Synthesis and Biological Evaluation of O-[3-18F-fluoropropyl]-α-methyl Tyrosine in Mesothelioma-Bearing Rodents

Radiolabeled tyrosine analogs enter cancer cells via upregulated amino acid transporter system and have been shown to be superior to 18F-fluoro-2-deoxy-D-glucose (18F-FDG) in differential diagnosis in cancers. In this study, we synthesized O-[3-19F-fluoropropyl]-α-methyl tyrosine (19F-FPAMT) and used manual and automated methods to synthesize O-[3-18F-fluoropropyl]-α-methyl tyrosine (18F-FPAMT) in three steps: nucleophilic substitution, deprotection of butoxycarbonyl, and deesterification. Manual and automated synthesis methods produced 18F-FPAMT with a radiochemical purity >96%. The decay-corrected yield of 18F-FPAMT by manual synthesis was 34% at end-of-synthesis (88 min). The decay-corrected yield of 18F-FPAMT by automated synthesis was 15% at end-of-synthesis (110 min). 18F-FDG and 18F-FPAMT were used for in vitro and in vivo studies to evaluate the feasibility of 18F-FPAMT for imaging rat mesothelioma (IL-45). In vitro studies comparing 18F-FPAMT with 18F-FDG revealed that 18F-FDG had higher uptake than that of 18F-FPAMT, and the uptake ratio of 18F-FPAMT reached the plateau after being incubated for 60 min. Biodistribution studies revealed that the accumulation of 18F-FPAMT in the heart, lungs, thyroid, spleen, and brain was significantly lower than that of 18F-FDG. There was poor bone uptake in 18F-FPAMT for up to 3 hrs suggesting its in vivo stability. The imaging studies showed good visualization of tumors with 18F-FPAMT. Together, these results suggest that 18F-FPAMT can be successfully synthesized and has great potential in mesothelioma imaging.


Introduction
Numerous studies have demonstrated that growing cancer cells have higher metabolism of glucose and amino acids than other cells in the body. One well-known modality for imaging the metabolic activity of cancers is positron emission tomography (PET) using 18 F-2-fluoro-2-deoxy-Dglucose ( 18 F-FDG), the current gold standard for cancer diagnosis [1]. However, 18 F-FDG has limitations such as poor differentiation between low-grade tumor and normal tissues in brain [2] and between tumor and inflamed or infected tissues [3]. Radiolabeled amino acids offer higher specificity in characterizing tumors than 18 F-FDG does. In particular, radiolabeled aromatic amino acids are attractive alternatives to 18 F-FDG because of easier chemistry alteration and their ability of detection of upregulated amino acid transporters [4], which indirectly reveal cell proliferation. Therefore, 11 C-and 18 F-labeled amino acid analogs were developed as alternative metabolic imaging tracers for PET. 11 C-methyl methionine ( 11 C-MET) and L-1-11 C tyrosine ( 11 C-TYR) have been commonly used for clinical research and practices. Unfortunately, the half-life of 11 C is only 20 min, and, therefore, 11 C-labeled amino acid analogs require an inconvenient on-site synthesis which reduces their broad clinical usages. 18 F has a half-life of 110 min, and it can be used at a centralized remote facility to synthesize radiolabeled compounds which can then be delivered to different hospitals simultaneously. Moreover, low + -energy of 18 F causes a short positron linear range in tissue, thereby providing high resolution in PET images. A number of 18 F-labeled amino acid analogs in PET have been investigated, including L-2-18 F-fluorotyrosine ( 18 F-TYR) [5], O-2-18 F-fluoroethyl-Ltyrosine ( 18 F-FET) [6], and L-3-18 F-fluoro--methyl tyrosine ( 18 F-FAMT). Recently, Wiriyasermkul et al. found that, unlike 18 F-TYR, 18 F-FET, and other 18 F-labeled amino acids, 18 F-FAMT is transported into cells through L-type amino transporter 1, which contributes to its highly tumor-specific accumulation [4]. 18 F-FAMT was first studied as a brainimaging probe [7]; later, its use in detecting oral squamous cell carcinoma [8], nonsmall cell lung cancer [9], and esophageal squamous cell carcinoma [10] was investigated. However, the yield of 18 F-labeled amino acids by an electrophilic fluorination reaction is low (17% for 18 F-TYR [5]; 20% ± 5.1% for 18 F-FAMT [11]). Wester et al. synthesized O-2-18 F-fluoroethyl-L-tyrosine ( 18 F-FET) by a nucleophilic fluorination reaction in about 50 min with an overall radiochemical yield of 40% and evaluated it as a PET tracer for cerebral and peripheral tumors [6]. Hamacher and Coenen synthesized 18 F-FET using one-pot reaction, and the radiochemical yield obtained within 80 min was about 60% [12]. However, both methods require high-performance liquid chromatography (HPLC) for purification, which limits the possibility of automated synthesis. Wang et al. obtained 18 F-FET by direct nucleophilic fluorination reaction of the protected precursor N-butoxycarbonyl-(O-(2-tosyloxyethyl))-L-tyrosine methyl ester, followed by a rapid removal of the protecting group, and a labeled intermediate was separated out with Sep-Pak silica plus cartridge [13]. The radiochemical yield was about 40% at the end of synthesis (50 min). Bourdier et al. used this method for automated radiosynthesis of 18 F-FET, and the yield was about 35% within 63 min [14]. 18 F-FET was widely used in clinical studies in patients with high-grade or lowgrade glioma [15,16]. Despite the very promising clinical results of 18 F-FAMT, existing methods for synthesizing 18 F-FAMT produce a low chemical yield, which limits the availability of the compound for clinical use, and they require high-performance liquid chromatography (HPLC) for purification, which precludes the use of an automated module to synthesize 18 F-FAMT. Therefore, it is desirable to develop an 18 F-FAMT analog with high chemical yield that can be applied clinically in most major medical facilities. In the present study, we synthesized unlabeled O-[3-19 F-fluoropropyl]--methyl tyrosine ( 19 F-FPAMT) and 18 F-labeled O-[3-18 F-fluoropropyl]--methyl tyrosine ( 18 F-FPAMT) by using nucleophilic substitution to place a fluorine atom on the aliphatic chain of -methyl tyrosine and solid-phase extraction (SPE) column to purify the products. We then used our customized, fully automated synthesis module to synthesize 18 F-FPAMT. Finally, we used a rat mesothelioma model to investigate the feasibility of using 18 F-FPAMT as a tumor-seeking imaging agent.

