Animals
All experiments were conducted in accordance with FELASA and GV-SOLAS standards for animal welfare. Experiments were approved by the local ethics committee of Freiburg University and the regional council of Freiburg, Baden-Wuerttemberg, Germany (licence number 35-9185.81/G-21/008).
In total, experiments were performed on 7 juvenile (3 months of age) domestic landrace pigs (body weight 50–70 kg). For premedication, the pigs received an intramuscular injection of midazolam (0.5 mg/kg body weight (bw)) and ketamine (20 mg/kg bw). After preoxygenation, anaesthesia was induced by propofol injection (2–4 mg/kg bw) via a peripheral vein catheter and maintained with a mixture of isoflurane (1.5–2%) and oxygen/air (FiO2 > 0.3) as well as intravenous (iv) administration of vecuronium (0.2–0.4 mg/kg bw per hour). Fluid loss was compensated at a dose of 5–10 ml/kg bw ringer’s solutionper hour. Analgesia was maintained by intravenous application of fentanyl at a dose of 0.002–0.004 mg/kg bw per hour. Mechanical ventilation (IPPV) was adjusted to keep parameters within physiological range. Oxygen level, electrocardiogram and concentration of carbon dioxide were monitored.
Closed chest model of myocardial ischemia/reperfusion injury
For angiography, a 10 French arterial access sheath was introduced into the right femoral artery using ultrasound needle-guidance. To reduce the risk of sudden fatal arrhythmias, potassium and magnesium was supplemented and amiodarone (10 mg/kg bw) was intravenously administered prior to the procedure. Coronary angiography was performed using a C-arm x-ray (Philips Medical) and standard coronary catheters. The left coronary artery was depicted with bolus injection of diluted iodinated contrast agent (Accupaque™ 300mg, GE Healthcare). A coronary wire was thus mounted with a balloon catheter and inserted into the circumflex coronary artery. The balloon was then inflated in the mid-segment of the coronary artery to trigger ischemia, and contrast agent was injected to ensure tight sealing of the vessel. After 40 min, the balloon was deflated and removed from the coronary vessel.
CMR
Cardiac MR imaging and MR-guided coronary catheterization was performed at a clinical 3 T system (PrismaFit, Siemens) with the animals in head-first supine position and the heart at magnet iso-center. A 32-channel spine coil and an anterior 18-channel thorax coil array were used for signal reception. An ECG with four leads was attached and to enable cardiac gating of the imaging sequences. All acquisitions except the cine and real-time sequences were gated to end-diastole. Real-time images for catheter guidance were displayed on an in-room monitor (BOLD Screen 24, Cambridge Research Systems Ltd) positioned close to the patient table and communication between the cardiologist and the system operator was established via the conventional headphones and in-room microphone of the MRI system. A custom-made active coronary guiding catheter equipped with a single-loop receive coil at the tip was used for catheterization of the left coronary artery. The catheter was connected to the MRI system via a custom-made tuning/matching circuit with variable signal attenuation.
Functional Imaging
Cardiac MRI started with the acquisition of a set of localizer images in the three orthogonal standard views and definition of the main axes of the hearts. Then, a multi-slice 2D cine bSSFP sequence was acquired in short-axis view for functional and volumetric imaging (TE/TR = 1.5/3.0 ms, flip angle (FA) = 42°, BW = 970 Hz/px, FoV = 340x273 mm², matrix: 224x126, slice thickness (SL) = 8 mm, number of slices: 6–8, retrospective cardiac gating with 20 reconstructed phases, multiple breath-holds).
