Synthesis and Evaluation of 18F-labeled Pyridaben Analogues for Myocardial Perfusion Imaging in Mice, Rats and Chinese mini-swine

This study reports three novel 18F-labeled pyridaben analogues for potential myocardial perfusion imaging (MPI). Three precursors and the corresponding nonradioactive compounds were synthesized and characterized. The radiolabeled tracers were obtained by substituting tosyl with 18F. The total radiosynthesis time of these tracers was 70–90 min. Typical decay-corrected radiochemical yields were 47–58%, with high radiochemical purities (>98%). Tracers were evaluated as MPI agents in vitro, ex vivo and in vivo. In the mouse biodistribution study, all three radiotracers showed high initial heart uptake (34–54% ID/g at 2 min after injection) and fast liver clearance. In the microPET imaging study, [18F]Fmpp2 produced heart images with good quality in both mice and rats. In the whole-body PET/CT images of mini-swine, [18F]Fmpp2 showed excellent initial heart standardized uptake value (SUV) (7.12 at 5 min p.i.) and good retention (5.75 at 120 min p.i.). The heart/liver SUV ratios were 4.12, 5.42 and 5.99 at 30, 60 and 120 min after injection, respectively. The favorable biological properties of [18F]Fmpp2 suggest that it is worth further investigation as a potential MPI agent.

chains. Herein, we report the synthesis of three novel 18 (Fig. 1). The yields of the precursors mpp1-OTs, mpp2-OTs, mpp3-OTs were 64%, 68% and 58%, respectively. The yields of non-radioactive references [ (Fig. 2). The RCPs of three tracers calculated from radio-HPLC chromatogram were over 98% after purification. The specific activities were 20-40 GBq/μ mol at the end of synthesis as determined by HPLC analysis.
Physicochemical properties study. The  , respectively, at 60 min p.i. compared with that at 2 min p.i.), which affected the heart/liver ratios at subsequent time points. Among the three radiotracers, [ 18 F]Fmpp2 showed the best biological properties. Its heart/ liver, heart/lung and heart/blood ratios were higher than 3.5 at all time points, encouraging further evaluation. All three radiotracers exhibited distinct uptake in muscle, which might be due to the high mitochondrial expression in calf muscle. Low bone uptake was observed in [ 18 F]Fmpp2 and [ 18 F]Fmpp3, indicating no defluorination of those tracers in vivo. The relative increase in [ 18 F]Fmpp1 in bone uptake might be due to its instability in vivo, which is consistent with the stability study. All three radiotracers had notable uptake in the kidneys, indicating that these tracers were excreted via the renal system. Whole-body PET/CT imaging in Chinese mini-swine. The behavior of tracers in Chinese mini-swine is different from that in mice. All three tracers showed high initial heart uptake in Chinese mini-swine (Fig. 3 Fmpp3 was 59% and 48% lower than that at 5 min p.i. Furthermore, there was low uptake in other non-target organs and tissues, except the gallbladder, kidneys and bladder. Among these three tracers, [ 18 F]Fmpp2 provides images with the best quality. Its heart/liver SUV ratios were 1.72, 4.12, 5.42 and 5.99 at 5, 30, 60 and 120 min, respectively (Table S1 of Supplementary file). In addition, the low uptakes and fast clearance were observed for non-target organs. The SUVs of [ 18 F]Fmpp2 in most of the non-target tissues were negligible after 30 min p.i., including the liver, lung, kidneys and blood. As shown in Fig. 3, the heart can be observed clearly from 5 min to 120 min p.i., almost without any interruption from the nearby tissues. Thus, we could obtain quality images within a wide range of time (from 30-120 min p.i.), which is convenient for diversified clinical imaging protocols.
The images of [ 18 F]Fmpp3 were also commendable. Its early heart uptake was not as outstanding as [ 18 F] Fmpp2. However, it had preferable retention in the heart (from 4.51 at 5 min p.i. to 4.53 at 120 min p.i., SUVs) and faster liver clearance (from 4.10 at 5 min p.i. to 1.12 at 120 min p.i., SUVs). The heart/liver SUV ratios were 2.56, 3.98 and 4.04 at 30, 60, and 120 min p.i., respectively (Table S1 of Supplementary file). Accordingly, the images obtained during 30-120 min p.i. had good signal-noise ratios. However, although the other tracer [ 18 F]Fmpp1 also had high initial heart SUV (6.82 at 5 min p.i.), the clearance from the heart (2.19 at 120 min p.i., SUV) was much faster than the other two tracers. Thus, the outline of the heart was not as clear as the other tracers of [ 18 F] Fmpp2 and [ 18 F]Fmpp3 at subsequent time points.
All three tracers had notable initial uptake in the kidneys. The radioactivities were quickly washed out to the bladder, indicating that they were excreted mainly via the renal system. This result was consistent with that of the mouse biodistribution. No obvious bone uptake was observed in all three tracers, indicating that there was no