Materials and Methods
2.1. General. All chemicals and solvents were obtained from Sigma-Aldrich (St. Louis, MO, USA). Nuclear magnetic resonance (NMR) spectra were obtained using a Bruker 300 MHz Spectrometer (Bruker BioSpin Corporation, Billerica, MA, USA), and mass spectra were recorded on a Waters Q-TOF Ultima mass spectrometer (Waters, Milford, MA, USA) at the Chemistry Core Facility at The University of Texas MD Anderson Cancer Center (Houston, TX, USA). An HPLC system (Waters) was integrated with an ultraviolet detector and a flow-count radio-HPLC detector (BioScan Inc., Washington, DC, USA). The analyses of radio-thin layer chromatography (TLC) were performed on radio-TLC Imaging Scanner (BioScan, Inc.). The scintigraphic imaging studies were processed on microPET (Siemens Medical Systems, Inc., Malvern, PA, USA).

Synthesis of N-t-butoxycarbonyl-O-[3-tosylpropyl]-methyl Tyrosine Ethyl Ester. N-t-butoxycarbonyl-O-[3-
hydroxypropyl]--methyl tyrosine ethyl ester, which we used as the precursor compound for synthesis of 19 F-FPAMT and 18 F-FPAMT, was prepared as described previously [17]. Briefly, N-t-butoxycarbonyl-O-[3-hydroxypropyl]--methyl tyrosine ethyl ester (490 mg; 1.28 mmol) in anhydrous pyridine (32 mL) was cooled to 0 ∘ C. Paratoluenesulfonyl chloride (1015 mg; 5.32 mmol) was added to this solution, and the solution was stirred for 30 min. The reaction mixture was then stored in a refrigerator overnight. The mixture was filtered, and the filtrate was poured into an ice and water mixture and extracted with diethyl ether. The ethereal solvent was washed with 30 mL of hydrochloric acid and water (1 : 1, v/v) to remove pyridine, and the solvent was dried over anhydrous MgSO 4 . After filtration and solvent evaporation, N-t-butoxycarbonyl-O-[3-tosylpropyl]--methyl tyrosine ethyl ester was purified by column chromatography using a silica gel column and eluted with hexane and ethyl acetate (2 : 1, v/v) to yield 430 mg (62.5%). NMR and mass spectrometry were performed to confirm the structures. 19 F-FPAMT. We used a three-step procedure to synthesize 19 F-FPAMT ( Figure 1). The first step was a displacement reaction. Kryptofix 222 (253.9 mg; 0.67 mmol) and K 19 F (40.5 mg; 0.69 mmol) were added to a vial containing Nt-butoxycarbonyl-O-[3-tosylpropyl]--methyl tyrosine ethyl ester (compound 1; 390 mg; 0.75 mmol) in acetonitrile (1 mL). The reaction vial was heated under reflux at 90 ∘ C for 40 min. After heating, the solution was evaporated to dryness. The mixture was reconstituted in 0.5 mL of ethyl acetate. Nt-butoxycarbonyl-O-[3-19 F-fluoropropyl]--methyl tyrosine ethyl ester (compound 2) was purified by column chromatography using a silica gel column and eluted with hexane and ethyl acetate (4 : 1, v/v) to yield 120.0 mg of the compound.  Step 1

