High-resolution (11.7 T) cardiac magnetic resonance imaging (MRI) and histological approaches have been employed in tandem to characterize the secondary damage suffered by the murine myocardium following the initial insult caused by ischemia-reperfusion (I/R). I/R-induced changes in the myocardium were examined in five separate groups at the following time points after I/R: 1 h, day 1, day 3, day 7, and day 14. The infarct volume increased from 1 h to day 1 post-I/R. Over time, the loss of myocardial function was observed to be associated with increased infarct volume and worsened regional wall motion. In the infarct region, I/R caused a decrease in end-systolic thickness coupled with small changes in end-diastolic thickness, leading to massive wall thickening abnormalities. In addition, compromised wall thickening was also observed in left ventricular regions adjacent to the infarct region. A tight correlation (r2 = 0.85) between measured MRI and triphenyltetrazolium chloride (TTC) infarct volumes was noted. Our observation that until day 3 post-I/R the infarct size as measured by TTC staining and MRI was much larger than that of the myocyte-silent regions in trichrome- or hematoxylin-eosin-stained sections is consistent with the literature and leads to the conclusion that at such an early phase, the infarct site contains structurally intact myocytes that are functionally compromised. Over time, such affected myocytes were noted to structurally disappear, resulting in consistent infarct sizes obtained from MRI and TTC as well as trichrome and hematoxylin-eosin analyses on day 7 following I/R. Myocardial remodeling following I/R includes secondary myocyte death followed by the loss of cardiac function over time.
acute myocardial damage caused by a surge in reactive oxygen species following ischemia-reperfusion (I/R) has been demonstrated and extensively studied (25, 26, 39). The secondary loss of myocytes subsequent to the initial insult is a progressive phase reported by numerous laboratories (2, 3, 12, 15, 18, 31, 32) and remains poorly understood. In addition, the functional significance of this secondary death process remains to be characterized.
The current state of magnetic resonance (MR) imaging (MRI) represents a versatile noninvasive tool for measuring cardiac function and structure (7, 8, 24, 28, 34, 37). Cardiac MRI has been applied to study I/R injury (8–11, 23, 24, 35–37), left ventricular (LV) remodeling (13), and myocardial metabolism (5, 19, 33) in canine as well as rodent models. Contrast-enhanced MRI monitors infarct damage after myocardial infarction (9, 10, 21, 30, 34). Because these observations are obtained from rodent as well as canine models, it is important to recognize that the rodent and canine models do have some contrasting features, primarily because of a higher abundance of collateral circulation in dogs. Unlike in rodents where complete ischemia is achieved, most dog models represent an uncontrolled mixture of incomplete and complete ischemia, which gives rise to a delay in cell death in those areas that were never totally ischemic. Murine MRI has been hampered by the lack of widespread availability of high-resolution scanners that provide requisite spatial and temporal resolution. The high (400–600 beats/min) heart rate of the anesthetized mouse coupled with limited dimensions (5 mm in diameter and 8 mm in length) of the heart poses significant challenges for routine high-resolution cardiac MRI. The ability to noninvasively and accurately perform the repeated examination of cardiac parameters without operator bias over a period of time after a myocardial insult is a major advantage of MRI over other techniques including echocardiography, hemodynamic measurement techniques, and histology. In this study, we have simultaneously employed histological analyses and 11.7-T high-resolution cardiac MRI (2 to 3 time points/mouse) to study LV function, structure, and infarct volume in mice after I/R injury.
MATERIALS AND METHODS
Animal and experimental protocol.
All animal procedures were approved by the Institutional Laboratory Animal Care and Use Committee of the Ohio State University. Thirty-one young adult (10–12 wk of age) male C57BL/6 mice (Harlan Technologies) were used in this study. The vast majority (94%) of the animals survived the study protocol. Baseline MR images were obtained 5–7 days before surgery to ensure the normalization of all cardiovascular parameters. The mice were divided into five groups (n = 4–9 animals) and were euthanized for harvesting heart tissue accordingly. The mice were euthanized either at 1 h, day 1, day 3, day 7, or day 14 postreperfusion immediately after MRI.
Histological determination of infarct size.
