Assessing mitochondrial respiration in isolated hearts using ¹⁷O MRS

Abstract

The application of ¹⁷O MRI and MRS for the evaluation of cardiac mitochondrial function has been limited because of the challenge of detecting metabolic ${\mathrm{H}}_2^{17}$ in the vast background of naturally abundant ${\mathrm{H}}_2^{17}$. In this study, we have developed a direct ¹⁷O MRS approach to examine the feasibility and sensitivity of detecting metabolically produced ${\mathrm{H}}_2^{17}$ in isolated rat hearts perfused with ¹⁷O₂-enriched Krebs-Henseleit buffer containing normal (1.5 mM) and high (2.5 mm) calcium (Ca²⁺) concentrations to induce high workload. Consistent with increased workload at high Ca²⁺ concentration, the measured myocardial oxygen consumption rate (MVO₂) increased by 82%. Dynamic ¹⁷O MRS showed an accelerated increase in the ${\mathrm{H}}_2^{17}\mathrm{O}$ signal at high Ca²⁺ concentration, suggesting increased mitochondrial production of ${\mathrm{H}}_2^{17}\mathrm{O}$ in concordance with the increased workload. A compartment model was developed to describe the kinetics of ${\mathrm{H}}_2^{17}\mathrm{O}$ production as a function of MVO₂. The myocardial ¹⁷O₂ consumption rate (MV¹⁷O₂) was determined by least-squares fitting of the model to the NMR-measured ${\mathrm{H}}_2^{17}\mathrm{O}$ concentration. Consistent with the measured MVO₂, the model-determined MV¹⁷O₂ showed a 92% increase at high Ca²⁺ concentration. The increase in metabolic activity at high workload allowed the balance between ATP production and utilization to be maintained, leading to a similar phosphocreatine to ATP ratio. These results demonstrate that dynamic ¹⁷O MRS can provide a valuable tool for the detection of an altered metabolic rate associated with a change in cardiac workload.

INTRODUCTION

The heart is the most energy-demanding organ because of its nonstop contraction. There is increasing evidence that abnormal mitochondrial function and energy metabolism are characteristic of several prevalent cardiac diseases, such as heart failure (1–3), and ischemic and diabetic cardiomyopathy (4–6). However, our fundamental understanding of cardiac energetics in diseased hearts remains incomplete because of the limited means to investigate myocardial metabolism in intact hearts. To date, most of the studies on mitochondrial function have been performed in isolated mitochondria or with in vitro tissue samples. As a result of the lack of nondestructive methods to quantify mitochondrial respiration in vivo, an integrative understanding of the mechanisms underlying metabolic dysfunction in diseased hearts is still incomplete.

In the last 20 years, MRS has become a valuable tool for the evaluation of cardiac metabolism under various pathophysiological conditions, largely because of its noninvasive nature and the exquisitely detailed information it can offer (7,8). While ¹³C and ³¹P MRS allow the quantification of substrate metabolism and ATP synthesis, ¹⁷O MRS provides the unique opportunity to evaluate mitochondrial respiration in intact, viable tissues through the observation of metabolic H₂¹⁷O production via the reduction of ¹⁷O₂. The use of ¹⁷O MRS for the assessment of mitochondrial H₂¹⁷O production was first demonstrated, independently, in 1991 by Mateescu et al. (9) in mouse brain and Pekar et al. (10) in cat brain. Recently, ¹⁷O MRS methods for the quantification of cerebral oxygen consumption have been developed in animals through the monitoring of dynamic changes of metabolically generated H₂¹⁷O from inhaled ¹⁷O₂ (11–16). However, cardiac applications of ¹⁷O MRS have been limited. Only a few studies have explored the potential of detecting mitochondrial function in the heart with ¹⁷O MR techniques (17–20).

