Synthesis, Biodistribution, and Radiation Dosimetry of a Novel mGluR5 Radioligand: 18F-AZD9272

The metabotropic glutamate receptor subtype mGluR5 has been proposed as a potential drug target for CNS disorders such as anxiety, depression, Parkinson’s disease, and epilepsy. The AstraZeneca compound AZD9272 has previously been labeled with carbon-11 and used as a PET radioligand for mGluR5 receptor binding. The molecular structure of AZD9272 allows one to label the molecule with fluorine-18 without altering the structure. The aim of this study was to develop a fluorine-18 analogue of AZD9272 and to examine its binding distribution in the nonhuman primate brain in vivo as well as to obtain whole body radiation dosimetry. 18F-AZD9272 was successfully synthesized from a nitro precursor. The radioligand was stable, with a radiochemical purity of >99% at 2 h after formulation in a sterile phosphate buffered solution (pH = 7.4). After injection of 18F-AZD9272 in two cynomolgus monkeys, the maximum whole brain radioactivity concentration was 4.9–6.7% of the injected dose (n = 2) and PET images showed a pattern of regional radioactivity consistent with that previously obtained for 11C-AZD9272. The percentage of parent radioligand in plasma was 59 and 64% (n = 2) at 120 min after injection of 18F-AZD9272, consistent with high metabolic stability. Two whole body PET scans were performed in nonhuman primates for a total of 231 min after injection of 18F-AZD9272. Highest uptakes were seen in liver and small intestine, followed by brain and kidney. The estimated effective dose was around 0.017 mSv/MBq. 18F-AZD9272 shows suitable properties as a PET radioligand for in vivo imaging of binding in the primate brain. 18F-labeled AZD9272 offers advantages over 11C-AZD9272 in terms of higher image resolution, combined with a longer half-life. Moreover, based on the distribution and the estimated radiation burden, imaging of 18F-AZD9272 could be used as an improved tool for quantitative assessment and characterization of AZD9272 binding sites in the human brain by using PET.


■ INTRODUCTION
Glutamate is the brain's main excitatory neurotransmitter primarily located on the membranes of neuronal and glial cells and is present in over 50% of nervous tissue. 1 It is particularly abundant in the human nervous system and mostly prominent in the human brain. Glutamate acts through ionotropic (NMDA, kainite, and AMPA) and metabotropic glutamate (mGlu) receptor subtypes and affects cells through a signal transduction cascade. 2 Metabotropic glutamate receptors (mGluRs) are a family of G protein-coupled receptors. In 1992, mGluR5 was first cloned in animals and followed by humans several years later. 3 Even though its actions are mostly excitatory, 4 there are strong links and receptor interactions between mGluR5 and the NMDA receptor. 5 It has also been reported that the activation of mGluR5 enhances GABA, especially in the nucleus accumbens. 6 The density of mGluR5 is high primarily in the forebrain regions, striatum, and limbic regions. 7,8 Moreover, the density of mGluR5 is much higher in younger animals than in adults suggests that early intervention targeting the mGluR5 may lead to prevent the neurodevelopmental disorders. 9 It is believed that presynaptic mGluR5 receptors participates in the regulation of synaptic plasticity and changes in neuronal excitability to maintain homeostasis. 10 Because of its functions in different neuronal processes, mGluR5 has been suggested as a therapeutic drug target for several neurological disorders such as Parkinson's diseases, 11,12 anxiety, 13,14 depression, 15 schizophrenia, 16 seizure disorder, 17 addiction 18 as well as various chronic pain states. 19 Positron emission tomography (PET) has been widely utilized in visualizing the localization of mGluR5. Imaging brain mGluR5 with PET in humans has been useful for studying neurodegenerative disorders and addiction. 20 Several radioligands ( Figure 1) [ 18 F]FPEB, 21 11 C-ABP688, 22 18 F-FPECMO, 23 11 C-MPEP, 24 11 C-M-FPEP, 25 11 C-M-MTEB, 25 18 F-SP203, 26 18 F-MTEB, 27 11 C-SP203, 28 and 18 F-FE-DABP688 29 have been developed and applied preclinically or clinically to image brain mGluR5. Most of those are diaryl alkynes, structural analogues of the prototype mGluR5 NAM MPEP. A noncompetitive mGluR5 antagonist AZD9272 which does not depend on the presence of alkyne moiety was developed by AstraZeneca. 