T1, T2 and T2* mapping
Parametric mapping of the relaxation times T1, T2, and T2* was performed in 3 or 4 mid-ventricular short axis slices depending on the size of the heart. T1 maps were acquired with an inversion recovery bSSFP sequence (TE/TR = 1.2/2.7 ms, FA = 35°, BW = 1085 Hz/px, FoV = 360x307 mm², matrix: 192x132, SL = 5 mm, 8 different TI values from 100 to 4500 ms depending on the heart rate, single breath-hold). T2 maps were acquired with a T2-prepared FLASH sequence with varying T2 preparation times (TE/TR = 1.3/3.1 ms, TET2prep: [0, 30, 40] ms, FA = 12°, BW = 1185 Hz/px, FoV = 360x247 mm², matrix: 192x132, SL = 6 mm, single breath-hold). The inline motion correction and calculation of T1/T2 values provided by the vendor were used for both T1 and T2 mapping. T2* mapping was performed with a multi-echo FLASH sequence (TE/TR = [2.4, 6.0, 9.5, 13.0]/16.2 ms, FA = 12°, BW = 590 Hz/px, FoV = 260² mm², matrix: 256², SL = 5 mm, single breath-hold) and T2* maps were calculated offline in Matlab by a pixel-wise linear fit of the logarithm of the signal intensities. The T2* mapping sequences was acquired again after injection of MPIOs under real-time guidance.
Real-time imaging
After functional and parametric imaging a non-contrast 3D compressed-sensing accelerated prototype whole heart FLASH sequence (TE/TR = 2.3/5.2 ms, FA = 15°, BW = 250 Hz/px, FoV = 320x310x139 mm³, matrix: 256 x 248 x 104, navigator gating to end-expiration) was acquired for coronary angiography and planning of the imaging planes for the MR-guided catheterization. Therefore, three planar views were extracted from the 3D dataset to cover the aortic arch and the left coronary ostium in a short-axis and long-axis view. Catheterization of the LCA was performed under imaging with a real-time FLASH sequence (TE/TR = 1.3/3.4 ms, FA = 10°, BW = 790 Hz/px, FoV = 289² mm², matrix: 192x144, SL = 8 mm). Successful intubation was verified by imaging the perfusion of a small amount of 1% Gd-solution injected via the catheter in a single short-axis slice with an inversion recovery FLASH sequence (TE/TR = 1.0/2.0 ms, TI = 95 ms, FA = 10°, BW = 1185 Hz/px, FoV = 360x270 mm², matrix: 192x106, SL = 8 mm, single breath-hold). The MPIO contrast agent was injected after successful intubation and the injection was imaged with a FLASH sequence in short-axis view (TE/TR = 4.0/5.0 ms, FA = 12°, BW = 1185 Hz/px, FoV = 280² mm², matrix: 128x102, SL = 6 mm, single breath-hold).
LGE imaging
Late gadolinium enhancement (LGE) image data were acquired after the coronary catheterization and 10 min after intravenous injection of 2.5 mmol/kg Gd. A TI scout was first acquired in short axis view to determine the optimal inversion time. The LGE sequence was then acquired with phase sensitive inversion recovery FLASH sequence in short-axis view (TE/TR = 1.4/3.7 ms, FA = 20°, BW = 465 Hz/px, FoV = 360² mm², matrix: 144x141, SL = 10 mm, single breath-hold).
Ex vivo T2* mapping
Ex vivo imaging was performed after the hearts were removed and stored in the fixative for at least 7 days to allow for the fixative to fully diffuse through the heart. The hearts were imaged at the same 3T MRI system as the in vivo experiments. For imaging, the hearts were placed in a plastic container containing the fixative positioned at iso-center inside a 64 channel head/neck coil. Three-dimensional R2* maps were acquired via a multi-echo spoiled gradient echo sequence with 0.58 mm isotropic resolution covering the whole heart (TE/TR = [3.4, 9.8, 17.1]/23 ms, FA = 12°, BW = 260 Hz/px, FoV = 129×129×84 mm³, matrix: 224×224×144, averages = 2). R2* maps were calculated offline in the same way as the in vivo R2* maps.
Real-time MRI catheter guidance
Using an in-room video monitor (BOLD Screen 24, Cambridge Research Systems Ltd, Rochester, UK) next to the patient table of the MRI system, the real-time images were presented to the interventionalist during MRI. Conventional headphones and the in-room microphone of the MRI system were used for communication between the interventionalist and the system operator. The animal was positioned on the MR table in supine position with the heart located in the magnet’s isocenter. For a full coverage of the vascular system, a posterior 32-channel spine coil and an anterior 18-channel thorax coil array were used. The wireless ECG system supplied by the vendor was attached with hydrogel electrodes.
