SV2A PET Imaging Is a Noninvasive Marker for the Detection of Spinal Damage in Experimental Models of Spinal Cord Injury

Animals and Experimental Design

All experiments complied with the European Communities Council Directive (2010/63/EU) and were authorized by the local animal ethic committees (University of Namur [UN19-339] and University of Antwerp [ECD2019-49]). Experiment outcomes were determined at 1 wk after SCI (subacute phase) and 6 wk after SCI (chronic phase), as outlined in Supplemental Figure 1 (supplemental materials are available at http://jnm.snmjournals.org). First, we investigated in vitro both mouse and rat models to assess whether SV2A levels were affected differentially between species. For the longitudinal in vivo study, the rat model was preferred given its favorable body and spinal cord size compared with mice. Nonetheless, investigation in mice is relevant given the abundance of genetic mouse models with relevant spinal cord diseases (23,28).

Adult male C57BL/6J mice (n = 7; body weight, 20–25 g) (internally bred at the University of Namur) and female Sprague–Dawley rats (n = 8; body weight, 225–275 g) (Charles River Laboratories) underwent SCI and were euthanized 1 wk after SCI for histologic and autoradiographic assessment of SV2A. An additional 15 female Sprague–Dawley rats (225–275 g) were subjected either to SCI (n = 8) or laminectomy only (sham, n = 7) and were included in the longitudinal PET study and euthanized at 6 wk after SCI. The animals were group-housed under a 12-h light/dark cycle in a temperature- and humidity-controlled environment with food and water ad libitum.

Contusion SCI

The animals underwent a cervical C5 unilateral right-sided contusion as previously described for mice (26) and rats (27). The animals were anesthetized using a mix of ketamine, 100 mg/kg (ketamine 1000; Ceva), and xylazine, 5 mg/kg (XYL-M 2%; VMD). Briefly, dorsal skin and underlying muscles were cut to access the spine between the spinous processes C2 and T1. The paravertebral muscles overlying C3–C5 were carefully removed. The animals were then subjected to laminectomy at the C5 vertebra before receiving a computer-controlled impact of 51.3 ± 1.7 kdyne for mice and 407 ± 6.1 kdyne for rats (dwell time of 0 s) using an IH Spinal Cord Impactor (Precision Systems and Instrumentation, LLC). The full procedure included unilateral laminectomy on the right side, clamping of the spinous processes of C2 and T2 using toothed Adson forceps to stabilize the whole spinal column, raising of the impactor tip (1.0-mm-diameter impactor for mice and 1.5 mm for rats) 2.5 mm above the dura, and contusion of the right hemicord (bathed in 0.9% sterile saline for rats). Muscles were sutured in layers using silk 2.0 (mice) or 4.0 (rats), and the skin was closed using surgical clippers. The animals received buprenorphine subcutaneously (Vetergesic; Ceva) at a dose of 0.05 mg/kg for analgesia after surgery and again 12 and 24 h after surgery. Saline was subcutaneously administered for rehydration immediately after surgery and 24 h later. One SCI rat did not recover from the surgical procedure and was euthanized. Sham animals underwent the full laminectomy procedure without contusion.

Radiosynthesis

¹¹C-UCB-J synthesis was performed on an automated synthesis module (Carbosynthon I; Comecer) as previously described (20,23), with an average molar activity of 52.9 ± 12.4 GBq/μmol. Radiochemical purity was greater than 99% for both ¹¹C-UCB-J and ¹⁸F-FDG. ³H-UCB-J (Novandi Chemistry AB) was synthetized with a molar radioactivity of 2,997 MBq/μmol and radiochemical purity of more than 99%.

Small-Animal PET Imaging

Small-animal PET/CT was performed on 2 Siemens Inveon PET/CT scanners (Siemens Preclinical Solutions). The animals were prepared as previously described (29). For the ¹⁸F-FDG PET scan, the animals were kept fasting starting the evening before the scan, for an average of 17.7 ± 1.6 h. Glucose strips (OneTouch Ultra 2; LifeScan) were used to measure the glucose concentration (81.2 ± 15.1 mg/dL) in duplicate from blood samples obtained via the tail vein of the rats before radioligand injection. The rats were anesthetized using isoflurane (Forane; Ohio Medical Products) in medical oxygen (induction, 5%; maintenance, 2%–2.5%). The respiration rate and body temperature were monitored and kept constant during the scan. ¹¹C-UCB-J PET images were acquired dynamically (50-min scan), whereas static (20-min) ¹⁸F-FDG PET scans were obtained for glucose metabolism after 30 min of radioligand uptake in animals that were awake (29,30). Into a catheterized tail vein over a 30-s interval (1 mL/min) using an automated pump (Pump 11 Elite; Harvard Apparatus), a bolus of ¹¹C-UCB-J (6.7 ± 4.0 MBq) or ¹⁸F-FDG (25.1 ± 2.0 MBq) was injected. Data were acquired in list-mode format. After the small-animal PET scan, a 10-min 80 kV/500 μA CT scan was performed for coregistration and attenuation correction.