30 AZD9272 was previously labeled with carbon-11 and evaluated as a potential mGluR5 radioligand in nonhuman primates (NHPs) and human subjects. 31,32 However, due to the short half-life of carbon-11 (20.4 min), the use of such a radioligand is restricted to imaging facilities that are close to the site of production. In addition, a three step synthesis of 11 C-AZD9272 using palladium mediated 11 C-cyanation was associated with difficulties in automation. 31 Since AZD9272 has two fluorine functional groups, it allows the radiolabeling with 18 F from the corresponding nitroprecursor without altering the structure. Fluorine-18 labeled PET tracers may provide advantages for use in PET imaging. Relatively long half-life (109.8 min) facilitates imaging at later time-points and lower positron energy (0.635 MeV) allows higher intrinsic resolution in the PET images. Fluorine-18  The objectives of the present PET-study were to (i) to develop a fast and efficient synthetic method for labeling of a mGluR5 ligand AZD9272 with fluorine-18, (ii) to evaluate its binding in vivo in the NHP brain, and (iii) to provide dosimetry estimates for 18 F-AZD9272 based on NHP wholebody PET measurements.
Brain PET. The injected radioactivity of 18 F-AZD9272 was 155 and 155 MBq for NHP1 and NHP2, respectively. The MA at the time of injection was 45 and 31 GBq/μmol, and the injected mass was 0.98 and 1.46 μg. Fusion images of MRI and summated PET of NHP2 are shown in Figure 3A. The uptake was highest in the striatum, lower in the thalamus and frontal cortex, and lowest in the cerebellum and occipital cortex. The corresponding regional brain time−activity curves (TACs) of NHP2 are shown in Figure 3B. The total distribution volume (V T ) obtained by the two tissue compartment (2TC) model and Logan graphical analysis (GA) using metabolite corrected plasma radioactivity are shown in Table 2. V T 's by GA are well correlated to those by 2TC although with slightly lower values (11% and 6%) ( Figure 4A). Bias is quite small as within 3% at 75 min for both 2TC and GA ( Figure 4B). The brain uptake of 18 F-AZD9272 is consistent with that previously reported for 11 C-AZD9272. 33 Also, the V T values obtained by 2TC are similar between 18 F-AZD9272 and 11 C-AZD9272, e.g., 17.3 vs 18.0 mL/cm 3 (ventral striatum) and 5.7 vs 6.5 mL/cm 3 (occipital cortex).
V T values estimated by GA were well correlated with those by 2TC, although slightly lower values were obtained. Additionally, a short length of measurement showed a small bias in V T values obtained with both 2TC and GA. This indicates that GA with a short duration of measurement (e.g., 75 min) could be used for the estimation for V T of 18 F-AZD9272.
The consistent binding pattern observed with 18 F-and 11 Clabeled AZD9272 further confirms that the regional brain distribution for radiolabeled AZD9272 differs from that for other mGluR5 radioligands. 33 Observations reported elsewhere 31 suggest that AZD9272 displays additional, non-mGluR5-related binding that may represent specific binding to monoamine oxidase-B (MAO-B). Based on our previous observations, 31 close to 90% of specific binding for 18 F-AZD9272in NHP brain can be inhibited by administration of a MAO-B-selective ligand. Moreover, the MAO-B component of 11 C-AZD9272 binding has been estimated to be approximately 90−95% in human brain stem, thalamus, amygdala, cerebellum, and ventral striatum. The significant contribution MAO-B binding to the signal for AZD9272 suggests that 18 F-AZD9272could serve as a radioligand for assessment of MAO-B availability in these regions of the primate brain.  Nevertheless, characterization as a tracer for imaging brain MAO-B would require comparison of regional AZD9272 binding with that of established MAO-B radioligands in the same subjects. Taken together, 18 F-AZD9272 shows favorable properties in terms of improved stability and image resolution and offers advantages over 11 C-AZD9272 for further identification and characterization of AZD9272 binding to MAO-B.
