Imaging the Cytokine Receptor CXCR4 in Atherosclerotic Plaques with the Radiotracer ⁶⁸Ga-Pentixafor for PET

Radiosynthesis of ⁶⁸Ga-Pentixafor

⁶⁸Ga was obtained by elution of a ⁶⁸Ge/⁶⁸Ga generator with SnO2 matrix (iTHEMBA LABS) with 1 M HCl (5.5 mL) and immobilized on a strong cationic exchanger cartridge (SCX-Chromafix, size M; Macherey-Nagel).

For animal studies, ⁶⁸Ga-pentixafor was prepared on a Gallelut+ system in analogy to a previously published ⁶⁸Ga-labeling procedure (SCINTOMICS GmbH) (13). Briefly, ⁶⁸Ga generator eluate fractions (1.25 mL, 600–800 MBq, buffered to pH 3.3 with 900 μL of a solution of 14.4 g N-(2-hydroxyethyl)piperazine-N′-(2-ethanesulfonic acid) in 12 mL of water) were reacted with 3.5 nmol of pentixafor (SCINTOMICS GmbH) for 5 min. The radiochemical purity was always greater than 99% as confirmed by thin-layer chromatography and high-performance liquid chromatography (HPLC). Addition of 1 mL of phosphate-buffered saline and concentration in vacuo to 1 mL of total volume yielded solvent-free formulations with specific activities ranging from 100 to 150 MBq/nmol.

For patient studies, the synthesis of ⁶⁸Ga-pentixafor was performed in a fully automated, good manufacturing practices–compliant procedure using a GRP module (SCINTOMICS GmbH) equipped with a disposable single-use cassette kit (ABX) including a standardized labeling sequence (14). Upon being washed with water, ⁶⁸Ga3+ was eluted into the reaction vessel (containing 20 μg of pentixafor in 3 mL of 1.5 M N-(2-hydroxyethyl)piperazine-N′-(2-ethanesulfonic acid)) using 1.7 mL of 5 M NaCl. The reaction mixture was heated to 140°C for 10 min and, upon cooling, transferred onto a C18 light cartridge (Waters). The cartridge was washed with water to remove unreacted ⁶⁸Ga activity, and ⁶⁸Ga-pentixafor was eluted using 2 mL of EtOH/water (50/50). The eluate was diluted with 12 mL of phosphate-buffered saline and passed through a sterile filter. The radiochemical purity of the ready-to-inject formulation (14 mL) was always greater than 99% as confirmed by radio-HPLC and thin-layer chromatography, and the specific activity was in the range of 30–65 GBq/μmol.

Radiosynthesis of ¹²⁵I-Pentixafor

For radioiodination, 100–200 μg of pentixafor (cold gallium(III) complex) were dissolved in 0.5 mL of TRIS iodination buffer (25 mM Tris·HCl, 0.4 M NaCl, pH 7.5) and transferred to an Eppendorf reaction tube coated with 150 μg of IodoGen. Upon addition of ¹²⁵I-Na (18–20 MBq; Hartmann Analytik), the reaction vessel was briefly vortexed and the labeling reaction was allowed to proceed for 15 min at room temperature. The peptide solution was then removed from the insoluble oxidizing agent. Separation of ¹²⁵I-pentixafor from unlabeled precursor was achieved using gradient reversed-phase HPLC (column: Nucleosil 100 C18 [5 μm, 125 × 4.0 mm; CS GmbH]; gradient: 22%–42% ethanol [0.5% acetic acid] in water [0.5% acetic acid] within 20 min; flow: 1mL/min). The high-performance liquid chromatography product fraction was used as such and diluted to the required concentration using phosphate-buffered saline.

Animal Model

Atherosclerotic plaques were induced in male New Zealand White rabbits (n = 7; mean age, 4 mo; mean weight = 3.5 ± 0.2 kg) by a combination of 4 mo of a high-cholesterol diet (0.3% cholesterol-enriched diet; Altromin Spezialfutter GmbH) and a double balloon injury of the aorta and the right carotid artery (2 and 6 wk after starting the high-cholesterol diet). Endothelial abrasion was performed in the aorta with a 4-French Fogarty embolectomy catheter (Edward Lifescience GmbH) introduced through the femoral artery as previously described (2,15) and in the right carotid artery with a 3-French Fogarty embolectomy catheter (Edward Lifescience GmbH). All procedures were performed under general anesthesia by continuous intravenous injection of propofol and analgesia with fentanyl. Nonoperated New Zealand White rabbits, fed a normal chow diet, were used as controls (n = 5). Experiments were approved by the local animal care committee and were in accordance with the German Animal Welfare Act.