The mice were euthanized, and the hearts were excised and encased in 2% agarose solution. After the agarose was settled for 2 min, the heart was laterally cut just below the left auricle, and 1-mm-thick sections were made. The agarose was removed, and the sections were incubated with 2% 2,3,5-triphenyltetrazolium chloride (TTC) at 37°C for 20 min, with rotation every 2 min to allow uniform tissue staining. Each slice was photographed with a digital camera mounted on a dissecting microscope. Following photography, the sections were fixed in 10%-buffered Formalin solution for 2 wk after which they were processed for hematoxylin-eosin staining. ImageJ software (National Institutes of Health) was used for the planimetry of the LV border and infarct region on TTC images. Infarct volume was calculated as a percentage of total LV volume (including septum). The exact correlation between the TTC and MRI slice was not always possible because of slice thickness mismatch and partial voluming effects. Masson's trichrome and hematoxylin-eosin staining were done on the same heart sections used for TTC staining. Standard paraffin embedding and tissue processing protocols were used. Sections of 5-μm thickness obtained from each 1-mm-thick slice were mounted on glass slides and stained with hematoxylin-eosin and standard Masson's trichrome staining protocol. The stained sections were photographed using a digital camera (model D30; Hitachi) mounted on a microscope (Axiovert 200M; Zeiss). Manual planimetry was performed on the microscope using PALM RoboSoftware v2.2 to calculate infarct volume using hematoxylin-eosin-stained slices. The myocyte-silent space in hematoxylin-eosin sections was interpreted as the infarct area in this study.
MR image acquisition.
In the animal preparation for MRI, the mice were anesthetized with 1.5% isoflurane mixed with 1.5 l/min carbogen. Subdermal ECG leads were fixed on the right forelimb and right leg of the animal. A respiratory sensor was placed under the abdomen and secured to the animal. Gadolinium-diethylenetriaminepentaacetic acid-bismethylamide (Gd-DTPA-BMA) contrast agent (Omniscan) was injected intraperitoneally at a dose of 0.6–0.9 mmol/kg of body wt just before placing the animal in the magnet. Core body temperature was maintained at 37.0 ± 0.5°C using circulating warm air. Core body temperature, respiration, and ECG were monitored with the SAII model 1025 monitoring and gating system (Small Animal Instruments, Stony Brook, NY). The MRI protocol included localizer FLASH scans, short-axis cine loops (6–8 slices), and contrast-enhanced infarct imaging (6–8 slices). MRI was performed on a Bruker 11.7-T vertical MR scanner with a maximum gradient strength of 1,000 mT/m. A 30-mm birdcage radiofrequency coil was used. Shim currents were initialized to previously saved values from earlier imaging experiments, and automatic shimming was then done. The proton line width of <600 Hz (1.2 ppm) was routinely obtained in <30 s and yielded good signal to noise ratio (SNR) on the MR images. After localizer scans were completed, an ECG-triggered two-dimensional FLASH pulse sequence was used to generate cine loops of cardiac motion. For baseline images, a flip angle of 20° was used since it generated good contrast between blood and the myocardium at heart rates from 375 to 500 beats/min (R-wave-R-wave interval from 160 to 120 ms). The flip angle was increased to 60° (to increase T1 weighting) for postcontrast images in an attempt to obtain contrast-enhanced images concurrently with cine images (38). Sixteen frames were obtained in each cardiac cycle that led to a reception time between 7.5 and 10 ms. Echo time was kept constant at 1.43 ms. A spatial resolution of 117 μm in plane (slice thickness, 1 mm; field of view, 3 cm × 3 cm; and matrix size, 256 × 192) was obtained. To obtain contrast-enhanced images, an ECG-triggered inversion recovery (I/R) sequence with an inversion time of 150–200 ms was used. Double inversion preparation pulses were used as required to null blood signal and improve image quality. This sequence was used to obtain 6–8 1-mm-thick sections at 117 μm in-plane resolution. Contrast-enhanced images were acquired 30–45 min postcontrast administration to accurately measure infarct volume (21, 34). The contrast-enhanced images from cine loops were later correlated with inversion recovery images to provide additional frames of reference for the measurement of infarct area. The full imaging protocol was completed within 1 h after anesthetizing the animal.
Determination of cardiac functional parameters.
