At-risk but viable myocardium in a large animal model of non ST-segment elevation acute coronary syndrome: cardiovascular magnetic resonance with ex vivo validation

An overview of the study protocol is shown in Figure 1. All the animal procedures were approved by The Ohio State University Institutional Animal Care and Use Committee and conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health. Thirteen heartworm-free, random-source, mixed-breed dogs (aged 1-3 years, weight 11.4-15.9 kg) were anesthetized using ketamine (5-8 mg/kg, IM) and inhaled isoflurane (1-2% in 95% O₂) and remained anesthetized throughout all experimental procedures. In 2 randomized dogs, no surgical operations were performed (unoperated shams). ECG was monitored via surface electrodes. After left thoracotomy in the remaining 11 dogs, a catheter was placed in the left atrial appendage for microsphere injection. In a random selection of 4 dogs, no further operations were performed prior to closing of the chest wall (shams). In the 7 model dogs, an 18-gauge needle was laid flat on the left circumflex coronary artery (LCx) then removed once a suture was secured around the artery and needle, leaving a lumen constricted to the nominal outer diameter of the needle (~1.27 mm). Copper pacing wires were then threaded through the epicardium of the right ventricle and connected to an external electrical stimulator for later myocardial pacing. Standard chest wall closing procedures and post-operative anesthesia and analgesia protocols were then followed [10].

Serum isolated from peripheral blood drawn at 3 time points (0-baseline prior to surgery, 1-post-stenosis and post-pacing in model dogs or at the end of CMR examination in sham dogs, and 2-just before euthanasia) was examined for troponin I (TnI) content using a high-sensitivity assay according to the manufacturer’s recommendations (TnI-Ultra assay and ADVIA Centaur XP immunoanalyzer, Siemens Healthcare Diagnostics; Malvern, PA).

Based on the anatomical perfusion territory of the LCx and preliminary results from pilot studies, the lateral LV myocardium at the level of the posterior papillary muscle was denoted as the affected region while a section of the anterior LV wall near the anterior papillary muscle was designated as the remote region. At the conclusion of each imaging experiment, the animal was euthanized and the heart immediately harvested. Myocardium from affected and remote regions was excised using the papillary muscles as anatomical landmarks. Tissue samples from each myocardial region were obtained for immediate tissue respiration, wet-to-dry weight and tissue necrosis analyses on fresh tissue. Additional samples were acquired and either fixed or frozen (-80°C) for later light and electron microscopy or gene expression analyses, respectively, as discussed below and in the provided online supplement (see Additional file 1).

CMR was initiated within 30 minutes of chest wall closure. All CMR scans were performed on a 1.5 Tesla system with a 12-element phased-array cardiac coil (MAGNETOM Avanto, Siemens Medical Solutions, Inc.; Erlangen, Germany). Identical imaging protocols were used in both groups of animals, NSTE-ACS model and sham. In the NSTE-ACS model dogs, scans were performed immediately after 15 minutes of epicardial pacing at 180-210 beats/min (60-70% of normal canine maximum heart rate vs. average stable post-anesthesia heart rate of 100 beats/min). We acquired multiplane cines to measure LV systolic function, T2 maps to identify areas of acute injury, first-pass perfusion imaging of microcirculatory myocardial blood flow, and LGE images to visualize areas of necrosis. Further detailed information is provided in Additional file 1.

Two investigators analyzed all images by consensus review (HC and TT). For analysis of T2 maps, myocardial regions of interest (ROIs) were drawn encompassing the affected region and remote region for each dog (typically in the regions of segments 9 and 12 by standardized LV segmentation nomenclature [11]), and the mean T2 in each ROI was recorded. Perfusion images were visually scored by counting the number of LV segments that showed abnormalities. LGE images were similarly evaluated using consistent display settings. LV ejection fraction was calculated using Simpson’s rule and manual segmentation of the endocardial border on end-systolic and end-diastolic frames of contiguous short-axis cines.

