| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Experimental Studies |
1 From the Department of Radiology, Division of MR Research (P.A.B.), and the Department of Medicine, Division of Cardiology (R.G.W.), Johns Hopkins University, JHOC-4221, 601 N Caroline St, Baltimore, MD 21287-0843. Received June 22, 2000; revision requested August 8; revision received September 6; accepted September 11. Supported by National Institutes of Health grants RO1 HL56882-01, HL52315-05, and HL61912-01. Address correspondence to P.A.B. (e-mail: bottoml@mri.jhu.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Water-referenced localized phosphorus and proton MR spectroscopy were combined in a single protocol to noninvasively measure phosphocreatine (PCr), adenosine triphosphate (ATP), and total of phosphorylated and unphosphorylated creatine (CR) concentrations and pH in the myocardium in six normal dogs and six dogs with surgically induced myocardial infarction. Unphosphorylated creatine and adenosine diphosphate (ADP) levels were calculated. The results were compared with biochemical measurements at postmortem biopsy.
RESULTS: Significant reductions in PCr-to-ATP ratios (1.7 ± 0.3 [SD] vs 1 ± 0.4; P < .001), PCr (10.3 ± 2.1 vs 4.3 ± 2.0 µmol/g wet weight; P < .0001), ATP (6.4 ± 1.4 vs 3.7 ± 1.4 µmol/g wet weight; P < .001), and CR (24.7 ± 6.1 vs 6.3 ± 3.7; P < .0001) were measured noninvasively in infarcted, as compared with normal, tissue. Biopsy measurements confirmed infarct-related reductions observed at MR spectroscopy, although high-energy phosphate concentrations were lower at biopsy. ADP calculated from noninvasive MR spectroscopic measurements was 0.11 ± 0.07 µmol/g wet weight in normal myocardium.
CONCLUSION: This combined phosphorus and proton MR spectroscopic approach provides a near-complete picture of in vivo myocardial CK metabolism in normal and diseased heart and a tool for noninvasively measuring metabolite reductions associated with the loss of viability.
Index terms: Animals • Heart, MR, 511.121411, 511.121415 • Heart, MR spectroscopy, 511.12145 • Magnetic resonance (MR), phosphorus studies, 511.121411, 511.121415, 511.12147 • Magnetic resonance (MR), spectroscopy, 511.12145 • Metabolism, 511.91 • Myocardium, infarction, 511.814
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In human MI, the appearance of myocardial CK isozyme in the serum is used clinically for diagnosis, and biochemical analyses of tissue obtained at autopsy also show significant reductions in tissue PCr, Cr, and total creatine (CR) (2). In addition, [Cr], [PCr], and [ATP] are reduced in noninfarcted myocardium sampled at surgery in patients with heart failure and cardiomyopathy (3–5), which potentially links energy deprivation with contractile dysfunction in the failing heart (5,6). These invasive observations underscore the value of measurements of tissue CK metabolite concentrations to probe the role of energy deprivation in the failing heart and to assess local viability in ischemic injury. However, robust noninvasive methods for quantifying CK metabolite concentrations are essential to extend these observations to the clinical setting.
Localized phosphorus 31 [31P] magnetic resonance (MR) spectroscopy uniquely provides noninvasive measurements of the relative myocardial PCr and ATP levels in normal and injured myocardium (7), including human MI (8–12). To adapt such methods for the measurement of metabolite concentrations requires both a concentration reference and accurate knowledge of the volume of tissue present in the localized MR spectroscopic volume element, or voxel. These two requirements can be met by including measurements from an external concentration reference and MR imaging–based tissue volumetric measurements as part of the MR spectroscopic protocol. By such means, the high-energy phosphate metabolite concentrations in normal human myocardium can be measured noninvasively (10,13,14). Quantitative measurements of PCr and ATP concentrations reveal depletion in patients with fixed thallium radionuclide imaging defects (9,10), which links loss of viability with phosphate metabolite depletion in infarction.
