Author + information
- Received February 4, 1998
- Revision received May 11, 1998
- Accepted May 20, 1998
- Published online September 1, 1998.
- Jens Bremerich, MDa,
- Michael F Wendland, PhDa,
- Håkan Arheden, MD, PhDa,
- Rolf Wyttenbach, MDa,
- Dong W Gao, MDa,
- John P Huberty, BSa,
- Michael W Dae, MDa,
- Charles B Higgins, MD, FACCa,* ( and )
- Maythem Saeed, DVM, PhDa
- ↵*Address for correspondence: Dr. Charles B. Higgins, Department of Radiology, MRI-Section, University of California San Francisco, 505 Parnassus Avenue, L308, San Francisco, California 94143-0628
Objectives. The purpose of this study was to measure the accumulation of labeled albumin and to visualize its distribution pattern in reperfused infarcted myocardium as a function of time between onset of reperfusion and administration of the tracer.
Background. Myocardial microvascular injury leads to leakage of albumin from the intravascular space. Quantitative measurements of GdDTPA-albumin with inversion recovery echoplanar imaging (IR-EPI) may allow noninvasive monitoring of microvascular injury.
Methods. After 1 h of coronary artery occlusion, 56 rats were injected with GdDTPA-albumin or 123I-GdDTPA-albumin either immediately before reperfusion or ½, 1 or 24 h after reperfusion. GdDTPA-albumin in blood, normal myocardium and reperfused infarction was dynamically measured with IR-EPI during 1 h postinjection (PI). Autoradiograms were obtained at 15 min PI. Accumulation of labeled albumin in myocardium was expressed as the ratio of myocardial to blood content.
Results. In normal myocardium, the ratio of changes of relaxation rate-ratio (ΔR1-ratio) was 0.12 ± 0.01 and did not change over 1 h. In reperfused infarction, however, the ΔR1-ratio increased after administration. Animals given GdDTPA-albumin before reperfusion exhibited fastest accumulation (ΔR1-ratio 15 min PI: 0.56 ± 0.03) and essentially homogeneous distribution. The accumulation was slower when administered at ½, 1 and 24 h after reperfusion (ΔR1-ratios 15 min PI: 0.39 ± 0.03; 0.31 ± 0.04; 0.16 ± 0.01; p < 0.001 compared to administration before reperfusion). Moreover, the tracer accumulated predominantly in the periphery of the injury zone.
Conclusions. Amount and distribution pattern of labeled albumin in reperfused infarction are modulated by duration of reperfusion. The accumulation of GdDTPA-albumin can be quantified by IR-EPI. Thus, IR-EPI may be useful to noninvasively monitor myocardial microvascular injury in reperfused infarction.
Myocardial ischemia causes injury to myocytes and to the vasculature which supports myocardial perfusion. It is known that myocyte necrosis precedes damage to the microvasculature, as defined on histology (1). However, injury to the microvasculature as defined by leakage of plasma macromolecules can be observed even after mild ischemic insult (2,3). Moreover, this elevated permeability or microvascular injury seems to progress during reperfusion (4). A noninvasive method to monitor microvascular injury may be useful in evaluating myocardial injury.
Microvascular integrity has been studied in vivo with magnetic resonance imaging (MRI) using macromolecular gadolinium chelates like GdDTPA-albumin or GdDTPA-polylysine (5–7). Macromolecules such as GdDTPA-albumin are retained in the vasculature in normal myocardium, but leak into the interstitium in reperfused infarction; this causes a relatively increased quantity of GdDTPA-albumin within the injured region (6). On conventional T1-weighted MR images, the signal intensity in the abnormal region is increased (5,8,9), but since increased signal intensity on conventional MRI is not proportional to changes of GdDTPA-albumin concentration, quantification is complicated (10). However, a proportional relationship exists between changes in longitudinal relaxivity (ΔR1) and changes of GdDTPA-albumin concentration in the injury zone. Thus, direct ΔR1 measurements with inversion recovery echoplanar imaging (IR-EPI) may provide quantification of GdDTPA-albumin accumulation in reperfused infarcted myocardium (11,12).
