Synthesis and in vivo characterization of 18F-labeled difluoroboron-curcumin derivative for β-amyloid plaque imaging

Positron emission tomography imaging of β-amyloid (Aβ) plaques has proven useful in the diagnosis of Alzheimer’s disease. A previous study from our group showed that 4′-O-[18F]fluoropropylcurcumin has poor brain permeability, which is thought to be due to its rapid metabolism. In this study, we synthesized difluoroboron complexes of fluorine-substituted curcumin derivatives (1–4) and selected one of them based on the in vitro binding assays. The selected ligand 2 was found to distinctively stain Aβ plaques in APP/PS1 transgenic mouse brain sections. Radioligand [18F]2 was synthesized via a two-step reaction consisting of [18F]fluorination and subsequent aldol condensation. Biodistribution and metabolism studies indicated that radioligand [18F]2 was converted to polar radioactive products and trapped in the normal mouse brain. In contrast, optical images of mice acquired after injection of 2 showed moderate fluorescence signal intensity in the mouse brain at 2 min with a decrease in the signal within 30 min. In the ex vivo optical images, the fluorescence signals in major tissues disappeared within 30 min. Taken together, these results suggest that [18F]2 may be converted to polar 18F-labeled blue-shifted fluorescent products. Further structural modifications are thus needed to render the radioligand metabolically stable.


Results
Synthesis of non-radioactive ligands. Ligands 1 and 2 were synthesized from condensation reactions of difluoroboron compounds 7 and 8 with 4-(2-(2-fluoroethoxy)ethoxy)benzaldehyde (11) in the presence of piperidine, respectively (Fig. 2). Compounds 7 and 8 were prepared by the aldol condensation reaction of 2,4-pentanedione and benzaldehyde derivatives, followed by difluoroboron complex formation 3,6,11 . Compound 11 was prepared from 4-hydroxybenzaldehyde in three steps in high yield; the final fluorination step was carried out by reacting 10 with CsF in t-BuOH, which provided 11 in higher yield than in acetonitrile 12 . Ligand 3 was synthesized from condensation reaction of 14 with difluoroboron compound 16 in the presence of piperidine (Fig. 3). Compounds 14 and 16 were prepared from vanillin in 3 and 2 steps, respectively. Ligand 4 was synthesized from condensation reaction of 8 with 4-(3-fluoro-2-hydroxypropoxy)benzaldehyde (20) in the presence of piperidine (Fig. 4). Compound 20 was prepared from 4-hydroxybenzaldehyde in five steps; reaction of  4-hydroxybenzaldehyde and (±)-3-chloro-1,2-propanediol, tosylation of the 3-hydroxy group, protection of the 2-hydroxy group, fluorination of the tosylate group, and then deprotection of the THP group. Ligands 1, 3, and 4 were not completely pure after flash column chromatography. This result is likely due to the presence of the monomethylamino group or hydroxy group in those ligands. Therefore, all the ligands were purified by HPLC after flash column chromatography.
Radiochemical synthesis. Radioligand [ 18 F]2 was synthesized by [ 18 F]fluorination of 10, followed by aldol condensation with 8 (Fig. 5). A small amount of the precursor (1.4 μmol) was used for radiolabeling to increase the molar activity of the radioligand. Displacement of the base from piperidine to n-butylamine and reduction in amount of the base increased the yield of the aldol condensation reaction from 9% to 37% 11 . The following HPLC purification gave [ 18 F]2 in overall 17-22% decay-corrected radiochemical yield with a molar activity of 39-46.8 GBq/μmol. Excitation and emission spectra. Ligand 2 showed a red-shifted emission and a larger Stokes shift (excitation 550 nm, emission 650) compared to those of curcumin (excitation 510 nm, emission 560) ( Supplementary  Fig. S15) 6 . CRANAD-2 displayed far more red-shifted emission (760 nm) than did ligand 2 because of the presence of dimethylamino groups at both ends (Fig. 1). Despite this, ligand 2 has sufficient red-shifted emission for optical imaging.