Current in vivo ¹⁷O MRS methods are complicated by several factors that can potentially affect the accuracy of the measurement of mitochondrial respiration. These include the efflux of metabolic H₂¹⁷O via normal tissue water exchange and the input of H₂¹⁷O from other metabolizing tissues. These issues can be addressed using an isolated perfused heart model. The advantages of using an isolated heart model are as follows: (i) it allows precise control of the substrates and other experimental conditions; (ii) it enables easy manipulation of mitochondrial respiration via changes in cardiac workload; (iii) it allows the simultaneous recording of cardiac function to investigate the energy–function relationship; and (iv) it eliminates systemic ${\mathrm{H}}_2^{17}\mathrm{O}$ input from other metabolizing organs. Therefore, ¹⁷O MRS on isolated perfused hearts provides the opportunity to validate quantitative ¹⁷O MRS methods for the accurate assessment of mitochondrial respiration in living tissues. These ex vivo MRS data may serve as the benchmarks for in vivo studies.

In this study, we aimed to develop a ¹⁷O MRS method on isolated perfused hearts to examine the feasibility and sensitivity of ¹⁷O MRS for the detection of metabolically produced H₂¹⁷O in hearts at different workloads. A closed-loop perfusion system specifically designed for economic and efficient ¹⁷O MRS studies was developed. Metabolically produced H₂¹⁷O was detected by rapid, dynamic ¹⁷O MRS on a 9.4-T vertical-bore spectrometer. A compartment model that describes the kinetics of H₂¹⁷O production was also developed for the estimation of the myocardial oxygen consumption rate (MVO₂). Our results demonstrate that a high-temporal-resolution ¹⁷O MRS method provides an opportunity to detect an altered mitochondrial respiration rate associated with a change in cardiac workload.

MATERIALS AND METHODS

Perfusion system

A major challenge in performing ¹⁷O MRS experiments on perfused hearts is the large system volume that may reduce the detectability of metabolic H₂¹⁷O because of the dilution of metabolic H₂¹⁷O by natural abundance ${\mathrm{H}}_2^{17}\mathrm{O}$. To improve the detection sensitivity, we developed a closed-loop perfusion system with a total volume of only 26 mL (Fig. 1A). The perfusion system has two gravitational reservoirs for ¹⁷O₂- and ¹⁶O₂-oxygenated perfusate, respectively. The perfusion pressure was maintained at a constant level by adjusting the level of the overflow port in the reservoir. Re-oxygenation of the recycling perfusate was achieved via ¹⁷O₂-filled silicone tubing inserted into the perfusion line.

Figure 1.

(A) Schematics of the perfusion system. Two gravitational reservoirs with overflow ports supply the heart with ¹⁶O₂- and ¹⁷O₂-oxygenated buffer. Re-oxygenation of the buffer during ¹⁷O MRS acquisition is accomplished via ¹⁷O₂-filled silicone tubing embedded in the water-jacketed perfusion column. Pressurized chamber (B) and closed-loop gas-exchange system (C) for efficient oxygenation of the perfusate with ¹⁷O₂.

Buffer preparation

The formidable cost of ¹⁷O₂ requires a low-loss system for efficient oxygenation of the perfusate with ¹⁷O₂. Two approaches were attempted in the current study—a pressurized chamber (Fig. 1B) and a gas-exchange circuit (Fig. 1 C)—with the goal of maximizing ¹⁷O₂ usage for each experiment. In the gas-exchange system, oxygenation with ¹⁷O₂ was achieved by circulating the perfusate through a cartridge of ¹⁷O₂-filled silicone tubing. The high gas permeability of the silicone tubing allowed the efficient transfer of ¹⁷O₂ to the perfusate. In the second approach, we used a pressurized chamber for buffer oxygenation. On introduction of ¹⁷O₂ (70% enrichment; Rockland Technimed Ltd., Airmont, NY, USA), the pressure of the chamber was increased to about 8 psi and a magnetic stirrer operating at 700 rpm allowed the gas to quickly dissolve into the perfusate. To maximize the ¹⁷O₂ content in the perfusate, the perfusate was degassed at atmospheric pressure by bubbling helium gas to reduce the content of ¹⁶O₂ prior to oxygenation with ¹⁷O₂.