[CR] can now be measured noninvasively in the myocardium, as demonstrated in humans and dogs, with MR imaging–guided hydrogen 1 (1H, photon) MR spectroscopy (15). In those studies (15,16), CR was indexed by using the signal from its N-methyl, or N-CH3, moiety and was quantified by using the myocardial tissue water signal from the same voxel as a concentration reference, thereby eliminating the need for MR imaging–based tissue volumetric measurements. In patients with prior MI, tissue [CR] was significantly reduced by nearly threefold, as compared with noninfarcted tissue in both the same subjects and with that of healthy volunteers. Because water referencing can also be applied to the quantification of human myocardial phosphate metabolites with 31P MR spectroscopy (16), the possibility of combining 1H and 31P MR spectroscopy in a single noninvasive examination to measure both phosphorylated and the total of phosphorylated and unphosphorylated CK metabolites in viable, nonviable, and dysfunctional myocardium now exists (17).
The goal of this study was to develop an MR image–guided localized MR spectroscopic method that combines quantitative 1H and 31P approaches to provide a near-complete picture of CK metabolite concentrations in the heart. The specific aim was to use the method to measure [PCr], [ATP], and [CR] and to calculate [Cr] and the concentration of adenosine diphosphate [ADP] in normal and infarcted canine myocardium, and thereby demonstrate significant depletion of metabolites in infarcted nonviable tissue.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Multisection cardiac-gated spin-echo 1H MR images were acquired to ensure correct positioning of the animals relative to the detection coils and to permit graphic prescription of voxels for [CR] measurements. Automatic shimming was performed. A fully relaxed cardiac-gated 1H one-dimensional phase-encoded data set was then acquired with the 31P coil for use as a water reference for the 31P data set. We used adiabatic 35° excitation pulses to provide a uniform excitation field (18) with a repetition time of 1.4–2.0 seconds, two signals acquired (averaged) per gradient step, and 1-cm resolution. A cardiac-gated 31P one-dimensional phase-encoded data set was then acquired in 11–16 minutes with a repetition time of about 0.8–1.0 second and 24–32 signals acquired per gradient-encoding step, all other parameters being the same.
Next, 1H MR spectroscopy was performed for [CR] measurements by using image-guided stimulated-echo acquisition mode (STEAM; 1.5–2.0/15 [repetition time seconds/echo time msec]; mixing time, 14 msec; 64–128 signals acquired per spectrum) localization. Voxels were located within the anterior myocardium in the same regions studied by means of 31P MR spectroscopy, near the reference vial, when visible. Voxel volumes of 4–9 mL were chosen to yield useful signal-to-noise ratios in 2–4-minute acquisitions, so that multiple 1H MR spectroscopic acquisitions in normal and infarcted regions, along with the MR imaging and 31P MR spectroscopic acquisitions, could all be accommodated in a reasonable total study time of 1–1.5 hours. The size and location of voxels were graphically prescribed by the authors on the corresponding spin-echo images to fall within the myocardium and to avoid substantial areas containing the ventricular chamber, as well as pericardial fat that can generate intense lipid resonances that can potentially overwhelm the CR signal intensity (Figs 1, 2). As in brain 1H MR spectroscopy, the location and size of selected voxels can be confirmed with MR imaging (Fig 1, C).
|
|
After completion of canine MR spectroscopic studies, the animals were removed and a calibration experiment was performed to determine the ratio of the MR spectroscopic signal per proton to the signal per phosphorus nucleus, CPH. For this purpose, 1H and 31P fully relaxed one-dimensional phase-encoded data sets were acquired with the same 31P receiver coil and the same spatial resolution that was used for the in vivo studies, from a bag containing a 0.2 mol/L inorganic phosphate (Pi) solution. Meanwhile, animals were taken to surgery and thoracotomy was performed with anesthesia. Tissue samples from infarcted and remote regions of the heart were quickly frozen and later fluorometrically assayed for PCr, CR, and ATP. The animals were then sacrificed, and the myocardium was removed and incubated in triphenyltetrazolium chloride (TTC) solution to permit classification of the tissue samples into normal and infarcted tissue. The fractional water content, w, was measured in MI and in noninfarcted tissue by means of weighing samples before and after dessication.
