¹⁸F-ASEM PET/MRI targeting alpha7-nicotinic acetylcholine receptor can reveal skeletal muscle denervation

A total of 30 C57BL/6 mice (10-weeks-old; male) were used for this study (Additional file 1: Fig. S1). Complete sciatic nerve transections were performed to establish skeletal muscle denervation. The mice were anesthetized with 2% isoflurane inhalation and the gluteal muscles were dissected to expose the left sciatic nerve. A 5 mm long nerve segment was resected in the proximal part of the nerve before it bifurcated into the common peroneal and tibial nerves. The experimental protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of the Asan Medical Center (IACUC approval no. 2020–12-081).

The ASEM precursor (3-(1,4-diazabicyclo [3.2.2] nonan-4-yl)-6-nitrodibenzo [b, d] thiophene 5,5-dioxide) and the radiotracer ¹⁸F-ASEM were prepared as described in a previously reported study with minor modifications [20]. Briefly, ¹⁸F-fluoride (F⁻) (85.7 ± 14.3 GBq) produced by a GE PETtrace cyclotron was trapped on a Sep-Pak Accell Plus QMA Plus Light Cartridge (Waters Corporation, Milford, MA, USA) and eluted using a solution of Kryptofix 222 (22 mg), 0.2 M potassium methanesulfonate (0.1 mL), and methanol (1 mL) [21]. The eluted ¹⁸F-F⁻ solution was transferred to a reaction vial of Trasis All-In-One synthesizer (Ans, Belgium) and heated at 100 °C under an N₂(g) stream. For azeotrope evaporation, 1 mL of acetonitrile was added to the reactor during the heating process. After drying, a solution of ASEM precursor (1 mg) dissolved in anhydrous dimethyl sulfoxide (1 mL) was added to the reactor and heated at 150 °C for 10 min. The reaction mixture was purified and reformulated with high-performance liquid chromatography and a Sep-Pak C18 Plus Short Cartridge (Waters Corporation, Milford, MA, USA). The average radiochemical purity and the specific radioactivity of the ¹⁸F-ASEM were determined to be > 99% and 260.3 ± 67.6 GBq/μmol (7.0 ± 1.8 Ci/μmol) calculated at the end of the synthesis, respectively (n = 3). The average radiochemical yield was 6.2 ± 0.6% (non-decay corrected), and the total preparation time was 60–70 min.

The mice underwent small animal PET/MRI using a nanoScanPET/MRI system (1 T, Mediso, Hungary) 1 week (n = 13; 2 for the dynamic study and 11 for the static study) and 3 weeks (n = 6) after the denervation. Under anaesthesia (1.5% isoflurane in 100% O₂ gas), whole-body MRI was used to obtain T1 weighted with gradient-echo three-dimensional (3D) sequence (TR = 25 ms, TE_eff = 3.4, FOV = 64 mm, matrix = 220 × 220) images approximately 12 min before the injection of ¹⁸F-ASEM. After completion of the MRI, 7.3 ± 1.4 MBq of ¹⁸F-ASEM (dissolved in 0.2 mL of normal saline) was administered intravenously via the tail vein and 60 min of dynamic PET and 20 min of static PET images were acquired in a 1–5 coincidence range in a single field of view with the MRI. Body temperature was maintained by heating the air on the animal bed (Multicell, Mediso, Hungary), and a pressure-sensitive pad was used for respiratory triggering. PET images were reconstructed by Tera-Tomo 3D, in full detector mode, with all of the corrections on, high regularization, and eight iterations. A 3D volume of interest (VOI) analysis of the reconstructed images was performed using the InterView Fusion software package (Mediso, Hungary) and applying standardized uptake value (SUV) analysis. VOI, fixed with a diameter of a 2 mm sphere, was first drawn for the left tibialis anterior muscle to identify the highest SUV lesion in the muscle. After that, another VOI was drawn for the corresponding contralateral right tibialis anterior muscle. The SUV of each VOI site was calculated using the formula: SUV = (tumor radioactivity in the tumor VOI with the unit of Bq/mL × body weight [g]) divided by the injected radioactivity.