In vitro binding assays using Aβ(1-42) aggregates. Binding constants of the ligands for Aβ  aggregates were determined using the intrinsic fluorescence of the ligands. The binding constant of CRANAD-2 was also measured for comparison. Its K d value measured in this study was 19.40 nM, while the reported value   was 38.69 nM using Aβ(1-40) aggregates 6 . The ligand with a dimethylamino group (2) exhibited higher binding affinity than did the ligand with a monomethylamino group (1)  Staining of Aβ plaques in double transgenic mouse brain sections. Fluorescent staining of Aβ plaques by 2 was performed using the brain sections from a double transgenic mouse (Tg APP/PS-1). Localization of the Aβ plaques was confirmed by immunofluorescent staining the adjacent brain sections from the transgenic mouse with an anti-Aβ antibody (Fig. 6). The Aβ plaques in the cerebral cortex and hippocampus regions of the transgenic mouse brain sections were stained with 2 ( Fig. 6a-c). The staining pattern was consistent with that observed with an Aβ-specific antibody ( Fig. 6d-f). In contrast, there was no notable positive staining by either 2 or anti-Aβ antibody in wild-type mouse brain sections ( Fig. 6g-l).
Partition coefficient measurement. Partition coefficient of [ 18 F]2 was measured using a mixture of 1-octanol and water, and its log P value was 2.58 ± 0.04. This result demonstrates that the radioligand can cross the blood-brain barrier. Difluoroboron complex formation lowered the lipophilicity of the curcumin derivative based on TLC analysis.
In vitro stability study. Stability study was performed by incubating [ 18 F]2 in phosphate-buffered saline (PBS) and in fetal bovine serum (FBS) at 37 °C for 120 min. Radio-TLC data showed that [ 18 F]2 remained 89% in PBS and 83% in FBS at the end of the incubation (Fig. 7). These results indicate that radioligand [ 18 F]2 is fairly stable in both PBS and FBS.
Biodistribution study. Biodistribution study of [ 18 F]2 in ICR mice demonstrated high radioactivity accumulation in the liver (26.92 ± 1.58% ID/g) 2 min after injection, which decreased by 60 min (9.36 ± 0.18% ID/g at 60 min). In contrast, radioactivity uptake in the small intestine increased significantly over time from 0.92 ± 0.13% ID/g at 2 min to 14.92 ± 4.06% ID/g at 60 min (Table 1). Radioactivity uptake in the spleen decreased slightly over time (12.59 ± 2.42% ID/g at 2 min to 10.72 ± 2.45% ID/g at 60 min). The radioligand did not appear to undergo metabolic defluorination because of low level of femur uptake (0.58 ± 0.10% ID/g at 2 min to 1.43 ± 0.14% ID/g at 60 min) (Table 1). However, there was poor brain uptake at 2 min after injection with increased uptake over time, Figure 6. A double transgenic mouse brain section (a-c) and an age-matched wild-type mouse brain section (g-i) stained with 2. The adjacent double transgenic mouse brain section (d-f) and age-matched wild-type mouse brain section (j-l) stained with an Aβ-specific antibody. Magnification of the regions in white boxes (c,f,i and l). Scale bars: 100 μm (c,f,i and l) and 200 μm (the rest). www.nature.com/scientificreports www.nature.com/scientificreports/ as follows: 0.49 ± 0.02% ID/g at 2 min, 1.02 ± 0.08% ID/g at 30 min, and 1.19 ± 0.06% ID/g at 60 min (Table 1). This result indicates that there is generation and retention of polar radioactive products in the mouse brain.
