myocardial swelling and reduced contractility in response to different types of stress: metabolic inhibition, exposure to hyperkalemic cardioplegia (CPG), and hypo-osmotic stress in animal^6–8 and in human cardiac myocytes.⁹ The exact mechanism of protection provided by K_ATP channel openers is incompletely defined and is the focus of intensive research.

Mouse Langendorff models with 20-45 min of global ischemia are known to induce myocardial infarction.^10–14 The benefit of K_ATP channel openers over conventional, hypothermic, hyperkalemic CPG is yet to be demonstrated. Previous models failed to emulate the additive injury induced during cardiac surgery, which is composed of prolonged global ischemia and exposure to hyperkalemic CPG. We sought to develop a murine model translatable to this setting and use it to assess the protective effect provided by K_ATP channel openers as a component of hypothermic, hyperkalemic CPG.

Methods

All animals received humane care in compliance with the “Principles of Laboratory Animal Care” formulated by the National Society for Medical Research and the “Guide for the Care and Use of Laboratory Animals” published by the US National Institutes of Health.¹⁵ The Animal Care and Use Committee from Washington University approved of this study.

Animals and anesthesia

Adult, male, C57BL/6 mice (25 g) were anesthetized by intraperitoneal injection of 0.4 mL Avertin (3% 2,2,2-tribromoethanol, 1.86% 2-methyl-2-butanol) and 100 units of unfractioned heparin. Sternotomy and rapid cardiectomy were performed. The heart was submersed in an ice-cold Krebs-Henseleit buffer (KHB) while aortic cannulation with a blunted 22-gauge needle was performed using a dissection microscope. Hearts were then perfused with KHB (NaCl 118.5, NaHCO₃ 25.0, KCl 3.2, KH₂PO₄ 1.2, MgSO₄ 1.2, CaCl₂, and D-glucose 35.0 [concentrations expressed in mmol/L]) at a pH range between 7.40 and 7.48 by bubbling with a gas mixture of 95% O₂ and 5% CO₂. Perfusate passed through 45-mm polyvinylidene fluoride filters (EMD Millipore Corp, Billerica, MA) and was not recirculated. Hearts were excluded if aortic cannulation time exceeded 5 min. The left atrium was opened by amputating the pulmonary veins to allow visualization of the mitral apparatus. A balloon was introduced through the mitral valve. All hearts were epicardially paced at 170 beats per minute and surrounded by a warm (37°C) water-jacketed beaker. Coronary perfusion pressure was maintained at 80 mm Hg, determined by the height of the column (see Fig. 1). This pressure has been identified as the zone of autoregulation for coronary vasculature in murine hearts.¹⁶

Fig. 1 –

Schematic representation of mouse Langendorff apparatus. Apparatus used for prolonged (90 min) global ischemia Langendorff model.

Experimental protocol

The left ventricular (LV) balloon was connected via a 15-cm long polyethylene tube to a HP1290C pressure transducer (Hewlett Packard, Andover, MA) and a 20-4 amplifier (Gould Instrument Systems Inc, Valley View, OH). The balloon volume was adjusted to an end-diastolic pressure (EDP) of 2.5 mm Hg. Isolated hearts underwent a 30-min baseline period permitting stabilization. Then a pressure–volume curve was obtained by 1.4 μL balloon volume increments with 90-s intervals between these for stabilization and data collection. The LV balloon volume was returned to the initial volume. Volume increments were precisely administered using a threaded plunger syringe (Hamilton Company, Franklin, MA). Hearts were arrested with test solution and pacing interrupted. A 90-min ischemic injury was sustained with a continuous epicardial drip of KHB maintained at a temperature of 27°C. The test solutions included were as follows: hyperkalemic CPG alone (Plegisol, Pfizer, New York City, NY) (in mmol/L: NaCl, 110; NaHCO₃, 10; KCl, 16; Mg, 32; and CaCl₂, 2.4) or CPG + DZX (Sigma Aldrich, St. Louis, MO) (7-chloro-3-methyl-1,2,4-benzothiadiazine-1,1-dioxide) at a dose of 100 μmol/L. The DZX dose was chosen based on previous experiments.^6–9 A stock solution of DZX was made by dissolving in 0.1% dimethyl sulfoxide, a concentration at which the antioxidant effect of dimethyl sulfoxide remains inconsequential.¹⁷ Pressure measurements were continuously recorded throughout the experiment. At the end of the ischemic period, retrograde coronary perfusion and epicardial pacing were resumed, providing reperfusion during 30 min. A second pressure–volume curve was recorded at identical balloon volumes as previously described.