The relative areas of phosphate metabolite, SP, CR, SCR, water, SW, and calibration signal peaks in MR spectroscopic data were quantified by means of Gaussian curve fitting of baseline-corrected spectra by using a semiautomated simplex algorithm that fitted amplitude, frequency, and line width. Application of the curve-fitting routine to overlapping peaks in a STEAM spectrum from the heart is exemplified in Figure 1. Phosphate metabolite concentrations in micromoles per gram of wet weight were calculated as follows:
Because acquisition delays for one-dimensional phase-encoded experiments were short (1.5 msec), EP/EW was taken as 1.0 (16). Fp was calculated from prior mean 31P myocardial metabolite data and was consistent with measurements performed in our laboratory (16,18). All other parameters in Equation (1) were measured individually for each study animal. This method of measuring phosphate metabolite concentrations has been validated with blinded phantoms and in the normal human leg and chest, which yielded an accuracy and precision of about 10%–15% of the measurements in conditions comparable with those used here and has been used to measure [PCr] and [ATP] in the normal human heart (16).
CR was determined from the curve-fitted area, SCR, of the 1H MR spectroscopy N-methyl peak at 3.0 ppm in the water-suppressed spectrum (Fig 1, A). The procedure was recently introduced for 1H MR spectroscopic experiments to noninvasively quantify myocardial [CR] by using the ratio of the N-methyl resonance to the water signal from the same voxel in a spectrum acquired without water suppression (15,19). Application of this noninvasive technique to canine and human MI demonstrated significant reductions in CR in MI (15). The complete abolition of the 3.0-ppm peak after treatment of canine myocardial tissue extracts with enzymes specific for creatine degradation (CK, creatinase, and sarcosine dehydrogenase) was evidence that this MR spectroscopic peak represents creatine, at least in extracts, and is uncontaminated with other moieties (15).
[CR] in micromoles per gram of wet weight was calculated from an expression similar to Equation (1), also validated in phantom, muscle, and cardiac studies (15,19):
[Cr] was determined from the difference between [CR] and [PCr]. Intracellular ADP and adenosine monophosphate (AMP) concentrations (subscript i) were calculated assuming
[H+]i was determined from the pH, which was measured from the chemical shift of Pi resonances that could be identified at chemical shifts of 
(Pi) 
5.2 ppm via (22)
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Figure 4 shows the variation of [PCr] and [ATP] in all of the normal and infarcted animals, as a function of depth through the anterior myocardial wall, as measured with quantitative 31P MR spectroscopy. The metabolite concentrations in normal and infarcted animals were the same in the chest wall but were significantly reduced in the anterior myocardial wall in animals with infarction. The infarctions produced by the surgical protocol were relatively large, often extending beyond the anterior myocardium to the lateral wall and septum, as exemplified in Figure 2, d. Thus, the spectra in 1-cm sections acquired as a function of depth typically intersected infarcted tissue identified with TTC at several depths, which depressed the phosphate metabolites for 2–3 cm, as is evident in Figure 4. At greater depth, noninfarcted tissue is encountered, and the [PCr] and [ATP] levels increase toward normal values. [PCr] and [ATP] measurements were therefore available in myocardial tissue at depths beyond the infarcted anterior wall, except in one animal in which all of the regions sampled by means of both 31P and 1H MR spectroscopy were infarcted at TTC staining. Consequently, MR spectroscopic metabolite measurements in noninfarcted tissue were also available in five of the six infarcted animals. The mean [PCr] and [ATP] in deeper noninfarcted tissue in animals with infarction were 11.5 µmol/g wet weight ± 2.5 [SD] and 7.0 µmol/g wet weight ± 1.2 (five animals), respectively. This was not significantly different from the values of 9.1 µmol/g wet weight ± 0.5 and 5.8 µmol/g wet weight ± 1.4 in the six normal control animals, which indicates that the surgical procedure had no significant effect on phosphate metabolite levels in uninvolved myocardium.