Radiometabolite Analysis. The recovery of radioactivity from plasma into acetonitrile after deproteinization was higher than 95%. The parent compound was more abundant at 5 min representing approximately 95% and it decreased to 65% at 120 min ( Figure 5A). The identity of the radio metabolite 18 F-AZD9272 was confirmed by coinjection with the authentic nonradioactive AZD9272. The similar time-course for radio metabolism was observed for 11 C-AZD9272 ( Figure 5B). 33 HPLC analysis of plasma following injection of 18 F-AZD9272, parent compound and one radioactive metabolite was detected which eluted at 6.7 and 6.3 min ( Figure 5C). The abundance of the radiometabolite at 6.3 min increases with time and representing approximately 4% at 5 min and it increased to 17% at 120 min. The identity of the radiometabolite was not analyzed. In principle, it is a possibility that the metabolite is sufficiently lipophilic to enter the brain. However, according to the time stability analysis, the bias to the duration of measurements was small ( Figure 4B. This observation provides indirect support for the view that the radiometabolite is less likely to impact the quantitative estimates. Whole Body PET. The injected radioactivity of 18 F-AZD9272 was 217 and 199 MBq for the two NHPs (NHP3 and NHP4) respectively. The MA at the time of injection was 41 and 58 GBq/μmol, and the injected mass was 1.53 and 0.97 μg. CT and PET images of maximum intensity projection (MIP) over time of NHP4 are shown in Figure 6 and the time activity curves of NHP4 are shown in Figure 7. High uptakes were seen in liver and small intestine, followed by brain and kidney. Human radiation dose estimates indicate that most organs appear to receive around 0.01−0.02 mSv/MBq (Table  3). The upper large intestine (ULI) wall appears to receive the highest dose, around 0.068 mSv/MBq. The estimated radiation dose of 18 F-AZD9272 was 0.017 mSv/MBq which is similar to that of other 18 F-labeled radioligands such as 18 F-FDG (0.015−0.029 mSv/MBq) and 18 F-AV-45 (0.016−0.020 mSv/MBq). 34 It means that 570 MBq can be injected to humans if the limit is 10 mSv for the total radiation exposure. It would allow multiple PET examinations on the same research subject in studies of the human brain.

■ CONCLUSION
The present study demonstrated that the radioligand 18 F-AZD9272 can be efficiently labeled with fluorine-18. A PET measurement in two cynomolgus monkey showed high brain uptake which is similar to the previously developed radioligand 11 C-AZD9272. The estimated radiation burden of 18 F-AZD9272 was 0.017 mSv/MBq based on the distribution which may allow repeated scans in the same human subject. 18 F-AZD9272 offers advantages over 11 C-AZD9272 for further identification and characterization of potential AZD9272 binding to MAO-B.