After ¹⁸F-ASEM PET/MRI for the 1-week model, the animals were processed for autoradiography. Target tissues were rapidly isolated, submerged in Tissue-Tek O.C.T. compound with Tissue-Tek cryomold (Sakura Finetek USA, Inc., CA, USA), and frozen slowly in isopentane at − 20 to − 30 °C. Serial 10 μm tissue sections were cut with a cryostat (CM3050S, Leica Biosystems, Wetzlar, Germany) and placed in a cassette under phosphor imaging plates (BAS-SR, Cytiva, MA, USA) for an exposure period of 24 h, which were read on an Amersham Typhoon plate reader (Cytiva, USA) at a 25 μm pixel resolution. Quantitative analysis was performed using ImageQuant TL Toolbox (v8.2.0, Cytiva, USA) with the circle region of interest (ROI) at the centre region, covering all conceivable tissue areas. To determine the relative uptake levels, quantitative data from the denervated group were normalized to the corresponding values of the control group.

A total of six denervation 1-week mice were anesthetized with 2% isoflurane for nEMG. A two-channel portable electrodiagnostic system (UltraPro S100, Natus Neurology Inc., RI, USA) was used for the nEMG. A monopolar electrode (Disposable Monopolar Needles, Natus Neurology, USA) was inserted into the lower leg muscles to confirm membrane instability after anesthetization. At four points in the lower leg muscles, fibrillation and positive sharp waves were checked. After nEMG, the mice were sacrificed by cervical dislocation and the tibialis anterior muscles of both sides were harvested and placed in 4% paraformaldehyde overnight for fixation. The fixed tissues were rinsed with tap water for approximately 2 h to remove the formalin. The tissues were then dehydrated in graded ethanol, cleared in xylene using a tissue processor (Excelsior ES, Thermo C Shandon, UK), and embedded in paraffin blocks using a paraffin embedding station (EG1150H, Leica Biosystems, USA). Paraffin-embedded tissues were sectioned into 3 µm thick segments on a rotary microtome (RM2255, Leica Biosystems, USA) and mounted onto CREST glass slides (Matsunami, Japan). The immunohistochemistry staining was performed using a Ventana BenchMark XT immunostainer and UltraView Universal DAB Detection Kit (Ventana Medical Systems, AZ, USA). In brief, the tissue sections were warmed to 60 °C and incubated for 4 min. Slides were dewaxed in xylene and rehydrated through graded alcohols. Sections were rinsed twice with Tris buffer (pH 7.6) and incubated with Tris/borate/ EDTA buffer (pH 8.4) for antigen retrieval for 60 min at 100 °C. After rinsing the slides for 10 min in Tris buffer three times, endogenous peroxidase activity was blocked by incubating with 0.3% H₂O₂ in methanol for 4 min. The sections were blocked off and covered with the purified anti-alpha7-nAChR rabbit polyclonal primary antibody (#ANC-007, Alomone Labs, Israel), which was diluted at 1:100 in antibody diluent solution (GBI Labs, WA, USA) for 36 min at room temperature. Sections were washed with Tris buffer and incubated with the UltraMap anti-rabbit HRP conjugated secondary antibody (#760-4315, Ventana, USA) for 16 min at 37 °C. The incubated section was washed with Tris buffer, and Ultraview DAB and DAB H₂O₂ solution was used for 8 min at room temperature for the development of the brown colour. The slides were washed with Tris buffer and Ultraview Copper solution was applied for 4 min at room temperature. The specimens were counterstained with hematoxylin. A bluing reagent was used for post-counterstaining for 4 min at room temperature. Immunohistochemistry sections were digitally captured for quantitative analysis. Ten random locations were selected within the entire muscle area at a 200 × magnification. The quantification of alpha7-nAChR expression was performed using Image J software (NIH, MD, USA). The digital images were converted to grayscale and the percentage of the area stained with DAB was calculated. The threshold for staining positivity was uniformly set at 174 across all images.

