Synthesis and Preclinical Characterization of [18F]FPBZA: A Novel PET Probe for Melanoma

Introduction. Benzamide can specifically bind to melanoma cells. A 18F-labeled benzamide derivative, [18F]N-(2-diethylaminoethyl)-4-[2-(2-(2-fluoroethoxy) ethoxy)ethoxy]benzamide ([18F]FPBZA), was developed as a promising PET probe for primary and metastatic melanoma. Methods. [18F]FPBZA was synthesized via a one-step radiofluorination in this study. The specific uptake of [18F]FPBZA was studied in B16F0 melanoma cells, A375 amelanotic melanoma cells, and NB-DNJ-pretreated B16F0 melanoma cells. The biological characterization of [18F]FPBZA was performed on mice bearing B16F0 melanoma, A375 amelanotic melanoma, or inflammation lesion. Results. [18F]FPBZA can be prepared efficiently with a yield of 40–50%. The uptake of [18F]FPBZA by B16F0 melanoma cells was significantly higher than those by A375 tumor cells and NB-DNJ-pretreated B16F0 melanoma cells. B16F0 melanoma displayed prominent uptake of [18F]FPBZA at 2 h (7.81 ± 0.82 %ID/g), compared with A375 tumor and inflammation lesion (3.00 ± 0.71 and 1.67 ± 0.56 %ID/g, resp.). [18F]FPBZA microPET scan clearly delineated B16F0 melanoma but not A375 tumor and inflammation lesion. In mice bearing pulmonary metastases, the lung radioactivity reached 4.77 ± 0.36 %ID/g at 2 h (versus 1.16 ± 0.23 %ID/g in normal mice). Conclusions. Our results suggested that [18F]FPBZA PET would provide a promising and specific approach for the detection of primary and metastatic melanoma lesions.


Introduction
Melanoma, with a steadily increasing incidence rate for the past 30 years, has become a serious public health problem worldwide [1,2]. In the United States, the incidence rose from 7.5 cases per 100,000 inhabitants in 1973 to 21.9 cases per 100,000 inhabitants in 2002 [3].
Malignant melanoma is very aggressive and more likely than other skin cancers to spread to other parts of the body, including brain, lungs, liver, and lymph nodes [4]. The prognosis and the outcome of melanoma treatment heavily depend on the clinical stage of the disease at the time of diagnosis. Melanoma is highly curable if detected in its earliest stages and treated properly [4][5][6]. Early stage melanoma can be cured by surgical excision and the 5-year relative survival rates for persons with melanoma are >90% [1,5]. However, advanced melanoma with distant tumor spread is largely refractory to existing therapies and has a poor prognosis with 5-year survival rate of 15% [1,5,7,8]. Thus, early diagnosis and accurate disease staging are required for optimizing the treatment plan and can avoid unnecessary and potentially harmful surgical therapies that would yield no improvement in survival.
Positron emission tomography (PET) imaging technique offers unique advantages including high sensitivity as well as deep tissue tomographic and repetitive imaging capabilities.  [9,10]. However, it was reported that the sensitivity of [ 18 F]FDG PET was unacceptably low for the initial assessment of early-stage malignant melanoma [11,12]. [ 18 F]FDG PET scans also failed to detect the pulmonary and brain metastases and should be limited to staging patients with more advanced melanoma [13,14]. Moreover, [ 18 F]FDG displays relatively poor selectivity for distinguishing tumor from inflammatory tissue. Increased glucose metabolism in inflammatory tissues is often the main cause of false-positive [ 18 F]FDG PET findings in oncology.
Primary melanoma most commonly arises from the skin; however, it can also develop in other areas, including the eye and mucosa. A key feature of melanoma is the extensive melanin present in most tumor cells, making it a very attractive target for both diagnosis and treatment. Various radiolabeled benzamide analogues have been developed for targeting melanin with high selectivity and affinity [15][16][17][18]. Among these derivatives, N-(2-diethylaminoethyl)-4-[ 18 F]fluorobenzamide ([ 18 F]FBZA) exhibited high tumor uptake and was considered as a promising candidate for clinical study [17,18]. The synthesis of [ 18 F]FBZA is conducted via conjugation of a 18 F-synthon (N-succinimidyl-4-[ 18 F]fluorobenzoate, [ 18 F]SFB) to the primary amino group of N,N-diethylethylenediamine (DEDA). This method requires a multistep synthesis and is time-consuming and laborious. Furthermore, it is difficult to make the multistep radiosynthesis fully automatic, which in turn sets a high technical barrier for [ 18 F]FBZA in the clinical use.
One of the commonly used methods to append a F-18 atom involves conjugating a fluoroalkyl group to the compound; however, the results were not always promising. The 18 F-labeled tracers with an additional fluoroalkyl moiety have higher lipophilicity and tend to display a higher nonspecific binding to the normal organs. Polyethylene glycol (PEG) is a component that has been widely used as a modifier of drugs and has been approved by the food and drug administration of United States of America. This study developed an easyto-synthesize method to prepare a [ 18 F]fluoroPEGylated radiotracer for melanoma detection. We modified the phenol moiety of benzamide with a short chain PEG and then labeled with 18 F to afford [ 18 Figure 1). This study presented a preclinical assessment of [ 18 F]FPBZA as a PET probe for specific imaging of melanin. To achieve this purpose, biological characterizations of [ 18 F]FPBZA were performed in C57BL/6 mice bearing subcutaneous or pulmonary metastatic B16F0 murine melanoma with high melanin expression, in BALB/c nude mice bearing A375 human amelanotic melanoma, and in C57BL/6 mice bearing turpentine-induced aseptic inflammation.