Construction of LV balloons

LV balloons were constructed using silicone to be small, durable, and highly elastic.¹⁸ In brief, a hard candy mold was dipped into silicone dispersion gel (NuSil Tech, Carpinteria, CA) to create a thin layer around the mold, and the gel-covered mold was dried followed by submersion in deionized water to dissolve the candy. The silicone sleeve was removed from the mold by squeezing the bland candy out. The sleeve was then cannulated by a PE 60 tube (Becton Dickinson, Franklin Lakes, NJ) and tied with 4-0 silk to create a water-tight seal. Since presence of air in the circuit could dampen the pressure signal significantly, all the balloons were assessed under magnification for bubbles before use.

Data acquisition and analysis

Left ventricular end-systolic (ESP), EDP, and developed (LVDP) pressures were determined from digitalized data files with the use of LabVIEW 2014 (National Instruments, Austin, TX) set at a sampling rate of 1000 Hz. The software quantifies ESP as the maximum value of each waveform and EDP as the point at which the increasing slope of the pressure waveform exceeds a threshold of 0.5 mm Hg/ms. LVDP was defined as the difference between ESP and EDP for each data point. Average ESP, EDP, and LVDP were calculated by sampling 10 consecutive beats. Coronary flow rates were measured every 5 min by an inline N-series flow probe (Transonic Systems, Ithaca, NY) and a T206 flow meter (Transonic Systems, Ithaca, NY). Figure 2 illustrates LVDP over a series of balloon volumes in a single experiment from pressure–volume curves before and after injury. The percent recovery of LVDP was defined as the quotient of the area under the curve (AUC) for the reperfusion LVDP (blue) by the AUC for baseline LVDP (orange). AUC was calculated by the use of the trapezoidal rule as depicted in the following equation:

$$100*{\displaystyle \underset{{BV}_{min}}{\overset{{BV}_{max}}{\int }}\frac{LVDPpostreperfusiondBV/(B{V}_{max}-B{V}_{min})}{LVDPbaseline}}$$

where BV_max is the maximum postreperfusion-matched balloon volume, and BV_min is the minimum postreperfusion-matched balloon volume. The percent recovery of coronary flow rate was calculated in a similar manner to that described for LVDP. Raw values of ESP, EDP, and LVDP were standardized by the baseline measurements for each experiment and described as percent from baseline (see Figs. 3 and 4).

Fig. 2 –

Depiction of LVDP calculation (area under the curve) for a single experiment. Each isolated mouse heart underwent 90 min of global ischemia and 30 min reperfusion. Left ventricular pressure was measured using identical balloon volumes. Distribution of LVDP over an incremental series of intracavitary balloon volumes in a single experiment. In orange, LV developed pressure before 90 min of ischemia, and blue, after the injury. The quotient of the area under the reperfusion curve by the baseline curve provides the percentage recovery of LVDP after the injury.

Fig. 3 –

Representative pressure tracing during a single experiment. Each isolated mouse heart underwent 90 min of global ischemia and 30 min of reperfusion. Left ventricular pressure was measured using identical balloon volumes. Zero EDP was measured a minute before cardioplegic arrest. Ischemic EDP was measured immediately after cessation of ventricular activity from the arrest. Plateau EDP was measured at the point where pressure remained unchanged for 5 min. Reperfusion EDP was measured 1 min before the reinstatement of retrograde aortic perfusion. The third plateau EDP was measured at the highest point after reperfusion, usually occurring within 4-8 min of reperfusion. The final EDP was measured after 30 min of reperfusion. Time to plateau was measured from the start of the ischemic injury to beginning of the plateau in minutes. The t½ was measured as half of the duration of the slope to reach the plateau in minutes.

