In vivo opening of the mitochondrial permeability transition pore in a rat model of ventricular fibrillation and closed-chest resuscitation

Abstract

Opening of the mitochondrial permeability transition pore (mPTP) is considered central to reperfusion injury. Yet, most of our knowledge comes from observations in isolated mitochondria, cells, and organs. We used a rat model of ventricular fibrillation (VF) and closed-chest resuscitation to examine whether the mPTP opens in vivo and whether cyclosporine A (CsA) attenuates the associated myocardial injury. Two series of 26 and 18 rats each underwent 10 minutes of untreated VF before attempting resuscitation. In series-1, rats received 50 µCi of tritium-labeled 2-deoxyglucose ([³H]DOG) harvesting their hearts at baseline (n=5), during VF (n=5), during resuscitation (n=6), and at post-resuscitation 60 minutes (n=5) and 240 minutes (n=5). mPTP opening was estimated measuring the ratio of mitochondria to left ventricular intracellular [³H]. In series-2, rats received 10 mg/kg of CsA or vehicle before resuscitation, measuring mitochondrial NAD⁺ content to indirectly assess mPTP opening. In Series-1, the mPTP opening ratio vs baseline (10.4 ± 1.9) increased during VF (16.8 ± 2.4, NS), closed-chest resuscitation (20.8 ± 6.3, P<0.05), and at post-resuscitation 60 minutes (20.9 ± 4.7, P<0.05) and 240 minutes (25.7 ± 11.0, P<0.01). In series 2, CsA failed to attenuate reductions in mitochondrial NAD⁺ and did not affect plasma cytochrome c, plasma cardiac troponin I, myocardial function, and survival. We report for the first time in an intact rat model of VF that mPTP opens during closed-chest resuscitation consistent with previous observations in mitochondria, cells, and organs of mPTP opening upon reperfusion. CsA, at the dose of 10 mg/kg neither prevented mPTP opening nor attenuated post-resuscitation myocardial injury.

Introduction

The term mitochondrial permeability transition pore (mPTP) denotes the formation of a mega pore across the inner and outer mitochondrial membranes through the apposition of mitochondrial proteins enabling the passage of molecules up to 1.5 kDa [1]. The specific proteins involved in pore formation are an issue of continuing scientific debate. The most recent studies proposed that the pore is a heterooligomeric complex composed of three proteins; cyclophilin-D (Cyp-D; a mitochondrial matrix protein), the voltage dependent anion channel (VDAC; an outer mitochondrial membrane protein), and spastic paraplegia 7 (SPG7; an inner mitochondrial membrane protein) [2].

Under physiological conditions, the mPTP may open and close transiently [3,4] representing a phenomenon known as “flickering” and involved in the regulation of mitochondrial calcium homeostasis [5-7]. mPTP flickering stimulates generation of mitochondrial superoxide, the so-called “superoxide flashes”, which is a physiological event that results from functional coupling between transient mPTP opening and electron transport chain dependent reactive oxygen species production [8].

Under pathophysiological conditions, however, the mPTP may open in a sustained manner causing collapse of the proton motive force across the inner mitochondrial membrane required for ATP synthesis by F_o-F₁ ATP synthase [9-11], uncoupling oxidative phosphorylation and leading to energy depletion, cell injury, and eventual cell death [11-15]. Support for this mPTP effect stems from studies in cardiomyocytes subjected to oxidative stress [16-18], isolated perfused rat hearts subjected to elevated Ca²⁺ levels [14,19], and studies in an intact rat model of heart failure in which pore opening was measured ex vivo after heart removal [20].

However, measuring mPTP opening in isolated mitochondria, cells, and organs simulating ischemia and reperfusion injury may be subject to experimental artifacts limiting extrapolation to in vivo settings. Moreover, removal of organs from their natural environment eliminates the influence of adaptive responses affecting ischemia and reperfusion injury such as neuroendocrine response to stress with release of catecholamines and other stress hormones [21].

In vivo measurement of transient mPTP opening accompanied by “superoxide flashes” has been reported recently in mice expressing a superoxide biosensor [22]. However, to the best of our knowledge, measurement of mPTP opening in vivo in the setting of ischemia and reperfusion has not been reported. We therefore adapted a technique previously used to measure mPTP opening in isolated rat hearts based on tritium-labeled 2-deoxyglucose ([³H]DOG) [14] and measured mPTP opening in a rat model of ventricular fibrillation (VF) and closed chest resuscitation. We conducted additional experiments to examine whether cyclosporine A (CsA) a putative inhibitor of mPTP opening could elicit favorable myocardial effects in the same rat model, indirectly assessing mPTP opening by measuring mitochondrial NAD⁺ levels [23].

Materials and methods

The studies were approved by our Institutional Animal Care and Use Committee and conducted in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.