Study Design

In vivo imaging was performed with ⁶⁸Ga-pentixafor PET/MR imaging in a total of 12 rabbits, including 7 rabbits that developed injury-induced atherosclerotic plaques in the abdominal aorta and right carotid artery and 5 noninjured control rabbits (Fig. 1). In 2 atherosclerotic rabbits, ⁶⁸Ga-pentixafor PET/MR imaging was repeated after the blocking of CXCR4 obtained by the preinjection of an established CXCR4 inhibitor (AMD3100; plerixafor; Sigma-Aldrich) at a dose equivalent to 20 mmol (≈2000-fold concentration of pentixafor). Immediately after the last imaging session, animals were injected with ¹²⁵I-pentixafor, and biodistribution studies were performed 1 h after injection in 8 rabbits (5 atherosclerotic rabbits, including 2 with blocking of CXCR4 and 3 control rabbits).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Flow chart of experimental study.

PET/MR Imaging Protocols

Imaging was performed on a human whole-body simultaneous PET/MR (3T) scanner (Biograph mMR; Siemens). Rabbits were anesthetized with an intravenous infusion of Propofol (2%, ∼1.2–1.3 mg/kg/min; Propofol 2% MCT Fresenius, Fresenius Kabi Deutschland GmbH) and intubated, and catheters were placed in ear veins. Rabbits positioned prone in the scanner were ventilated (Servo-I; Maquet Critical Care AB) and monitored (pulse oxymeter and electrocardiogram) during the whole imaging session. PET and MR images were acquired simultaneously in all rabbits and started 45 min after injection of 15 MBq/kg of ⁶⁸Ga-pentixafor. Detailed PET and MR acquisition protocols are provided in the section “PET and MR Acquisition Protocols.”

Image Analysis

Images were analyzed by an experienced nuclear medicine physician masked to the rabbit group and to the results of histology. Image registration and fusion of the time-of-flight MR acquisitions and attenuation-corrected PET images were performed with an Osirix workstation (OsiriX version 3.5.1 64-bit; OsiriX Imaging Software), an open-source DICOM viewer and image manipulator using the Horn algorithm, a manual rigid registration algorithm based on multiple points (16,17). Matching of both datasets was considered as correct when rabbit contours of PET and MR images were perfectly aligned. If required (e.g., in the case of animal motion), alignment was corrected manually using the Fusion tool of the OsiriX software. For biodistribution studies, SUV_mean calculated as decay-corrected tissue radioactivity divided by body weight and injected dose was measured in circular regions of interest placed on each organ located using the MR images.

For the quantification of the vascular uptake, circular regions of interest were placed on axial adjacent PET images of the abdominal aorta and both carotid arteries (12 axial sections each) identified using the time-of-flight images. SUV_max was recorded in each region of interest. In addition, the SUV_mean of blood was measured by placing a circular region of interest in the right atria. The tissue-to-background ratio (TBR) was calculated as the ratio of SUV_max and the background venous activity measured as SUV_mean of blood (18). Mean TBRs were calculated for each arterial territory as the average of TBR values throughout the entire arterial segment.

MR Acquisition Protocol

MR imaging was performed using a clinical knee coil for the aorta and a bilateral 4-channel phased array coil for carotid arteries (4-channel, Machnet BV). MR acquisition first started with a localizer scan to define the bed position and was followed by a time-of-flight 3-dimensional angiographic sequence to visualize the abdominal aorta. Subsequently, a coronal 2-point Dixon 3-dimensional volumetric interpolated breath-hold T1-weighted (VIBE) MRI sequence was acquired for the generation of attenuation maps (μ-maps) and anatomic images in association with the PET acquisition. The Dixon MRI sequence was set as follows: integrated parallel acquisition technique; acceleration factor, 2; voxel size, 3.7 × 2.3 × 2.7 mm³ (in-plane resolution × slice thickness); acquisition time, 19 s; repetition time, 3.6 ms; first echo time, 1.23 ms; second echo time, 2.46 ms; matrix, 192 × 192; number of excitations, 1; field of view, 450 mm; phase field of view, 65.6%; 1 slab with 128 slices; slice thickness, 2.73 mm; flip angle, 10°; and bandwidth, 965 Hz/pixel. The software of the MRI scanner automatically used the raw images to generate 4 different images : T1-weighted in-phase, T1-weighted out-of-phase, water-only, and fat-only. The same MR acquisition protocol was repeated for the imaging of carotid arteries.