Fig. 4 –

Mean EDP for all experiments at each balloon volume. Each isolated mouse heart underwent 90 min of global ischemia and 30 min of reperfusion. Left ventricular pressure was measured using identical balloon volumes. Mean EDP, expressed as a percentage from the baseline measurement, over an incremental series of intracavitary balloon volumes. Error bars represent standard error of the mean. Any reperfusion value above its baseline value represents a loss in ventricular compliance (i.e., change in pressure at the same volume) secondary to the ischemic injury. The control group (CPG) consistently had higher EDP in comparison with the study group (CPG + DZX).

Ventricular compliance was assessed through change in EDP over a series of identical intracavitary balloon volumes as obtained from the pressure–volume curves (Fig. 4). After standardizing for baseline values, mean EDP at each volume was compared between groups. To quantify the impact of preservation of ventricular compliance provided by DZX, a linear regression model was fitted with EDP as the outcome variable using end-diastolic volume and treatment group as covariates.

The investigators were blinded to test groups at the time of performing the experiments. Results are reported as mean ± standard deviation unless stated otherwise. Univariate comparisons were conducted through Wilcoxon rank-sum test or unpaired Student’s t-test for continuous variables, as appropriate. P-value < 0.05 (two-sided) is used to define statistical significance. All statistical analyses were done with R software, version 3.3.2.¹⁹

Waveform morphology

Using intraventricular balloon-recorded pressure measurements during ischemia, a characteristic and reproducible signal became evident, as seen in Figure 3. Recurring landmarks in this waveform were consistently measured and labeled as: zero EDP, ischemic EDP, plateau EDP, third plateau EDP, reperfusion EDP, and final EDP. Zero EDP was measured 1 min before cardioplegic arrest. Ischemic EDP was measured immediately after cessation of ventricular activity from the arrest. Plateau EDP was measured at the point where pressure remained unchanged for 5 min. Reperfusion EDP was measured 1 min before the reinstatement of retrograde aortic perfusion. The third plateau EDP was measured at the highest point after reperfusion, usually occurring within 4-8 min of reperfusion. The final EDP was measured after 30 min of reperfusion. Time to plateau was measured from the start of the ischemic injury to beginning of the plateau in minutes. The t½ was measured as half of the duration of the slope to reach the plateau in minutes.

Results

In the present study, the effects of a 90-min period of global ischemia on myocardial mechanics were studied in an isolated murine heart Langendorff preparation. Twenty-three experiments were performed, 11 in the control group (CPG) and 12 in the test group (CPG + DZX). Three hearts in the control group and one in the test group failed to recover contractile activity after the ischemic injury (P = 0.518). Hearts with failure to recover from injury were included for the comparisons in Table 1. The percentage recovery of LVDP was similar between groups with 22.3 ± 20.8 for CPG and 28.3 ± 21.2 for CPG + DZX (P = 0.385). The percentage recovery of coronary flow was similar between groups with 42.5 ± 31.8 for CPG and 54 ± 37.3 for CPG + DZX (P = 0.559) (Table 1).

Table 1 –

comparison of hemodynamic outcomes following 90 min of global ischemia.

Endpoints	CPG	CPG + DZX	P-value
% Recovery of developed pressure	22.3 ± 20.8	28.3 ± 21.2	0.385
% Recovery of coronary flow	42.5 ± 31.8	54 ± 37.3	0.559
Time to plateau (min)	74.9 ± 9.7	77.2 ± 7.7	0.600
t½ to plateau (min)	61.6 ± 8.2	68.9 ± 9.9	0.097
Baseline EDP (mm Hg)	1.0 ± 2	0.7 ± 1.2	0.999
Ischemic EDP (mm Hg)	2.5 ± 1	2.5 ± 1.3	0.999
Plateau EDP (mm Hg)	11.5 ± 3.6	8.2 ± 2.6*	0.042*
Reperfusion EDP (mm Hg)	12.0 ± 3.7	8.4 ± 3.3*	0.016*
Third plateau EDP (mm Hg)	33.8 ± 9.3	18.5 ± 7.6*	0.042*
Final EDP (mm Hg)	24.9 ± 8.9	19.0 ± 6.3*	0.049*

Results reported as mean ± standard deviation.

t_½ = half-life; CPG = cardioplegia; DZX = diazoxide; EDP = end-diastolic pressure.

^*P-value < 0.05 when compared with CPG.