Reagents and Instruments.
All reagents and solvents were purchased from commercial suppliers and were used without further purification. Triethylene glycol di-p-tosylate was synthesized following the procedures described in literature [19]. Thin layer chromatography (TLC) was performed on silica gel F 254 aluminum-backed plates (Merck, Darmstadt, Germany) with visualization under UV (254 nm). NMR spectra were recorded on a NMR spectrometer (Bruker, Germany) operating at 400 MHz for 1 H NMR spectra and 100 MHz for 13 C NMR spectra at Instrumentation Resource Center of National Yang-Ming University. All chemical shift values were reported in ppm ( ). Electrospray ionization-mass spectra (ESI-MS) were acquired on a FINNIGAN LCQ mass spectrometer at Instrumentation Resource Center of National Taiwan University. Analytic as well as semipreparative high performance liquid chromatography (HPLC) was performed with a Waters 600E pump equipped with a Waters 2998 photodiode array detector and a flow count radio detector (Bioscan, Washington DC) for gamma ray detection. Radioactivity was assayed using a Capintec CRC-15R dose calibrator (Ramsey, NJ) or a -scintillation counter (Wallac 1470 Wizard automatic gamma counter, Perkin-Elmer, Waltham, MA).  (1). A solution of 4-hydroxybenzoic acid (500 mg, 3.6 mmol) in thionyl chloride (5 mL) was refluxed for 5 h. Excess thionyl chloride was removed under reduced pressure. The residue was then redissolved in THF (10 mL). To this solution was added DEDA (462 mg, 4.0 mmol) and K 2 CO 3 (1.0 g, 7.2 mmol). The reaction mixture was stirred at ambient temperature overnight, diluted with H 2 O and extracted with dichloromethane. The combined organic layers were washed with H 2 O, dried over MgSO 4 , and evaporated. Column chromatography of the crude product on silica gel eluting with MeOH/CH 2 Cl 2 (1/10) afforded compound 1 (790 mg, 3.34 mmol, 93% yield) as a yellowish solid. 1

