This study was designed as a single group of hyperlipidemic (HL) pigs for two purposes: to evaluate the extent and severity of atherosclerosis and RAGE expression in this large animal model and to see whether the signal from a radiolabeled antibody coming from this receptor can be visualized on in vivo imaging and correlated with quantitative histomorphometry. Four farm pigs were used as disease control. Specificity was determined in two HL pigs injected with isotype control antibody. All animal experiments were performed with the approval of the Institutional Animal Care and Use Committee of Columbia University. Nine juvenile male LDL-deficient (Rapacz) swine were sent to us from the University of Wisconsin-Madison. Seven arrived at age 3 months and were maintained in house for approximately 9 months or to 1 year of age, and two HL pigs arrived at age 11 months and were maintained in house for 1 month. Four age- and weight-matched farm pigs were also studied for blood pool clearance and as disease controls for imaging and ex vivo well counting.
All pigs received a high-fat swine diet (15% lard, 1.2% cholesterol) (Harlan Teklad, Madison, WI, USA). Weights were obtained monthly, and on the same day, the animals were sedated for venous blood samples for fasting glucose, blood chemistry profile, and lipid profile.
The anti-RAGE antibody is a murine monoclonal antibody against the V-domain of RAGE designed to display immunoreactivity in mice, pigs, and human. The peptide sequence and production of the murine hybridoma has been described []. The monoclonal anti-RAGE antibody was fragmented using pepsin digestion into F(ab′)2 fragments (approximately 110 kDa) and immunoreactivity tested by ELISA using soluble RAGE antigen. Direct coupling of diethylenetriaminepentaacetic acid (DTPA) (bicyclic anhydride) to anti-RAGE F(ab′)2 antibody fragments for 99mTc labeling was performed as previously described [,]. The mean specific activity was 8.14 ± 3.8 MBq per microgram of protein, and the mean radiopurity was 98% ± 0.83% by instant thin-layer chromatography.
Blood pool clearance
Ear vein catheters were placed in both ears of four farm pigs. Into one ear vein, an average dose of 16 mCi of 99mTc-anti-RAGE F(ab′)2 was injected, and from the opposite ear vein, 1-ml samples were withdrawn at 2, 5, 10, 15, 20, 30, 45, 60, 90, 120, 180, 240, 300, and 360 min. From each tube, 50 μl was pipetted into pre-weighed tubes and counted in a gamma well counter. The counts vs. time were averaged for each time point and plotted. These pigs were also used for in vivo imaging and ex vivo well counting and histology.
Radiotracer injection and imaging
At approximately 1 year of age, each HL pig was sedated, and angiocatheters were placed in both ear veins. A dose of 16.3 ± 3.5 mCi (0.3 ± 0.1 mCi/kg, 658.6 ± 148.0 MBq, 63.6 ± 13.3 MBq/kg) of 99mTc-anti-RAGE F(ab′)2 antibody in 5-ml saline with a flush was injected through an ear vein. Two of these nine HL pigs also received 99mTc non-immune IgG(Fab′)2 as well as 99mTc-anti-RAGE F(ab′)2, 1-week apart with SPECT imaging following both injections, but the tissue was obtained after the control antibody. Four weight-matched farm pigs received injection of 99mTc-anti-RAGE (17.5 mCi, 0.4 mCi/kg, 653.7 MBq, 12.9 MBq/kg) and were imaged.
After injection, the pigs were awakened and returned to their cages for 5-6 h to allow for blood pool clearance, then were re-sedated, intubated, and transported to the imaging laboratory where they first underwent CT angiography followed by SPECT/CT imaging (Philips Precedence 16 slice Hybrid SPECT/CT, Philips Healthcare, Andover, MA, USA). Each pig was injected with 100-ml non-ionic iodinated contrast agent Optiray 320 (Covidien, Mansfield, MA, USA) through the ear vein following test bolus. Scouts comprised neck to abdomen/pelvis. The following acquisition protocol was used: 120 kV, tube current 337 mA, collimation 16 × 0.75 mm, pitch 0.688, and slice thickness 0.8 mm. Acquisition trigger was set for 180 HU at aortic arch. Following CT angiogram, hybrid SPECT/CT imaging was performed with bed positions to include the chest, neck, and hind limbs. Each SPECT scan was set for two heads mounted at 180° for 64 steps over 360°. At completion of the imaging, the pigs were returned to the necropsy suite of the Institute of Comparative Medicine and euthanized with a bolus of Euthasol (100 to 120 mg/kg IV; Virbac Animal Health, St. Louis MO, USA).
