INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

APPLICATION OF ³¹p nmr spectroscopy has gained increasing popularity in recent years in providing a window into metabolism and cellular bioenergetics, but a number of concerns remain (2, 6, 14, 19, 20). Notwithstanding the problem of NMR sensitivity (>0.5 mM), a longstanding difficulty has been how to convert the number of "spins" into an absolute tissue concentration (12). Two popular methods for quantifying the absolute phosphorus concentration in tissues are using an external reference standard (9, 28) and using tissue ATP measured independently after the experiment and equating the content to the area under the ATP peak (8, 13, 14, 20, 29). Other methods of quantitation include internal [¹H]H₂O calibration (27) and the use of a synthesized NMR reference produced by the transmittance of signal from an electronic device received at the same time as the sample signal (2). The assumptions of all these methods are not trivial and involve knowledge of the radiofrequency and coil loading properties, inhomogeneity in the B₁ coil, volume sensitivities, the effect of motion, and the choice of spectral analysis program (28, 29).

Using differences between an external standard and measured ATP has led some investigators to propose an up to 40% invisibility of ATP in the heart (26, 30). Invisibility is a complex term but largely reflects line broadening from the highly decreased mobility of a phosphorus compound, e.g., from compartmentation. One of the greater difficulties, however, with invoking invisibility is that one has to be confident that all the technical assumptions mentioned above are not a contributing factor. The aim of this study is to take these factors into account and compare the two most common methods of an external standard and comparing the phosphorus concentrations with freeze-clamp ATP values measured from the same tissues in rat heart and gastrocnemius muscle in vivo. Our study showed that by positioning the external standard above the coil and providing symmetrical volume sensitivities, both methods agreed to within 5-10% at 7 T. The heart showed the greatest discrepancy due to the lower signal-to-noise ratio, primarily from motion effects.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Surgical Protocol

Heart. Seven male Sprague-Dawley rats weighing 300-350 g were obtained from the James Cook University (JCU) breeding colony and housed in the animal facility. Rats were supplied with unrestricted access to food and water. Animals were anesthetized with an intraperitoneal injection of pentobarbital sodium (60 mg/kg rat weight). The left femoral artery and vein were cannulated with polyethylene (PE)-50 tubing. The venous line was used for maintenance of anesthesia (15 mg/ml solution of pentobarbital sodium in 0.9% saline), whereas the arterial line was used for blood collection and continuous monitoring of blood pressure. The arterial line was flushed periodically with a 100 U/ml solution of heparin in 0.9% saline (David Bull Laboratories; Melbourne, Australia). Blood pressure was measured using a pressure transducer interfaced to a Maclab 200 (AD Instruments; Castle Hill, Australia) and Apple Macintosh computer.

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Configuration of the surface coil assembly in the heart (skeletal muscle not shown). The diagram is a cross-sectional view showing the symmetrical arrangement of the balloon standard [10 mM phenylphosphonic acid (PPA)] above a three-turn surface coil, with the apical region of the left ventricular wall positioned below the coil.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   31P NMR signal arising from consecutive 2-mm layers of a multilayered phantom (inset) using the three-turn surface coils used for the rat heart (solid bars) and rat skeletal muscle (open bars). The signal from each layer is presented as a percentage of the total signal from all four layers.

Skeletal muscle. Six male Sprague-Dawley rats weighing 350-400 g were obtained from the Animal Resources Centre (Canning Vale, Australia) and housed in the animal facility at JCU. Rats were supplied with unrestricted access to food and water. Animals were anesthetized with an intraperitoneal injection of pentobarbital sodium (60 mg/kg rat weight), tracheotomized, and ventilated using a Harvard Small Animal Ventilator to give a blood PO₂ of ~90 mmHg, a PCO₂ of ~40 mmHg, and a pH of ~7.4. Blood gases were analyzed with the Ciba-Corning 865 blood gas analyzer. Anesthesia was maintained by delivering 1% isoflurane (Abbott Australasia; Kurnell, Australia) in compressed air via the ventilator at a rate of 0.5 l/min. The right carotid artery was cannulated with PE-50 tubing and utilized for blood collection and continuous monitoring of blood pressure. The arterial line was flushed periodically with a 100 U/ml solution of heparin in 0.9% saline. Blood pressure was measured using a Statham P23 XL pressure transducer interfaced to a Maclab 200 and Apple Macintosh computer.