Animal preparation

Male retired breeder Sprague-Dawley rats (456 to 543 g) were anesthetized with 45 mg/kg of sodium pentobarbital given intraperitoneally followed by 10 mg/kg given intravenously every 30 minutes to maintain a surgical plane of anesthesia. A 5-F catheter was orally advanced into the trachea and used for positive pressure ventilation during chest compression and the post-resuscitation interval. Proper tracheal placement was verified using an infrared CO₂ analyzer (CO₂ SMO model 7100, Novametrix Medical Systems Inc., Wallingford, Connecticut). A conventional lead II ECG was recorded through subcutaneous needles. For pressure measurement and blood sampling, fluid-filled PE50 catheters were advanced from the right femoral artery into the abdominal aorta and from the left femoral vein into the right atrium. To assess left ventricular (LV) function, another fluid-filled PE50 catheter was advanced from the right carotid artery into the left ventricle. A 3-Fr catheter (model C-PUM-301J, Cook Inc., Bloomington, IN) was advanced through the right external jugular vein into the right atrium, and through its lumen a precurved guidewire was fed into the right ventricle for electrical induction of VF. The guidewire was removed before starting chest compression. For measurement of core temperature and thermodilution cardiac output, a thermocouple probe (0.64-mm diameter, IT-18, Physitemp Instruments, Clifton, NJ) was advanced from the left femoral artery into the thoracic aorta. A 3-F catheter (C-PUM-301J, Cook Inc., Bloomington, IN) was advanced from the left external jugular vein into the right atrium and used for injection of the thermal tracer (200 μl of normal saline at ~24°C) to measure cardiac output. Core temperature was monitored through the thermocouple and maintained between 36.5 and 37.5°C using a lamp.

VF and resuscitation protocols

VF was induced by delivering a 60-Hz alternating current to the right ventricular endocardium (0.1 to 0.6 mA) for 3 minutes after which VF was allowed to continue spontaneously for 7 additional minutes completing a 10 minute interval of untreated VF. Chest compression was then initiated using a pneumatically driven piston device (CJ-80623, CJ Enterprises, Tarzana, CA) centered on the mid-chest and programmed to deliver 200 compressions per minute with a 50% duty cycle. The depth of compression was adjusted to maintain a coronary perfusion pressure between 22 and 24 mmHg by the second minute of chest compression, exceeding the 20 mmHg threshold required for successful resuscitation in rats [24,25]. The piston travel was measured with a displacement transducer (DSPL, World Precision Instruments Inc., Sarasota, FL). Positive pressure ventilations were provided with a volume-controlled ventilator (Model 683, Harvard Apparatus, Holliston, MA) set to deliver 25 unsynchronized breaths per minute using a tidal volume of 6 ml/kg and 100% oxygen. Transthoracic defibrillation was attempted after 8 minutes of chest compression by delivering up to two 5-Joules (J) biphasic waveform electrical shocks across the chest wall (Heartstream XL, Philips Medical Systems, Andover, MA) 5 seconds apart. If VF persisted or an organized rhythm with a mean aortic pressure ≤25 mmHg ensued, chest compression was resumed for 30 seconds. The defibrillation-compression cycle was repeated up to three additional times, increasing the energy of individual shocks (if VF persisted) to 7-J for the subsequent two cycles. Successful defibrillation was defined as the return of an organized rhythm with a mean aortic pressure >60 mmHg for >5 minutes. After return of spontaneous circulation, rats were monitored for a maximum of 240 minutes in series 1 and 360 minutes in series 2 (see Experimental Series).

In vivo measurement of mPTP opening

We adapted a technique previously developed in isolated perfused rat hearts [14,26,27] to measure in vivo mPTP opening in our rat model. The technique in the isolated heart involves: i) retrograde perfusion of the heart with a [³H]DOG solution mounted in a Langendorff apparatus to enable [³H]DOG uptake by cardiomyocytes through glucose transporters (i.e., facilitated diffusion), ii) phosphorylation of uptaken [³H]DOG by hexokinase to [³H]DOG-6-phosphate a form not able to enter the glycolytic pathway trapping [³H]DOG-6-phosphate in the cytosol, iii) entrance of [³H]DOG-6-phosphate to the mitochondrial matrix upon mPTP opening, iv) flushing the coronary circuit to remove [³H]DOG from the extracellular space, v) trapping of [³H]DOG-6-phosphate in the mitochondria by calcium chelation during tissue processing, and vi) measuring mitochondrial and LV intracellular (LV_IC) [³H] radioactivity (activity) expressing mPTP opening as the ratio of mitochondria to LV_IC [³H] activity [14]. For the present experiments, we loaded the hearts in vivo during spontaneous circulation 30 minutes before inducing VF by intravenous injection of [³H]DOG and avoided flushing the coronary circuit to minimize additional ex vivo manipulation of myocardial tissue before processing.