Image Processing and Analysis

¹¹C-UCB-J data were reconstructed into 31 frames of increasing length (12 × 10 s, 3 × 20 s, 3 × 30 s, 3 × 60 s, 3 × 150 s, and 7 × 300 s), whereas a static reconstruction was applied to ¹⁸F-FDG data using 8 iterations and 16 subsets of the 3-dimensional ordered-subset expectation maximization algorithm with spatially variant modeling (31) and normalization, dead-time, and CT-based attenuation corrections. Images were reconstructed on a 128 × 128 × 159 grid with 0.776 × 0.776 × 0.796 mm³ voxels. PET/CT images were processed and analyzed using PMOD software (version 3.6; PMOD Technologies).

To align all brains to a common space and use the same volume of interest to extract the whole-brain activity, brain PET images were rigid-matched to the Schiffer rat ¹⁸F-FDG template (32) or the ¹¹C-UCB-J PET template (generated in the same space as the Schiffer one). Cervical spinal cord volumes of interest were manually delineated using the individual coregistered PET/CT images from C3 to C7, with C5 subdivided into right and left segments. Since we anticipated a potential partial-volume effect on the contralateral C5 segment, comparisons were made against the C3 section. The volumes of interest were kept at the same volume across subjects (0.016 cm³ at each C3–C7 level and 0.008 cm³ for left and right C5). Across animals, volumes of interest were manually reoriented to follow the profile of the spine. Supplemental Figure 2 shows the CT image of the spine in a representative sham rat with laminectomy.

For ¹¹C-UCB-J analysis, the interval from 25 to 45 min after injection was selected to determine radioligand uptake measured using SUV, calculated as brain activity multiplied by body weight and divided by the injected dose. No difference in cerebral uptake was observed for either ¹¹C-UCB-J (SUV) (Supplemental Fig. 3) or ¹⁸F-FDG (SUV corrected for glucose levels) (Supplemental Fig. 4); therefore, spinal uptake was measured as the ratio of regional spinal uptake over whole-brain uptake.

Histologic Analysis

At 1 or 6 wk after injury, the animals were deeply anesthetized with sodium pentobarbital (50 mg/kg, intraperitoneally) and transcardially perfused with ice-cold phosphate-buffered saline (pH 7.4). The spinal cords were dissected, and a cervical segment containing the contusion injury was snap-frozen in 2-metylbuthane at −35°C for 2 min and preserved at −80°C until use. Coronal sections (30 μm thick) were collected on Superfrost Plus slides (Thermo Fischer Scientific) using a cryostat (Leica) and were distributed into a series of 12 slices so that each slide covered rostral, trauma, and caudal sections.

SV2A immunostaining was performed as previously described (23), using the primary antibody anti-SV2A (rabbit IgG; 1:400; catalog number 66724 [Cell Signaling Technologies]) followed by the secondary fluorescent antibody (donkey antirabbit; 1:100; Alexa Fluor 488 [Jackson ImmunoResearch]). Images were acquired at ×10 magnification for quantification with a fluorescence microscope (Olympus) using CellSens software. Images were analyzed using Fiji (version 2.1.0) ImageJ software (National Institutes of Health). SV2A signal intensity was measured in regions of interest (right and left gray matter spinal cord) manually delineated on each section. Adjacent slides were stained with hematoxylin and eosin to visualize tissue morphology.

Autoradiography

In vitro ³H-UCB-J (Novandi Chemistry AB) autoradiography was performed as previously reported (23). Mouse and rat SV2A selectivity in the spinal cord was validated using a blocking solution (1 nM of ³H-UCB-J plus 1 mM of levetiracetam in binding buffer) (Supplemental Fig. 5).

Regional quantification was performed using Fiji (version 2.1.0) ImageJ software. Specific binding of ³H-UCB-J was measured by converting the mean gray values into radioactivity density (Bq/mg), calculated using commercial tritium standards (American Radiolabeled Chemicals). Next, from ³H-UCB-J molar activity on the experimental day, radioactivity density was converted into binding density (pmol/mg) for each region.

Statistical Analysis

In vivo spinal cord PET data, as well as postmortem analyses, were investigated using 2-way ANOVA (experimental group and region as variables). A decrease in spinal PET signal compared with C3 and cerebral binding was assessed using the Mann–Whitney test. Autoradiographic blocking experiments were analyzed using paired t tests. Correlations between in vivo and postmortem SV2A signals were evaluated using a Spearman r test. P values were corrected for multiple comparisons using the Šídák multiple-comparisons test. Statistical analyses were performed with GraphPad Prism (version 9.1) statistical software. Data are represented as mean ± SD. All tests were 2-tailed, and statistical significance was set at a P value of less than 0.05.