Perfusion protocol

Isolated hearts from 2-month-old male Sprague-Dawley rats were perfused with oxygenated Krebs-Henseleit buffer containing (in mm): NaCl, 118; NaHCO₃, 25; KCl, 5.3; MgSO₄, 1.2; CaCl₂, 1.5 or 2.5; glucose, 10. Two different calcium (Ca²⁺) concentrations, normal at 1.5 mm (NCA, n = 7) and high at 2.5 mm (HCA, n = 7), were used to induce normal and high workloads, respectively. The left ventricular developed pressure (LVDP) and heart rate (HR) were measured via a water-filled balloon connected to a pressure transducer (ADInstruments, Colorado Springs, CO, USA). The rate–pressure product (RPP) was used as an index of ventricular workload.

At the start of the perfusion protocol, ¹⁶O₂-oxygenated perfusate was supplied for 15 min to establish metabolic equilibrium. A ³¹P spectrum was collected to verify the metabolic status of the heart. A natural abundance ¹⁷O spectrum was then collected for background subtraction. The perfusate was then switched to ¹⁷O₂-oxygenated buffer. Dynamic acquisition of ¹⁷O spectra began immediately for 5 min. To reduce the negative inotropic effect caused by adenosine accumulation, theophylline (Sigma-Aldrich, St. Louis, MO, USA), an adenosine antagonist, was infused at 0.5 mL/min once the perfusate was switched to ¹⁷O₂-oxygenated buffer. At the end of the perfusion protocol, hearts were taken off the perfusion column, blotted dry and weighed.

To validate the ¹⁷O MRS results, the oxygen content (pO₂)of the perfusion medium in the supply line and in coronary effluent was measured with a blood-gas analyzer (Radiometer ApS, Copenhagen, Denmark). MVO₂ was calculated experimentally from the measured coronary flow rate and the difference in pO₂ between the perfusate and the coronary effluent (21). These measurements were performed during the acquisition of the ³¹P spectrum. All procedures were approved by the Institutional Animal Care and Use Committee at Case Western Reserve University.

³¹P and dynamic ¹⁷O MRS

³¹P and ¹⁷O MR spectra were acquired on a Bruker AVANCE-400 9.4-T vertical-bore spectrometer (Bruker Biospin Co., Billerica, MA, USA) with a 20-mm broad-band probe. Field homogeneity was adjusted by shimming on the proton signal to a linewidth of 30–50 Hz. ³¹P spectra were acquired at 162 MHz with 90° radiofrequency pulses, TR = 1.3 s, 12 K dataset and 64 averages. Spectrum analysis was performed using NMR-dedicated software MestRec (Mestrelab Research SL, Santiago de Compostela, Spain). Raw free induction decay signals were converted into frequency domain spectra by Fourier transformation following the application of an exponential filter (15-Hz line broadening) to enhance the signal-to-noise ratio. Relative ATP and phosphocreatine (PCr) levels were quantified by fitting the β-ATP and PCr peaks with Lorentzian curves, respectively, followed by the integration of these two peak areas. The PCr to ATP ratio was calculated as an index of the high energetic status of the heart.

Sequential ¹⁷O spectra were acquired at 54 MHz with 90° radiofrequency pulses, TR = 300 ms, 1 K dataset and 64 averages. The temporal resolution for ¹⁷O acquisition was 20 s. For spectrum quantification, the natural abundance ¹⁷O signal was digitally subtracted. The raw MRS signal was processed by exponential filtering with a line broadening of 30 Hz, followed by Fourier transformation to obtain frequency domain spectra. H₂¹⁷O was quantified by fitting the resonance peaks with Lorentzian curves, followed by the integration of the peak area. Assuming an average water concentration of 55 M in the whole NMR tube, the H₂¹⁷O concentration was calculated using the natural abundance ¹⁷O spectrum as the reference.