Plasma and LV tissue processing

At the end of each experiment and immediately before removing the heart, whole blood was collected, centrifuged for 15 minutes at 1,000 g, and 200 µl aliquots of plasma were collected and stored at -80°C. Hearts were excised via a sternotomy and the LV separated, weighed, and homogenized in ice-cold MSH buffer (2.5 ml for 0.1 g tissue) containing (in mM) mannitol 210, sucrose 300, HEPES 5, EGTA 2 at pH 7.4, and BSA 5 mg/ml. EGTA was used to chelate Ca²⁺ sealing the mPTP and thus trapping [³H]DOG-6-phosphate inside the mitochondrial matrix. A 50 µl aliquot of the crude homogenate was used to measure LV [³H] activity and to estimate sample protein content by BCA protein assay (Thermo Scientific Inc., Bannockburn, IL). Mitochondria were isolated by differential centrifugation [28], suspended in 500 µl of MSH buffer and stored at -80°C in separate 100 µl aliquots. From each aliquot, 50 µl was used to measure [³H] mitochondrial activity and the other 50 µl to estimate protein content and quantify citrate synthase (CS) activity [29].

Measurement of [³H] activity in LV homogenate, mitochondrial fraction, and plasma

For [³H] activity measurement in LV homogenate and mitochondrial fraction, suspensions were thawed on ice, mixed with equal volume of 5% perchloric acid (HClO₄), and centrifuged at 10,000 g for 2 minutes to precipitate proteins. After centrifugation, 50 µl of the supernatant was mixed with 5 ml of scintillation fluid. For [³H] measurement in plasma, samples were thawed on ice and 10 µl was mixed with 5 ml of scintillation fluid. LV tissue homogenate, mitochondrial fraction, and plasma activity were measured as disintegrations per minute (dpm) in a scintillation counter (Beckman Instruments Inc., Irvine, CA) programmed to read each sample for 10 minutes. [³H] activity in LV homogenate, mitochondrial fraction, and plasma was expressed as dpm per ml of the original respective homogenate, mitochondrial fraction, and plasma sample volume (dpm/ml).

Determination of mPTP opening

mPTP opening was determined as the ratio between mitochondrial [³H] activity (normalized to mitochondrial CS; [³H] dpm/IU of CS) divided by LV_IC [³H] activity (normalized to wet weight; [³H] dpm/g wet-weight) ×10⁵.

LV_IC [³H] activity (dpm/g wet-weight)

Because we did not flush the coronary circuit, the LV extracellular activity was presumed equivalent to plasma [³H] (Pl [³H]) activity and subtracted it from the total LV [³H] (LV_T [³H]) activity. We assumed the extracellular fraction of LV tissue to be 0.21 based on previous measurements in the same rat model [30] and determined the LV_IC [³H] activity according to equation (i):

Next, the LV_IC [³H] (dpm/ml) was divided by LV density estimated to be 1.05 g/ml [31] to obtain LV_IC [³H] (dpm/g wet-weight).

Mitochondrial [³H] activity (dpm/IU CS)

Mitochondrial CS activity was measured according to Shepherd and Garland [29]. The assay is based on a colorimetric measurement of the molar extinction coefficient of 5-thio-2-nitrobenzoic acid (TNB) by coupling the enzymatic reaction (ii) catalyzed by CS (1) to the irreversible chemical reaction (iii). The thiol group reacts with 5, 5’-dithiobis-2-nitrobenzoic acid (DTNB) forming a yellow product whose absorbance correlates linearly with TNB molar concentration.

For the assay, mitochondrial suspensions thawed on ice were diluted 1:10 with MSH buffer and a 10 µl aliquot transferred to a 1000 µl cuvette with 920 µl of assay buffer composed of 50 mM Tris-HCl, pH 8, 30 mM acetyl-CoA, and 10 mM DTNB. Kinetic activity was measured at wavelength 412 nm (BioMate 3 UV-Vis Spectrophotometer, Thermo Fisher Scientific Inc., Bannockburn, IL) before and after adding 70 µl of 10 mM oxaloacetate (OAA) to the assay reaction. The activity measured before adding the OAA (background) was subtracted from the activity measured after adding the OAA.

Specific enzyme activity was calculated using equation iv:

Whereas IU denotes specific activity of CS expressed in international units, ∆A ₄₁₂/min denotes rate of absorbance change measured at 412 nm, ε^mM denotes extinction coefficient of TNB at 412 nm and pH 8.1 (13.6 mM^-1cm^-1), l denotes optical length using a 1000 µl cuvette (1 cm), v denotes stoichiometric number of TNB in the reaction iii (1), V_cuvette denotes volume of solution in the cuvette (1000 µl), V_sample denotes volume of sample in the cuvette (10 µl), and P_sample denotes protein concentration in the sample (mg protein/µl).