|
|
|
The results from the assays at sacrifice also demonstrated highly significant reductions in [PCr], [ATP], and [CR] in infarcted tissue compared with values in noninfarcted tissue. The tissue water contents, used as a concentration reference for both 31P and 1H MR spectroscopy, were the same in infarcted and noninfarcted tissue. Therefore, tissue water content was not a factor affecting the concentration measurements. In addition, the assay values for the ratio of PCr to ATP in noninfarcted tissue and [PCr] and [ATP] in both normal and infarcted tissue were significantly less than the in vivo 31P MR spectroscopic measurements. The mean depletion of [PCr] and [ATP] in infarcted tissue, when expressed as percentages of reduction relative to noninfarct values, was larger at 80%–90% of the assay measurements, compared with concentration reductions of 50%–60% measured with 31P MR spectroscopy (n = 6). However, [CR] levels measured in the biopsy specimens of normal and infarcted tissues did not differ significantly from the 1H MR spectroscopic values; the mean [CR] depletion in infarction was 80% ± 6 (n = 6).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Earlier MR spectroscopic methods of measuring metabolite concentrations often used external concentration references and MR imaging–based estimates of the volume of myocardial tissue present in a voxel (13), or they used volume approximations based on a spherical shell model of the myocardial geometry (10) or MR imaging–based postacquisition compartmentation (14). However, MR imaging–based tissue volumetric measurements must be performed in three dimensions and are subject to errors arising from local wall thinning and from inherent differences in both phase and sensitivity between the MR imaging and MR spectroscopic coils at the myocardium being interrogated and at the location of the concentration reference (13,16). The use of the tissue water signal acquired in the same voxel with the same coil and localization method (16) avoids these errors by canceling out the effects of nonuniformity in the coil sensitivity and the characteristics of the point spread function that are specific to the MR spectroscopic or MR imaging localization method. It also obviates tissue volumetric measurement inasmuch as the water and metabolite signals are derived from the same tissue volume (16). It does, however, require an estimate of the tissue water concentration, which in the present study was the same in normal and infarcted tissue (Table).
|
Although we were unable to reliably quantify [Pi], results of prior work (7,26) suggest an upper limit for normal myocardial Pi such that Pi/PCr 
0.14. This would result in a lower limit for the change in free energy due to ATP hydrolysis of -56.3 ± 1.1 kJ/mol for normal canine myocardium in vivo, with use of the equations in reference 23 and our in vivo MR spectroscopic data.
The results showing reduced [PCr], [ATP], and [CR] in infarcted tissue were consistent with the notion that dead myocytes are depleted of energy metabolites and/or metabolite levels are lower in scar tissue, which thereby reduces overall metabolite concentrations (1,2,7–10). The observation that pH is normal in MI 5–20 days after onset is consistent with results of prior studies (24) that show an evolution of initially acidic myocardial pH toward normal (slightly alkaline) values in the hours up to 5 days after infarction. Reductions in the ratio of PCr to ATP have been noticed in some (10,11) but not all (8,12,26) studies of chronic MI (7).
That depletion of [PCr], [ATP], and [CR] is incomplete indicates the presence of substantial numbers of living viable cells, likely present in normal tissue that surrounds and is interspersed with the infarcted tissue and that resides in the relatively coarsely resolved MR spectroscopic volume elements (7). Thus, for 31P MR spectroscopy, in which the voxels were largest, contributions to the PCr and ATP signals in surrounding normal tissue explain the smaller fractional depletion of [PCr] and [ATP] than was measured at biopsy.