Waveform morphology

Time to plateau was similar between the CPG and CPG + DZX groups: 75 versus 77 min (P = 0.600). t½ to plateau was similar between groups: 62 versus 69 min (P = 0.097). Baseline EDP was similar between groups: 1.0 versus 0.7 mm Hg (P = 0.999). Ischemic EDP was also similar between groups: 2.5 versus 2.5 mm Hg (P = 0.999). Plateau EDP was higher in the CPG group: 11.5 versus 8.2 mm Hg (P = 0.042). Reperfusion EDP was higher in the CPG group: 12.0 versus 8.4 mm Hg (P = 0.016). The third plateau EDP was higher in the CPG: 33.8 versus 18.5 mm Hg (P = 0.042). Final EDP was higher in the CPG group: 24.9 versus 19 mm Hg (P = 0.049) (Table 1).

Preservation of compliance and contractility

Compliance was quantified only in those hearts that recovered from the injury to obtain a second pressure–volume curve. The mean baseline EDP was similar between groups (P-values ≥ 0.05 for all balloon volumes). The mean increase from baseline EDP was greater in the CPG group than that in the CPG + DZX group for three of the five balloon volumes: 686 versus 362% at 0 μL (P = 0.007), 871 versus 509% at 1.4 μL (P = 0.004), 962 versus 631% at 2.8 μL (P = 0.062), 1236 versus 743% at 4.2 μL (P = 0.027), and 1391 versus 759% at 5.6 μL (P = 0.073) (Fig. 4). In a linear regression model for percent increase from baseline EDP, the coefficient for belonging to the CPG + DZX group was −414.7% (P < 0.001) (Table 2).

Table 2 –

Left ventricular compliance assessed using linear regression model for end-diastolic pressure, standardized to baseline performance.

Covariates	Coefficient	SE	P-value
Volume (μL)	98.3	20.1	<0.001
Group
DZX	Ref.	Ref.	Ref.
DZX + CPG	−414.7	77.2	<0.001

CPG = cardioplegia; DZX = diazoxide; Ref. = reference group; SE = standard error.

Contractility was also quantified only for those hearts that recovered from the injury. The mean LVDP was similar between the groups at baseline (all P-values ≥ 0.05). The mean percentage LVDP change from baseline was similar between CPG and CPG + DZX groups: 31 versus 28% at 0 μL (P = 0.968), 37 versus 38% at 1.4 μL (P = 0.840), 42 versus 48% at 2.8 μL (P = 0.967), 42 versus 74% at 4.2 μL (P = 0.360), and 39 versus 66% at 5.6 μL (P = 0.180) (Fig. 5).

Fig. 5 –

Mean percentage LVDP for all experiments at each balloon volume. Each isolated mouse heart underwent 90 min of global ischemia and 30 min of reperfusion. Left ventricular pressure was measured using identical balloon volumes. LVDP expressed as a percentage from the baseline measurement over an incremental series of intracavitary balloon volumes. Errors bars represent standard error of the mean. LVDP values were similar between groups both at baseline and after reperfusion (all P-values ≥ 0.05).

Discussion

This study builds on previous work from our laboratory which explored the cellular mechanisms involved in the cardioprotective effect of K_ATP channel openers in murine ventricular^6–8 and in human atrial myocytes.⁹ The present study suggests that the addition of DZX to hyperkalemic CPG provides reduced diastolic injury in hearts exposed to a prolonged, ischemic injury through preservation of the diastolic pressure–volume relationship. Other investigators have suggested that reduction in ventricular compliance is secondary to the severity of the ischemic injury.^20–22 A reduction in ventricular compliance was observed during ischemia in both groups and was more pronounced in the control group (experiencing significantly higher plateau, reperfusion, third plateau, and final EDP values) (Table 1). These findings were also evident in the upward shift in the EDP-volume relationship (Table 2 and Fig. 4).