After the acquisition of an initial set of localizer images, a 3D whole-heart ECG-triggered gradient echo (FLASH) data set was acquired with the following imaging parameters: fat saturation, TE/TR = 1.6/3.5 ms, FA = 16°, FoV = 282x282x102 mm³, matrix: 176x176x64, T2-preparation with TET2prep = 40 ms, GRAPPA acceleration factor R = 2. The ECG-triggered FLASH sequence was repeated later during the experiment to confirm the position of the interventional instruments.
First, a guidewire (standard Terumo 0.018") was inserted via the arterial sheath to guide a modified 6F guiding catheter (Terumo; Optitorque Radial Tig II 4.0) to the aortic root. After intubation of the LCA, proper position of the guiding catheter was confirmed by injection of 5 ml diluted gadolinium contrast agent (1:20, Gd-DTPA, Magnevist, Bayer, Germany) via the guiding catheter. A 2D real-time radial bSSFP sequence with the following imaging parameters was used to monitor the advancement of the instruments: TE/TR = 1.4/2.8 ms, #spokes = 105, FA = 40°, FoV = 275x275x7 mm³, matrix: 160x160, fat saturation. Real-time image slice orientations and positions were defined using the localizer and 3D FLASH images acquired prior to the catheter advancement. Molecular contrast agents were injected via the guiding catheter in a volume of 20 ml and flushed with saline.
Molecular contrast agent
Construction of the molecular contrast agent was performed using pelleted MyOne™ Tosylactivated MPIOs with a size of 1 µm. MPIO pellets were washed in 0.1M natrium borate buffer and resuspended in ammonium sulfate buffer containing 200 microgram of antibody to reach a final concentration of 1M 20 hours at 37°C. Constant rotation insured separation of pelleted beats. Afterwards, residual active tosyl-remnants were removed using a blocking buffer and contrast agent was resuspended in a storage buffer with constant rotation.
Post-processing
Dedicated MRI post-processing software (CVI 42, Circle Imaging) was used to quantify T2* times from 3 short axis slices of the basal, mid and apical left ventricle. Regions of interest (ROI) were defined in the area of LAD, LCX or RCA on each slice. For analysis, the mean T2* time in each area, i.e. LAD, LCX or RCA, was used.
FACS
For flow cytometry full blood was stained with antibodies targeting cell specific epitopes after red blood cell lysis. Neutrophils were identified according to SSC and FSC, classical-type monocytes were identified as SWC3+, CD163neg, CD14+, non-classical-type monocytes were identified as SWC3+, CD163+, CD14int. Analysis was performed using FACS Diva software.
In vitro flow chamber
Cell culture dishes (35mm; CytoOne) were coated with 1ml fibrinogen (100µg/ml) and stored overnight at 4°C. The following day human blood was collected in a citrate tube and separated in various blood components via centrifugation at 150G for 5 minutes. The supernatant (platelet rich plasma; PRP) was then transfered into a Falcon tube and 1ml of PRP was applied to each of the previously prepared fibrinogen coated dishes to allow platelet adhesion to the surface of the dish. Platelet activation was induced using a 1:10 dilution of 20 µg of adenosine diphosphate (ADP). The activation process was carried out for a precisely controlled duration of 30 min at room temperature.
The parallel plate flow chamber kit (GlycoTech, Rockville, Maryland, USA) consists of a transparent chamber with two parallel plates, one plate is coated with a monolayer of platelets, and the other plate is a continous fluid flow channel. A syringe pump connected to the inlet port is used to deliver a precise and constant flow of contrast agent through the chamber and thus over the platelets. The outlet port is used to allow the left over contrast agent to exit the chamber into a waste container. The chamber was observed using a microscope (Zeiss Vert. A1) at 20x magnification connected to a digital imaging system (AxioCam ICc1, Carl Zeiss AG, Feldbach, CH).
IgG-MPIO served as a non-specific control group, whereasanti-CD62P-MPIO was used as contrast agentwith specific binding properties for P-selectin. The binding properties of both contrast agents were evaluated on the platelet monolayer over a specific duration of 60 seconds. The video recording commenced upon the appearance of the initial MPIO within the designated field of view, measuring 450 µm x 350 µm. Only MPIOs that demonstrated adhesion for a minimum of 10 seconds were deemed to have established a bond with the platelets.