Compared with noninvasive MR spectroscopic measurements, the biopsy measurements had the advantage that verifiable sampling of normal and infarcted tissue and avoidance of heterogeneous regions containing infarcted and noninfarcted tissue was directly possible by reference to the TTC-stained sections. Even so, the relative reduction in CK metabolites with infarction was larger and comparable between biopsy measures and noninvasive 1H MR spectroscopy than were the changes detected with 31P MR spectroscopy. This suggests that the smaller 1H MR spectroscopic voxels may be better suited to detecting infarction-induced reductions in cardiac CK metabolites than are the intrinsically larger 31P MR spectroscopic voxels. Therefore, although noninvasive MR spectroscopy may provide better measures of [PCr] and [ATP] in normal myocardium than does biopsy, its application in focal disease may be limited to processes that affect regions of size comparable with or larger than the spatial resolution afforded by 31P MR spectroscopy.
Such observations emphasize the importance of efforts to improve the spatial resolution of cardiac 31P MR spectroscopy. Unfortunately, because the signal-to-noise ratio is directly proportional to voxel size and to the square root of averaging time, improving spatial resolution generally requires a sacrifice of either signal-to-noise ratio or imaging time, unless other gains can be realized, such as with new phased-array detectors (27), Overhauser enhancement (28), or higher magnetic field strengths (29–31), which offer the potential for imaging metabolite depletion in MI (32). In the present study, imaging time was constrained by the need to accommodate the entire 31P and 1H MR spectroscopic and MR imaging procedure in a single examination of about 1 hour, which would be tolerable for human studies. The limited sensitivity contributes to scatter in the results listed in the Table.
For 1H MR spectroscopy of CR, the much higher 1H signal-to-noise ratio per nucleus, the higher [CR], and the fact that three nuclei contribute to the CR 1H MR spectroscopic resonance for every one that contributes to a PCr or ATP 31P resonance, translates to an enormous signal-to-noise ratio advantage of 20–60-fold that enables acquisitions from smaller voxels. Nevertheless, the presence of intense residual water and lipid resonances (Figs 1, 2) and the sensitivity of spin-echo localization methods to motion (33,34) are additional factors contributing to scatter in the 1H MR spectroscopic measurements that are not present at 31P MR spectroscopy. The use of different localization strategies for 31P and 1H MR spectroscopy may also confound calculations of CK reaction kinetics for focal disease such as MI, to the extent that the measurements may sample different tissue volumes, although this is often a problem for biopsy measurements as well.
Although metabolite concentrations measured at biopsy show less scatter than those measured at noninvasive MR spectroscopy, they may fail to show important systematic changes that occur during sampling and analysis, as exemplified in the differences between the normal high-energy phosphate levels measured with MR spectroscopy and biopsy in the Table. Such differences, especially the rapid hydrolysis of PCr, may significantly affect calculated CK metabolite levels and reaction energetics. Moreover, biopsy is invasive and therefore impossible to perform as a source of high-energy phosphate measurements for assessing myocardial viability in the clinical setting.