Kinetic analysis

To quantify the mitochondrial respiration rate from ¹⁷O MRS data, we developed a compartment model to describe the transport and production of metabolic H₂¹⁷O during the perfusion protocol. The model included three compartments: the heart; the fluid in the NMR tube that surrounded the heart; and the recirculating perfusate (Fig. 2).

Figure 2.

Three-compartment model of ${\mathrm{H}}_2^{17}\mathrm{O}$ generation and recirculation for the estimation of mitochondrial respiration from ¹⁷O MRS data. FL, coronary flow rate; HW, wet heart weight.

According to the mass balance principle, the H₂¹⁷O content in the heart (C_h; μmol/g wet weight) is governed by:

$$\mathrm{HW}\cdot \frac{{\mathrm{dC}}_h}{\mathrm{dt}}=2{\mathrm{MV}}^{17}{\mathrm{O}}_2\cdot \mathrm{HW}+\mathrm{FL}\cdot C_r-\mathrm{FL}\cdot \frac{C_h}{\mathrm{p}}$$ [1]

where HW is the wet heart weight (g), MV¹⁷O₂ is the myocardial ¹⁷O₂ consumption rate (μmol/min/g wet weight), FL is the coronary flow rate (mL/min) measured during the experiment, C_r is the H₂¹⁷O concentration in the recirculating perfusate (μmol/mL) and p is the myocardium–blood partition coefficient of water. A myocardium–blood partition coefficient of 0.91 mL/g wet weight was used in the current study (22).

The mass balance of the H₂¹⁷O content in the NMR tube (C_t; μmol/mL) can be described as:

$$V_t\frac{{\mathrm{dC}}_t}{\mathrm{dt}}=\mathrm{FL}\left(\frac{C_h}{\mathrm{p}}-C_t\right)$$ [2]

where V_t is the fluid volume in the NMR tube. V_t was maintained at 6 mL in the current study.

Similarly, the change in H₂¹⁷O content in the recirculating perfusate (C_r; μmol/mL) can be described as:

$$V_r\frac{{\mathrm{dC}}_r}{\mathrm{dt}}=\mathrm{FL}(C_t-C_r)$$ [3]

where V_r is the volume of the circulating perfusate. The initial perfusate volume, including both the reservoir and the tubing, was 20 mL. With the infusion of theophylline at 0.5 mL/min, the total recirculating perfusate (V_r) can be described by the following equation:

$$V_r=20+0.5\mathrm{t}$$

The model output was the NMR signal arising from both the heart and the fluid in the NMR tube, i.e.:

$$S_{NMR}=\mathrm{HW}\cdot C_h+V_t{C}_t$$ [4]

The only unknown parameter in Equations [1–3] is the myocardial ¹⁷O₂ consumption rate (MV¹⁷O₂), which was determined from the least-squares fitting of the model output (Equation [4]) to the NMR-measured metabolic H₂¹⁷O accumulation.

Statistical analysis

All data are presented as means±standard deviation. Comparisons of datasets were performed using unpaired Student's t-test. Differences in mean values were considered to be significant at a probability level of <5%.

RESULTS

Perfusate oxygenation with ¹⁷O₂ gas

Both the closed-loop system and the pressurized chamber allowed efficient oxygenation of the buffer with ¹⁷O₂ gas. With the closed-loop system, the pO₂ of 40 mL of perfusate was increased to >400 mmHg within 2 min of circulation and at the cost of only 30 mL of ¹⁷O₂ gas. The efficiency of the pressurized chamber was slightly higher, with the pO₂ of 50 mL of perfusate increased to >500 mmHg within 2 min and at the cost of 30 mL of ¹⁷O₂ gas.

