Comparison of transient and permanent LAD ligation in mice using 18F-FDG PET imaging

Male C57/BL6 mice were purchased from Charles River (Sulzfeld, Germany). Animal care and all experimental procedures were performed according to the Guideline for the Care and Use of Laboratory Animals published by the U.S. National Institutes of Health (NIH publication no. 85–23, revised 1996). Myocardial infarction and ischemia–reperfusion injury were induced in 12–16 week-old C57BL/6 mice by surgical permanent (n = 10) or transient (n = 9) ligation of the left anterior descending artery, as described previously [3, 4]. The ECG of one mouse (MI group) was erroneous during the scan, and thereby no cardiac function parameters could be assessed. Healthy animals served as a control group (n = 8). In brief, mice were anaesthetized using intraperitoneal injection of fentanyl (0.05 mg/kg), midazolam (5.0 mg/kg), and medetomidine (0.5 mg/kg), and mechanically ventilated (MiniVent Ventilator model nr. 845, Harvard Apparatus^®). After left lateral thoracotomy, the proximal LAD was either ligated permanently (inducing MI) or transiently for 60 min, inducing IR injury. Study protocols complied with the institution’s guidelines and were approved by the Government’s animal ethics committee (ROB-55.2-2532.Vet_02-19-17 and ROB-55.2-2532.Vet_02-19-1).

ECG-gated 18F-FDG-PET scans were performed on post-operation day 6 and day 30, using a dedicated small-animal PET scanner (Inveon Dedicated PET, Preclinical Solutions, Siemens Healthcare Molecular Imaging, Knoxville, TN, USA) as described previously [13, 14]. The animals had free access to food (standard laboratory chow from Ssniff Spezialdiäten GmbH, Soest, Deutschland) and water, as described previously [9, 12‐15]. Anesthesia was induced (2.5%) and maintained (1.5–2%) with isoflurane delivered in pure oxygen at a rate of 1.5 L/min via a face mask. The core body temperature was maintained within the normal range using a heating pad and monitored by a rectal thermometer. Neonatal ECG electrodes (3 M, St. Paul, MN, USA) were placed on both forepaws and the left hind paw. Vital parameters were monitored and recorded using a dedicated physiological monitoring system (BioVet; Spin Systems Pty Ltd., Brisbane, Australia) [16]. After placing an intravenous catheter into a tail vein, approximately 20 MBq of 18F-FDG was injected in a volume of ~ 0.1 ml. The catheter was then flushed with 0.05 ml of saline solution. Animals remained anaesthetized during the entire scan and were placed in a prone position within the PET tomograph. A three-dimensional PET recording was obtained in list mode from 30 to 45 min after injection of the tracer. For attenuation and scatter correction, a 7-min transmission scan was performed with a rotating [57Co] source immediately after each PET scan, as described previously [12]. Recovery from anesthesia and the PET scan was monitored closely in the home cage by a veterinarian. The recorded data were processed with the Inveon Acquisition Workplace (Siemens Medical Solutions, Knoxville, TN, USA). FDG list-mode acquisitions were reconstructed, as described previously [12]. Reconstruction was performed using an OSEM 3D algorithm with 4 iterations (16 subsets) and a MAP 3D algorithm with 32 iterations (16 subsets) in a 128 × 128 matrix with a zoom of 211%. Data were either reconstructed as a static image or as gated images with 16 bins, normalized, corrected for randoms, dead time, and decay, as well as attenuation and scatter.

Analysis of PET images was performed by the Inveon Research Workplace (Siemens Medical Solutions) described previously [17, 18].

Inveon Research Workplace was used for assessing the percentage of the cardiac injected dose per gram (%ID/g) and left ventricular metabolic volume (LVMV) from static PET images. A cubic volume of interest (VOI) was drawn around the left ventricle, and a threshold value excluding 30% of the least hot voxels was applied. Correct VOI placement was verified in three projections (axial, sagittal, and coronal) [13]. In addition, ECG trigger signal accuracy was retrospectively verified using in-house software programmed in MATLAB (The Mathworks, Natick, USA) and C programming language [19].

The defect parameter was calculated with QPS^® (Cedars-Sinai, Los Angeles, CA, USA) using a normative database, as described previously [20]. In brief, infarct sizes were estimated with commercially available software (QPS^® 2012) from static FDG PET scans by creating polar maps of the left ventricle. In this approach, the FDG uptake in a given region is quantified as the percentage of the maximum tracer uptake in the polar map. The individual polar map is then compared to the average polar map from our normative database consisting of gender, strain, and age-matched animals (18 in total). The extent of the abnormally viable myocardium is calculated as a percentage of the left ventricle surface area.

Furthermore, the severity of the injured cardiac tissue is calculated in units of standard deviations below the normal tracer uptake. Extent and severity are combined into a single parameter corresponding to the infarct size expressed as a percentage of the left ventricular surface area. We used a similar approach as described by Slomka et al. [21] for myocardial perfusion data.

Left ventricular function parameters: end-diastolic (EDV), end-systolic (ESV) and the stroke volume (SV), and the left ventricular ejection fraction (EF), were calculated from ECG-gated images using QGS^® (Cedars-Sinai, Los Angeles, CA, USA), as described previously [12, 17].

On day 30, mice were sacrificed, and the hearts were excised. After fixation in 4% phosphate-buffered formalin, hearts were cut into 2 mm thick slices and embedded in paraffin. Five micrometer thick sections were cut and mounted on positively charged glass slides. Standard histological procedure (Sirius Red/Fast green Staining) was performed as described previously [22].