Synthesis of N-(2-Diethylaminoethyl)-4-[2-(2-(2-fluoroethoxy)ethoxy)ethoxy]-benzamide (FPBZA).
To a solution of compound 2 (104 mg, 0.2 mmol) in THF (4 mL) was added tetrabutylammonium fluoride (TBAF) solution in THF (1 M, 0.4 mL). The resulting mixture was refluxed for 12 h. The solvent was evaporated and the crude product was purified by column chromatography on silica gel eluting with MeOH/CH 2 Cl 2 (1/10) afforded FPBZA (compound 3, 48 mg, 0.13 mmol, and 65% yield) as an yellowish oil. 1   solution in the reaction vessel was heated at 100 ∘ C under reduced pressure. The azeotropic drying was repeated twice by further addition of anhydrous MeCN (2 × 0.8 mL). To the dry residue was added compound 2 (2.0 mg) in anhydrous dimethyl sulfoxide (DMSO) and the resulting mixture was heated at 100 ∘ C for 5 min. After cooling to room temperature, the reaction mixture was diluted with HPLC mobile phase and filtered with a 0.45 m membrane filter (Millipore, Bedford, MA). The crude product was injected into a C18 semipreparative HPLC column (Merck Purospher STAR RP-18e, 10 × 250 mm). The mobile phase started from 95% solvent A (0.1% trifluoroacetic acid in water) and 5% solvent B (0.1% trifluoroacetic acid in MeCN) and then ramped to 35% solvent A and 65% solvent B at 20 min at a flow rate of 4 mL/min. The desired fractions were combined, evaporated under reduced pressure, redissolved in normal saline, and filtered through a 0.22 m aseptic membrane filter (Millipore, Bedford, MA) to afford the final product. Radiochemical purity of the final product was analyzed by analytic HPLC using a Phenomenex Luna C18 column (4 × 250 mm) with a mobile phase consisting of MeCN/10 mM ammonium formate buffer (21/79) at a flow rate of 1 mL/min.

Serum Stability.
[ 18 F]FPBZA (7.4 MBq) was incubated in 0.5 mL of human serum. At designated time points (30,60, and 120 min), 75 L of the sample was taken and mixed with MeCN (250 L). The resulting mixture was vortexed intermittently for 1 minute, then followed by centrifugation at 1,000 g for 10 min to pellet the precipitated serum proteins. The radioactivity of the supernatant and the precipitate was measured with a -scintillation counter. The radioactive components in the supernatant were assayed on an analytic HPLC as described above.

Lipophilicity.
[ 18 F]FPBZA (74 kBq) was added to a mixture of 1-octanol (1 mL) and phosphate buffer saline (PBS; pH 7.4, 1 mL); then, the mixture was vigorously vortexed for 5 min. After subsequent centrifugation at 3,000 rpm for 5 min, aliquots (100 L) were taken from each phase and the radioactivity was determined with a -scintillation counter. The partition coefficient was expressed as log = log 10 (counts in 1-octanol/counts in PBS).

Cellular Uptake Study. B16F0 melanoma cells and A375
amelanotic melanoma cells were obtained from Bioresource Collection and Research Center (Taiwa,) and were cultured in Dulbecco's modified Eagle high-glucose medium (Gibco, Carlsbad, CA) containing 10% fetal bovine serum at 37 ∘ C in a humidified atmosphere of 5% CO 2 .
For cellular uptake assay, B16F0 cells and A375 cells were trypsinized and grown overnight in 6-well culture plates (5 × 10 5 cells/1.0 mL/well), and the medium was changed before experiment. [ 18 F]FPBZA (37 kBq/1.0 mL/well) was added to each well and incubated at 37 ∘ C for 5, 15, 30, 60, and 120 min. Triplicates were carried out for each time point. One milliliter of ice-cold PBS was used to intercept the uptake of tracer. The supernatants were aspirated and the cells rinsed twice with 1 mL of ice-cold PBS. Then, cells in each well were harvested with 0.5 mL of trypsin-EDTA and washed twice with 1 mL of ice-cold PBS. A 10 L of cell suspension was taken to access cell viability with trypan blue and to count the number of viable cells under the microscopy. The radioactivity of the cell suspensions was measured with a -scintillation counter and normalized to the number of viable cells. Cellular uptake of [ 18 F]FPBZA was expressed as the percentage of the administered dose that had accumulated per million cells (%AD/10 6 cells).