MicroPET/CT studies in mice and rats. To visualize the distribution of [ 18 F]Fmpp2 in different animal
species, microPET imaging studies were performed in mice and rats (Fig. 4). [ 18 F]Fmpp2 had notable heart uptake in the mice at 5 and 30 min p.i. However, most of the radioactivities were washed out from the heart at 120 min p.i., which was coincident with the result of the biodistribution study. The liver clearance of [ 18 F]Fmpp2 was faster than that of the heart (Table S2 of Supplementary file). Thus, we could obtain clear images at 30 min p.i., with a good heart/liver ratio (1.52). The lower heart/liver ratios in microPET images compared to the biodistribution study may be due to different injection doses and quantification methods.
From the microPET imaging results of rats (Fig. 4), [ 18 F]Fmpp2 had much higher heart uptake and better retention than that of mice (Table S2 of Supplementary file). At 120 min p.i., only approximately 43% radioactivity was cleared from the heart compared with that at 5 min p.i. The heart clearance of rats was much slower than that of mice (74% at the same condition), resulting in higher heart/blood, heart/liver, heart/lung and heart/muscle ratios. The outline of the heart was still visible at 120 min p.i.
In general, microPET imaging of [ 18 F]Fmpp2 in both mice and rats could provide legible images at 30 min p.i. There was specific uptake in the kidneys of both mice and rats, indicating that the kidneys were the main organs for excretion.
Metabolic stability. The metabolic stabilities of [ 18 F]Fmpp2 were investigated in mice, rats and Chinese mini-swine. As shown in Fig. 5, [ 18 F]Fmpp2 had much better stability in the hearts of rat and Chinese mini-swine than that of mice. According to the HPLC analysis of the extracts from hearts of mouse, rat and Chinese mini-swine, approximately 45%, 13% and 13% activities were metabolized into more hydrophilic compounds at   60 min p.i., respectively. In blood and urine samples, the peak of [ 18 F]Fmpp2 almost disappeared at 60 min p.i. in all animal models using HPLC analysis, suggesting that the tracer was transformed into more hydrophilic compounds in both the blood and urine of mouse, rat and Chinese mini-swine.  Fig. 3, the radioactivities were washed out extremely quickly from non-target tissues, including the kidneys, which can definitely lower the irradiation dose.

Discussion
The different behaviors of [ 18 F]Fmpp2 in mice, rats and Chinese mini-swine may be due to its metabolic stability in different animal species. In the heart of mice, 45% activities of [ 18 F]Fmpp2 were metabolized at 60 min p.i., accompanied by 57% activities of [ 18 F]Fmpp2 washed out from the heart at the same time. Fortunately, the heart clearance of [ 18 F]Fmpp2 was much slower in rats and Chinese mini-swine. This may due to the better stabilities in the hearts of rat and Chinese mini-swine. Only 13% activities of [ 18 F]Fmpp2 were metabolized in both the hearts of rat and Chinese mini-swine at 60 min p.i., and 43% and 19% activities washed out from the hearts of rat and Chinese mini-swine, respectively, at 120 min p.i. (compared to the data at 5 min p.i.). Further studies, such as imaging with ischemia animal models and evaluation with microspheres, are needed in the future.  Other reagents and solvents were purchased from commercial suppliers. The reversed-phase HPLC analysis for metabolic stability studies was performed on a Waters 2535Q system with a Gabi-star flow-counter (Raytest Inc. Germany). HPLC for other studies were performed on an SHIMADZU system with LC-20AT pumps and a B-FC-320 BIOSCAN flow-counter. A C-18 reverse-phase semi-preparative HPLC column (10 × 250 mm, 5-μ m particle size, Venusil MP-C18) was purchased from Agela Technologies Inc. (USA). A Labgen 7 homogenizer was purchased from Cole-Parmer Instrument (USA). 1 H NMR spectra were recorded on a Bruker (400 MHz) spectrometer, while 13 C NMR spectra were recorded on a Bruker (100 MHz) spectrometer. Chemical shifts are reported in δ (ppm) values. Infrared spectra were measured on a Nicolet 360 Avatar instrument, scanning from 400 to 4000 cm −1 . Mass spectra were recorded using a Brucker Apex IV FTM instrument.