Mitochondrial [³H] activity (dpm/ml) was first normalized to mitochondrial protein concentration (dpm/mg protein) and then to CS activity by dividing the mitochondrial [³H] dpm/mg protein by CS activity (IU/mg protein) and expressed as dpm/IU of CS. Pore opening was calculated using equation (v) and expressed as the ratio between mitochondrial and the left ventricular [³H] activity [14].

Cardiac troponin I measurement

Cardiac troponin I (cTnI) was measured using a commercially available one step “sandwich” enzyme immunoassay method developed for human cTnI (Dimension® clinical chemistry system using Cardiac Troponin-I Flex® reagent cartridge, Dade Behring Inc., UK). The method has excellent reactivity and specificity for rat cTnI, which has 92.8% homology with human cTnI [32-35]. Measurements were made in plasma samples collected at predetermined intervals that had been stored at -80°C.

Plasma cytochrome c measurement

Cytochrome c was measured by reversephase high-performance liquid chromatography (HPLC) as previously reported [28]. Briefly, arterial blood samples (200 µl) were collected into heparinized syringes and centrifuged at 5,000 rpm (2,320 g) for 10 minutes at 4°C (Sorvall Biofuge Stratos, Heraeus, Thermo Fisher Scientific Inc., Bannockburn, IL). Plasma was collected and frozen at -80°C until analysis by HPLC. Absorbance was measured at 393 nm. Standard curves (0.2-20 µg/ml) were prepared using rat heart cytochrome c dissolved in plasma from sham rats and used for quantification.

Mitochondrial NAD⁺ measurement

Mitochondrial NAD⁺ was measured according to Correa et al [23]. Briefly, hearts were harvested at the end of the post-resuscitation monitoring period and mitochondria were isolated as described above for measuring [³H] activity. Samples were treated with perchloric acid to release NAD⁺ and then neutralized with potassium hydroxide (KOH). The amount of NAD⁺ was determined fluorimetrically by measuring the amount of NADH generated in the NAD⁺ dependent lactate dehydrogenase (LDH) reaction shown below (vi) at excitation 340 nm and emission 460 nm using a microplate reader (Gemini XPS, Molecular Devices, Sunnyvale, CA). The reaction mixture contained (in mM) lactate 2, hydrazine sulfate 300, glycine 870, EDTA 0.19, and LDH 1.2 U, at pH 9.5. The hydrazine was used to drive lactate oxidation forward to completion by removing pyruvate. NAD⁺ was expressed as pmol/mg mitochondrial protein.

Chemicals

[³H]DOG was purchased from MP Biomedicals. Mannitol, sucrose, HEPES, EGTA, EDTA, BSA, oxaloacetate (OAA), hydrazine sulfate, glycine, EDTA, pyruvate, lactate dehydro-genase, NADH, and lactate were purchased from Sigma-Aldrich, St. Louis, MO. 2-Amino-2-(hydroxyme-thyl)-1, 3-propanediol, hydrochloride (Tris-HCl), acetyl coenzyme A (acetyl-CoA), 5, 5’-dithiobis-(2-nitrobenzoic acid) (DTNB), Acetonitrile (ACN), trifluoroacetic (TFA) acid, potassium hydroxide (KOH), lactic acid, and perchloric acid were purchased from Thermo Fisher Scientific Inc., Bannockburn, IL. Bradford dye reagent was purchased from Bio-Rad Laboratories, Hercules, CA.

Preparatory experiments

A 2.5 nmol/kg [³H]DOG dose was chosen to achieve a plasma concentration similar to that used by others measuring glucose metabolism in rats without causing adverse metabolic effects [36]. The achieved plasma [³H]DOG concentration has been shown adequate to load the heart [37] given that glucose transporters favor deoxyglucose over glucose. A formulation of [³H]DOG with a specific activity of 40 Ci/mmol (27088S, MP Biomedicals, Santa Ana, CA) was used, delivering 1.25 nmoles (50 μCi) dissolved in 200 µl of 0.9% saline as bolus dose into the right atrium (~100 μCi/kg or 2.5 nmol/kg). In one preparatory experiment, we serially measured arterial plasma [³H] activity during spontaneous circulation at 2, 4, 8, 16, and 32 minutes after injection. As shown in Figure 1, plasma [³H] levels decreased over time following a mono-exponential decay function with a half-life of 18.6 minutes attaining a [³H] plasma activity of ~1.3×10⁵ dpm/ml at 30 minutes post injection. No adverse hemodynamic or functional effects were observed during this interval. The [³H]DOG dose used was significantly below the dose reported to affect cell glucose transport (~10 mM) [14,36] but sufficient to load the heart given the exposure for 30 minutes to a concentration enabling facilitated diffusion to occur. Moreover, the subsequent LV tissue analysis in the formal experiments demonstrated quantifiable intracellular [³H] activity.

Figure 1.