Effect of Melanin Content on Cellular Uptake of
To further demonstrate the specific binding of [ 18 F]FPBZA to cellular melanin, an in vitro study using B16F0 cells, with and without pretreatment of Nbutyldeoxynojirimycin (NB-DNJ, a tyrosinase inhibitior that can retard melanin synthesis), was conducted [20]. Briefly, B16F0 cells were seeded into 6-well plates at a density of 2 × 10 5 cells per well and incubated overnight at 37 ∘ C. After cells had adhered, the growth medium in each well was replaced with fresh medium containing NB-DNJ (0, 0.5 and 1 mM) and replenished every 24 h thereafter. After 48 h of treatment, cells were washed with PBS and [ 18 F]FPBZA (37 kBq/1.0 mL/well) was added to each well and incubated at 37 ∘ C for 120 min. One milliliter of ice-cold PBS was used to intercept the uptake of tracer. The supernatants were removed and the cells were rinsed twice with 1 mL of ice-cold PBS. The cells in each well were harvested with 0.5 mL of trypsin-EDTA and washed twice with 1 mL of ice-cold PBS. The radioactivity of the cell suspensions was measured with a -scintillation counter and normalized to the number of viable cells. The cellular uptake of [ 18 F]FPBZA was expressed as %AD/10 6 cells.

Animal Model.
The animal experiments were approved by the Institutional Animal Care and Use Committee of the National Yang-Ming University. For the subcutaneous tumor models, male C57BL/6 mice were injected in the right flank with 0.1 mL of B16F0 melanoma cell suspension of 5 × 10 6 cells/mL, and BALB/c nude mice were inoculated in the right flank with 2 × 10 6 A375 amelanotic melanoma cells. Biological studies were conducted when the tumor burden reached 50∼ 150 mm 3 . For the pulmonary metastasis animal model, mice were injected via the tail vein with 2 × 10 5 B16F0 melanoma cells in 0.2 mL of PBS as previously described [21]. Animal studies were performed 15 days following tumor cell inoculation intravenously. The inflammation was induced by intramuscular injection of 0.1 mL of turpentine into the right flank of male C57BL/6 mice at 4 days before the biological experiment was conducted [22].

Biodistribution Studies.
Mice bearing tumor or inflammation lesion were injected with 3.7 MBq of [ 18 F]FPBZA in 0.1 mL of normal saline through lateral tail veins. Five mice were sacrificed with CO 2 at each designated time point following injection. Blood samples were obtained by cardiac puncture. Tumor, inflammation lesion, and major organs were excised, and parts of these tissues/organs were weighed and assayed for radioactivity with a -scintillation counter. The results were presented as percentage injected dose per gram of tissue (%ID/g). Values were expressed as mean ± SD for a group of five animals.  (30,60, and 120 min after injection), with the long axis of the animal parallel to the long axis of the scanner. All images were reconstructed with the OSEM method, with a 128 × 128 pixel image matrix, 16 subsets, 4 iterations, and use of a Gaussian filter. For data quantitative analysis, a region of interest (ROI) was placed on each tumor, inflammatory lesion and normal organs. The average radioactivity was obtained from the average pixel value within the multiple ROI volume. AsiPro software (Concorde Microsystems) was used for viewing microPET images and for data analysis. The counts in each ROI were converted to radioactivity per gram of tissue (nCi/g), assuming a tissue density of 1 g/mL and were then normalized to percentage of injected dose per gram of tissue (%ID/g).

Statistical Analysis.
Results were expressed as mean ± SD. Statistical analysis was performed using Student's -test for unpaired data. A 95% confidence level was chosen to determine the significance of differences between groups, with a value of less than 0.05 indicating a significant difference.