Methods
Kunming mice (18-20 g) and rats (approximately 200 g) were obtained from the Animal Center of Peking University for biodistribution and metabolic stability studies. Kunming mice (approximately 23 g) and rats (approximately 220 g) were obtained from Xiamen University Laboratory Animal Center for microPET imaging. Healthy Chinese mini-swine (approximately 15 kg) were obtained from the Animal Center of Fu Wai Hospital. All animal studies were performed under a protocol approved by the Beijing Administration Office of Laboratory Animal (BAOLA).
The solution of tert-butylammonium fluoride (1 mmol in 1 mL tetrahydrofuran) was stirred in a stream of nitrogen at 110 °C to remove the solvent. Next, mpp1-OTs (or mpp2-OTs, mpp3-OTs) (0.30 mmol in 3 mL anhydrous CH 3 CN) were added and refluxed for 2 h. After concentration under reduced pressure, the residue was chromatographically separated over a column of silica gel and eluted with a mixture of hexane and ethyl acetate = 1:2 (v/v) for [ 19 (Fig. 1)  The final RCPs were determined by re-injection of the product onto a radio-HPLC. The radioactive fraction was collected and measured in a dose calibrator for specific activity calculation. The mass of the product was calculated by comparing the area under the UV curve at 254 nm with that of a standard reference.
Physicochemical properties studies. The octanol/water partition coefficient was measured using a shake-flask method as previously described 28 . In the study of vitro stabilities 29 , radiotracers were incubated in water at room temperature for 3 h, or in 0.5 mL murine plasma at 37 °C for 2 h, respectively. Plasma proteins were precipitated by the addition of 100 μ L acetonitrile and removed by centrifugation. Next, the RCPs were assayed using radio-HPLC.
Biodistribution study. Biodistribution studies were carried out in accordance with the approved guidelines. MicroPET/CT studies in mice and rats. PET/CT imaging studies of mice and rats were carried out in accordance with the approved guidelines. For microPET imaging studies, mice were injected with approximately 10 MBq of [ 18 F]Fmpp2 via the tail vein (n = 4 for each group). For static scans, the mice were anesthetized with isoflurane (5% for induction and 2% for maintenance in pure oxygen gas) using a knockdown box at 5, 30 and 120 min p.i. With the help of a laser beam attached to the scanner, the mice were placed in the prone position near the center of the FOV of the scanner. Static microPET images were obtained within 5 min after whole body CT scan (6 min). The images were reconstructed using ordered subset expectation maximization with three-dimensional resolution recovery (OSEM 3D) with CT-based attenuation correction, and scatter correction. For data analysis, the region of interest (ROI) was manually drawn and covered the entire target on the CT images. This ROI was copied to the corresponding PET images. The mean SUV of the heart, liver and lung in the ROIs was recorded.
For the dynamic scan, the acquisition began at 30 s after injection and continued for 30 min. The micro-PET and CT images were generated separately and then fused using Inveon Research Workplace (Siemens). The images were reconstructed using ordered subset expectation maximization with three-dimensional resolution recovery (OSEM 3D) with CT-based attenuation correction, and scatter correction. The reconstruction frames were 12 × 5, 8 × 30, 5 × 300 s for dynamic imaging. For data analysis, the region of interest (ROI) was manually drawn and covered the entire target on the CT images. This ROI was copied to the corresponding PET images. The mean SUVs of the heart, liver and lung in the ROIs were recorded.
MicroPET/CT imaging in rats were also performed according to the same procedure described above with more activities (approximately 35 MBq per rat).

Metabolic stability.
Metabolic studies were carried out in accordance with the approved guidelines.
Wildtype Kunming mice, rats and Chinese mini-swine were intravenously injected with 3.7 MBq, 18.5 MBq and 74 MBq [ 18 F]Fmpp2, respectively. Mice and rats were sacrificed at 60 min p.i. Blood, urine and heart samples were collected as previously described 30 . Chinese mini-swine was anesthetized 23 . At 60 min p.i., blood was collected from the superior vena cava. Urine was collected from the bladder. Heart sample was collected from the anterior wall of the heart. All the samples were treated as previously described 30 . HPLC analysis was performed on a Waters 2535Q system with a Gabi-star flow-counter (Raytest Inc.) to improve the signal-noise ratios. The retention time of [ 18 F]Fmpp2 was confirmed by co-injecting with [ 19 F]Fmpp2 onto a HPLC.