Practical application: Thus, results of the present study show that myocardial [CR], [PCr], [ATP], and calculated [Cr] and [ADP] can all be quantified noninvasively in dogs by combining 1H and 31P MR spectroscopy in a single study with an MR spectroscopy–equipped 1.5-T clinical MR imager. Results of recent separate quantitative studies of normal and infarcted human myocardium with 31P MR spectroscopy (9,10,13,14) and 1H MR spectroscopy (15) lend credence to the practicality of extending our combined MR spectroscopic approach to human clinical studies. Because the presence of an adequate energy supply is central to cellular viability, such quantitative MR spectroscopic measurements have potentially important applications to the noninvasive assessment of myocardial viability in clinical MI (35). It may also provide new insight into the role of CK metabolism in nonfocal conditions, including cardiomyopathy and heart failure, in which energy deprivation is thought to play a key role (5,6) and in which the technical limitations of spatial resolution are more relaxed.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, P.A.B.; study concepts and design, P.A.B., R.G.W.; literature research, P.A.B.; experimental studies, P.A.B., R.G.W.; data acquisition, P.A.B., R.G.W.; data analysis/interpretation, P.A.B.; statistical analysis, P.A.B.; manuscript preparation, definition of intellectual content, editing, revision/review, and final version approval, P.A.B., R.G.W.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. F. A. Jansen, W. H. Backes, K. Nicolay, and M. E. Kooi 1H MR Spectroscopy of the Brain: Absolute Quantification of Metabolites. Radiology, August 1, 2006; 240(2): 318 - 332. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. C. Chen and O. Ogut Decline of contractility during ischemia-reperfusion injury: actin glutathionylation and its effect on allosteric interaction with tropomyosin Am J Physiol Cell Physiol, March 1, 2006; 290(3): C719 - C727. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. C. Wu and J. A.C. Lima Noninvasive Imaging of Myocardial Viability: Current Techniques and Future Developments Circ. Res., December 12, 2003; 93(12): 1146 - 1158. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Nakae, K. Mitsunami, T. Omura, T. Yabe, T. Tsutamoto, S. Matsuo, M. Takahashi, S. Morikawa, T. Inubushi, Y. Nakamura, et al. Proton magnetic resonance spectroscopy can detect creatine depletion associated with the progression of heart failure in cardiomyopathy J. Am. Coll. Cardiol., November 5, 2003; 42(9): 1587 - 1593. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. M. Pohost and J. R. Forder From the atomic nucleus to man: Nuclear magnetic resonance spectroscopy, the next horizon in diagnostic cardiology J. Am. Coll. Cardiol., November 5, 2003; 42(9): 1594 - 1595. [Full Text] [PDF]
|
|
|
|
|
|
O. Ogut and F. V. Brozovich Creatine Phosphate Consumption and the Actomyosin Crossbridge Cycle in Cardiac Muscles Circ. Res., July 11, 2003; 93(1): 54 - 60. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Pucar, P. Bast, R. J. Gumina, L. Lim, C. Drahl, N. Juranic, S. Macura, E. Janssen, B. Wieringa, A. Terzic, et al. Adenylate kinase AK1 knockout heart: energetics and functional performance under ischemia-reperfusion Am J Physiol Heart Circ Physiol, August 1, 2002; 283(2): H776 - H782. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
1.8 seconds; echo time, 15 msec; 128 signals acquired; acquisition time, 4 minutes). Hatching denotes the region of water suppression around 5 ppm. The insert (gray box) shows the fitting of overlapping peaks in the 3-ppm region with Gaussian functions (top) and the difference between the fitted and the actual spectrum (bottom). B, Transverse spin-echo MR image (0.86-1.00/15 [repetition time sec/echo time msec]; gated to heart rate; two signals acquired per gradient step; 128 x 256 voxels; 32-cm field of view; 1-cm-thick section) shows the location of the STEAM voxels in A and a STEAM voxel deliberately chosen to intersect the ventricular chamber. C, Spin-echo MR image of the latter voxel (signals acquired per gradient step increased to four) shows no signal intensity in the chamber. The integrated signal intensity in the larger portion of the STEAM voxel that intersects the blood-filled chamber is less than 7% of the total owing to the "black blood" properties of the sequence. In practice, the voxel size and position are carefully adjusted with MR imaging guidance to minimize contamination from nonmyocardial tissue and to exclude large portions of the ventricular chamber.
) and animals with infarction (
) as a function of depth through the chest wall and myocardium as measured at water-referenced 31P one-dimensional phase-encoded MR spectroscopy and calculated from Equation (1). Error bars = mean ± SEM (*P < .05 vs normal control animals).
= mean ± SD (error bars). Note that points overlap in two animals in the normal group (with [CR] = 29.75 and 29.84 µmol/g wet weight) and three animals in the MI group (with [CR] = 4.45, 4.48, and 4.72 µmol/g wet weight). In one animal in the MI group, all regions studied at MR spectroscopy were infarcted at TTC staining, so it does not have a line connected to its MI point.