In the blocking group (n = 5), methyllycaconitine (MLA; Abcam, ab120072), an antagonist of alpha7-nAChR, was used [22]. MLA was diluted in 0.9% saline and intraperitoneally administered to the mice at a concentration of 10 mg/kg/day [23]. The volume of a single daily intraperitoneal injection was 100 µL. The surgery was performed on the 3rd day and ¹⁸F-ASEM PET/MRI was performed on the 10th day after the first administration of the blocking agent. The last administration was performed 10 min before imaging. The mice were sacrificed after imaging by cervical dislocation and both tibialis anterior muscles were harvested for autoradiography.

Statistical analysis was performed using the Wilcoxon signed-rank test, with MedCalc Version 19.1.7 for Windows (Mariakerke, Belgium). A P-value less than 0.05 was considered statistically significant. We compared the relative uptake of autoradiography, and the relative uptake with 95% confidence intervals not including the value 1.0 were statistically significant (which means a P-value less than 0.05) [24]. Additionally, we drew Box-and-Whisker plots for visual comparisons of the study results. Boxes were drawn from the 1st and 3rd quartile values and a horizontal line was drawn at the median value. The highest/lowest value just below the upper/lower inner line was the upper/lower adjacent value and a horizontal line was drawn at this value. Vertical lines were drawn from the 3rd quartile to the upper adjacent value and the 1st quartile to the lower adjacent value. The small points represent the actual values.

A dynamic study of two mice showed increased ¹⁸F-ASEM uptake in the denervated left tibialis anterior muscle relative to the control muscle (Fig. 1A). As ¹⁸F-ASEM uptake in the denervated left tibialis anterior muscle slowly decreased after radiotracer injection, we set the static PET imaging time at 15–35 min after radiotracer injection (Fig. 1B). Static ¹⁸F-ASEM PET/MRI also showed increased uptake in the denervated left tibialis anterior muscle relative to the control muscle of the denervation model at one week (Fig. 1C). The lesion maximum SUV (SUVmax) and mean SUV (SUVmean) of ¹⁸F-ASEM PET/MRI were significantly higher than the control SUVmax and SUVmean (1.08 ± 0.31 vs. 0.66 ± 0.25, P = 0.0033; 0.89 ± 0.23 vs. 0.54 ± 0.20, P = 0.0033; n = 11, respectively) (Fig. 1D).

Autoradiography demonstrated a significantly increased uptake in the denervated tibialis anterior relative to the control (95% confidence interval 1.07–1.62, n = 11) (Fig. 2A, B). In addition, six 1 week denervation model mice showed abundant membrane instability in the left lower leg muscles by nEMG, which indicated skeletal muscle denervation after left sciatic nerve transection. Immunohistochemistry revealed a significantly larger alpha7-nAChR positive area in the denervated tibialis anterior muscle than in the control (25.3 ± 2.17% vs. 8.13 ± 3.19%, n = 6, P = 0.0277) (Figs. 2C, D).

We next investigated whether an inhibitor could antagonize the increased ¹⁸F-ASEM uptake after skeletal muscle denervation. After blocking with MLA, ¹⁸F-ASEM uptake in the denervated left tibialis anterior muscle was not evident compared to the control side (Fig. 3A). The lesion SUVmax and SUVmean of ¹⁸F-ASEM PET/MRI were not significantly different from the control SUVmax and SUVmean (0.84 ± 0.10 vs. 0.72 ± 0.02, P = 0.0796; 0.68 ± 0.10 vs. 0.62 ± 0.03, P = 0.1380; n = 5, respectively) (Fig. 3B). Autoradiography also demonstrated no significantly increased uptake in the denervated tibialis anterior relative to the control (95% confidence interval 0.83–1.22, n = 5) (Fig. 3C, D).

We established a denervation model using left sciatic transection. Three weeks later, ¹⁸F-ASEM PET/MRI showed increased uptake in the denervated left tibialis anterior muscle relative to the control muscle (Fig. 4A). The lesion SUVmax and SUVmean of ¹⁸F-ASEM PET/MRI were significantly higher than those of the control (1.02 ± 0.20 vs. 0.56 ± 0.12, P = 0.0277; 0.79 ± 0.18 vs. 0.56 ± 0.12, P = 0.0277; n = 6, respectively) (Fig. 4B). There was no obvious difference in the ¹⁸F-ASEM PET/MRI results between the denervation mouse models at 1 week and 3 weeks.