Plasma levels after administration of 50 µCi [³H]DOG into the right atrium. dpm = disintegrations per minute.

Experimental series

Series 1

A total of 26 rats (456 to 543 g) received a bolus of 50 μCi [³H]DOG at baseline 30 minutes before the planned induction of VF and were randomized to have their hearts removed immediately before inducing VF (control group; n=5), at the end of untreated VF (n=5), at the end of chest compression (n=6), at 60 minutes post-resuscitation (n=5), or at 240 minutes post-resuscitation (n=5) (Figure 2).

Figure 2.

Diagram depicting the experimental protocol. A: Series 1. In vivo measurement of mPTP opening using [³H]DOG. Black arrows represent the times at which the hearts were harvested. B: Series 2. In vivo effect of cyclosporine A during cardiac resuscitation.

Series 2

A total of 18 rats (470 to 516 g) were randomized to receive a bolus of cyclosporine A (10 mg/kg) [39-42] five minutes before inducing VF (n=6), immediately before starting chest compression (n=6), or to receive equal volume vehicle (cremophor EL) control before inducing VF (n=3) or before starting chest compression (n=3) with the investigators blind to the treatment assignment. Resuscitated rats were monitored for up to 360 minutes. Four rats undergoing the same surgical preparation but without inducing VF were monitored for 360 minutes and served as a sham group controlling for the effects of instrumentation and prolonged anesthesia.

Statistical methods

SigmaPlot 11.0 (Systat Software, Point Richmond, CA) was used for statistical analysis. Differences in continuous variables among groups were analyzed by one-way ANOVA using the Holm-Sidak’s method for multiple comparisons if overall differences were detected. Differences in continuous variables within groups were analyzed by repeated-measures one-way ANOVA applying Holm-Sidak’s correction for multiple comparisons relative to baseline. Alternative nonparametric tests were used if the data failed tests for normality or equal variance. Non-linear regression analysis was performed between variables of interest based. The data was presented as mean ± SD in the text and tables and mean ± SEM in figures. A two-tail p value<0.05 was considered significant.

Results

Series 1

Hemodynamics and LV function

The administration of [³H]DOG into the right atrium had no effects on hemodynamic or LV function (Table 1). In addition, the baseline measurements were comparable among the various experimental groups (Table 2). During chest compression, adjustments of compression depth successfully maintained an average coronary perfusion pressure above 22 mmHg between minute 2 and minute 8. After return of spontaneous circulation, rats developed characteristic post-resuscitation myocardial dysfunction evidenced by persistently low mean aortic pressure, left ventricular stroke work index, and cardiac index (Table 2). Myocardial dysfunction accounted for three rats dying before post-resuscitation minute 240 yielding an average survival time in the 240 minute group of 202 ± 57 minutes.

Table 1.

Hemodynamic and left ventricular effects of [³H]DOG injection

	MAP, mmHg	HR, beats/min	CI, ml/min∙kg-1	SVRI, mmHg/ml∙min-1∙kg-1	LVSWI, gm-m/kg∙beat-1
Before	132 ± 11	324 ± 25	143 ± 15	0.92 ± 0.12	0.91 ± 0.15
After	136 ± 11	337 ± 25	144 ± 17	0.95 ± 0.13	0.92 ± 0.15

Measurements were obtained at baseline before and 30 minutes after bolus injection of [³H]DOG (1.25 nmoles) into the right atrium in the 26 rats used for measuring mPTP opening. No differences were observed by one-way repeated-measures ANOVA. [³H]DOG, Tritium-labeled 2-deoxyglucose; MAP, Mean aortic pressure; HR; Heart rate; CI, Cardiac index; SVRI, Systemic vascular resistance index; LVSWI, Left ventricular stroke work index. Mean ± SD.

Table 2.