Optical imaging. Optical images of hairless Balb/C nude mice were acquired after injection of 2. Moderate fluorescence signal intensity was detected in mouse brain at 2 min after injection, which significantly decreased within 30 min (Fig. 8a). Ex vivo optical images of the mouse brain was consistent with the result of in vivo optical imaging (Fig. 8a-c). Ex vivo images also showed moderate to strong fluorescence signals in the muscle, kidney, intestines, and liver at 2 min after injection, all of which decreased to the background level within 30 min (Fig. 8c). In particular, there was rapid disappearance of the fluorescence signals from the organs of elimination. Considering the log P value of 2, this ligand would have not been eliminated within 30 min after injection, suggesting metabolism/degradation of 2.
Metabolism study. Radio-TLC data of the brain homogenates showed that most of [ 18 F]2 remained in the brain 2 min after injection (84%; Fig. 9a,b). However, the radioligand disappeared almost entirely, and the new polar radioactive peaks appeared at the origin of radio-TLC at 30 and 60 min after injection (82% and 100%, respectively; Fig. 9c,d). The radioligand also disappeared in the blood samples, but at a rate slower than that of the brain samples ( Fig. 9f-h). The polar radioactive products were further analyzed by radio-TLC using a polar developing solvent system, and the polar products moved toward the solvent front. This showed that the polar products did not contain free [ 18 F]fluoride ion, which would remain at the origin of the TLC, even in a polar developing solvent system (Fig. 9e,i) 13,14 . Moreover, the polar products were not adsorbed onto calcium phosphate (7% in calcium phosphate pellet vs. 93% in supernatant), which is a component of the bone matrix. This finding  www.nature.com/scientificreports www.nature.com/scientificreports/ indicated that [ 18 F]2 did not undergo metabolic defluorination in vivo 13,14 . This result was consistent with low bone uptake in mice (Table 1).
In order to further investigate the metabolites of [ 18 F]2, metabolism study of 2 was also performed; the brain homogenates were analyzed by TLC, followed by fluorescence detection. TLC data showed that a fluorescent spot was detected at the same position as that of 2 in the 2-min sample, which did not appear in the 30-and 60-min samples (Fig. 9j). This result was consistent with those of optical imaging, but not with radio-TLC data of [ 18 F]2 (Figs 8 and 9b-d).

Discussion
Difluoroboron-curcumin derivatives, such as CRANAD-2, have been shown to label Aβ plaques in transgenic mouse brain and have potential for NIR imaging of Aβ plaques 6 . In order to extend this class of ligands to radioligands for PET imaging, we attempted a 18 F/ 19 F exchange reaction on one of the fluorine atoms of difluoroboron-curcumin. This is a well-established 18 F/ 19 F exchange reaction on BODIPY dye [15][16][17] . The isotope exchange reaction on the difluoroboron complex of 2,4-pentanedione, a center fragment of curcumin, only afforded 18 F-labeled product when a methyl or ethyl group was substituted at C3, but with high instability (data not shown). Therefore, we synthesized four difluoroboron-curcumin derivatives for the development of 18 F-labeled ligands (Fig. 1). These ligands (1)(2)(3)(4) have substituents, such as 2-(2-fluoroethoxy)ethoxy or 3-fluoro-2-hydroxypropoxy group at the para-position of one of the phenyl rings (Figs 2-4). These substituents have been introduced to radioligands for improvement of in vivo properties as well as radiolabeling with 18 F, as shown in the studies of radioligands for Aβ plaque imaging [18][19][20] .
For synthesis of [ 18 F]2, a one-step [ 18 F]fluorination was performed using the tosylate precursor of the difluoroboron-curcumin derivative. However, this synthesis had very low yield (0-16% based on radio-TLC).
As an alternative, we synthesized a precursor, which was obtained from several steps including the reaction with (±)-epichlorohydrin. However, the epoxide ring opening by n-Bu 4 N[ 18 F]F was not successful 20,21 . Therefore, a two-step reaction consisting of [ 18 F]fluorination of 10, followed by aldol condensation with 8 was conducted and gave [ 18 F]2 in relatively high yield (Fig. 5).