The distribution pattern of GdDTPA-albumin in reperfused infarcted myocardium is not always uniform (5–7). An increased tracer accumulation or hyperenhancement of the periphery of the infarct zone and a hypoenhancement of the infarct core have been demonstrated (3,5,6). This pattern of differential enhancement was more pronounced when the duration of ischemia increased (7). However, the direct effect of duration of reperfusion on the amount of GdDTPA-albumin in reperfused infarction is unclear. The accumulation of GdDTPA-albumin as a function of delay time between onset of reperfusion and tracer administration has not been studied.
Therefore, the specific aims of this study were to 1) validate the use of IR-EPI and ΔR1 assessment for quantitative evaluation of microvascular dysfunction by comparison with a radionuclide based approach as the gold standard and 2) define the time course of pathophysiologic changes in myocardial microvasculature and regional blood volume early after reperfusion.
GdDTPA-albumin, a blood pool MRI contrast agent, was prepared by the method of Ogan et al. (13), in which the chelate DTPA was covalently linked to human serum albumin residues. This preparation yields GdDTPA-albumin with a molecular weight of approximately 92 kDa and T1, T2 relaxivities of 13.7, 18.4 mM−1s−1, respectively. This agent has been well characterized as an intravascular blood pool agent for MR imaging, with a plasma half-life of approximately 90 min (8). In normal myocardium this material is confined to the intravascular volume.
123I-GdDTPA-albumin was prepared by adding 123iodine (Cyclotron, Vancouver, Canada) to GdDTPA-albumin taken from the same stock solution as was used for contrast-enhanced MRI animals. The labeled product was purified by placing it over an AG-1X8 anion exchange column.
Care and maintenance of experimental animals were in strict accordance with the National Institutes of Health guidelines. The experimental protocol received prior approval from the Committee on Animal Research at this Institution. Female Sprague–Dawley rats (n = 56; 240 to 310 g; Simonsen Labs Inc., Gilroy, CA) were divided into four groups (n = 14 per group) for preparation of myocardial infarction with varied reflow duration (Fig. 1). Animals were prepared as described previously (6). Briefly, animals were anesthetized with sodium pentobarbital (50 mg/kg; IP). Catheters were placed in a jugular vein for administration of contrast agent or radiotracer and in a carotid artery for measurement of blood pressure (Gould Transducer, Gould Inc., Cleveland, OH) and withdrawal of blood. The chest was opened and the anterior branch of the left coronary artery was occluded for 1 h followed by reperfusion. This rat model of 1 h ischemia followed by reperfusion is known to produce substantial myocardial infarction and microvascular injury (6,14). GdDTPA-albumin (n = 8) or radiotracer (n = 6) was administered 3 min before reperfusion, at ½ h, or 1 h after reperfusion for groups 1 to 3, respectively. Animals in group 4 were anesthetized with short-acting anesthesia (8 mg/100 g xylazine and 0.2 mg/100 g ketamine; IP). Surgical preparation and coronary occlusion duration were identical to groups 1 to 3. Reperfusion was maintained for 24 h. A total of 87 rats were prepared in the present study. Thirty-one rats died after occlusion (15)or reperfusion (13), but before administration of the contrast agent. No subjects were lost during the measurements. All completed experiments were included in the analysis.
Magnetic resonance imaging was conducted on a Bruker Omega CSI 2 Tesla system (Bruker Instruments Inc., Fremont, CA) as previously described (16). Each animal was placed supine in a home-built bird cage resonator; the electrocardiogram signal was obtained from two subcutaneous copper needles. The field of view contained a phantom with known T1.
IR-EPI was conducted in the transverse plane to repeatedly measure T1 of left ventricular chamber blood, normally perfused myocardium and reperfused infarction. EPI sequence parameters were: matrix 64 × 64 points; field of view 50 × 50 mm; slice thickness 2 mm; acquisition time = 32.7 ms; echo time (TE) = 10 ms; pulse delay repetition time minus inversion time (TR-T1) > 4 × T1 (fully relaxed).