Chemistry and Radiochemistry.
The preparation of all the intermediates and final product can be achieved in a straightforward manner (Figure 1). Compound 1 was prepared from 4-hydroxybenzoic acid by the treatment of thionyl chloride in THF, followed by addition of DEDA to provide >90% yield of the desired product. Compound 2, the tosyl precursor for radiofluorination, was prepared successfully by coupling the free phenolic hydroxyl groups of compound 1 with triethylene glycol di-p-tosylate in 51% yield. Reaction of compound 2 with TBAF in THF produced the authentic FPBZA in 65% yield. All compounds were fully characterized by 1 H NMR, 13 C NMR, and mass spectra. The radiofluorination of compound 2 with K[ 18 F]F/K2.2.2 in dry DMSO at 100 ∘ C for 5 min produced [ 18 F]FPBZA with an average radiolabeling efficiency of >90% ( = 5). After purification by semipreparative HPLC, [ 18 F]FPBZA final product was obtained with a radiochemical yield of 40-50%, a radiochemical purity of >97%, and a specific activity of 30-40 GBq/ mol. The synthesis time was ∼40 min from the end of bombardment, including the process of purification and formulation. The identity of [ 18 F]FPBZA was confirmed by coinjection with the authentic FPBZA ( Figure 2).

Characterization of [ 18 F]FPBZA.
After treating the [ 18 F]FPBZA-added serum with acetonitrile followed by centrifugation, only a small portion of radioactivity was found in the serum protein precipitate and >90% of the radioactive components retained in solution. Intact [ 18 F]FPBZA accounted for >95% of the radioactivity in solution within 2 h of incubation (see Figure S1 in  (Figure 3). Continued exposure led to an appreciably increased cellular uptake up to 2 h. Accumulation of [ 18 F]FPBZA in B16F0 melanoma cells was significantly higher than that in amelanotic A375 cells  Figure 4(a)). The NB-DNJ-pretreated B16F0 cells contained less melanin and showed a significantly reduced pigmentation compared with that of the control (Figure 4(b)).  The tumor-to-muscle ratio was 3.72 at 0.5 h p.i., and significantly increased to 6.60 at 2 h p.i.. The tumor-toliver and tumor-to-blood ratios also reached 4.49 and 3.52 at 2 h p.i., respectively. The bony uptake increased slightly over time and reached 3.21 ± 0.16 %ID/g at 2 h p.i.. The accumulation of [ 18 F]FPBZA in inflammation lesion was low (1.67 ± 0.56 %ID/g) and was only slightly higher than that in muscle (1.28 ± 0.07 %ID/g; > 0.1).

Biodistribution Studies.
In A375 amelanotic melanoma-bearing nude mice, the tumor uptake was low (3.00 ± 0.71 %ID/g) and the tumor-tomuscle ratio was just 2.72 at 2 h p.i.. Although high uptake of [ 18 F]FPBZA was observed in the pigmented eyes of C57BL/6 mice (7.69 ± 1.56 %ID/g), there was a minor uptake in the unpigmented eyes of nude mice (2.91 ± 0.45 %ID/g).

MicroPET and
MicroPET/CT Imaging. [ 18 F]FPBZA microPET imaging of B16F0 melanoma-bearing C57BL/6 mice showed excellent tumor-to-background contrast ( Figure 5(a)). Tumor lesion can be clearly visualized early at 30 min p.i. and exhibited tenacious radioactivity retention up to 2 h p.i., compared with that of muscle, thorax, liver, and other normal tissues. MicroPET imaging of mice bearing A375 human amelanotic melanoma and turpentine-induced inflammation revealed low radioactivity accumulation in tumor and inflammation lesions, which is consistent with those observed in biodistribution studies ( Figure 5(a)). Quantitative analysis of tumor ROIs revealed a significantly different radioactivity accumulation between B16F0 tumor and A375 tumor at 2 h p.i. (7.03 ± 1.05 and 2.91 ± 0.41 %ID/g, resp., < 0.05; Figure 5(b)). Owing to a rapid clearance of [ 18 F]FPBZA from most of normal tissues, the tumor-tomuscle, and tumor-to-liver ratios increased with time and reached 6.53 ± 0.55 and 4.18 ± 0.35 at 2 h p.i., respectively. An appreciable radioactivity accumulation in bladder and feces was noticed at 30 min p.i., indicating that [ 18 F]FPBZA was rapidly excreted through both renal and hepatobiliary routes.