Hemodynamic and left ventricular function

	Baseline	VF	Chest Compression		Post-Resuscitation
	0 min	10 min	4 min	8 min	15 min	60 min	240 min
Mean Aortic Pressure, mmHg
Baseline	132 ± 11
VF	143 ± 3	7 ± 1
CC	143 ± 11	7 ± 2	37 ± 3	38 ± 3
PR-60	132 ± 12	8 ± 1	38 ± 2	37 ± 4	83 ± 10†	95 ± 13†
PR-240	129 ± 12	7 ± 1	37 ± 2	39 ± 3	66 ± 20†	94 ± 15†	119 ± 4 [2]
Cardiac Index, ml/min∙kg-1
Baseline	149 ± 18
VF	136 ± 20
CC	143 ± 11
PR-60	146 ± 20				105 ± 46	67 ± 20*
PR-240	147 ± 18				118 ± 25*	80 ± 13‡	62 ± 6 [2] ‡
Systemic Vascular Resistance Index, mmHg/ml∙min-1∙kg-1
Baseline	0.89 ± 0.08
VF	1.05 ± 0.15
CC	1.00 ± 0.12
PR-60	0.91 ± 0.12				0.87 ± 0.35	1.47± 0.47
PR-240	0.88 ± 0.11				0.54 ± 0.25*	1.17 ± 0.24*	1.90 ± 0.13 [2]†
Left Ventricular Stroke Work Index, gm-m/kg∙beat-1
Baseline	0.91 ± 0.21
VF	0.93 ± 0.16
CC	0.93 ± 0.13
PR-60	0.93 ± 0.15				0.56 ± 0.20‡	0.30 ± 0.12†
PR-240	0.92 ± 0.14				0.62 ± 0.29*	0.40 ± 0.11*	0.31 ± 0.00 [2]*
Coronary Perfusion Pressure, mmHg
Baseline	112 ± 11
VF	124 ± 4	2 ± 1
CC	125 ± 13	1 ± 0	23 ± 1	24 ± 1
PR-60	115 ± 8	2 ± 1	24 ± 1	23 ± 2	69 ± 14†	84 ± 12*
PR-240	112 ± 10	2 ± 1	24 ± 1	24 ± 1	50 ± 21‡	83 ± 16*	106 ± 2 [2]
Coronary Perfusion Pressure/Depth Ratio (mmHg/mm)
Baseline
VF
CC			1.67 ± 0.21	1.57 ± 0.29
PR-60			1.68 ± 0.28	1.51± 0.32
PR-240			1.68 ± 0.06	1.53 ± 0.15

Comparison within each of the 5 groups (Baseline, n=5; VF, n=5; CC, n=6; PR-60, n=5; PR-240, n=5) over the various time events. Numbers in brackets indicate when the sample size decreased from the preceding sample size. VF, Ventricular fibrillation; CC, Chest compression; PR, Post-resuscitation. Mean ± SD.

^* P<0.05;

^‡ P<0.005;

^† P<0.001.

PR-15, PR-60 and PR-240 vs baseline by one-way repeated-measures ANOVA using Holm-Sidak method for multiple comparisons.

mPTP opening

Intact mitochondrial yield measured by CS activity in each experimental group was comparable to that of baseline hearts except for mitochondria isolated from hearts during VF which demonstrated less CS activity (170 ± 18 vs 131 ± 7 IU/mg protein, P<0.05). Mitochondrial [³H] activity increased during VF but decreased during chest compression, demonstrating further reduction during the post-resuscitation interval (Figure 3A). The LV_IC [³H] activity remained unchanged during VF but decreased during chest compression (NS) with further reduction post-resuscitation that was statistically significant (Figure 3B). The mPTP opening ratio (Figure 3C) measured after 10 minutes of untreated VF increased relative to baseline but without attaining statistical significance. During chest compression, the mPTP opening ratio significantly increased and remained elevated during the post-resuscitation interval at 60 minutes with further increase at 240 minutes (Figure 3C). Regression analysis assessing the relationship between mPTP opening and both, left ventricular stroke work index and cardiac index, showed a highly significant negative correlation conforming to a mono-exponential decay curve (Figure 4).

Figure 3.

(A) Mitochondrial [³H], (B) left ventricular intracellular [³H] (LV_IC [³H]), and (C) their ratio as a measurement of mPTP opening. Mean ± SEM. *P<0.05, †P<0.01 vs baseline using one-way ANOVA. CS, Citrate synthase; VF, Ventricular fibrillation; CC, Chest compression; PR, Post-resuscitation.

Figure 4.

Non-linear regression analysis (exponential decay, 2 parameter) showing the relationship between mitochondrial permeability transition pore opening (mPTP) and left ventricular stroke work index (LVSWI, left) and cardiac index (CI, right) for measurements obtained at baseline (BL), and post-resuscitation at 60 minutes (PR-60) and at 240 minutes (PR-240).

Series 2

Hemodynamic and LV function

All rats that were treated with cyclosporine A (6 at baseline and 6 during VF) were pooled for the analyses after noticing no differences between subgroups. Similarly, the vehicle treated rats (3 at baseline and 3 during VF) were also pooled. Baseline hemodynamic measurements were comparable between groups (Table 3). Treatment with cyclosporine A had no beneficial effects on myocardial distensibility during chest compression as demonstrated by a comparable (cyclosporine A vs vehicle) depth of compression (13.9 ± 1.0 vs 13.7 ± 0.9 mm) and coronary perfusion pressure to depth of compression ratio (1.7 ± 0.1 vs 1.7 ± 0.2 mmHg/mm) averaged over the 2^nd and 8^th minute of chest compression. All rats were successfully resuscitated with no differences in the energy level of the shock required to terminate VF (cyclosporine A vs vehicle; 1.3 ± 0.6 vs 1.0 ± 0.6 Joules). However, after return of spontaneous circulation, episodes of recurrent VF requiring post-resuscitation electrical shocks were observed only in vehicle-treated rats (0.0 ± 0.0 vs 1.2 ± 1.6 shocks, P<0.05). During the subsequent post-resuscitation interval both groups displayed comparable hemodynamic and LV function (Table 3) with comparable survival times (cyclosporine A vs vehicle; 321 ± 67 minutes vs 331 ± 67 minutes).