In each rat, 10 sets of 20 images each with different TI were acquired before and at 4, 9, 14, 19, 24, 29, 39, 49 and 59 min after administration of GdDTPA-albumin (0.03 mmol Gd/kg). These image sets were used to measure regional T1 values and calculate ΔR1 as described previously (16). At the conclusion of T1 measurements, a conventional T1-weighted spin-echo image (TR ∼ 300 ms; TE = 25 ms; raw data matrix 256 × 128 zero filled to 256 × 256) was acquired at the same midventricular level.
After MR measurements, hearts were excised, rinsed, sliced at the midventricular level and stained by soaking in a 2% solution of triphenyltetrazolium chloride (TTC) at 37°C for 8 min to define the infarction. This staining procedure provides brick-red coloration of normal myocardium, whereas infarcted myocardium is unstained (14).
Six rats per group were examined for distribution of 123I-GdDTPA-albumin (500 μCi). The tracer was allowed to circulate for 15 min since MRI data demonstrated maximal differences of ΔR1-ratios among groups at that time. A sample of blood (0.5 ml) was then withdrawn from the carotid artery for radioactivity measurements. The coronary artery was reoccluded and phthalocyanine blue dye (particle size of 7.4 ± 3.2 μm) was administered intravenously to define the area at risk (17). The heart was excised and sliced at the midventricular level. Samples from both the unstained area at risk and the blue-stained, normally perfused myocardium were taken from the apical portion of the transsected heart. Samples were weighed and radioactivity counted in a gamma-counter was measured (Searle-Analytic Inc., Tampa, Florida). The basal portion of the transsected heart was placed in embedding medium (Tissue-Tek, Sakura Finetechnical Co. LTD, Tokyo), immediately frozen in dry ice and sectioned (20 μm). Sections were placed on a photostimulable storage phosphor imaging plate (Molecular Dynamics, Sunnyvale, CA) to obtain digital autoradiograms. These were evaluated using software (ImageQuant, Molecular Dynamics, Sunnyvale, CA) to determine the distribution of radioactivity in these slices.
MRI data analysis
The inversion time with nil signal (TInull) was determined for normally perfused myocardium, reperfused infarcted myocardium, ventricular blood and phantom using an image analysis program (NIH-image). T1 values and ΔR1-ratios were calculated as described previously (16): (1)(2)
ΔR1-ratios were plotted vs. time. Regional blood volume was estimated by fitting the ΔR1-ratios of reperfused infarcted and of normal myocardium and by extrapolating the fitted curves to the time of GdDTPA-albumin injection. Myocardial blood volumes of the infarcted regions were expressed as a percentage of normal myocardium.
Radioisotope data analysis
The radioactivity ratio of normally perfused myocardium was calculated as: (3)
The injury zone was defined by a count density of ≥2 SD higher than the mean of normally perfused myocardium. The radioactivity ratio of injured myocardium was calculated as: (4)
The same slices were scanned with a flatbed scanner (Epson, Seiko Epson Corp., Japan); data were processed on a Macintosh computer with image analysis software (NIH-image). The area at risk defined as unstained tissue was measured and expressed as a percentage of the total cross-sectional area of the left ventricular slice.
All values are expressed as mean ± SEM. Comparisons among groups at 15 min after contrast administration were made with a one-way analysis of variance (ANOVA). If the analysis showed an overall p < 0.05 value, Scheffé’s test was implemented as post hoc test. For single comparisons a paired Student’s ttest was used. A p value of <0.05 was considered significant.
Administration of 123I-GdDTPA-albumin caused no significant changes in heart rate (309 ± 12 vs. 298 ± 12 beats/min) or in systolic (112 ± 6; 120 ± 5 mm Hg), diastolic (87 ± 5; 92 ± 5 mm Hg) or mean (98 ± 6; 105 ± 5 mm Hg) blood pressures. Heart rate and blood pressure did not differ among the groups.