To verify the changes in coronary blood flow due to stenosis and pacing, we injected fluorescent microspheres (BioPAL, Inc.; Worcester, MA) into the left atrium of 5 animals (2 NSTE-ACS model and 3 sham) at three time points, each with a different fluorescent label: 1) prior to surgical stenosis of the LCx (gold), after stenosis and pacing (samarium), and just prior to euthanasia (lutetium). The disintegrations per minute (dpm) of each label was quantified (BioPAL, Inc.) in tissue sections (~1 gram) obtained from both the affected and remote regions. After calibration, the dpm of each label was converted to flow in the affected region relative to the flow in the remote region at each of the three time points.

Examinations were made on myocardial tissue sections from both the affected and remote regions in all animals to determine the presence of tissue necrosis and edema using established 2,3,5-triphenyltetrazolium chloride (TTC) staining and wet/dry weight ratio measurement, respectively, as described further in Additional file 1.

Tissue sections harvested from the remote and affected myocardial regions were routinely processed and stained [12] (further detailed in Additional file 1). Tissue histology was graded blindly for evidence of structural and/or cellular injury/abnormalities by pathologist (PBB) review. High-resolution electron microscopy digital image analysis of myocardial mitochondrial ultrastructure was performed [13], and values obtained for the affected myocardial mitochondria were normalized to those averaged from the corresponding remote myocardial mitochondria for the sake of comparison between the groups.

To evaluate the effect of ischemia upon cellular respiration in our NSTE-ACS model, O₂ consumption by myocardial tissue (~0.2-0.3 grams) obtained from both the affected and remote regions was measured in all animals employing standard techniques as further described in Additional file 1. The rate of mitochondrial respiration in the affected region was normalized to that of the remote region in each animal and compared between the model and sham groups.

Frozen harvested tissue samples from both the affected and remote myocardial regions obtained from each animal were processed according to standard techniques to isolate high quality total RNA, as detailed in Additional file 1. Gene expression for select canine genes in both the affected and remote myocardial regions was determined using quantitative real-time PCR. Relative copy numbers and expression ratios for all genes of interest were normalized to that of the housekeeping gene, β-actin, and then compared within the model group.

Results are reported as mean ± standard deviation (SD). In vivo and ex vivo measurements for affected myocardial tissue respiration in model vs. sham groups were compared by use of unpaired t-tests. Delta T2 (affected minus remote) was also compared between the two groups using unpaired t-tests. For serial microsphere and TnI data, a two-way analysis of variance (ANOVA) test was performed to examine the individual contributions of animal group and sampling time, as well as their interaction effects. The T2 data was also analyzed in this way using animal group and region (remote and affected) as the main effects. Comparisons of selective gene expression between the affected and remote regions within the model group were made using the Wilcoxon Rank-Sum test. Statistical significance was based upon a value of p < 0.05.

LV ejection fraction by CMR acquired during anesthesia was mildly reduced in the model dogs with LCx stenosis post-pacing compared to the sham dogs (40 ± 2 vs. 48 ± 8%, p = 0.03). Myocardial T2 was increased in the LCx-territory myocardium relative to remote myocardium in NSTE-ACS model dogs after pacing (Figure 3), averaging 53.2 ± 4.9 ms in the affected region vs. 43.6 ± 2.8 ms in the remote region. A smaller regional difference in myocardial T2 was also present in sham animals (48.6 ± 2.1 vs. 43.4 ± 2.9 ms; Figure 4), somewhat larger in those that were unoperated (46.3 ± 1.6 vs. 40.0 ± 0.6 ms) vs. operated (49.7 ± 1.1 vs. 45.2 ± 1.4 ms). Two-way ANOVA showed significant contributions to T2 due to myocardial region (p < 0.001), animal group (p = 0.02), and region-group interaction (p = 0.03). Perfusion and LGE images showed no first-pass defect, microvascular obstruction or enhancement (Figure 3) in any model or sham-operated animals.