Table 3.

Baseline and post-resuscitation hemodynamic and LV function

	Baseline	Post-Resuscitation
	-2 min	15 min	120 min	240 min	360 min
Mean Aortic Pressure, mmHg
Cyclosporine A	141 ± 14	74 ± 21	102 ± 14	110 ± 31 [10]	75 ± 30 [9]
Vehicle	136 ± 11	84 ± 11	105 ± 17	112 ± 14 [5]	97 ± 35
Cardiac Index, ml/min∙kg-1
Cyclosporine A	155 ± 19	100 ± 37	63 ± 15	56 ± 9 [10]	46 ± 15 [9]
Vehicle	154 ± 16	96 ± 33	62 ± 20	58 ± 13 [5]	52 ± 10
Left Ventricular Stroke Work Index, mmHg/ml∙kg-1
Cyclosporine A	1.08 ± 0.22	0.58 ± 0.24	0.31 ± 0.10	0.33 ± 0.12 [10]	0.21 ± 0.12 [9]
Vehicle	0.97 ± 0.17	0.54 ± 0.19	0.31 ± 0.13	0.29 ± 0.08 [5]	0.23 ± 0.02
+dP/dtmax, mmHg·sec-1
Cyclosporine A	3964 ± 826	2695 ± 389	2959 ± 554	3138 ± 1042 [10]	2057 ± 971 [9]
Vehicle	3416 ± 681	2504 ± 472	2986 ± 373	3107 ± 232 [5]	2536 ± 803

N=12 in cyclosporine A group and 6 in vehicle treated rats. Numbers in brackets indicate when sample size decreased from the initial or from the preceding sample size. There were no statistically significant differences between groups. Mean ± SD.

Myocardial injury

Treatment with cyclosporine A did not attenuate the release of cytochrome c into the bloodstream (Figure 5). Plasma cTnI levels were negligible in sham rats but increased after VF and resuscitation with no differences in levels between cyclosporine A and vehicle treated rats. No difference in NAD⁺ levels was observed between cyclosporine A and vehicle treated rats. The pooled NAD⁺ levels of both treatment groups that underwent VF and resuscitation were ~34% lower than sham rats (P<0.001).

Figure 5.

Levels of mitochondrial NAD⁺, plasma cytochrome c, and plasma cardiac troponin I in sham rats (open bars), and after treatment with vehicle (gray bar) or cyclosporine A (black bars) at 360 minutes post-resuscitation or before death in rats subjected to VF and resuscitation. Mean ± SEM. †P<0.001 vs baseline.

Discussion

We have documented in vivo opening of the mPTP in rat hearts while undergoing VF-induced cardiac arrest and closed-chest resuscitation, confirming previous in vitro and ex vivo observations showing mPTP opening upon reoxygenation [43,18] or reperfusion [14,9]. We also documented failure of cyclosporine A to prevent mPTP opening based on mitochondrial NAD⁺ levels and failure to ameliorate mitochondrial and myocardial injury and thereby improve short-term survival.

Our method for in vivo measurement of mPTP opening circumvented technical and physiological shortcomings of in vitro or ex vivo mPTP measurements and enabled for the first time to document mPTP opening during resuscitation from cardiac arrest. With our approach, the key initial step in which [³H]DOG-6-phosphate enters the mitochondria upon mPTP opening occurred in the intact in situ heart. In addition, in isolated rat heart experiments, the coronary circuit is flushed to remove extracellular [³H] before tissue processing. We eliminated this step to minimize processing time and technical confounders and instead calculated and excluded the extracellular [³H] contribution to the total tissue [³H] by concomitant measurement of plasma [³H] and calculating the intracellular [³H] as shown in equation (i). The subsequent key step, in which the uptaken [³H]DOG-6-phosphate is trapped in mitochondria by calcium chelation during tissue processing and the ensuing additional steps required for mitochondrial isolation and measurement of the ratio between mitochondrial and intracellular [³H] were the same as previously reported for the isolated heart [27].

Our in vivo approach included the influence of the neuroendocrine stress response to cardiac arrest and resuscitation. As part of this response, there is an intense adrenergic surge with increased circulating catecholamines that also contributes to myocardial injury during cardiac resuscitation [44-46]. Increased catecholamine levels can augment myocardial injury by promoting intracellular Ca²⁺ entry and reduce the threshold for mPTP opening [47].