Based on the binding constants of the ligands for Aβ(1-42) aggregates, ligand 2 was selected for in vitro and in vivo studies. To examine whether 2 could label Aβ plaques in vivo, fluorescent staining of Aβ plaques was performed using the double transgenic mouse brain sections. The Aβ plaques in the cerebral cortex and hippocampus regions of the mouse brain sections were stained with ligand 2, and the plaques in the same regions of the adjacent brain sections were also stained with an Aβ-specific antibody ( Fig. 6a-f). This result indicated that ligand 2 distinctively stained Aβ plaques in transgenic mouse brain. In order to predict brain permeability of [ 18 F]2, partition coefficient of the radioligand was measured. Optimal lipophilicity is one of important properties required for brain imaging ligand. Radioligands with measured log P values of 0.1-3.5 are known to penetrate the blood-brain barrier 22 . Therefore, [ 18 F]2 appeared to have favorable brain permeability (log P = 2.58 ± 0.04). The radioligand was further evaluated to predict its in vivo stability. Radio-TLC data showed that [ 18 F]2 was www.nature.com/scientificreports www.nature.com/scientificreports/ stable in both PBS and FBS (>83%) (Fig. 7). The result of [ 18 F]2 in FBS was consistent with the reported result of curcumin, in that curcumin was stable in a cell culture medium containing 10% fetal calf serum and in human blood (>80% at 1 h) 23 . However, the result of [ 18 F]2 in PBS is unlike curcumin, which was degraded to trans-6-(4′-hydroxy-3′-methoxyphenyl)-2,4-dioxo-5-hexenal, vanillin, ferulic acid, and feruloyl methane in phosphate buffer (pH 7.2) 23 .
The in vitro studies demonstrated that [ 18 F]2 satisfied the criteria required for development of Aβ imaging radioligand, and thus, the radioligand was evaluated in vivo. Brain imaging radioligand should have desirable brain pharmacokinetics in normal mice with high brain uptake at early time points after injection (>5% ID/g at 2 min) and fast wash-out from the brain by 30 min (<30% of initial uptake), as there are no Aβ plaques in the normal mouse brain 22,24 . However, the biodistribution study of [ 18 F]2 in normal mice showed poor brain uptake 2 min after injection, which increased over time (Table 1). These unusual brain pharmacokinetic data suggest that [ 18 F]2 may be converted to polar radioactive products in the mouse brain.
In order to investigate whether 2 exhibits the same brain pharmacokinetics as that of [ 18 F]2, optical images of Balb/C nude mice were acquired. Optical images of the mouse brain showed moderate fluorescence signal intensity at 2 min after injection with the background signal intensity within 30 min (Fig. 8a,b). Ex vivo optical images displayed strong signal intensity in the liver and moderate signal intensities in the organs of elimination, such as small and large intestines at 2 min after injection. However, the fluorescence signals decreased significantly within 30 min (Fig. 8c). This result, therefore, suggests that ligand 2 may be rapidly converted to blue-shifted fluorescent products relative to emission of the ligand, and thus the excitation and emission wavelengths used for optical imaging may be no longer optimal for detection of the products formed from 2.
The optical imaging data were not consistent with biodistribution data of mice injected with [ 18 F]2 ( Fig. 8 and Table 1). Radioligand [ 18 F]2 and its non-radioactive equivalent (2) are the same compound with the difference of one neutron, and thus they would have the same metabolites in vivo. Therefore, metabolism study of [ 18 F]2 and 2 was performed, and the data were analyzed using radioactivity and fluorescence, respectively. Samples of the brain and blood of ICR mice were homogenized in acetonitrile to extract out the polar radioactive products, as well as non-polar products, and the supernatants were analyzed. The metabolism study performed using radioactivity showed that most of [ 18 F]2 remained in the brain and blood 2 min after injection, but it was converted to the polar radioactive products within 30 min after injection (Fig. 9). This result was consistent with the biodistribution data of [ 18 F]2 (Table 1). In contrast, the results obtained using fluorescence showed that 2 remained intact in the brain at 2 min after injection but disappeared within 30 min (Fig. 9j). This discrepancy between the optical images and biodistribution data is likely due to the formation of the metabolites from a hybrid PET/fluorescent ligand and their two different detection methods, such as radioactivity and fluorescence.