PET Acquisition Protocol

Static emission scans were acquired during 10 min in a single bed position 50 min after the injection of ⁶⁸Ga-pentixafor in the abdominal aorta and 60 min after the injection for carotid arteries. Data were corrected for randoms, dead time, scatter, and attenuation based on the μ-maps generated from the Dixon images (19). Two-mm-thick axial PET images were reconstructed with a 50 × 50 cm² field of view and a 344 × 344 matrix using a 3-dimensional attenuation-weighted ordered-subsets expectation maximization iterative reconstruction algorithm with 3 iterations and 21 subsets, a zoom of 1, and no postreconstruction filter. Spatial resolution of the reconstructed PET images was previously determined to be 4.3 mm at 1 cm from the scanner's isocenter and 5.0 mm at 10 cm in the transverse direction in full width at half maximum (20).

Tracer Biodistribution in Tissues

At the end of the imaging session, 8 rabbits (atherosclerotic rabbits, n = 5; control rabbits, n = 3) were injected intravenously with 7.5 MBq/kg of ¹²⁵I-pentixafor into the ear vein under isoflurane anaesthesia for biodistribution studies. Because of the long half-life of ¹²⁵I (60 d), ¹²⁵I-pentixafor allowed for more accurate biodistribution studies and autoradiographies in comparison to ⁶⁸Ga-pentixafor (half-life, 68 min). The animals were sacrificed 60 min after injection by intravenous injection of 150 mg/kg of pentobarbiturate (Narcoren; Merial), and the organs of interest including the aorta and carotid arteries were dissected. The radioactivity was measured in weighted tissue samples using a γ-counter (1480 Wizard; PerkinElmer Wallac). Data are expressed in %IA/g.

Autoradiographies were performed on 20-μ-thick cryosections. After an overnight exposure to image plates (Fuji TR; Fuji Photo Film Co.), the plates were scanned with an image plate reader (CR 35 BIO; Dürr Medical, Raytest Isotopenmeßgeräte GmbH). The mean activity was quantified in regions of interest encompassing tissues and corrected for background signal with an image analysis program (AIDA Image Analyser; Raytest Isotopenmeßgeräte). Radiotracer uptake in sections was expressed as the mean activity corrected for background and adjusted to the mean activity measured in the vascular wall of the left carotid artery used as a reference for each plate.

Histology of Rabbit Tissues

Five-μm-thick cryosections were fixed for 10 min in acetone. CXCR4 expression and macrophages were detected by immunohistochemistry using mouse monoclonal biotinylated antibodies directed against CXCR4 (clone MAB170 [R&D Systems]; dilution, 1:500) and RAM-11 (M0633 [Dako]; dilution, 1:3,000), respectively. Primary antibodies were visualized using a peroxidase–diaminobenzidine mouse detection kit (Vectastain; Vector Laboratories). Unspecific primary immunoglobulin G antibodies were used as negative controls. Total CXCR4-positive area was measured digitally on each section using an automated, contrast-based, area analysis function of the Image Scope software (Aperio; Vista). The percentage of CXCR4-positive cells was calculated as the ratio between pixels with high contrast and pixels with low contrast using the automated analysis tool of the software Image Scope.

Histology of Human Carotid Plaques

Atherosclerotic plaque samples (n = 10) obtained from carotid endarterectomy were fixed in formalin, embedded in parrafin, and sectioned in 5-μm-thick sections. CXCR4 expression and macrophages were detected by immunostaining with the use of the following primary antibodies: anti-CXCR4 (ab124824 [Abcam Inc.]; dilution, 1:300); anti-CD68 (clone KP1 [Dako]; dilution, 1:2,000); anti-CD31 (clone JC70A [Dako]; dilution, 1:40); anti-CD45 (clone 2B11 [Dako]; dilution, 1:100). After primary antibody incubation, visualization was performed using a peroxidase–diaminobenzidine ChemMate Detection Kit (Dako). Biotinylated antibody against CXCR4 was detected by a streptavidin–horseradish peroxidase complex (Jackson Immunoresearch Laboratories), visualized with diaminobenzidine (Sigma-Aldrich), and counterstained with Mayer hematoxylin. Unspecific primary immunoglobulin G antibodies were used as negative controls.

Expression of CXCR4 on Macrophage Cell Surface

Human peripheral blood mononuclear cells were isolated using density-gradient centrifugation with polymorphprep (Axis Shield). Isolated cells were plated for attachment in monocyte attachment medium, according to the manufacturer's protocol (C28051; PromoCell). Subsequent differentiation was performed using Macrophage Generation Media DXF according to the manufacturer's protocol (C28055; PromoCell). After 10 d of differentiation, cells were detached and replated as 200,000 cells/well (48-well plates). Cells were stimulated with statins (Fluvadin, 5 μM, 48 h; Selleckchem), TNF (25 ng/mL, 24 h; Sigma), oxLDL (25 μg/mL, 48 h; KB Kalen), or a combination. Afterward, CXCR4 expression was measured using FACS analysis (CXCR4-eF450 with CD45-FITC and CD14-APC as controls, all eBioscience).