All results were expressed as means with standard deviation. One-way ANOVA analysis with Tukey’s multiple comparisons, unpaired Student’s t tests were used where appropriate. For groups without normal distribution, the Wilcoxon signed-rank or the Mann–Whitney U test was applied. The differences were considered statistically significant at a P value of 0.05.

The mouse hearts were evaluated using 18F-FDG PET imaging at days 6 and 30 post-transient or permanent LAD ligation. The defect assessment revealed a substantial defect in both models.

Permanent LAD ligation results in larger defect areas than transient ligation (illustrated in polar Bulls-eye map Fig. 1A, histology Fig. 1B and Fig. 1C, IR d6 vs MI d6, p <  0.001). Figure 1 B illustrates the histological evaluation after transient and permanent LAD ligation. The Sirius Red/Fast green staining was used to illustrate the fibrotic scar after injury. The transient LAD ligation leads to a smaller defect in the LV, due to the re-established perfusion of the infarct area. The defect after the permanent LAD ligation is far more severe and more cardiac mass is lost by the persistent obstruction of the LAD.

The defect among the two surgical approaches remained the same at 30 days post-operation (IR d30 vs MI d30, p <  0.001, Fig. 1C).

Neither after permanent nor after transient LAD ligation, a relevant change in the defect could be detected from day 6 to day 30 (MI d6 vs MI d30, ns; IR d6 vs IR d30, ns). Figure 1D and E is depicting the correlation of the histological defect assessment towards the PET defect in the IR model (r =   0.807, p =  0.008) and the MI model (r =   0.931, p <  0.001).

Serial cardiac PET imaging enabled evaluating the left ventricular metabolic volume (LVMV), which is an 18F-FDG derived parameter [13] and the cardiac percentage of injected dose (%ID/g).

The cardiac %ID/g increased in both models compared to the control mice (control vs IR d6, p = 0.019; control vs MI d6, p <  0.001, Fig. 2A) and remained elevated at day 30 after permanent LAD ligation (control vs MI d30, p = 0.004). Regarding the %ID/g change, we could not detect any significant alterations from day 6 to day 30 (IR d30-d6 vs MI d30-d6, ns, Fig. 3A).

LVMV in both IR and MI models was decreased at day 6 compared to the control group and LVMV was reduced in IR compared to MI at day 6 (control vs IR d6, p <  0.001, control vs MI d6, p <  0.001, IR d6 vs MI d6, p = 0.042, Fig. 2B).

We observed increased LVMV in both models at day 30 (IR d6 vs IR d30, p =  0.046; MI d6 vs MI d30, p <  0.001). No difference in LVMV in the two injury models after 30 days was observed (IR d30 vs MI d30, ns).

We further calculated the change of LVMV from day 6 to day30 (Fig. 3A). Here we detected an increase in LVMV after permanent compared to transient LAD ligation (IR d30-d6 vs MI d30-d6, p =  0.049). Of note, we could not detect any difference in the maximum injected dose in the heart (Fig. 3B).

Comparing the defect and the change in LVMV displayed a significant correlation after permanent LAD ligation (day 6: r =   0.749, p =  0.013; and day 30: r =   0.832, p =  0.003; Fig. 3C). There was no correlation among the defect after transient LAD ligation and the change of LVMV from day 6 to day 30 (Data not shown). Figure 3D illustrates the correlation of the histological defect with the change in LVMV in both model (r =   0.500, p =  0.029).

Next, we assessed the alterations in cardiac volumes and function parameters to further enable the correlative analysis to the LVMV after surgical LAD ligation.

We detected an increased EDV on day 6 after permanent LAD ligation compared to the control and the transient ligation (Fig. 4A showing the 3D-reconstruction and quantification from QGS software, Fig. 4B, control vs MI d6, p =  0.037).

At day 30, the EDV further increased after MI, demonstrating ongoing LV dilatation (control vs MI d30, p <  0.001; MI d6 vs MI d30, p =  0.005; IR d30 vs MI d30, p =  0.002). The EDV in the reperfusion model increased but not achieve statistically significance over time (IR d6 vs IR d30, ns).

Additionally, we could note changes in the ESV (control vs MI d6, p <  0.001; control vs MI d30, p <  0.001; IR d6 vs MI d6, p =  0.003; IR d30 vs MI d30, p <  0.001; Fig. 4B).

The SV decreased after MI at day 6 (control vs MI d6, p =  0.012; Fig. 4B) and increased over time (MI d6 vs MI d30, p =  0.0061).

In line with the defect after MI, the EF was significantly lower after permanent LAD ligation than IR injury (control vs MI d6, p <  0.001; control vs MI d30, p <  0.001; IR d6 vs MI d6, p <  0.001; IR d30 vs MI d30, p <  0.001, Fig. 4B). After transient ligation, the EF was preserved after day 6 (control vs IR d6, ns; control vs IR d30, ns).

Furthermore, we could detect no improvement in the EF after 30 days of follow-up (IR d6 vs IR d30, ns; MI d6 vs MI d30, ns). Correlations of the cardiac function to the histological defect such as positive correlation to EDV (r =   0.618, p =  0.006), ESV (r =   0.626, p =  0.005), and negative correlation to EF (r =   -0.563, p =  0.015) underlining the methods itself are displayed in the supplement.

Finally, we evaluated the correlation of LVMV to cardiac volume and function estimated by the QGS software for both models.

We could detect a positive correlation after permanent LAD ligation to EDV (r =   0.609, p =  0.007), ESV (r =   0.522, p =  0.026) and SV (r =   0.688, p =  0.002) (Fig. 5).    

After transient LAD ligation, we as well detected several positive correlations: EDV (r =   0.714, p <  0.001), ESV (r =   0.592, p =  0.009), and SV (r =   0.540, p =  0.021).