The results of the in vivo studies suggest that the polar products may have been formed from [ 18 F]2 in the mouse brain via two proposed pathways; either formation of polar 18 F-labeled blue-shifted fluorescent products or formation of polar 18 F-labeled metabolites derived from 2-(2-[ 18 F]fluoroethoxy)ethoxy group with rapid wash-out of the rest non-radioactive product. In both cases, the radioactivity in the brain would increase, but the fluorescence signal would disappear over time. Although the polar radioactive products were not identified, it is unlikely that 2-  Table S1). Moreover, the polar radioactive products did not contain the proposed degradation products, such as 4-(2-(2-fluoroethoxy)ethoxy) benzoic acid and 4-(2-(2-fluoroethoxy) ethoxy)cinnamic acid (Supplementary Table S1). These characteristics are unlike those of similar degradation products that were shown in curcumin 23 . In addition, there are quite a few radioligands substituted with 2-(2-[ 18 F]fluoroethoxy)ethoxy or 2-(2-(2[ 18 F]fluoroethoxy)ethoxy)ethoxy group 22,[25][26][27] . These polyethylene glycol groups are known to control lipophilicity and improve bioavailability of radioligands, and thus they have been introduced to the radioligands for Aβ imaging 22,26,27 . This result led us to propose that [ 18 F]2 may be converted to polar 18 F-labeled blue-shifted fluorescent products.
The in vivo characteristics of [ 18 F]2 are distinct from those of 18 F-labeled curcumin derivatives. In our previous studies, 4′-O-[ 18 F]fluoropropylcurcumin had poor brain permeability at 2 min after injection in normal mice, but showed rapid wash-out from the brain within 30 min. Moreover, the radioligand was intact in the mouse brain once it was taken up by the brain 3 . Similarly, ligand 2 may have different in vivo properties from CRANAD-2, which has dimethylamino groups at both phenyl rings 6 . Therefore, further studies are warranted to identify the polar products formed from [ 18 F]2 in vivo.

Conclusion
Of the four difluoroboron-curcumin derivatives, ligand 2 was selected for radiolabeling, in vitro and in vivo evaluation. Although the ligand was able to distinctively stain Aβ plaques in transgenic mouse brain sections and had suitable lipophilicity, the in vivo studies of [ 18 F]2 did not show favorable brain pharmacokinetics in normal mice. Although the polar radioactive products formed from [ 18 F]2 need to be identified, the results of this study would serve as a starting point for the design of metabolically stable 18 F-labeled difluoroboron-curcumin derivatives for Aβ imaging.

2-(2-(4-Formyl-2-methoxyphenoxy)ethoxy)ethyl toluenesulfonate (13).