The MR images were converted to DICOM format using Paravision 4.0 and processed with ImageJ software. After the end-diastolic (ED) and end-systolic (ES) phases were identified on a slice-by-slice basis, the endocardial and epicardial borders were traced and LV ES volume (ESV), LV ED volume (EDV), LV stroke volume (SV), cardiac output (CO), and LV ejection fraction (EF) were computed from the traced borders (24). LV mass was calculated by multiplying the LV myocardial volume with density (34).
Determination of infarct volume.
The contrast-enhanced MR images in DICOM format were processed in ImageJ. Appropriate contrast enhancement of the images was done to maximize the signal from the hyperintense region and the null signal from the nonenhanced region. Manual planimetry was performed on images obtained from the inversion recovery imaging sequence. To help exclude artifacts during planimetry, contrast-enhanced images obtained with the FLASH cine sequence were examined as required to provide an additional reference. For each slice, the hyperenhanced region and total LV myocardial area were calculated. Slice hyperintense areas were then summated to generate infarct volume as a percentage of LV myocardial volume.
Determination of LV wall thickening.
A slice from the midventricular section of the heart, 4.2 mm from apex, was chosen for the analysis of segmental wall thickness to determine abnormalities in LV wall motion compared with that of the control group. Typically, this slice had both infarcted and noninfarcted regions. Wall thickness analysis was conducted employing software developed in our laboratory using Matlab. After epicardial and endocardial borders in end systole and end diastole were contoured, the LV wall thickness was calculated at each 2° angular distance from the center of the LV. Papillary muscles were not included in the wall thickening analysis. Twenty-two contiguous wall thickness measurements were averaged to segment the LV into eight equiangular sectors of 45° each (Fig. 8A). The averaged wall thickness was indexed to the right ventricular insertion point in the anterior wall and was plotted against the corresponding angle at each time point after I/R injury. Fractional shortening (FS) was calculated as FS = (wall thickness at end systole/wall thickness at end diastole) − 1.
Planimetry and the analysis of segmental LV wall thickness were performed by a trained investigator blinded to the study groups. All values are reported as means ± SD. Paired and unpaired two-tailed Student's t-tests were used where appropriate. One-way ANOVA followed by Tukey's pairwise comparisons test was used to compare parameters between different groups. A value of P < 0.05 was considered statistically significant.
ECG recordings during surgery were helpful in objectively discriminating between successful and failed LAD occlusions (supplementary Fig. S1; note: all supplementary material can be found with the online version of this article). All mice that developed myocardial infarction, as later proved by MRI and histology, exhibited a strikingly pathological, infarct-typical ECG pattern during and after the surgical procedure. The pathological ECG pattern was observed for the complete duration of the study. Early manifestations of infarction included reduced R-wave amplitude, marked ST-segment elevation, and the development of a large Q-wave (supplementary Fig. S1). The MRI protocol used in the current study involved the intraperitoneal administration of a contrast agent outside the MR scanner. The time course of the hyperenhancement of the LV was followed to obtain a time window where MRI reflected true infarct size as later confirmed with TTC staining. The earliest contrast-enhanced images were obtained 20 min following contrast administration. A thin layer of viable tissue was observed between blood and a region of hyperintense myocardium, which allowed for an easy delineation of the infarct region (arrows in Fig. 1A). In a mouse subjected to serial analysis over the first 2 h after contrast injection, infarct size measured with MRI remained constant from 20 to 60 min postcontrast injection (Fig. 1B), thus providing a wide window of time for acquiring MR images. In some regions of the heart, hyperintensity was noted even up to 2 h after contrast injection. Sections of the myocardial tissue were stained with TTC after MRI (Fig. 2, A and B). Infarct volumes calculated from MRI obtained 30–45 min after contrast administration and TTC were plotted against each other (Fig. 2C). A tight correlation (r2 = 0.85, slope = 1.07, and intercept = −0.87) between measured MRI and TTC infarct volumes was noted. The correlation was especially strong in infarcts smaller than 20% of the LV.