Contractile function

The body weight and heart weight were similar between the two experimental groups. The average heart weights were 1.71±0.27 g and 1.91±0.14 g for the NCA and HCA groups, respectively (p = not significant). The average HR was also similar between the two groups: 339±43 beats/min for the NCA group and 321±24 beats/min for the HCA group (p = not significant). The HR remained constant during data acquisition (Fig. 3A).

Figure 3.

Contractile function during ¹⁷O₂-enriched buffer perfusion. (A) Heart rate. (B) Left ventricular developed pressure (LVDP). (C) Rate–pressure product (RPP). *p<0.05. HCA, high calcium concentration group; NCA, normal calcium concentration group.

LVDP decreased slightly from 59±24 mmHg at the beginning of ¹⁷O perfusion to 48±23 mmHg at the end of 5 min of ¹⁷O MRS data acquisition in the NCA group. This reduction in LVDP was even more pronounced in the HCA group: from 92±25 mmHg to 59±26 mmHg (Fig. 3B). As a result, RPP changed from 16 800±5900 mmHg/min to 14 300±5400 mmHg/min for the NCA group, and from 28 200±9400 mmHg/min to 19 400±8700 mmHg/min for the HCA group (Fig. 3 C). LVDP and RPP were statistically higher in the HCA group for the first 2 min of ¹⁷O MRS data acquisition.

³¹P MRS and measured oxygen consumption rate

A representative ³¹P spectrum is shown in Fig. 4A. Both groups showed normal and similar PCr to ATP ratios: 1.49±0.16 for the NCA group and 1.45±0.20 for the HCA group (Fig. 4B, p = not significant). The measured MVO₂ in the HCA group (6.23±0.19 μmol/min/g wet weight) increased by 82% relative to that of the NCA group (3.42±0.50 μmol/min/g wet weight; p < 0.05; Fig. 4 C).

Figure 4.

(A) Representative ³¹P NMR spectrum. (B) Phosphocreatine (PCr) to ATP ratio. (C) Measured oxygen consumption rate (MVO₂). HCA, high calcium concentration group; NCA, normal calcium concentration group.

Dynamic ¹⁷O MRS and model-fitted ¹⁷O₂ consumption rate

Representative dynamic ¹⁷O spectra are shown in Fig. 5A. Both groups showed a progressive increase in the H₂¹⁷O resonance peak (Fig. 5B), suggesting steady production of mitochondrial H₂¹⁷O. As shown in Fig. 5B, hearts perfused with higher Ca²⁺ concentration demonstrated an accelerated increase in the H₂¹⁷O signal.

Figure 5.

Production of mitochondrial ${\mathrm{H}}_2^{17}\mathrm{O}$ measured by ¹⁷O MRS. (A) Representative dynamic ¹⁷O spectra. (B) Average time course of ${\mathrm{H}}_2^{17}\mathrm{O}$ accumulation and results of kinetic analysis. HCA, high calcium concentration group; NCA, normal calcium concentration group.

Model-fitted ${\mathrm{H}}_2^{17}\mathrm{O}$ curves are also shown in Fig. 5B. The myocardial ¹⁷O₂ consumption rate (MV¹⁷O₂), determined from the least-squares fitting of the model to group-averaged MRS data, increased from 1.15 μmol/min/g wet weight in the NCA group to 2.21 μmol/min/g wet weight in the HCA group, a 92% increase, which is comparable with the overall increase in MVO₂ (82%).

DISCUSSION

Oxygen utilization in the mitochondrial respiratory chain is an important index of cardiac metabolic status. Although ¹⁵O positron emission tomography (PET) has been the standard imaging modality for the assessment of oxygen consumption for more than 20 years, its application has been limited (23). In ¹⁵O PET, all ¹⁵O-labeled oxygen contributes to the measured signal, regardless of its chemical form. Hence, data interpretation is complicated, and requires complex compartmental modeling to determine the oxygen consumption rate. Frequently, a multitracer study is necessary to account for blood volume and blood flow in modeling (22). By contrast, ¹⁷O MRS provides specificity by measuring only nascent mitochondrial H₂¹⁷O generated from the reduction of ¹⁷O₂. Therefore, it allows more direct and straightforward quantification of the mitochondrial respiration rate.