The SPECT scans were reconstructed using filtered back projection, and the CT and SPECT images blended using Syntegra software (Philips, Andover, MA, USA). Focal regions of tracer uptake were localized to the territory of the carotids on sagittal and coronal blended SPECT/CT scans, and regions of interest were drawn around the uptake on the coronal slices to comprise the area of focal uptake excluding the most cephalad and caudad slices to reduce partial volume effect, and the counts from these regions were summed and converted to %ID using measured camera efficiency values and decay times. Because of the very low myocardial uptake, focal hotspots in coronary artery territories were visualized and localized on the blended chest CTA/SPECT images (Syntegra). For the hind limbs, the ROIs were drawn around sequential 1-voxel thick transverse slices using the blended SPECT/CT to define limb boundaries from the proximal femur to the distal tibia/fibula, and the activity from all the slices were summed for each limb.
Necropsy and tissue preparation
The chest and abdomen were opened, the blood was drained from the vasculature, the heart was removed, and the coronaries from the opening of the coronary sinus to distal vessel were dissected out and rinsed with PBS. The aorta was dissected and removed from the aortic valve to beyond the iliac bifurcation to include the proximal femoral arteries and the major arch vessels (the brachiocephalic trunk (innominate artery) and carotid arteries). The tissue was cleaned and rinsed carefully and placed in a plastic tray kept moist in 10% formalin solution and placed on one of the detectors used for the in vivo scanning and imaged for 20 min. After imaging, representative cross-sections of the arteries were cut and labeled for well counting and sectioning. Each coronary was cut in four segments. The aorta was sampled as full-width cross-sectional segments from the aortic root through the abdominal aorta for a total of 12 segments per animal. The samples were taken from the proximal, mid, and distal innominate, from the proximal and mid right and left carotids, and from the proximal and mid femoral arteries. In addition, three samples were taken from each gastrocnemius muscle (six samples per pig). Each sample was weighed and counted in the gamma well counter (Wallac Wizard 1470, PerkinElmer, Waltham, MA, USA) along with an aliquot of the injected dose as standard and the %ID/g calculated for each sample and subsequently embedded in paraffin for sectioning. In selected animals, the lungs, liver, and heart were removed, weighed, and counted in the gamma well counter and then embedded in paraffin for sectioning.
For immunohistochemical analyses, serial sections were deparaffinized in xylene, treated with 0.3% hydrogen peroxide for 20 min, and incubated in protein-free block (Dako Inc., Carpinteria, CA, USA) for 10 min to inhibit the non-specific binding of primary antibody. All sections were stained with hematoxylin and eosin (H&E). Staining for RAGE was performed using monoclonal anti-RAGE antibody (50 μg/ml). Macrophages were identified using marker Mac-3 (1:20; BD Pharmingen, San Diego, CA, USA). Smooth muscle cells were identified using monoclonal mouse anti-human smooth muscle actin (1:50; Dako Inc.). Secondary staining was performed with HRP-conjugated respective secondary antibody, followed by diaminobenzidine (DAB substrate kit for peroxidase; Vector Laboratories, Burlingame, CA, USA), and counterstaining with Gill's hematoxylin solution.
Morphometric and immunohistochemical analyses of the arterial segments were performed using a Nikon microscope (Tokyo, Japan) and Image-Pro Plus software (Media Cybernetics Inc., Silver Spring, MD, USA). The lesion was measured as percent lesion area per total area of the aorta. RAGE staining was quantified as percent RAGE staining in the lesion area per total area of the aorta. Lesion morphology was classified according to the American Heart Association (AHA) criterion from class I to class IV [].
All data are presented as mean ± standard deviation. Correlation was assessed using the Pearson product–moment correlation.