Time-dependent changes in I/R-induced myocardial infarction were examined in mice with MRI over 14 days after surgery (Fig. 3). Separate groups of mice were imaged at five time points after I/R: 1 h, day 1, day 3, day 7, and day 14. Infarct volumes at each time point were calculated as described in materials and methods. Maximum infarct size was observed in the group assessed at day 3 after I/R surgery, and infarct size in this group was significantly larger than that in the 1-h group and the group assessed at day 14 post-I/R. TTC staining on heart sections was performed immediately after MRI (Fig. 4). Infarct sizes measured with TTC closely matched those measured with MRI. Both data sets consistently demonstrated infarction until day 7 after I/R. Of note, the hematoxylin-eosin staining of tissue revealed very little structural loss of myocytes at 1 h (0.3% LV), day 1 (1.03%), and day 3 (2% LV) following I/R. Infarct size measured by hematoxylin-eosin staining on the same heart sections on which MRI and TTC staining were done increased to a maximum at day 7 following I/R (Figs. 5 and 6). Infarct volume over time measured with the three different techniques is plotted in Fig. 6. An apparent disagreement among the techniques was observed at acute time points of 1 h, day 1, and day 3 after I/R injury.
I/R insult caused a statistically significant worsening of all global LV functional parameters (Fig. 7). There was a significant increase in EDV between baseline and day 7 and a significant increase in ESV between baseline and day 1. This resulted in significant decreases in SV, CO, and EF at every time point relative to baseline. The worst cardiac function was measured at day 7, where I/R caused a 46% increase in EDV, 340% increase in ESV, 25% decrease in SV, 32% decrease in CO, 47% decrease in EF, and 38% increase in LV mass. The analysis of regional wall thickening indicated a trend toward worsening wall motion in the groups assessed at later time points after I/R (Fig. 8). Statistically significant differences in LV wall thickness were observed across the infarct zone and adjacent regions at end systole and end diastole, which led to severe differences in FS during the cardiac cycle. A small improvement in FS was observed in remote LV regions 180° away from the infarct zone. The differences in wall thickening that reached statistical significance are indicated in Table 1. LV ES thickness, ED thickness, and FS were compared at the following five time points: baseline, day 1, day 3, day 7, and day 14 following I/R. For each myocardial sector, these wall motion parameters were serially compared with a different group of mice assessed at the previous time point. Such analysis revealed significantly worsened wall motion post-I/R in several myocardial segments. This observation is consistent with the noted loss of cardiac function following I/R over time.
Timely coronary reperfusion of the ischemic myocardium is aimed at reducing myocardial dysfunction and improving tissue survival. Reperfusion, however, also poses the threat of reperfusion injury. Reperfusion injury may be broadly viewed as having two major components: oxidant-induced myocardial tissue damage during the reperfusion process and progressive tissue damage secondary to the initial insult (3, 32). When compared with the first component, the process of progressive tissue damage after reperfusion remains poorly understood. Acute myocardial infarction in humans is associated with the activation of myocyte death in the surviving peri-infarct portion of the heart (20). It has been proposed that the secondary progressive myocardial tissue damage is aimed at clearing nonsalvageable cells (12). Others have proposed that excessive mechanical forces associated with increases in ventricular loading following I/R cause the secondary progressive tissue damage (15). Yet others have implicated post-I/R inflammation to cause the secondary progressive myocardial tissue damage (14, 31). In contrast, products of post-I/R inflammation, e.g., TNF-α, have been shown to induce cytoprotective signals that prevent and/or delay the development of cardiac myocyte death secondary to acute ischemic injury (18). In this study, we employed MRI and histological approaches to characterize post-I/R myocardial tissue damage from both functional as well as structural standpoints. High-resolution imaging performed on the 11.7 T MRI scanner with large gradient strength and fast slew rate allowed a superior SNR ratio and contrast between hyperintense and nonischemic regions. Voxel volumes of ∼0.01 mm3 and a temporal resolution of 6 ms were routinely obtained to generate cine loops, which were utilized to analyze cardiac parameters. The use of a double inversion preparation pulse to null blood signal proved critical to enhance contrast with the noise ratio in certain cases. A tight positive correlation was found between infarct volumes measured with MRI and TTC, consistent with previous studies (9, 10, 16, 22, 23). It has been previously suggested that contrast-enhanced MRI overestimates the size of myocardial infarcts when images are acquired much less than 20 min after contrast injection (11, 21). We observed that infarct volume measured with MRI remains practically constant from 20 to 60 min after contrast administration, thus providing a wide window of time where MR images could be acquired.