An average of 38,395 microspheres was collected for each reading. In the NSTE-ACS model dogs, the flow through the affected region relative to the flow through the remote region decreased after stenosis and pacing with a mean end-experiment flow ratio of 0.53 ± 0.36. Lack of stenosis and pacing in the sham dogs resulted in no change in myocardial blood flow over the course of the experiment where end-experiment flow ratio averaged 1.0 ± 0.10 (Figure 5A). Two-way ANOVA showed significant contributions to this flow ratio due to time (p < 0.001), animal group (p < 0.001), and time-group interaction (p < 0.001). In six sham animals, the myocardial respiration rate from the affected and remote regions was 0.44 ± 0.10 nmol O₂/mg/min and 0.43 ± 0.08 nmol O₂/mg/min, respectively, yielding an average respiration rate ratio of 1.04 ± 0.07. In four NSTE-ACS model dogs, the myocardial respiration rate from the affected and remote regions was 0.44 ± 0.07 nmol O₂/mg/min and 0.61 ± 0.05 nmol O₂/mg/min, respectively, resulting in a significantly lower respiration rate ratio of 0.72 ± 0.07 when compared to the sham group (p < 0.001, Figure 5B). These findings are in keeping with previous investigations wherein tissue hypoxia was shown to limit pacing-induced aerobic respiration [14], indicating that the affected myocardium in the NSTE-ACS model was operating at or near its maximal aerobic capacity at baseline.

Immediate post-mortem analysis of fresh tissue revealed comparable wet-to-dry ratios in the affected region in the NSTE-ACS dogs compared to the matching remote regions or compared to matching myocardial tissues from sham animals. In addition, the results demonstrated no change in wet-to-dry ratios related to the surgical preparation (Additional file 1: Table S3). Staining of fresh tissues obtained from the affected and remote myocardial regions with TTC was negative indicating an absence of non-viable myocardium in tissue samples (2-3 mm slice thickness) corresponding to the abnormal T2 signal by this technique (Additional file 1: Figure S1).

In keeping with the TTC findings and negative LGE results on CMR, light microscopy generally showed no increase in contraction band necrosis, a marker of irreversible cellular injury, in the affected region. Small areas with contraction band-like patterns were seen in both sham and model groups, suggesting an artifact of the tissue cutting and preparation process or transient myocardial fasciculations in the immediate post-mortem period. Unexpectedly, mitochondrial size was statistically greater in the myocardium of the lateral wall compared to the anterior wall in all animals, but normalized data indicated no mitochondrial swelling relative to controls or other manifestations of mitochondrial damage in the NSTE-ACS dogs (Figure 6).

Enhanced expression of CaMKIID, a pleiotropic kinase that is responsive to changes in intracellular calcium [15], was observed within 3 hours in the affected myocardium in the NSTE-ACS model (Figure 7A). There was no significant change in HIF-1α gene expression in the affected myocardium (Figure 7B). In keeping with previous studies showing that myocardial mitochondrial biogenesis is promoted in response to ischemia [16], T2 changes within the affected myocardium of the NSTE-ACS animals correlated with increased mitochondria-related gene expression (Figure 7C).

Despite the advantages of this large-animal model (e.g., compared to rodents [36]) in terms of modeling human ischemic heart disease, we recognize certain model limitations. First, we did not extend the period of ischemia beyond a 3-4 hour window; as such, the consequence of the observed T2 alterations in terms of longer-term tissue viability remains unclear. Due to concerns regarding adequate myocardial tissue availability after histopathological and oxygen respiration analyses, and the lack of invasive catheterization in unoperated shams, microspheres were only delivered in 5 animals. While we did not directly measure coronary artery stenosis or absolute coronary blood flow rates (only relative flow rates were obtained), this model produces mechanical restriction of the cross sectional coronary diameter (as dictated by the surgical protocol) and was shown to reliably reduce tissue perfusion, particularly following demand pacing, as reflected by tissue microsphere quantification. Absolute flow measurements may be incorporated into future studies. In terms of tissue bioenergetics, while it would be optimal to measure in vivo myocardial VO₂ and ATP levels, the demonstrated decrease in tissue VO₂ measured ex vivo under conditions of excess oxygen supply and associated changes in CaMKIID and mitochondrial biogenesis are in keeping with established bioenergetics adaptations to tissue hypoxia [15, 37]. It is postulated that a lower metabolic state (stunning or hibernation) and a transition from aerobic towards anaerobic respiration is an adaptation that prevents lethal ATP depletion in the context of limited oxygen supply; further investigation using techniques such as phosphorus spectroscopy would help corroborate or refute this proposition.