NaOH
Step 3 the mixture was heated at 90 ∘ C for 15 min to remove ethyl ester group. The mixture was passed through a 0.22 M filter to yield 19 F-FPAMT (compound 4). NMR and mass spectrometry were used to confirm the structure of this compound. . The reaction mixture was heated at 90 ∘ C for 15 min to allow the displacement to occur. After the reaction mixture cooled, it was passed through a 500 mg silica gel packed SPE column (Whatman Lab., Clifton, NJ, USA) and eluted with acetonitrile (2 mL). The acetonitrile was then evaporated in vacuo at 85 ∘ C. The resulting mixture was hydrolyzed with trifluoroacetate (0.2 mL) in dichloromethane (0.2 mL) at room temperature for 10 min to deprotect BOC. After the solvent was evaporated to dryness in vacuo, sodium hydroxide (1 N; 0.2 mL) in methanol (0.2 mL) was added and heated at 90 ∘ C for 15 min to remove ethyl ester group. After methanol evaporated, hydrochloric acid (0.1 N; 0.2 mL) was used to adjust the pH of the final product to 6.5. Radio-TLC and HPLC were performed to assure the purity and identity of the product.   For the nucleophilic substitution, [ 18 F]fluoride-kryptofix complex (0.2 mL; 29.36 mCi) was manually injected into the RV1 through the injection hole, and additional acetonitrile (0.35 mL) was manually injected into the RV1 to flush the residual [ 18 F]fluoride-kryptofix complex inside the flow channel. Following this step, the infrared (IR) heater automatically heated the RV1 at 90 ∘ C for 15 min. For free fluoride separation, the mixture in the RV1 was automatically passed through a silica gel packed column (SPE 500 mg; Whatman Lab., Clifton, NJ, USA) to the reaction vial 2 (RV2) via nitrogen flow. Additional acetonitrile (2.0 mL) was then added to RV1, and the residual mixture was filtered through a SPE column to remove the free fluoride. The solution inside RV2 was evaporated in vacuo at 90 ∘ C for 15 min before deprotection of BOC was performed. Trifluoroacetate in dichloromethane (0.4 mL) was loaded into RV2, and the solution was set under room temperature for 10 min to allow the reaction to finish. The solvent was then evaporated to dryness in vacuo for 15 min. For deesterification, sodium hydroxide in methanol (0.6 mL) was loaded into RV2. The reaction mixture in RV2 was heated at 90 ∘ C for 15 min. Once deesterification was completed, the solvent in RV2 was evaporated in vacuo, and the radioactivities of the solvent in the column, RV1, and RV2 were measured upon the completion of 18 F-FPAMT. Radio-TLC and HPLC were performed to assure the purity and identity of the final product.

Manual Radiosynthesis of
2.6. In Vitro Cellular Uptake Studies. Rat mesothelioma IL-45 cells were maintained in the mixtures of Dulbecco's modification of Eagle's medium, F-12 (GIBCO, Grand Island, NY, USA), and 10% phosphate-buffered saline at 37 ∘ C in a humidified atmosphere containing 5% CO 2 . Cells were plated onto 6-well tissue culture plates (2 × 10 5 cells/well) and incubated with 18 F-FPAMT (8 Ci/well) or 18 F-FDG (Cyclotope, Houston, TX, USA; 8 Ci/well) for 0-2 h. After incubation, the cells were collected, and their radioactivity was measured using a gamma counter. Data were expressed as the mean percent ± the standard deviation of the cellular uptake of 18 F-FPAMT or 18 F-FDG. Nine rats were inoculated with mesothelioma IL-45 cells (1 × 10 5 cells/rat) at the hinged leg. Twelve days after being inoculated with the mesothelioma cells, the rats were anesthetized with ketamine (10-15 mg/rat). 18 F-FPAMT dissolved in saline (0.5 mCi/5 mL) was injected intravenously into 9 rats ( = 3 rats/group, 30 Ci/rat,). For comparison, the clinical BioMed Research International 5 standard, 18 F-FDG (Cyclotope,), was injected intravenously into 9 rats ( = 3 rats/group; 30 Ci/rat). The distribution of 18 F-FPAMT or 18 F-FDG in various tissues was assessed at 30 min, 1.5 hrs, and 3 hrs after injection by COBRA. Percent of injected dose per tissue type was then calculated, and the data were expressed as the mean percent ± the standard deviation of the injected dose. 18 F-FPAMT and 18 F-FDG. Dosimetric calculations were performed from 30 to 180 min after the administration of 18 F-FPAMT and 18 F-FDG, and time-activity curves were generated for each organ. Analytic integration of the curves was used to determine the area under the curve (AUC), which was divided by the injected dose to yield the residence times of 18 F-FPAMT and 18 F-FDG in each organ. Residence times were then used to calculate target organ absorbed radiation doses based on the medical internal radiation dosimetry methodology for the normal adult male using the Olinda software package (Oak Ridge, TN, USA).