TTC staining of tissue is coupled with the state of mitochondrial metabolism. Tissue with dysfunctional mitochondria does not stain with TTC, resulting in a pale appearance (4). In this study, infarct volumes measured in separate groups of mice showed significant differences from 1 h to day 1 but no statistical difference in size from day 1 to day 7 post-I/R. With the assumption that infarct volumes were similar between groups, these data suggest a progressive increase in infarct volume between 1 h and day 1 post-I/R. Moreover, this observation must be tempered by the fact that no statistical difference in infarct size was found between the day 1 compared with the day 3 and day 7 post-I/R results. The significant decrease in infarct size observed between day 3 and day 14 post-I/R can be attributed to the wound contracture that accompanies scar maturation. Myocyte-deficient regions of the infarct site are known to be richly populated with fibroblasts that differentiate to myofibroblasts over time (17, 27, 29). Such myofibroblasts may account for the wound contraction noted between day 7 and day 14 after I/R. I/R injury compromises LV function (8, 24, 35, 37). Concurrent with the changes in infarct volume observed using all three techniques after I/R, a time-dependent change in LV function was also noted in this study. Increased LV, EDV, and ESV and reduced CO and EF were consistently observed until day 7 following I/R. A marginal gain of LV function was observed on day 14, but it was still lower than baseline values, suggesting a more permanent change in cardiac function because of the injury. The segmental LV wall motion analysis performed in the current study points to a time-dependent worsening of wall motion after I/R. In the infarct region (20–200° on most slices), I/R caused a decrease in ES thickness coupled with small changes in ED thickness, leading to massive wall thickening abnormalities. This observation must be tempered by the fact that this was not a serial study. Furthermore, no attempt was made to distinguish between infarcted and noninfarcted sectors in the current study. The loss of wall thickening was associated, albeit not significantly, with the loss of cardiac function and increased infarct volume. In addition, compromised wall thickening was also observed in LV regions adjacent to the infarct region consistent with previous findings (8, 24, 34). Interestingly, an improvement in FS was observed over baseline in remote (250–360°) regions of LV. The recovery of cardiac contractility in remote regions at day 14 after I/R occurred simultaneously with morphological adaptations and was paralleled by a significant increase in ventricular weight. Part of this hypertrophic response may be attributed to normal physiological growth of the tissue. However, the enlargement of EDV and the relative thickening of remote regions of LV suggest an eccentric hypertrophic response to I/R injury. De Celle et al. (6) studied LV remodeling in mice until 8-wk post-I/R and also reported a positive correlation between infarct size and ventricular weight. This long-term remodeling response can be explained as a compensatory mechanism for reduced cardiac contractility caused by time-dependent functional and structural damage caused by I/R. Findings of the current study would be best interpreted in light of the practical limitations in execution. MRI was performed at two to three time points per group. All conclusions drawn regarding progressive changes in structure and function are based on the assumption that infarct size was similar between the five groups of mice. With respect to the segmental analysis of FS, it should be noted that any conclusions drawn regarding progressive changes in wall motion are based on the assumptions that both the size and spatial distribution of infarct were equivalent between the different groups of mice.
In summary, we present the first evidence from tandem high-resolution cardiac MRI and histological studies aimed at characterizing the temporal changes in LV function, infarct size, and regional wall motion in mice after I/R. A time-dependent loss of myocardial function was observed, which was associated with increased infarct volume from 1 h to day 1 post-I/R and worsened regional wall motion. The histological determination of infarct size did not match with MRI and TTC at acute time points after I/R, demonstrating the presence of a small number of structurally intact but functionally compromised myocytes at the infarct site after I/R injury. This observation is consistent with the literature (1). With the progression of time, such affected myocytes were noted to structurally disappear, resulting in consistent infarct sizes obtained from MRI and TTC as well as trichrome and hematoxylin-eosin analyses on day 7 following I/R. Specific molecular mechanisms implicated in such myocyte death in the peri-infarct region following I/R remain to be characterized.
This study was supported by National Heart, Lung, and Blood Institute Grant RO1-HL-073087 (to C. K. Sen).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 by the American Physiological Society