A High-Affinity 64Cu-Labeled Ligand for PET Imaging of Hepsin: Design, Synthesis, and Characterization

Hepsin, a cell surface serine protease, is a potential biomarker for the detection of prostate cancer due to its high expression in prostate cancer but not in normal prostate. This study aimed to develop a radioligand for positron emission tomography (PET) imaging of hepsin. Six leucine–arginine (Leu–Arg) dipeptide derivatives (two diastereomers for each of three ligands) were synthesized and evaluated for their binding affinities and selectivity for hepsin. Based on the binding assay, a natCu-1,4,7,10-tetraazacyclododecane-N,N′,N″,N‴-tetraacetic acid (DOTA)-conjugated ligand (3B) was selected for the development of a PET radioligand. [64Cu]3B was synthesized by labeling the DOTA-conjugated compound 11B with [64Cu]CuCl2 at 80 °C for 20 min. The radioligand was evaluated for prostate cancer cell binding and PET imaging in a prostate tumor mouse model. The results demonstrated that [64Cu]3B exhibited high binding to LNCaP cells, intermediate binding to 22Rv1 cells, and low binding to PC3 cells. PET studies of [64Cu]3B in mice, implanted with 22Rv1 and PC3 cells on each flank, revealed that the radioligand uptake was high and persistent in the 22Rv1 tumors over time, whereas it was low in PC3 tumors. The results of this study suggest that [64Cu]3B is a promising PET radioligand for hepsin imaging.


Introduction
Prostate cancer is the second most common cancer following lung cancer and represents the sixth leading cause of cancer-related death among men worldwide [1]. In the United States, prostate cancer is the most common cancer and the second leading cause of cancer-related death among men [2]. Prostate cancer that is still confined within the organ can be treated by surgery and radiation therapy [3]. However, metastasis of this cancer into other organs may increase the mortality rate [4]. Therefore, early detection of prostate cancer is essential for an improved prognosis after therapy. The serum prostate-specific antigen (PSA) test is widely used for the early screening of prostate cancer [5]. However, the serum PSA levels are not specific for prostate cancer, because PSA is expressed not only in prostate cancer but also in the normal prostate. This makes discrimination between normal prostate or benign prostatic hyperplasia (BPH) and prostate cancer challenging [6]. Therefore, specific molecular biomarkers for prostate cancer are required for early and accurate diagnosis of the disease.
Hepsin is a type II transmembrane serine protease composed of 413 amino acids. Tissue microarray-based expression profiling demonstrated that hepsin expression was upregulated in prostate cancer specimens, but low or no hepsin expression was detected in normal prostate and BPH specimens [7][8][9][10][11]. Additionally, increased expression of hepsin Hepsin is a type II transmembrane serine protease composed of 413 amino acids. Tissue microarray-based expression profiling demonstrated that hepsin expression was upregulated in prostate cancer specimens, but low or no hepsin expression was detected in normal prostate and BPH specimens [7][8][9][10][11]. Additionally, increased expression of hepsin has been observed in metastatic tumors compared to primary tumors [7,11]. Moreover, the crystal structure of soluble human hepsin exhibits that two domains are located extracellularly and that the larger domain contains the active site cleft for the binding of hepsin inhibitors [12]. These characteristics render hepsin a potential biomarker for the diagnosis of prostate cancer.
Over the past few decades, ligands have been specifically developed for molecular imaging of hepsin. Hepsin-binding peptides with high affinity (190 nM) and selectivity were identified using phage display [13]. When analyzed by fluorescence-activated cell sorting, multiple FITC-peptides conjugated to fluorescent nanoparticles (Cy5.5) (IPL-NP) exhibit a markedly increased signal in hepsin-expressing prostate cancer cells compared with the monomeric FITC-peptide. Using fluorescence-mediated tomography, small tumors were detected in mice injected with IPL-NP and its accumulation was 2.8-fold higher in hepsin-positive LNCaP tumors than in hepsin-negative PC3 tumors. In another study, a heterobivalent ligand, targeting both hepsin and prostate-specific membrane antigen (PSMA), was synthesized. Its SulfoCy7-conjugated version was evaluated in vitro and in vivo [14]. The ligand displayed inhibitory activities for hepsin (IC50 = 2.8 μM) and PSMA (IC50 = 28 nM). The near-infrared (NIR) images of mice showed significantly stronger signals in PC3/ML-PSMA-hepsin tumors than in low hepsin expressing PC3/ML tumors [14]. Moreover, stronger signals were observed in PC3/ML-PSMAhepsin tumors compared to those in PC3/ML-PSMA tumors. More recently, BODIPYand SulfoCy7-conjugated Leu-Arg dipeptide derivatives were developed for the NIR imaging of hepsin [15]. Despite their high binding affinities for hepsin (Ki = 21 and 22 nM), the ligands were not evaluated for NIR imaging. As such, small molecule hepsin inhibitors have been developed for optical imaging. However, none have been applied to PET imaging so far.
In the present study, we developed a PET radioligand that was derived from a Leu-Arg dipeptide-based hepsin inhibitor [15]. We synthesized six Leu-Arg dipeptide derivatives (two diastereomers for each of three ligands) (