Quality Control of 18 F-AZD9272. The radiochemical purity, identity, and stability of 18 F-AZD9272 was determined by using an analytical HPLC system which included a ACE RP column (C18, 3.9 Ø × 250 mm, 5 μm particle size), Merck-Hitatchi L-7100 Pump, L-7400 UV detector, and GM tube for radioactivity detection (VWR International). The mobile phase CH 3 CN/0.1% TFA with a gradient HPLC method (10−90% in 8 min) and flow rate of 3 mL/min was used to elute the product. The effluent was monitored with a UV absorbance detector (λ = 270 nm) coupled to a radioactive detector (b-flow, Beckman, Fullerton, CA). The retention time (Rt) of 18   ACS Chemical Neuroscience pubs.acs.org/chemneuro Research Article column (C18, 3.9 Ø × 250 mm, 5 μm particle size) using mobile phase CH 3 CN/50 mM H 3 PO 4 (35/65) and flow rate of 2 mL/min. MA was calibrated for UV absorbance (λ = 270 nm) response per mass of ligand and calculated as the radioactivity of the radioligand (GBq) divided by the amount of the associated carrier substance (μmol). Each sample was analyzed three times and compared to a reference standard also analyzed three times. PET Measurements in a Cynomolgus Monkey. The study was approved by the Animal Ethics Committee of the Swedish Animal Welfare Agency (Dnr N185/14) and was performed according to "Guidelines for planning, conducting and documenting experimental research" (Dnr 4820/06-600) of Karolinska Institutet. The NHPs were housed in the Astrid Fagraeus Laboratory (AFL) of the Swedish Institute for Infectious Disease Control, Solna, Sweden.
Brain PET. Two male cynomolgus monkeys (NHP1: 5460 g and NHP2: 7080 g) were used in this study. Parts of the data have been reported. 32 Anesthesia was induced by intramuscular injection of ketamine hydrochloride (approximately 10 mg/kg) at AFL and maintained by intravenous infusion of ketamine (4 mg/kg/h) and xylazine (0.4 mg/kg/h) with a pump. The NHP head was immobilized with a fixation device. Body temperature was maintained by using a Bair Hugger system (model 505, Arizant Healthcare, MN) and monitored with an esophageal thermometer. Heart rate, blood pressure, respiratory rate, and oxygen saturation were continuously monitored throughout the experiments. Fluid balance was maintained by continuous infusion of saline.
PET measurements were conducted using a High Resolution Research Tomograph (HRRT) PET scanner (Siemens Molecular Imaging, Knoxville, TN). A transmission scan of 6 min using a single 137 Cs source was performed before the 18 F-AZD9272 injection. Listmode data were reconstructed with a series of 34 frames (20 s × 9, 1 min × 3, 3 min × 5, and 6 min × 17), using the ordinary Poisson-3Dordered subset expectation maximization (OP-3D-OSEM) algorithm with 10 iterations and 16 subsets including modeling of the point spread function (PSF). Only a 105 min image of NHP1 was obtained due to the technical issue.
An automated blood sampling system (ABSS) was used to measure the continuous radioactivity for the first 3 min after the radioligand injection. Then, blood sampling was performed manually for the measurement of metabolism and radioactivity at 5, 15, 30, 60, 90, and 120 min after the injection. A reversed-phase radio-HPLC method was used to determine the amount of unchanged 18 F-AZD9272 and its radioactive metabolites in NHP plasma. 35  The regions of interest (ROIs) were delineated manually on the MRI images of each NHP for the anterior cingulate cortex, amygdala, caudate, cerebellum, hippocampus, insular cortex, occipital cortex, prefrontal cortex, putamen, temporal cortex, thalamus, ventral midbrain, and ventral striatum. The MRI of the individual NHP was co-registered to summed PET images of the whole measurement. By applying the co-registration parameters to ROIs, the time−activity curves of brain regions were generated from dynamic PET data.
As the main outcome measure, the total distribution volume (V T ) defined as (K 1 /k 2 )(k 3 /k 4 + 1) by the two tissue compartment (2TC) model and Logan graphical analysis (GA) was calculated with the metabolite corrected plasma radioactivity as the input function. The acquisition data was truncated to 63 min from 105 min (NHP1) or 123 min (NHP2) to check the time stability of V T . The procedure has been described in detail in a previous publication on [ 11 C]AZD9272 binding. 33 Whole Body PET. Two female cynomolgus monkeys (NHP3: 7420 g and NHP4: 6360 g) were used in this study. Anesthesia was induced by intramuscular injection of ketamine hydrochloride (approximately 10 mg/kg) at AFL and maintained by intravenous infusion of ketamine (4 mg/kg/h) and xylazine (0.4 mg/kg/h) with a pump. The body of the NHP was immobilized using a vacuum pad. Body temperature was maintained by using a Bair Hugger system (model 505, Arizant Healthcare, MN) and monitored with an esophageal thermometer. Heart rate, blood pressure, respiratory rate, and oxygen saturation were continuously monitored throughout the experiments. Fluid balance was maintained by continuous infusion of saline.