PET Imaging of Mesothelioma-Bearing Rats.
Mesothelioma-bearing rats cells were imaged when their tumors were 1-2 cm in diameter. The rats were anesthetized with 2% isoflurane and administered with 500 Ci of 18 F-FDG or 500 Ci of 18 F-FPAMT. Four serial 15-minute transaxial PET images of each rat were obtained using microPET (Siemens Medical Systems, Inc., IL, USA). respectively. The decay-corrected yield for hydrolysis (deprotection of BOC and deesterification) was 89%. At the end-of-synthesis (88 min), 10 mCi of 18 F-FPAMT was obtained, and the decay-corrected yield was 34%. The specific activity of this compound was 0.32 Ci/ mol. For the automated synthesis of 18 F-FPAMT, the decay-corrected yield was 15%, the end-of-synthesis time was 110 min, and the specific activity was 0.16 Ci/ mol.

In Vitro Cellular Uptake
Studies. The uptake of 18 F-FPAMT reached saturation at 60 min ( Figure 3). 18 F-FDG uptake continued to increase throughout the period, and the percentage uptake of 18 F-FDG was higher than that of 18 F-FPAMT at each time point.  Tables 1 and 2, respectively. Both compounds showed no marked increase in bone uptake, representing their in vivo stability. High kidney and pancreas uptake of 18 F-FPAMT was observed, and this phenomenon was also observed from other tyrosine-based radiotracers [18]. Unlike 18 F-FDG, 18 F-FPAMT had poor uptake in brain tissue.

Dosimetry of 18 F-FPAMT in Rats.
The estimated absorbed radiation dose of 18 F-FPAMT is shown in Table 3. According to the US Food and Drug Administration Regulations, human exposure to radiation from the use of "radioactive research drugs" should be limited to 3 rem per single administration and 3 rem per year to the whole body, blood-forming organs (red marrow, osteogenic cells, and spleen), the lens of the eye, and gonads (testes and uterus); the limit for other organs is 5 rem per single administration and 15 rem annually. The total rem of 18 F-FPAMT absorbed by each organ was below these limits at the proposed injection of 30 mCi per patient.

Imaging of Mesothelioma-Bearing Rats.
Scintigraphic images of mesothelioma-bearing rats administrated 18 F-FPAMT or 18 F-FDG showed that tumors could be clearly detected, and bone uptake was low (Figure 4). The standardized uptake value (SUV) curve of 18 F-FPAMT for tumor and muscle reached the plateau at 30 min after injection, but the SUV curve of 18 F-FDG for tumor continued increasing during the imaging. The SUV ratios of tumor to muscle for 18 F-FPAMT and 18 F-FDG were 2.82 and 8.26, respectively. There was extremely low uptake of 18 F-FPAMT in the brain and spinal cord when compared with 18 F-FDG ( Figure 5).