Chemical Synthesis
Non-radioactive ligands 1-3 were designed and synthesized for the development of a PET imaging agent for hepsin ( Figure 1). Mini-polyethylene glycol linkers were introduced to the Leu-Arg dipeptide derivatives in order to increase the hydrophilicity and
The two diastereomers, A and B isomers, were not identified. Therefore, we designated the peaks eluted at the earlier and later retention times on HPLC as A and B isomers, respectively.

In Vitro Binding Assay
The binding affinities of all six ligands for hepsin were measured ( Figure S1). The isomers exhibited higher binding affinities for hepsin compared to A isomers (Table Binding affinities of the ligands for matriptase, a well-characterized serine protease [2 were also measured to investigate the selectivity of the ligands for hepsin. Similar bin ing patterns of the ligands were observed; B isomers showed higher binding affinities f matriptase as compared to A isomers. The majority of the ligands displayed significant higher binding affinities and selectivity for hepsin compared with the Ac-Leu-Arg- (Figure 1), which is a known hepsin inhibitor [23]. The Ki value and hepsin selectivi (ratio of Ki values for matriptase and hepsin) of Ac-Leu-Arg-kt measured in this stud were 7.8 ± 2.8 nM and 7.2, while the previous reported values were 22.4 ± 0.50 nM an 14.9, respectively [23]. Fluorine-substituted ligands 1A and 1B exhibited the highe binding affinities (Ki = 0.9 ± 0.1 nM and 0.5 ± 0.1 nM, respectively), whereas nat Ga-an nat Cu-DOTA-conjugated ligands showed binding affinities comparable to each other (2 and 3A: Ki = 14.3 and 15.0 nM; 2B and 3B: Ki = 5.7 and 5.1 nM, respectively). Hepsin s lectivity over matriptase was also higher for 1 (60.6 for A isomer and 40 for B isome than that for 2 (11.6 and 12 for A and B isomers, respectively) and 3 (16.6 and 23.5 for and B isomers, respectively) (

In Vitro Binding Assay
The binding affinities of all six ligands for hepsin were measured ( Figure S1). The B isomers exhibited higher binding affinities for hepsin compared to A isomers (Table 1). Binding affinities of the ligands for matriptase, a well-characterized serine protease [22], were also measured to investigate the selectivity of the ligands for hepsin. Similar binding patterns of the ligands were observed; B isomers showed higher binding affinities for matriptase as compared to A isomers. The majority of the ligands displayed significantly higher binding affinities and selectivity for hepsin compared with the Ac-Leu-Arg-kt ( Figure 1), which is a known hepsin inhibitor [23]. The K i value and hepsin selectivity (ratio of K i values for matriptase and hepsin) of Ac-Leu-Arg-kt measured in this study were 7.8 ± 2.8 nM and 7.2, while the previous reported values were 22.4 ± 0.50 nM and 14.9, respectively [23]. Fluorine-substituted ligands 1A and 1B exhibited the highest binding affinities (K i = 0.9 ± 0.1 nM and 0.5 ± 0.1 nM, respectively), whereas nat Ga-and nat Cu-DOTA-conjugated ligands showed binding affinities comparable to each other (2A and 3A: K i = 14.3 and 15.0 nM; 2B and 3B: K i = 5.7 and 5.1 nM, respectively). Hepsin selectivity over matriptase was also higher for 1 (60.6 for A isomer and 40 for B isomer) than that for 2 (11.6 and 12 for A and B isomers, respectively) and 3 (16.6 and 23.5 for A and B isomers, respectively) ( Table 1). Ligand 3B was selected for radiolabeling and for further evaluation, because [ 64 Cu]3B had the potential to be developed as a theranostic pair with the ligand labeled with a therapeutic radionuclide, such as 67 Cu.