Precontrast T1 of infarction (1.13 ± 0.03 s) was longer than that of normal myocardium (0.92 ± 0.02 s) (p < 0.05; Student’s paired ttest). After GdDTPA-albumin administration, T1 of blood, reperfused infarction and normal myocardium decreased by ninefold, twofold to fourfold and twofold, respectively. However, regional variation of T1 shortening occurred, which reflects regional variations of GdDTPA-albumin content. Representative IR-EPI images obtained at identical anatomical level but with different TI settings demonstrate regional differences in GdDTPA-albumin content (Fig. 2). The shortest TInullwas found in the blood pool, indicating greatest concentration of GdDTPA-albumin. The TInullof reperfused infarction decreased steadily with time after injection of GdDTPA-albumin, reflecting accumulation of the tracer.
The accumulation of GdDTPA-albumin in reperfused infarction was fastest when administered immediately before reperfusion (ΔR1-ratio at 15 min PI: 0.56 ± 0.03). The accumulation was slower when administered at ½, 1 and 24 h after reperfusion (ΔR1-ratios at 15 min PI: 0.39 ± 0.03; 0.31 ± 0.04; 0.16 ± 0.01; p < 0.001 compared to administration immediately before reperfusion; one-way ANOVA with post hoc Scheffé’s test).
The ΔR1-ratio of normal myocardium (0.12 ± 0.01) did not differ among the experimental groups and did not change significantly during the observation period of 1 h after GdDTPA-albumin administration. Reperfused infarction, however, was characterized by increasing ΔR1 ratios after GdDTPA-albumin administration (Fig. 3), reflecting accumulation of the tracer in the injury zone. Accumulation was maximal when GdDTPA-albumin was injected immediately before reperfusion. The accumulation declined as the time of tracer injection was delayed during the reperfusion period (Fig. 3).
The radioactivity ratios in normal and injured myocardium, as measured after 123I-GdDTPA-albumin administration, were similar to the ΔR1-ratios of IR-EPI (Fig. 4). Both MR contrast media and radiotracer accumulated in injured myocardium, but accumulation declined with increasing delay between onset of reperfusion and injection of the tracers (Fig. 4). The distribution of 123I-GdDTPA-albumin in the injury zone was essentially uniform when administered before reperfusion (Fig. 5, A). However, in hearts subjected to delayed tracer administration, 123I-GdDTPA-albumin was distributed in a heterogeneous fashion in the injury zone (Fig. 5, B to D). As the delay increased, less 123I-GdDTPA-albumin was seen in the core of infarction (Fig. 5, D).
Regional blood volume
At onset of reperfusion the apparent blood volume of infarcted myocardium was substantially increased (280% ± 14% of normally perfused myocardium), consistent with hyperemia (Fig. 6). The regional blood volume progressively declined during reperfusion. At ½ h and 1 h there was no significant difference in blood volume between normal and postischemic myocardium (105% ± 11% and 106% ± 16%, respectively). At 24 h after reperfusion the regional blood volume was 69% ± 5% of that of normal myocardium.
The ΔR1 ratio of normal myocardium was 0.12 ± 0.01 and did not change significantly during 1 h of observation, which is consistent with a myocardial blood volume of approximately 12 ml/100 g (15,18). This value is similar to previously reported results as measured with various methods (19,20).
The area at risk defined by phthalocyanine blue was 47.1% ± 1.5% of the left ventricular surface area and did not differ significantly among the groups. In MR experiments, the shape and location of infarction on T1-weighted spin echo images was similar to the TTC negative region (Fig. 7). The current model of 1 h of ischemia followed by reperfusion produces substantial myocardial infarction evident as TTC negative area, which is in agreement with previous work (7). All animals subjected to 24-h reperfusion exhibited substantial hemorrhage (Fig. 7). Intramyocardial hemorrhage was less frequent when reflow was 2 h or less.