Discussion
Mesothelioma is an asbestos-related neoplasm generating from mesothelial cells in the pleural, peritoneal, and pericardial cavities, and its incidence increased in several countries [19]. The diagnostic tools and treatment regimens for these tumors are disappointing, and median survival time is 12 months after initial diagnosis [20]. The initial diagnoses of mesothelioma are based on patient's medical history and physical examination. After that, computed tomography scans and magnetic resonance imaging are used to screen patients, and then biopsy test is needed to confirm the incidence of mesothelioma. 18 F-FDG/PET scan is the tool to determine whether a suspicious area is malignant mesothelioma or a benign condition such as pleural scarring, and the result can identify the best area for an accurate biopsy. PET scans are also effective for highlighting mesothelioma metastases that may not appear on other conventional imaging scans. However, 18 F-FDG/PET scans have limitations in differential diagnosis between cancerous cells and inflammation tissues which metabolize glucose with abnormally high rates. In this case, radiolabeled amino acids are the alternative methods to detect malignant pleural mesothelial and other cancerous cells which overexpress unregulated amino acid transporters [21][22][23]. Mesothelioma rat model was then selected because rat model provided better anatomical differentiation than mouse model in imaging studies. It is more accurate to determine radiation dosimetry from biodistribution data. 18 F-FET and 18 F-FAMT are radiolabeled amino acids, and they are useful in imaging cancers. However, existing 8 BioMed Research International methods for synthesizing these compounds result in low yields, thus limiting the availability of 18 F-FET and 18 F-FAMT in the clinic. In the present study, we synthesized 18 F-FPAMT, an 18 F-FAMT analog, and used a mesothelioma rat model to preliminarily evaluate it as a tumor-imaging compound. We used NMR and mass spectrometry to confirm the structure of 19 F-FPAMT. The yield of 19 F-FPAMT was 46.71%. Nt-butoxycarbonyl-O-[3-tosylpropyl]--methyl tyrosine ethyl ester was used as the starting material for manual and automated syntheses of 18 F-FPAMT. The quality control of 18 F-FPAMT was evaluated by radio-TLC and HPLC. Manual synthesis of 18 F-FPAMT resulted in the decay-corrected yield of 34%, radiochemical purity of >95%, the specific activity of 0.32 Ci/ mol, and pH value of 5 to 6; the manual synthesis time was 88 min. Automated synthesis of 18 F-FPAMT resulted in the decay-corrected yield of 15%, radiochemical purity of >95%, the specific activity of 0.16 Ci/ mol, and pH value of 5 to 6; the manual synthesis time was 110 min.
The traditional method of radiosynthesizing 18 F-labeled tyrosine analogs such as 18 F-FET and 18 F-FAMT was through electrophilic substitution reaction which has low synthetic yield. Besides, the reaction uses 18 F-F 2 gas, and HPLC separation makes it even difficult to use this method in automated modules. Although a nucleophilic reaction could result in a high yield of 18 F-FET (40%), this method still requires HPLC for purification, and, thus, it is not ideal to use this synthesis method in automated synthesis modules. In the present study, we obtained 18 F-FPAMT by a nucleophilic reaction, but we completed the purification process without HPLC. Therefore, our method of synthesizing 18 F-FPAMT can be applied to the customized automated synthesis module.
For the in vitro studies, although the result showed that 18 F-FPAMT had lower cellular uptake than that of 18 F-FDG, the uptake mechanism of these two compounds is different. Malignant cells utilize 18 F-FDG as glucose for upregulated aerobic glycolysis and 18 F-FPAMT as an amino acid for proliferation. The results indicate that 18 F-FPAMT has the potential to become a tumor detecting tracer. Biodistribution studies showed that 18 F-FPAMT and 18 F-FDG were rapidly cleared from blood and distributed in other tissues. Compared with 18 F-FDG, the accumulation of 18 F-FPAMT was significantly lower in heart, lungs, thyroid, spleen, and brain. High accumulation of 18 F-FPAMT was observed in the kidneys and pancreas after administration. This could be due to the high expression of the amino acid transporters in the kidneys and pancreas [4]. These results were consistent with those of other radiolabeled amino acid analogs such as 18 F-FAMT [24] and 77 Br-BAMT [18], although 18 F-FET showed only higher uptake in kidneys [25]. The bone uptakes of 18 F-FPAMT and 18 F-FDG at 180 min after administration increased slightly, suggesting defluorination of both compounds. In the microPET studies of 18 F-FDG and 18 F-FPAMT, the lesions could be observed clearly at 45 min after administration (Figure 4), and the accumulation of 18 F-FPAMT in the brain and spinal cord was significantly less than that of 18 F-FDG ( Figure 5), suggesting that 18 F-FPAMT has great potential in imaging brain tumors.

Conclusion
In this study, we manually synthesized 18 F-FPAMT with high yielding and radiochemical purity, and we used the customized automated synthesizer for the proof of concept of automated manufacturing of 18 F-FPAMT. Both in vitro and in vivo studies suggested that 18 F-FPAMT can be a good PET agent for detecting mesothelioma, and it might have great potential in brain tumor imaging. In the future, we will focus on optimization of the automated processes for a better yield and a higher specific activity.