In Vitro Serum Stability
An incubation mixture of the radioligand in fetal bovine serum (FBS) was analyzed using HPLC at the indicated time points ( Figure S2). An unidentified radioactive peak appeared at 4.5 min and increased slightly over time: 0.2% at 0 h, 0.7% at 1 h, 1.0% at 3 h, 5.8% at 21 h, and 6.3% at 24 h ( Figure 2 (Table 1). Moreover, the unidentified polar radioactive peak at 4.5 min was less than 6.3% at 24 h. Therefore, the [ 64 Cu]3B was further studied in vitro and in vivo.

In Vitro Serum Stability
An incubation mixture of the radioligand in fetal bovine serum using HPLC at the indicated time points ( Figure S2). An unidentifi appeared at 4.5 min and increased slightly over time: 0.2% at 0 h, 0.7% 5.8% at 21 h, and 6.3% at 24 h ( Figure 2 (Table 1). Moreover, the unidentified polar radioactive peak than 6.3% at 24 h. Therefore, the [ 64 Cu]3B was further studied in vitro ; an unidentified radioactive dentified radioactive peak at 4.5 min was 6.3% at 24 h.

Cell Binding
Cell binding of [ 64 Cu]3B was measured using the three prostate shown in Figure 3A, cell binding of [ 64 Cu]3B increased in the order LNCaP. The binding also increased in a time-dependent manner fro 1.64-3.33% ID/mg; 22Rv1: 1.17-2.46% ID/mg; PC3: 0.70-1.92% ID/m blocking study at 6 h demonstrated that binding of [ 64 Cu]3B to LNCa by 45.7% (*** p) and that to 22Rv1 cells was reduced by 50.5% (*** p) in ( Figure 3B), suggesting relatively higher specific binding of the radiol to LNCaP cells. Furthermore, binding to PC3 cells was reduced by 23 statistically significant. The results demonstrated that the hepsin lev was higher than in PC3 but lower than in LNCaP cells (Figure 3).

Cell Binding
Cell binding of [ 64 Cu]3B was measured using the three prostate cancer cell lines. As shown in Figure 3A, cell binding of [ 64 Cu]3B increased in the order of PC3, 22Rv1, and LNCaP. The binding also increased in a time-dependent manner from 1-24 h (LNCaP: 1.64-3.33% ID/mg; 22Rv1: 1.17-2.46% ID/mg; PC3: 0.70-1.92% ID/mg) ( Figure 3A). A blocking study at 6 h demonstrated that binding of [ 64 Cu]3B to LNCaP cells was reduced by 45.7% (*** p) and that to 22Rv1 cells was reduced by 50.5% (*** p) in the presence of 3B ( Figure 3B), suggesting relatively higher specific binding of the radioligand to 22Rv1 than to LNCaP cells. Furthermore, binding to PC3 cells was reduced by 23.9%, which was not statistically significant. The results demonstrated that the hepsin level in the 22Rv1 cells was higher than in PC3 but lower than in LNCaP cells ( Figure 3).

MicroPET Imaging
MicroPET images were acquired using mice in which PC3 and 22Rv1 cells were in oculated in the left and right flanks, respectively ( Figure 4A). Region of interest (ROI analysis demonstrated that the radioligand uptakes in the 22Rv1 tumors (1.67 ± 0.06 at h, 2.07 ± 0.15 at 14 h, and 2.10 ± 0.26 at 17 h) were higher than in the PC3 tumors (1.23 0.22 at 1 h, 1.07 ± 0.14 at 14 h, and 1.02 ± 0.17 at 17 h) at all time points ( Figure 4B). Th radioligand uptake was high and persistent over time in 22Rv1 tumors, whereas i slowly decreased in PC3 tumors over 17 h after injection. Relatively high uptake wa noted in the liver, suggesting that the radioligand was excreted through the hepatobili ary system.