Whole-body PET scans were conducted using a GE Discovery PET/CT 710 system (GE Healthcare, Waukesha, WI). One low-dose CT scan was performed before intravenous administration of 18 F-AZD9272 for attenuation correction. Then five series of PET acquisitions, each covering five axial fields of view (AFOV), were conducted. The five PET series consisted of two 20 s × 5 AFOV, three 40 s × 5 AFOV, three 80 s × 5 AFOV, three 160 s × 5 AFOV, and eight 240 s × 5 AFOV. The total duration of the whole-body PET measurement was around 4 h. PET images were reconstructed with a 3D ordered-subset expectation maximization (OSEM) algorithm with three iterations and 18 subsets, including the time-of-flight information (VUE Point FX) and the point spread function correction (Sharp IR). A 2D Gaussian filter with 5.5 mm cutoff was used.
ROIs were drawn for the brain, heart, liver, lung, kidney, gall bladder, bone (lumbar vertebrae), urinary bladder, stomach, small intestine, spleen, and pancreas with the help of the CT images for anatomic landmarks. The time−activity curve was expressed as the percentage of the injected dose (%ID) calculated as follows: decaycorrected radioactivity (Bq/cc) × ROI volume (cc)/injected dose (Bq) × 100.
Estimates of the absorbed radiation dose in humans were calculated with OLINDA/EXM 1.1 (Organ Level INternal Dose Assessment Code) software, using the adult male (70 kg) reference model. 36 The fractional uptake in NHP organs was assumed to be equal to the uptake in human organs.
Radiometabolite Analysis. Radiometabolite analysis was performed following a previously published method. 35 In short, a reverse-phase HPLC method was used for the determination of the percentages of radioactivity corresponding to unchanged radioligand 18 F-AZD9272 and its radioactive metabolites during the course of a PET measurement. Venous blood samples (2 mL) were obtained from the monkey at different time point such as 5, 15, 30, 45, 60, 90, and 120 min after injection of 18 F-AZD9272. Collected blood (2 mL) was centrifuged at 2000g for 2 min to obtain the plasma (0.5 mL). The plasma obtained after centrifugation of blood at 2000g for 2 min was mixed with a 1.4 times volume of acetonitrile. The mixture was then centrifuged at 2000g for 4 min, and the extract was separated from the pellet and was diluted with water before injecting it into the HPLC system coupled to an online radioactivity detector. An Agilent binary pump (Agilent 1200 series) coupled to a manual injection valve (7725i, Rheodyne), 1−3.0 mL loop and a radiation detector (Oyokoken, S-2493Z) housed in a shield of 50 mm thick lead was used for metabolite measurements. Data collection and control of the LC system was performed using chromatographic software (Chem-Station Rev. B.04.03; Agilent). The accumulation time of the radiation detector was 10 s. Chromatographic separation was achieved on an ACE C18 column (250 mm × 10 mm I.D) by gradient elution. Acetonitrile (A) and 10 mM ammonium format (B) were used as the mobile phase at 5.0 mL/min, according to the following program: 0− 8.5 min, (A/B) 50:50 → 95:5 v/v; 8.5−11.0 min, (A/B) 95:5 v/v. Peaks for radioactive compounds eluting from the column were integrated, and their areas were expressed as a percentage of the sum of the areas of all detected radioactive compounds (decay-corrected to the time of injection on the HPLC).
To calculate the recovery of radioactivity from the system, an aliquot (2 mL) of the eluate from the HPLC column was measured and divided with the amount of total injected radioanalytes.