K_ATP channel openers such as DZX have been shown to mimic ischemic preconditioning in multiple animal models.^1–5 Findings of the present study demonstrate preservation of diastolic function after a prolonged global ischemic period. These two beneficial effects are likely related to separate and independent mechanisms that may or may not include a K_ATP channel and require further research. Post-ischemic diastolic dysfunction occurs from uncoupled relaxation-repolarization and is usually associated with myocardial stunning. Contributing factors to stunning may include generation of oxygen-free radicals, intracellular calcium overload, and decreased responsiveness of contractile filaments.²³ In a canine model of cardioplegic arrest and global ischemia, it was determined that these cellular mechanisms lead to an overall increase in ventricular wall stiffness secondary to increased myocardial edema.²⁴ In addition, exposure to hypothermic hyperkalemic CPG in the absence of ischemia is known to induce myocyte swelling.²⁵ We previously determined that DZX prevented myocyte swelling secondary to CPG in isolated rabbit⁷ and human²⁶ myocytes. Finally, Murata et al²⁷ determined that preischemic exposure to DZX attenuated mitochondrial calcium overload during ischemia-reperfusion injury. Hence, we hypothesize that the mechanisms of reduced diastolic injury with DZX may be provided a reduction in free radicals, a reduction in cellular edema, and a reduction in intracellular calcium. However, the specific site and mechanism of action for DZX remain elusive. Myocardial stunning²⁸ and diastolic dysfunction^29,30 have become ever-more prevalent disease processes in patients undergoing increasingly complicated cardiac surgery procedures, emphasizing the need for interventions to ameliorate these phenomena.

Rationale for prolonged injury

Previously reported murine global ischemia-reperfusion injury models are significantly shorter, ranging from 20 to 45 min followed by varying reperfusion periods.^11–14 Prolongation of ischemic time was necessary to establish a significant injury given the presence of highly effective hyperkalemic CPG. This study describes a novel global ischemic injury tailored to simulate the pathophysiology of prolonged global cardioplegic arrest during cardiopulmonary bypass. In 2014, the European Society of Cardiology working group issued a position paper highlighting the importance of using complex translational models that simulate realistic clinical scenarios at the benchtop³¹ such as this one. Nonetheless, the clinical implications of these findings should be extrapolated cautiously.

The significant decline of LVDP in both groups (percent recovery of only 22%-28% LVDP) indicates a severe injury during the prolonged global ischemic period despite protection of hyperkalemic CPG with or without DZX. The characteristic waveform produced by a prolonged ischemic injury (Fig. 3) initially corresponds to depletion of phosphocreatine and ATP, leading to ischemic contracture and increased intracavitary pressure.³² Prior studies suggest that by 60 min of ischemia, anaerobic glycolysis has been significantly reduced, and 94% of ATP has been depleted.^33,34 We theorize that this corresponds to the stabilization of intracavitary pressure and onset of a plateau during ischemia. On reperfusion, a transient increase in EDP becomes evident (third plateau phase) that may be triggered by a combination of reperfusion injury and the “no-reflow” phenomenon.³⁵

The safety of DZX during cardiac surgery has been briefly explored in human studies. DZX has been investigated as a preconditioning agent³⁶ and as an additive to intermittent warm blood antegrade CPG.³⁷ However, these clinical trials included a relatively healthy group of patients. DZX has not been explored as an additive to hypothermic CPG in high-risk patients (who often require repeat surgery or prolonged cardiopulmonary bypass times). We propose that pharmacologic K_ATP channel openers have the potential to prevent stunning during prolonged periods of ischemia, and their use in these situations requires further investigation.

Limitations

This study was performed in isolated murine hearts perfused with an ex-viuo perfusion apparatus using crystalloid buffers for oxygen delivery. The absence of hemoglobin and plasma proteins requires providing high levels of partial pressure of oxygen while exposing capillaries to a low oncotic pressure. Coronary disease and myocardial ischemia are multifactorial processes that become more prevalent with older age, whereas the hearts used in this study were from healthy mice. It is likely that these hearts would require prolonged periods of ischemia to achieve a significant injury. Specific indices of ventricular relaxation such as the relaxation time constant (tau) or dP/dt min during isovolumetric relaxation were not measured.

Future directions

Further studies are needed to elucidate the precise mechanism of cardioprotection provided by DZX. Myocardial implications of K_ATP channel variation have not been extensively explored, and the precise mechanism of action by which K_ATP channel openers provide myocardial protection remains unclear. This model provides an ex-viuo, whole organ environment to systematically investigate the role of known channel subunits using targeted gene deletion or subunit-specific antagonists.

Conclusion

The addition of DZX to conventional, hypothermic, hyperkalemic CPG maintained ventricular compliance and provided reduced diastolic injury compared with hyperkalemic CPG alone in a novel, prolonged ischemia mouse Langendorff model.