The chronology of mPTP opening in the present study suggests that mPTP opening could have started during the interval of untreated VF when we observed an increase in the mPTP opening ratio (~60% higher than baseline) but without attaining statistical significance. Consistent with the concept that reperfusion after ischemia triggers mPTP opening, gasping that typically occurs after onset of cardiac arrest may generate blood flow in the absence of resuscitation efforts [48,49]. With the start of chest compression, a statistically significant increase in the mPTP opening ratio was observed that essentially doubled baseline values. The myocardial blood flow during chest compression in our rat model corresponds to ~50% of baseline [50]. The mPTP opening ratio remained high post-resuscitation attaining a level 2.5 fold baseline at 240 minutes, suggesting the possibility of additional opening upon further reperfusion during the post-resuscitation phase.

In previous studies using the same rat model of VF and closed-chest resuscitation, we reported release of cytochrome c to the bloodstream detected during chest compression with further increase post-resuscitation, attaining levels that were inversely proportional to myocardial function and survival [28,51,52]. These previous observation are consistent with the present ones as mPTP opening would be expected to injure mitochondria resulting in cytochrome c release [53-55]. Moreover, in the present study we documented an inverse relationship between the mPTP opening ratio and left ventricular function, again pointing to mitochondrial injury being at the center of post-resuscitation myocardial dysfunction, providing a solid rationale for examining the effect of preventing mPTP opening during cardiac resuscitation [56].

Current efforts to pharmacologically prevent (or increase the threshold for) mPTP opening are focused on cyclophilin-D, which is the only recognized integral protein of the mPTP [2,57]. Cyclosporine A inhibits cyclophilin-D activity increasing the threshold for mPTP opening [58-61]. Through this mechanism, cyclosporine A has been shown to be protective in isolated cardiomyocytes [16], in situ perfused hearts [14], and in intact animal models of myocardial infarction [62,63,42] and cardiac arrest [64,65]. Yet, in the present study cyclosporine A failed to attenuate post-resuscitation myocardial injury with indirect evidence suggesting that it failed to prevent mPTP opening based on mitochondria NAD⁺ levels [66]. mPTP opening allows efflux of NAD⁺ from the mitochondrial matrix such that mitochondrial NAD⁺ levels may be used as indirect marker of mPTP opening. The reduction in mitochondrial NAD⁺ after cardiac arrest in our study was ~34% and similar to the 30% reduction reported in isolated rat hearts subjected to global ischemia by Di Lisa et al [66]. However, in contrast to our study, the administration of cyclosporine A in the study by Di Lisa et al attenuated reperfusion injury and preserved mitochondrial NAD⁺.

The failure of cyclosporine A to prevent mPTP opening and ameliorate reperfusion injury in our study was intriguing at the time but consistent with a recent large multicenter, double-blind, randomized clinical trial by Cung et al in which 2.5 mg/kg of cyclosporine A delivered to patients with acute anterior ST-segment elevation myocardial infarction undergoing percutaneous coronary intervention failed to elicit protection from reperfusion injury [67].

There are several possible explanations. Cyclosporine A has a narrow therapeutic index [68] and dosing to achieve the intended effect if challenging. It was also shown by Griffith and Halestrap in isolated cardiac mitochondria that pore opening could become CsA-insensitive under conditions of adenine nucleotide depletion and high matrix [Ca²⁺] [14], conditions which are present during ischemia [30,69].Cyclosporine A has also been reported to induce production of reactive oxygen species and increase cytosolic Ca²⁺ levels potentially counteracting its beneficial effects during ischemia and reperfusion [17,70]. Moreover, the lipid emulsifying vehicle, cremophor, can accumulate in the inner mitochondrial membrane decreasing the oxidative capacity of respiratory chain complexes I and IV [71-73].

Despite the lack of myocardial protection, cyclosporine A in our study prevented episodes of recurrent VF during the early post-resuscitation period. A similar effect was reported in rats following reperfusion after coronary occlusion [74].

Extrapolation of the present findings to clinical settings is limited by several factors. VF was induced by electrical stimulation, whereas clinically VF typically occurs in patients with underlying coronary artery disease and is often precipitated by coronary occlusion. We cannot exclude that alterations in normal mitochondrial function occurred consequent to electrical stimulation. Inherent to animal models is the use of anesthesia, which exerts independent myocardial protective effects [75]. Despite these limitations, the rat model has been a highly effective translation model for understanding underlying cellular mechanisms during whole body ischemia.

In summary, this is the first study to report in vivo mPTP opening during VF and closed-chest resuscitation. These results are consistent with previous observations in isolated mitochondria, cells, and organs in which the mPTP opens during reperfusion. Cyclosporine A did not appear to limit mPTP opening or attenuate post-resuscitation myocardial dysfunction. Further studies are required to determine the therapeutic range and optimal formulation of cyclosporine A during whole body ischemia and reperfusion, or design alternative strategies to increase the threshold for mPTP opening during cardiac resuscitation.

Disclosure of conflict of interest

None.