Previously, in vivo cardiac ¹⁷O MRS/I studies have been reported which used either ¹⁷O spectroscopic imaging for the direct detection of H₂¹⁷O (19), or indirect H₂¹⁷O detection via T₂ or T_1ρ imaging (17,20). These studies demonstrated the feasibility of ¹⁷O MRS/I for the quantification of the oxygen consumption rate in the heart. Here, we report the first ¹⁷O MRS study on isolated perfused hearts. The use of an isolated perfused heart model eliminates the complications arising from the systemic circulation. It also allows better control of the physiological conditions and cardiac function. Such simplification can provide valuable benchmarks for future in vivo development.

Oxygen demand in the heart is primarily governed by its energy expenditure, i.e. workload. In the current study, the cardiac workload was altered by changing the Ca²⁺ concentration in the perfusate. A higher Ca²⁺ level caused a higher Ca²⁺ influx during contraction, leading to increased contractile force, but similar HR (Fig. 3). As a result, RPP was increased significantly in the HCA group. This increase in cardiac workload must be fueled by a corresponding increase in ATP generation. The ¹⁷O MRS-observed increase in the kinetics of H₂¹⁷O generation reflected an increase in the mitochondrial respiration rate (Fig. 5), which is coupled to ATP synthesis under normal physiological conditions. In addition, the increase in the measured oxygen consumption rate provides further evidence of increased mitochondrial respiration (Fig. 4). This increase in metabolic activity allowed the balance between ATP production and utilization to be maintained, as demonstrated by the ³¹P MRS data (Fig.4).

Several studies have reported a linear relationship between oxygen consumption and workload (RPP) both in vivo and ex vivo (24,25). In the current study, the ratios of oxygen consumption between normal- and high-workload hearts were 1.82 and 1.92, measured by the analytical method and by ¹⁷O MRS, respectively. The ratio of RPP between the two groups was 1.68 at the beginning of ¹⁷O MRS acquisition and decreased to 1.35 towards the end. These results were similar to those reported by Yamamoto et al.(24) in an in vivo PET ¹⁵O study on dogs. The authors reported ratios of 2.10 and 2.27 between isoproterenol-stimulated and baseline workload for RPP and MVO₂, respectively. The decrease in RPP ratio at the end of ¹⁷O MRS acquisition may be caused by reduced metabolic efficiency because of insufficient re-oxygenation.

A major challenge in performing ¹⁷O MRS studies in isolated hearts is the dilution of mitochondrial H₂¹⁷O, in combination with the low sensitivity of ¹⁷O MRS. This problem is less pronounced for in vivo studies as the total blood volume of a rat is less than 30 mL (26). In addition, the recirculation of metabolic H₂¹⁷O generated in other organs also contributes to the increase in ¹⁷O signal acquired from the heart. By contrast, a perfusion system designed for MRS studies typically has a perfusion volume of more than 1 L. The large system volume can quickly dilute the nascent H₂¹⁷O generated in the mitochondria, rendering the signal increase undetectable. To improve the detection sensitivity, we switched to a smaller system volume of only 26 mL on introduction of ¹⁷O-oxygenated perfusate. As a result, there were elevated levels of metabolic byproducts which may have an impact on the myocardial contractility. Among these metabolic byproducts is adenosine, which is a potent agent that produces a negative inotropic effect in both atrial and ventricular myocardium. To counteract the adenosine effect, theophylline, an adenosine antagonist, was employed to maintain normal ventricular function (27). Although there was still a trend of functional decrease, especially at high Ca²⁺ concentration, it was much less pronounced relative to perfusion without theophylline (data not shown).