The major findings of the current study are 1) accumulation of GdDTPA-albumin as an indicator of ischemic microvascular injury can be quantified by IR-EPI; 2) MR and radioisotope measurements agree; 3) accumulation of GdDTPA-albumin in reperfused infarcted myocardium decreases with duration of reperfusion; 4) distribution of macromolecules in the injury zone is essentially uniform early after reperfusion, but later during reperfusion, GdDTPA-albumin accumulates predominately at the border of the injury zone; 5) reperfusion causes progressive decline of blood volume in reperfused infarction.
Regional blood volume
Substantial elevation of the blood volume in infarction at onset of reperfusion suggests a hyperemic reaction, which subsided 30 min after reperfusion. Similar findings have been described by other investigators (21,22). During reperfusion the blood volume in the infarct declined progressively. At 24 h after reperfusion the blood volume in the infarct was less than that of normal myocardium, which is consistent with occurrence of the “no reflow” phenomenon as described by Kloner et al. (23,24). The progressive decline of the blood volume in reperfused infarction suggests a progression of the no reflow phenomenon during reperfusion. This finding is consistent with results from Ambrosio et al. (22), who monitored serial changes in regional blood flow in dogs subjected to prolonged reperfusion using radioactive microspheres. These investigators observed progression of the no reflow phenomenon during reperfusion; progressive impairment of flow in postischemic myocardium was associated with capillary plugging (22). Recent contrast-enhanced MR studies confirmed the presence of microvascular obstruction in acutely reperfused infarction in patients (25).
Role of edema, lymphatic clearance and capillary permeability
Edema associated with elevated tissue pressure evolves in reperfused infarction (26). Elevated tissue pressure may result in compression force on the vasculature, ensuing declining blood volume and aggravating circulation disturbances (27). Edema is promoted by extravasation of macromolecules into the extravascular space and by degradation of the extracellular matrix (28). An elevated tissue pressure can impede the accumulation of macromolecules because transport of large molecules is predominantly driven by motion of bulk fluid (convection), which is governed by pressure gradients (29). Thus, in postischemic myocardium, GdDTPA-albumin would encounter a low transcapillary pressure gradient, and extravasation driven by convective forces would be reduced. Lymphatic clearance in postischemic myocardium is elevated, evidenced as increased flow and protein content of lymphatic fluid in dogs by 53% and 67%, respectively (30). This initial elevation, however, is followed by a gradual decline in lymphatic clearance as reperfusion evolves (31). The route for extravasation of albumin into the extravascular space is likely interendothelial gaps which increase during hypoxia (32)and reoxygenation (4)or reperfusion (33,34).
Regional distribution of labeled albumin
On autoradiography, an essentially uniform distribution of 123I-GdDTPA-albumin was seen only when the tracer was administered before reperfusion. When administered after reperfusion, there was less labeled albumin in the core of the injury zone than in the periphery. This pattern was more pronounced when the administration of the tracer was delayed after reperfusion. These findings are in agreement with a previous study where hearts were subjected to 1 h of occlusion and 1 h of reperfusion before biotin-labeled GdDTPA-albumin was administered (6). These investigators compared histologic analysis and T1-weighted spin echo images and found dispersion of biotin-GdDTPA-albumin throughout the interstitium and the myocytes; however, less tracer was detectable with either method in the infarct core as opposed to the rim (6). Other investigators studied the effect of various duration of ischemia followed by 2 h of reperfusion before administration of GdDTPA-albumin (7). They described a similar pattern on T1-weighted spin echo MR images ex vivo with a bright border zone and a less enhanced central core which increased in size with increasing (30 to 75 min) duration of ischemia (7).
In vivo assessment of microvascular injury
In distinction to previous research, the current study employed IR-EPI to measure the accumulation of GdDTPA-albumin in ischemically injured myocardium. Since IR-EPI allows quantitative measurements, it may be used to assess the potential of various new therapeutic agents designed to protect the myocardial vasculature during ischemia and reperfusion. The current study, however, is the first to demonstrate noninvasively in a quantitative fashion that the accumulation of macromolecules in postischemic myocardium declines when the administration is delayed up to 24 h after reperfusion.