Incubation Assay
Porcine endothelial cells were cultured according to the manufacturer's protocol (Sigma Aldrich Chemie GmbH, P300-05). In short after thawing the cryovials in a 37°C water bath, the cells were resuspended a 10% dulbecco’s modified eagle medium (DMEM). The cells were centrifuged, washed three times with DMEM, and finally resuspended in Porcine Endothelial Growth Medium. The cell suspension was transferred to a T-75 cell culture flask, and daily medium changes were performed until the cells reached 60% confluency. The volume of the culture medium was then doubled, and after further incubation, the cells were split onto 12-well-dishes (Thermo Fisher Scientific, 168844). For proper comparison of the specifically targeting P-Selectin-MPIO contrast agent we compared the binding properties to MPIO with unspecific targeting. For each contrast agent six dishes were seeded with a total of 4000 cells each. Each contrast agent was prepared identically as described above. 10µl of each contrast agent was diluted with 990 ml of PBS and incubated for 30 seconds. After,, each dish was washed thoroughly with PBS. For subsequent analysis photographic documentation was performed to capture the different binding properties of each contrast agent. Per dish 10 photos of evenly distributed cells were taken randomly yielding a total of 60 photos per group. Each MPIO located exactly next to or on top of a cell was considered as bound to the endothelial cell layer.
Immunofluorescence staining
For each heart sample, three specimens of each supply areas of the RCA, LCX, and LAD (nine in total) were collected from porcine myocardium and cryo-embedded for further histologic processing. Using a microtome, cryosections of 6 µm were extracted and stained with hematoxylin (25%). 10 µm sections were cut for immunofluorescence imaging (for quantification, at least 9 sections per heart were used). Slices were permeabilized with 0.1%Triton X-100 (Invitrogen, Waltham, MA, USA) at RT for 15 minutes. Antigen retrieval was performed by boiling in 1× citrate buffer-based antigen retrieval solution (H-3300, Vector laboratories, USA). Unspecific antibody binding was blocked with 6% donkey serum (Sigma-Aldrich, Waltham, MA, USA), and 2.5% bovine serum albumin (BSA, Sigma-Aldrich, Waltham, MA, USA) at RT for one-hour. Samples were incubated with a primary antibody anti-P-selectin (NB100-65392, Novus Biologicals, C), USA) at 4 oC overnight. After washing, samples were incubated with Alexa Fluor 555 conjugated donkey anti-mouse secondary antibody at 1:400 for one-hour at RT. DAPI was applied at 1:500 concentration in PBS for 10 min. Samples were mounted with Fluoromount-G (ThermoFisher, Waltham, MA, USA). Samples were imaged with an automated slide scanner (AxioScan.Z1 Zeiss, Jena, Germany).
For image analysis, thresholds were manually selected in Zen Blue microscopy software (Zeiss, Jena, Germany). To acquire a baseline for fluorescence, a value eliminating approximately 95% of nonzero p-selectin pixels remote sample images was chosen and applied. Autofluorescence from the green channel was used to determine the total area of the tissue. A custom Python script using the packages czifile (Christoph Gohlke, University of California, Irvine) and NumPy was used to quantify fluorescent area. Positive pixels for autofluorescence were compared to positive pixels for p-selectin (red channel) in a given sample. Coverage was determined by dividing the sum of pixels both positive for p-Selectin and autofluorescence by the sum of all total membrane pixels positive for autofluorescence.
Hematoxylin staining (HE)
For analysis of MPIO, HE staining was performed using a 25% hematoxylin solution for 50 seconds. Further washing steps with saline were held to a minimum to avoid loss of MPIO binding. For analysis a bright-field microscopy at 100x magnification was used. 20 fields of view were recorded in a standardized fashion and analyzed to count individual MPIO. In total, 3 slides per region of interest (RCA, LCX, and LAD) were analyzed covering the epicardium, mid-section and subendocardium.
Statistics
GraphPad Prism software (GraphPad Software, Inc.) was used for statistical analyses. Results are depicted as mean ± standard error of mean. For two-group comparison a Mann-Whitney U test for nonparametric data or students-t test for parametric data was used. For a comparison of more than two groups an ANOVA, followed by a Bonferroni test for multiple comparison, was applied. P values of p < 0.05 indicate statistical significance.