To date, the MR detection of H₂¹⁷O has used two approaches, i.e. direct detection via ¹⁷O MRS/I and indirect detection via ¹H MRI. The indirect approach relies on ¹⁷O-induced changes in T_1ρ or T₂. It can be readily implemented on a clinical scanner at low field. However, the absolute quantification of H₂¹⁷O concentration can be complicated because of the sensitivity of the proton relaxation time to changes in the biochemical and physiological milieu in tissues (28). In addition, the relatively high radiofrequency power required by ¹H spin-locking or ¹⁷O decoupling can also limit its application to patients. In comparison, direct ¹⁷O MRS/I has the advantage of being highly specificfor H₂¹⁷O, especially at high magnetic field with significantly improved signal-to-noise ratio. Potentially, the natural abundance H₂¹⁷O signal can be used as an internal reference. However, because of the high water content in blood, changes in blood volume can lead to changes in the natural abundance H₂¹⁷O signal. We observed a >15% decrease in the natural abundance H₂¹⁷O signal when the perfusion pressure was reduced from 100 cm water column to 40 cm (data not shown). Such changes are probably caused by changes in vascular volume in response to perfusion pressure. Therefore, care must be taken when using the natural abundance H₂¹⁷O signal as a reference.

The current study has several limitations. First, a simplified compartment model was used to estimate the oxygen consumption rate. Because of our inability to measure the ¹⁷O enrichment of the oxygen gas in the perfusate, current parameter estimation can only determine the metabolic rate of ¹⁷O-labeled oxygen rather than the total oxygen consumption. Although model-determined ¹⁷O₂ consumption rates showed agreement with experimentally measured MVO₂, the robustness of parameter estimation needs to be evaluated further. As the acquired ¹⁷O signal comprised signals from both the heart and the perfusate in the MR tube, further validation of the model can be accomplished by measuring H₂¹⁷O in the heart and perfusate separately, either by MRI or with the assistance of other methods, such as mass spectrometry. In addition, given the short T₁ relaxation time (~5 ms) of ¹⁷O, a shorter TR (~20 ms) can be used to improve the signal-to-noise ratio and temporal resolution of data acquisition, which will lead to improved parameter estimation.

Second, the current system was not optimized for longer time durations. The need for cost-savings and for the maximization of detection sensitivity has rendered it a challenge to maintain normal cardiac function during the time course of data acquisition. The decrease in ventricular function at high Ca²⁺ concentration may have been caused by hypoxia because of increased oxygen extraction at high workload and the inability of the system to maintain adequate re-oxygenation of the perfusate. As a result, the model-fitted MV¹⁷O₂ reflects an average index of oxygen consumption during the time course of ¹⁷O MRS acquisition. Nevertheless, these preliminary results demonstrate that ¹⁷O MRS on an isolated perfused heart can provide a valuable approach for the validation of in vivo ¹⁷O studies, as well as for the investigation of metabolic regulation in cardiac mitochondria under various pathophysiological conditions.

In conclusion, this study has demonstrated the feasibility and sensitivity of dynamic ¹⁷O MRS for the detection of the altered metabolic rate associated with changes in cardiac workload. In combination with kinetic modeling, this high-temporal-resolution MRS method allows the quantification of the mitochondrial respiration rate in beating hearts. When combined with ³¹P magnetization transfer for the quantification of ATP synthesis, ¹⁷O MRS provides a unique opportunity to evaluate the coupling between mitochondrial oxidation and phosphorylation in a nondestructive manner.

Abbreviations used

HCA

high calcium concentration group

HR

heart rate

LVDP

left ventricular developed pressure

MVO₂

myocardial oxygen consumption rate

MV¹⁷O₂

myocardial ¹⁷O₂ consumption rate

NCA

normal calcium concentration group

PCr

phosphocreatine

PET

positron emission tomography

RPP

rate–pressure product