Postinfarct alterations of microvascular integrity have recently been shown by Wu et al. (25)in patients using GdDTPA-enhanced MRI. These investigators performed serial measurements after contrast injection and demonstrated that microvascular obstruction, defined as hypoenhancement at 1 to 2 min and hyperenhancement at 5 to 10 min after injection, is predictive of poorer outcome. Other imaging modalities like positron emission tomography (35)and contrast echocardiography (36,37)have been used to assess microvascular injury in reperfused infarction. Myocardial-contrast echocardiography (36)has correlated the presence of contrast defects with poor functional recovery despite restored flow in the infarct-related artery. Unlike MRI, however, this method currently requires intracoronary injection of microbubbles, necessitating cardiac catheterization.
The blood volume estimates in the current study were based on the assumption that GdDTPA-albumin is present only in the intravascular compartment at the moment of injection. Therefore, extrapolation of ΔR1 ratio data back to the moment of injection was assumed to reflect the blood volume. However, severe vascular damage in the injury zone may allow substantial extravasation of GdDTPA-albumin immediately during first pass. Thus, overestimation of the true blood volume may occur.
☆ This study was supported by NIH Grant #T32HL070570-05, and #R01HL52569-01, National Institute of Health, Bethesda, Maryland. Dr. J. Bremerich and Dr. R. Wyttenbach were supported by grants from the Swiss National Science Foundation, Bern, Switzerland. Dr. H. Arheden was supported by the Swedish Heart Lung Foundation, Stockholm, Sweden.
↵1 All editorial decisions for this article, including selection of referees, were made by a Guest Editor. This policy applies to all articles with authors from the University of California San Francisco.
- analysis of variance
- inversion recovery echoplanar imaging
- magnetic resonance imaging
- R1 and T1
- longitudinal relaxation rate and time
- inversion time
- inversion time with nil signal arising from tissue
- TR and TE
- repetition and echo times
- triphenyltetrazolium chloride
- Received February 4, 1998.
- Revision received May 11, 1998.
- Accepted May 20, 1998.
- American College of Cardiology
- Kloner R.A.,
- Rude R.E.,
- Carlson N.,
- Maroko P.R.,
- DeBoer L.W.,
- Braunwald E.
- Dauber I.M.,
- VanBenthuysen K.M.,
- McMurtry I.F.,
- et al.
- McDonagh P.F.,
- Roberts D.J.
- Schwitter J.,
- Saeed M.,
- Wendland M.F.,
- et al.
- Wolfe C.L.,
- Moseley M.E.,
- Wikstrom M.G.,
- et al.
- ↵Tweedle MF, Wedeking P, Telser J, et al. Dependence of MR signal intensity on Gd tissue concentration over a broad dose range. Magn Reson Med 1991;22:191–4; discussion 195–6.
- Ogan M.D.,
- Schmiedl U.,
- Moseley M.E.,
- Grodd W.,
- Paajanen H.,
- Brasch R.C.
- Fluck D.S.,
- Etherington P.J.,
- O’Hare D.,
- Winlove C.P.,
- Sheridan D.J.
- Salisbury P.F.,
- Cross C.E.,
- Rieben P.A.
- Judd R.M.,
- Levy B.I.
- Crystal G.J.,
- Downey H.F.,
- Bashour F.A.
- Ambrosio G.,
- Weisman H.F.,
- Mannisi J.A.,
- Becker L.C.
- Wu K.C.,
- Zerhouni E.A.,
- Judd R.M.,
- et al.
- Granger H.J.,
- Glen A.L.,
- Barnes G.E.,
- Lewis R.E.
- Feola M.,
- Glick G.
- Nayler W.G.,
- Elz J.S.
- Jeremy R.W.,
- Links J.M.,
- Becker L.C.
- Ito H.,
- Tomooka T.,
- Sakai N.,
- et al.