Compound 12 (500 mg, 2.08 mmol) was dissolved in 7 mL dichloromethane, and then p-toluenesulfonyl chloride (476 mg, 2.49 mmol) was added to this solution. After the addition of triethylamine (1.74 mL, 12.48 mmol) at 0 °C, the reaction mixture was stirred at rt overnight. The reaction was then quenched with saturated NH 4 Cl solution, and the reaction mixture was extracted with dichloromethane, washed with water, and then dried over Na 2 SO 4 . Flash column chromatography (1:1 hexane-ethyl acetate) gave 13 (700 mg, 85.3%) as a white solid. 1 Compound 13 (200 mg, 0.50 mmol) was dissolved in 10 mL t-BuOH, and CsF (231 mg, 1.52 mmol) was added to this solution. The reaction mixture was stirred at 100 °C overnight. At the end of reaction, the mixture was extracted with ethyl acetate and water, and the organic layer was washed with water and then dried over Na 2 SO 4 . Flash column chromatography (1:2 hexane-ethyl acetate) gave 14 (70 mg, 56.9%) as a colorless oil. 1 (4 mL) were added to this reaction solution. After stirring at 80 °C for 30 min, n-butylamine (0.66 mL, 6.70 mmol) was added dropwise to the mixture, which was then allowed to stir at 100 °C for 1 h. The reaction mixture was then treated with 0.4 N HCl (10 mL) at 50 °C for 30 min. After the reaction was then quenched with saturated NaHCO 3 solution, the reaction mixture was extracted with ethyl acetate, washed with water, and then dried over Na 2 SO 4 . Flash column chromatography (4:1 hexane-ethyl acetate) gave 15 (300 mg, 19.5%) as a yellow solid. 1  was dissolved in 1.5 mL ethyl acetate, and then (n-BuO) 3 B (59.2 μL,0.2 mmol) and compound 14 (21.4 mg, 0.08 mmol) were sequentially added to this solution. After stirring at rt for 10 min, piperidine (5.2 μL, 0.05 mmol) was added dropwise to the mixture. The reaction mixture was then allowed to stir at 110 °C for 20 min. At the end of reaction, the mixture was extracted with ethyl acetate and water, and the organic layer was washed with water and then dried over Na 2 SO 4 . Flash column chromatography (1:1 hexane-ethyl acetate) gave 3 (18.0 mg, 33.5%) as a red solid. Ligand 3 was further purified by HPLC using a semi-preparative column eluted with a 40:60 mixture of 0.1% TFA (aq.) and acetonitrile at a flow rate of 3 mL/min. 1  www.nature.com/scientificreports www.nature.com/scientificreports/ 4-(2,3-Dihydroxypropoxy)benzaldehyde (17). 4-Hydroxybenzaldehyde (500 mg, 4.09 mmol) was dissolved in 40 mL EtOH, and NaOH (246 mg, 6.14 mmol) in 4 mL water was added to this solution. After the solution was stirred at rt for 10 min, (±)-3-chloro-1,2-propanediol (0.479 mL, 5.73 mmol) was added dropwise. The reaction mixture was stirred at 100 °C overnight. At the end of reaction, the reaction mixture was filtered and the filtrate was extracted with ethyl acetate, washed with water, and then dried over Na 2 SO 4 . Flash column chromatography (1:2 hexane-ethyl acetate) gave 17 (500 mg, 62.3%) as a white solid. 1 Compound 17 (450 mg, 2.30 mmol) was dissolved in 30 mL dichloromethane, and triethylamine (1.92 mL, 13.77 mmol) was added to this solution at 0 °C. After stirring at rt for 10 min, p-toluenesulfonyl chloride (525 mg, 2.75 mmol) was added to the reaction mixture, which was then allowed to stir at rt overnight. The reaction mixture was extracted with dichloromethane and water, and the organic layer was washed with water and then dried over Na 2 SO 4 . Flash column chromatography (1:1 hexane-ethyl acetate) gave 18 (286 mg, 35.5%) as a white solid. 1

4-(3-Fluoro-2-hydroxypropoxy)benzaldehyde