Discussion
Early and accurate diagnosis of prostate cancer is vital for improving the prognosi of the disease. Hepsin is overexpressed in cancerous prostate but not in normal prostat [7][8][9][10][11]. Moreover, it is located on the cell surface [12], making it an ideal biomarker fo PET imaging. Although a few studies have examined the optical imaging of hepsin, PET imaging studies of hepsin specifically have not been conducted. In this study, we de signed, synthesized, and investigated a radioligand for PET imaging studies of hepsin Six ligands were synthesized and evaluated for their binding affinities and selectivity fo hepsin (Figure 1). Based on the binding assay results, all six ligands were found to b potent reversible hepsin inhibitors, and 1A and 1B had higher binding affinities and se lectivity for hepsin compared to the other ligands (Table 1). On the other hand, ligand 3B was the preferred candidate for the development of a radioligand because of the facil

MicroPET Imaging
MicroPET images were acquired using mice in which PC3 and 22Rv1 cells were inoculated in the left and right flanks, respectively ( Figure 4A) Figure 4B). The radioligand uptake was high and persistent over time in 22Rv1 tumors, whereas it slowly decreased in PC3 tumors over 17 h after injection. Relatively high uptake was noted in the liver, suggesting that the radioligand was excreted through the hepatobiliary system.

MicroPET Imaging
MicroPET images were acquired using mice in which PC3 and 22Rv1 cells were inoculated in the left and right flanks, respectively ( Figure 4A). Region of interest (ROI) analysis demonstrated that the radioligand uptakes in the 22Rv1 tumors (1.67 ± 0.06 at 1 h, 2.07 ± 0.15 at 14 h, and 2.10 ± 0.26 at 17 h) were higher than in the PC3 tumors (1.23 ± 0.22 at 1 h, 1.07 ± 0.14 at 14 h, and 1.02 ± 0.17 at 17 h) at all time points ( Figure 4B). The radioligand uptake was high and persistent over time in 22Rv1 tumors, whereas it slowly decreased in PC3 tumors over 17 h after injection. Relatively high uptake was noted in the liver, suggesting that the radioligand was excreted through the hepatobiliary system.

Discussion
Early and accurate diagnosis of prostate cancer is vital for improving the prognosis of the disease. Hepsin is overexpressed in cancerous prostate but not in normal prostate [7][8][9][10][11]. Moreover, it is located on the cell surface [12], making it an ideal biomarker for PET imaging. Although a few studies have examined the optical imaging of hepsin, PET imaging studies of hepsin specifically have not been conducted. In this study, we designed, synthesized, and investigated a radioligand for PET imaging studies of hepsin. Six ligands were synthesized and evaluated for their binding affinities and selectivity for hepsin ( Figure 1). Based on the binding assay results, all six ligands were found to be potent reversible hepsin inhibitors, and 1A and 1B had higher binding affinities and selectivity for hepsin compared to the other ligands (Table 1). On the other hand, ligand 3B was the preferred candidate for the development of a radioligand because of the facile radiolabeling of 11B with [ 64 Cu]CuCl2 and its potential use as a theranostic pair, in addi-