Discussion
Melanoma is an increasingly important public health problem around the world. Highly invasive melanoma is currently the fifth most frequently diagnosed cancer in men and the sixth in women in the United States [1]. The survival rate of melanoma has been steadily improved in decades, which can be greatly attributed to successful detection of melanoma in its early stage. Earlier diagnosis and treatment could promote survival rate and reduce healthcare expenditure. In this study, we sought to develop [ 18 F]FPBZA as a specific PET probe for melanoma and presented a preclinical assessment of [ 18 F]FPBZA in melanoma animal models. DMSO, a highly polar and aprotic solvent, is very suitable for conducting nucleophilic substitution reaction. In our previous study, preparing 9-(4-[ 18 F]-fluoro-3-hydroxymethylbutyl)guanine by radiofluorination of its tosyl precursor in DMSO gave higher radiochemical yield than that in acetonitrile [23]. In this study, conducting nucleophilic 18  F]MEL050 (6.57 %ID/g and 9.4 %ID/g, resp.,) [17,25]. Although the radioactivity in muscle and blood of [ 18 F]FPBZA was higher than those of [ 18 F]FBZA and [ 18 F]MEL050, high tumor-to-muscle and tumor-toblood ratios (6.60 and 3.52 at 2 h p.i., resp.,) suggested favorable pharmacokinetic profile for specific targeting of melanoma. [ 18 F]FPBZA PET would be a sensitive modality for noninvasive imaging of melanoma in vivo. However, the slightly increasing bone radioactivity implied that defluorination of [ 18 F]FPBZA in living subjects seems likely and may interfere the detection of bone metastasis of melanoma.
It has been reported that the density of sigma receptors are high in liver and kidneys [26], while the results of biodistribution study and microPET imaging did not show  [ 18 F]FPBZA PET was also employed to monitor metastatic lesions with an experimental pulmonary metastasis mouse model. [ 18 F]FPBZA microPET scan displayed remarkable radioactivity accumulation in lungs of pulmonary metastases-bearing mice compared with those of normal mice. The tracer uptake in muscle and liver was also significantly lower than that in pulmonary tumor lesions, rendering favorable tumor-to-normal tissues ratios. However, significant [ 18 F]FDG uptake in the heart and the interscapular brown fat could interfere with the delineation of pulmonary metastatic lesions [18,28,29]. The results of our study demonstrated that [ 18 F]FPBZA was a melaninspecific PET probe for imaging melanoma, either a subcutaneous xenograft or pulmonary metastases. Liver was another common site that readily developed distant metastases from the cutaneous malignant melanoma (15-20%) [4]. In comparison with [ 18 F]FBZA, the low liver uptake and favorable tumor-to-liver ratio of [ 18 F]FPBZA observed in subcutaneous melanoma animal models (1.74 %ID/g versus 2.35 %ID and 4.49 versus 2.80) suggested that it may be more suitable for imaging liver metastases of melanoma [17].
A possible explanation for high melanoma uptake of [ 18 F]MEL050 was the positive charge in the nicotinic ring of [ 18 F]MEL050. The binding of the cationic substances to the melanin can be reinforced by ionic interaction [30]. In contrast, the benzene ring of [ 18 F]FBZA or [ 18 F]FPBZA was electrically neutral. Thus, the further improvement of 18 Flabeled benzamide-derivative radiotracers could focus on the optimization of the pharmacokinetic properties by introducing the cationic moiety.
Several 18 F-labeled -melanocyte-stimulating hormone ( -MSH) analogues were developed via the conjugation of [ 18 F]SFB or 4-nitrophenyl 2-[ 18 F]fluoropropionate and proven to be as excellent PET imaging agents for melanoma [31,32]. However, the preparation of these 18 F-labeled -MSH analogues required multiple radiosynthetic steps and significantly longer reaction time and made them less attractive for clinical application [33].