NMR Spectroscopy

Heart. ³¹P NMR experiments were performed at 121.47 MHz in a 110-mm horizontal bore Oxford 7.05-T superconducting magnet. The magnet is the property of the School of Molecular and Biomedical Sciences (JCU). ³¹P spectra were obtained using a Varian Inova NMR spectrometer. Magnetic field homogeneity was maximized on the ¹H free induction decay (FID) measured off resonance using an Oxford Instruments 15-channel shim supply. Radiofrequency pulses of 8-µs duration at an ~40° flip angle were applied with a 1-s interpulse delay. FIDs were acquired over 0.4 s with a total of 1,024 FIDs averaged. An spectral width of 8,000 Hz was used, and 6,400 data points were obtained. An exponential line broadening factor of 10 Hz was applied to the ³¹P NMR spectra, which were then fitted using the Varian Fitspec software. After integration, all peaks were multiplied by a saturation correction factor specific for each peak. These factors were determined experimentally by comparing the peak integrals of partially relaxed spectra, obtained using the acquisition parameters described above, to the peak integrals of fully relaxed spectra (20-s interpulse delay). Mean values were used in NMR calculations of individual phosphorus compounds in the heart (and skeletal muscle) for a 1-s acquisition delay.

Skeletal muscle. ³¹P NMR experiments in muscle were performed using the same instrumentation as for the heart experiments. Radiofrequency pulses of 8-µs duration at an ~90° flip angle were applied with a 1-s interpulse delay. FIDs were acquired over 0.8 s with a total of 256 FIDs averaged. A spectral width of 6,000 Hz was used, and 9,600 data points were obtained.

Metabolic Analysis

Phosphorus Quantification

(1)

(2)

(3)

The second method employed PPA as an external standard. Notwithstanding issues of signal-to-noise ratios and peak integration methods, the following equations (Eq. 4-10) present the theory of using an external reference positioned symmetrically above the coil and heart tissue as follows.

The observed signal (S) (or integral of the phosphorus metabolite in heart) is given by

(4)

where  is the spin density distribution of the source signal, V is volume where signal is measured and B_{1 coil} is the sensitivity of the coil system, which is generally the transverse field at a point in space normalized to the amplitude per mole of phosphorus metabolite. For the reference signal,  = [PPA] × f_PPA (x, y, z), where f_PPA (x, y, z) is the spatial distribution. If we take f = 1 inside the sphere containing the standard PPA and f = 0 outside, therefore

(5)

The tissue phosphorus signal (e.g., PCr) from Eq. 4 above will therefore be

(6)

which leads to

(7)

and assumes that PCr is uniformly distributed. If f = 1 inside the heart and f = 0 outside, as for PPA in the sphere (Eq. 5), the volume integral is limited to the convoluted volume of the heart. Therefore, given these conditions

(8)

From these equations, PPA can be used as the external standard (Eq. 5) to determine tissue PCr (Eq. 8). The two basic assumptions of the method are that 1) [PCr] is uniformly distributed in the heart, and 2) the sensitivity of the coil system yields identical signal amplitude per mole of ³¹P between PPA and phosphorus compounds in the heart. Therefore

(9)

If these assumptions are correct, then

(10)

The concentration of PCr then becomes the equivalent of micromoles of metabolite per gram wet weight in equal volumes (taking into account that 1 ml of water is equivalent to 1 g, and 1 ml tissue is equivalent to 1.1 g).

The saturation-corrected phosphorus metabolite integrals were standardized from the saturation-corrected 10 mM PPA integral as follows

(11)

(12)

(13)

Intracellular pH

(14)

Free Magnesium

(15)

4

4

3

5

7

4

Statistical Analysis

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The sampling depths of the surface coils used for in vivo determination of phosphorus metabolite concentrations in heart and skeletal muscle are shown in Fig. 2. In the heart, ~70% of the signal came from the first 2 mm of tissue below the coil and ~90% came from within the first 4 mm. With the larger coil for skeletal muscle, 50% came from the first 2 mm and a total of 85% came from the first 6 mm below the coil. The thickness of the ventricle wall directly beneath the coil was measured in our study and found to be ~4 mm in thickness and 8 mm in rat gastrocnemius muscle below the site of coil placement. Given these distances, we conclude that over 90% of signal was received from both ventricular and skeletal muscle.