Discussion
Early and accurate diagnosis of prostate cancer is vital for improving the prognosis of the disease. Hepsin is overexpressed in cancerous prostate but not in normal prostate [7][8][9][10][11]. Moreover, it is located on the cell surface [12], making it an ideal biomarker for PET imaging. Although a few studies have examined the optical imaging of hepsin, PET imaging studies of hepsin specifically have not been conducted. In this study, we designed, synthesized, and investigated a radioligand for PET imaging studies of hepsin. Six ligands were synthesized and evaluated for their binding affinities and selectivity for hepsin ( Figure 1). Based on the binding assay results, all six ligands were found to be potent reversible hepsin inhibitors, and 1A and 1B had higher binding affinities and selectivity for hepsin compared to the other ligands (Table 1). On the other hand, ligand 3B was the preferred candidate for the development of a radioligand because of the facile radiolabeling of 11B with [ 64 Cu]CuCl 2 and its potential use as a theranostic pair, in addition to its potent binding affinity and selectivity for hepsin.
Radiochemical synthesis of [ 64 Cu]3B which involved the incorporation of 64 Cu into the DOTA-conjugated compound (11B) was straightforward (Scheme 2). However, a small quantity of [ 64 Cu]3A was also formed, even though the HPLC-purified 11B was used for radiolabeling. Therefore, [ 64 Cu]3B was purified by HPLC. The desired HPLC fraction was collected; however, the removal of the HPLC solvents at 80 • C under a gentle stream of N 2 was time-consuming. Although this procedure could be improved by using solidphase extraction cartridges, the water-solubility of the product would hamper the use of cartridges.
We performed in vitro and in vivo characterization of [ 64 Cu]3B using prostate cancer cells and a prostate tumor mouse model. LNCaP and PC3 cells were selected as the high and low hepsin expressing cell lines [13,24]. The 22Rv1 cell line was included in this study. The cell binding of [ 64 Cu]3B was high in LNCaP cells, intermediate in 22Rv1 cells, and lowest in PC3 cells ( Figure 3A). Blocking of cell binding using 3B resulted in 45.7-50.5% inhibition in LNCaP and 22Rv1 cells (*** p) at 6 h, but only a 23.9% inhibition in PC3 cells ( Figure 3B). The results revealed the hepsin level in the 22Rv1 cells, which was higher than in the PC3 but lower than in the LNCaP cells, and the specificity of [ 64 Cu]3B for hepsin.
Early detection of tumors is clinically important [13]. Therefore, we conducted PET imaging studies on small tumors in mice that were implanted with PC3 cells on the left flank and 22Rv1 cells on the right flank ( Figure 4A). The 22Rv1 cells lines were selected for the mouse tumor model owing to their fast growth rate and high self-blocking levels as compared to LNCaP cell lines. (Figure 3). In contrast, the PC3 was selected as the low hepsin-expressing cell lines. Based on the ROI analysis of tumor PET images, the radioligand uptake at 17 h after injection was two-fold higher in 22Rv1 tumors than that in PC3 tumors ( Figure 4B). Moreover, the radioligand uptake in 22Rv1 tumors increased over time, whereas the uptake in PC3 tumors decreased, suggesting that tumor uptake was due to the specific binding of [ 64 Cu]3B to hepsin. In this study, the radioligand was evaluated in mice implanted with two cell lines displaying varying levels of hepsin ( Figure 3). However, in vivo blocking studies may be needed to further support the hepsin specificity of [ 64 Cu]3B in vivo.
An in vitro serum stability study showed that the [ 64 Cu]3B was slowly epimerized to the corresponding A isomer when incubated with FBS at 37 • C ( Figure 2). Although the observed in vitro serum stability of the radioligands may not always reflect their in vivo stability, this in vitro study is widely used for predicting in vivo stability of radioligands. In this study, PET images were acquired at 1, 14, and 17 h after radioligand injection. Based on the in vitro serum stability study findings, the PET images obtained at 17 h could be attributed to the radioligand that was comprised of [ 64 Cu]3B (higher than 70.3%) and [ 64 Cu]3A (less than 23.9%) (Figure 2). The PET images acquired within a few hours using a radioligand with a shorter half-life will reveal an uptake only by [  Most of the reported hepsin inhibitors exhibit high binding affinities but relatively low selectivity for hepsin over matriptase and other proteases [15,23,[25][26][27][28][29][30]. The ligands developed in this study exhibited superior binding affinities with a moderate selectivity for hepsin (Table 1). To this end, the development of radioligands with both higher binding affinities and selectivity for hepsin is required, thus allowing for their improved uptake in hepsin-expressing prostate tumors.

(S)-4-Methyl-2-(pent-4-ynamido)pentanoic acid (5)
Trifluoroacetic acid (TFA) (10 mL) was slowly added to a stirred solution of compound 4 (2 g, 7.48 mmol) in CH 2 Cl 2 (50 mL) at 0 • C. The mixture was then stirred at room temperature for 2 h. At the end of the reaction, the solvent and TFA were removed from the reaction mixture. The residue was solidified with diethyl ether, filtered, and dried to yield compound 5 (1.58 g, 100%) as a white solid. 1 (7) Compound 6 (0.65 g, 1.0 mmol) was dissolved in TFA/thioanisole/water (95/2.5/2.5, v/v/v, 20 mL), and the solution was stirred at room temperature for 3 h. After the reaction mixture was cooled to 0 • C, isopropyl ether (60 mL) was added. The resulting white precipitates were filtered, washed with isopropyl ether, and dried to yield compound 7 (0.46 g, 83%) as a white solid. 1  Piperidine (1 mL) was added to a stirred solution of compound 8 (400 mg, 0.39 mmol) in CH 2 Cl 2 (10 mL). The reaction mixture was stirred at room temperature for 2 h. It was quenched with a saturated NaHCO 3 solution (aq.) and extracted with ethyl acetate. The organic layer was washed with brine, dried over MgSO 4 , and concentrated in vacuo. The crude residue was purified by flash column chromatography (9:1 dichloromethane:methanol) to yield compound 9 (300 mg, 96%) as a white solid. 1
11A: 1  The mixture was diluted with water and then purified by HPLC using a semi-preparative column eluted with an 80:20 mixture of 0.1% TFA-water and 0.1% TFA-acetonitrile at a flow rate of 3 mL/min. The desired fractions were collected, concentrated in vacuo, and lyophilized to yield 2A and 2B as white solids.