The residue was dissolved in 1.5 mL acetonitrile, and 0.2 mL 1 N HCl was added to the solution. The reaction mixture was stirred at 100 °C for 5 min. After the reaction was quenched with 0.4 mL saturated NaHCO 3 solution, the reaction mixture was extracted with ethyl acetate, washed with water, and then dried over Na 2 SO 4 . Flash column chromatography (1:1 hexane-ethyl acetate) gave 20 (20 mg, 72.1%) as a colorless oil. 1  After stirring at rt for 10 min, piperidine (14 μL, 0.14 mmol) was added dropwise to the mixture, which was then allowed to stir at 110 °C for 20 min. The reaction mixture was extracted with ethyl acetate and water, and the organic layer was washed with water and then dried over Na 2 SO 4 . Flash column chromatography (1:1 hexane-ethyl acetate) gave 4 (8 mg, 21.8%) as a violet solid. Ligand 4 was further purified by HPLC using a semi-preparative column eluted with a 35:65 mixture of 0.1% TFA (aq.) and acetonitrile at a flow rate of 3 mL/min. 1  Materials and equipment for synthesis of radioligand. [ 18 F]Fluoride was produced by the 18 O(p,n) 18 F reaction using a GE Healthcare PETtrace cyclotron (Uppsala, Sweden). Radioactivity was measured in a dose calibrator (Biodex Medical Systems, Shirley, NY, USA). TLC was performed on Merck F 254 silica plates and analyzed on a Bioscan radio-TLC scanner (Washington, D.C., USA). Purification and analysis of the radioligand were performed using HPLC equipped with a semi-preparative column (YMC-Pack C18, 10 × 250 mm, 5 µm) or an analytical column (YMC-Pack C18, 4.6 × 250 mm, 5 µm). The eluent was monitored simultaneously, using UV (254 nm) and NaI(T1) radioactivity detectors. Three azeotropic distillations were then performed using 200-300 μL aliquots of acetonitrile at 90 °C (oil bath) under a gentle stream of N 2 . The resulting n-Bu 4 N[ 18 F]F was then dissolved in acetonitrile (100 μL) and transferred to a reaction vial containing 10 (0.5 mg, 1.4 μmol). This reaction solution was allowed to stir at 110 °C for 10 min. The resulting mixture was extracted with ethyl acetate and water, and the organic layer was dried under a gentle stream of N 2 . The residue was re-dissolved in ethyl acetate (200 μL), which was then transferred to a reaction vial containing 8. The reaction mixture was heated at 110 °C for 20 min after addition of n-butylamine (13.6 μL, 1.4 μmol). The mixture was cooled, diluted with water (2 mL), and extracted with ethyl acetate (2 mL). The organic layer was passed through a 3-cm Na 2 SO 4 plug, and the solvent was removed under a stream of N 2 at (2019) 9:6747 | https://doi.org/10.1038/s41598-019-43257-9 www.nature.com/scientificreports www.nature.com/scientificreports/ Optical imaging. Ligand 2 was dissolved in 10% ethanol-10% Cremophor-saline and injected intravenously through a tail vein into 6-week-old Balb/C nude mice (male, 2 mg/kg), which had been fed an alfalfa-free diet for four days prior to this experiment. Optical images were acquired for 1 sec at 2, 30, and 60 min after injection using a Xenogen IVIS Spectrum (PerkinElmer, Waltham, MA, USA) (excitation, 570 nm; emission, 660 nm). After in vivo imaging, mice were sacrificed, and tissues of interest were excised and subjected to optical imaging (1-sec exposure). Data are expressed as photons/s/cm 2 /sr (sr: steradian).

Radioligand
Metabolism study. [ 18 F]2 (19.6 MBq) dissolved in 10% ethanol-saline was injected into ICR mice via a tail vein. At 2, 30, and 60 min after injection, mice were sacrificed, and samples of brain and blood were obtained. The samples were homogenized in 1 mL of acetonitrile and centrifuged. The supernatants were analyzed by radio-TLC using a 4:1 mixture of ethyl acetate-hexane or a 1:1:0.01 mixture of dichloromethane-methanol-triethylamine as the developing solvents.
After ex vivo optical imaging, the brain tissues were homogenized in 1 mL of acetonitrile and centrifuged. The supernatants were analyzed by TLC using a 4:1 mixture of ethyl acetate-hexane as the developing solvents, and the TLC plates were subjected to optical imaging for 1 sec (excitation, 570 nm; emission, 660 nm).