Table 1 summarizes the data obtained from the two methods of quantitation of ATP, PCr, and P_i in heart and skeletal muscle in vivo. With the use of the external standard (10 mM PPA), the mean concentrations of ATP, PCr, and P_i in the heart were 4.48, 9.21, and 2.25 µmol/g wet wt, respectively. With the use of the internal ATP standard measured enzymatically on the same tissue, the mean contents were 4.78, 9.83, and 2.51 µmol/g wet wt, respectively. There were no significant differences for each metabolite and the two methods of quantification in the heart. The P values for ATP, PCr, and P_i for the two methods are 0.41, 0.34, and 0.46, respectively (Table 1).

For skeletal muscle, using the external standard, the concentrations of ATP, PCr, and P_i were 6.09, 23.44, and 1.81 µmol/g wet wt, and, for the internal standard, the concentrations were 6.03, 23.30, and 1.82 µmol/g wet wt, respectively. In skeletal muscle, as in the heart, no significant differences were found between the two methods and each phosphorus compound. The P values for ATP, PCr, and P_i for the two methods are 0.83, 0.93, and 0.97, respectively (Table 1). Free [Mg²⁺]_i was assessed by the relative chemical shift of the ATP to ATP peaks in ³¹P NMR spectra and found to be 0.46 ± 0.04 mM (average chemical shift of 8.64 ± 0.03) in heart and 0.57 ± 0.04 mM (average chemical shift of 8.58 ± 0.01) in skeletal muscle. pH_i, determined from the chemical shift between P_i and PCr, was 7.32 ± 0.04 (average chemical shift of 5.17 ± 0.04) in heart and 7.18 ± 0.02 (average chemical shift of 5.03 ± 0.02) in skeletal muscle.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Our study shows that ³¹P NMR determination of absolute concentrations of ATP, PCr, and P_i in heart and skeletal muscle can be accurately measured using either external references or an internal ATP standard. No significant differences were found between the two methods in determining each phosphorus compound in heart or skeletal muscle. The higher P values in skeletal muscle demonstrate a tighter agreement (100-101% difference in means of metabolite concentrations between the two methods) compared with heart (89-94%).

External and Internal Standardization of NMR Spectra

On the other hand, the major advantage of the internal ATP enzymatic method is that the molar sensitivities and coil-sensitive volumes are identical. However, two potential sources of error require comment. First, it is known that a small fraction of the total tissue ATP resides in the mitochondria (~5 and 10% in muscle and the heart, respectively), and this would lead to higher levels of ATP, PCr, and P_i than reported in Table 1. Second, the internal method assumes that 100% of enzymatically measured ATP is equal to the total nucleoside triphosphate (NTP) pool under the ATP peak. We (8) have shown previously that ATP represents ~93% of the total NTP (ATP + UTP + CTP + ITP + GTP) in the rat heart. The net effect of underestimation of ATP from mitochondrial sequestration and overestimation from using the total NTP area appears to be of similar magnitude in opposing directions. Therefore, we conclude that equating tissue measurements of ATP with the area under the ATP peak is valid for heart and skeletal muscle given the excellent agreement between both methods.

Differences Between Heart and Skeletal Muscle

NMR Visibility of ATP

Our data agree with most other findings in the perfused rat heart (1, 7, 10, 13, 30) but are in contrast to those studies that have identified an invisible fraction of ATP ranging from 30 to 40% of total ATP (26, 30). Differences of this magnitude are difficult to reconcile given all the checks and balances required to quantify the concentrations as outlined earlier. ATP invisibility in the heart, but not skeletal muscle (4, 23), may also relate to differences in lower signal-to-noise ratios and incomplete myocardial occupancy of the coil sensitive volume through cardiac motion and/or chamber contamination. As discussed, these factors can introduce errors that significantly underestimate the ATP concentration.

While our study indicates that there is little or no NMR invisiblity of ATP in rat heart and skeletal muscle, it does not exclude the possibility of intracellular compartmentalization within the NMR "visible" volume. The possibility of microcompartmentation has been explored within the context of high-energy phosphate shuttling between the sites of ATP production and utilization (15, 17).

In conclusion, the present study demonstrates the use of a symmetrical external standard configuration positioned above the coil, which can provide accurate concentrations of ATP, PCr, and P_i in both heart and skeletal muscle in vivo. The method is valid to the extent that the sensitivity profile and homogeneity of the reference is identical (or similar) to the tissue. Our estimates of ATP, PCr, and P_i were not significantly different from the internal method of employing ATP measured on tissue extracts prepared from the same tissue. The main disadvantage of using enzymatic ATP as an internal standard is that larger numbers of animals are required at different times during an experiment compared with the external method. Our study further concludes that PCr, ATP, and P_i are nearly 100% visible in the normoxic heart and nonworking skeletal muscle given the errors of measurement.