Synthesis of Radioligand [ 64 Cu]3B
Compound 11B (20 µg, 0.018 µmol) was dissolved in 150 µL of sodium acetate buffer (0.1 M, pH 6), and [ 64 Cu]CuCl 2 (50 µL) was added to this solution. The reaction mixture was stirred at 80 • C for 20 min. After cooling to room temperature, the reaction mixture was diluted with water. It was then purified by HPLC equipped with a semi-preparative column using an 80:20 mixture of 0.1% TFA-water and 0.1% TFA-acetonitrile at a flow rate of 3 mL/min. The desired product was eluted between 21.8 and 22.6 min and the HPLC solvents were removed under a gentle stream of N 2 at 80 • C. The radioligand was re-dissolved in saline for in vitro and in vivo studies.
The molar activity was determined by comparing the UV peak area of the desired radioactive peak and those of the various concentrations of non-radioactive ligand using HPLC. This was performed using an analytical column eluted with a 75:25 mixture of water and acetonitrile, both containing 0.1% TFA, at a flow rate of 1 mL/min. The identity of [ 64 Cu]3B was determined by co-injecting the radioligand and the corresponding nonradioactive ligand into the HPLC system.
Ligands (0.03 nM-30 µM) were diluted in DMSO (2% final concentration) and mixed with either activated hepsin or matriptase in 96-well plates. The final assay concentration for both hepsin and matriptase in TNC buffer was 0.3 nM. After an incubation period of 30 min at 37 • C, BOC-QAR-AMC substrate (R&D Systems) was added to hepsin and matriptase assay mixtures. The final substrate concentration was 150 µM in a final reaction volume of 100 µL.
Changes in fluorescence (excitation at 380 nm and emission at 460 nm) were measured at room temperature over 2 h in a BioTek Synergy two-plate reader (Agilent Technologies). Using the GraphPad Prism version 6.02 software (San Diego, CA, USA), a non-linear curve fit was performed to determine the inhibitor IC 50 s from a plot of the mean reaction velocity versus the inhibitor concentration. The IC 50 values represent the average of three experimental determinations ( Figure S1). K i values were calculated using the Cheng and Prusoff equation (K i = IC 50 /(1 + [S]/K m )).

In Vitro Serum Stability
[ 64 Cu]3B (18.5 MBq) dissolved in saline was added to 50% FBS (Gibco, Brooklyn, NY, USA) and incubated at 37 • C with shaking. At the indicated time points (0, 1, 3, 21, and 24 h), an aliquot was taken, treated with the same volume of acetonitrile, and then centrifuged. The supernatants were analyzed by HPLC (Agilent Technologies) equipped with an analytical column using a 75:25 mixture of 0.1% TFA-water and 0.1% TFA-acetonitrile ( Figure S2). Eluates were monitored using a NaI(T1) radioactivity detector. The PC3 cell lines were purchased from the Korean Cell Line Bank (Seoul, Korea), and LNCaP cells were purchased from the American Type Culture Collection (ATCC; Manassas, VA, USA). The 22Rv1 cells (ATCC) were kindly provided by Dr. Yong Jin Lee (KIRAMS, Seoul, Korea). All three human prostate cancer cell lines were free of mycoplasma contamination. They were cultured in RPMI-1640 media supplemented with 10% FBS (ATCC), penicillin (100 U/mL), and streptomycin (100 µg/mL) and maintained at 37 • C in a humidified 5% CO 2 incubator. 4.6.2. Cell Binding PC3, LNCaP, and 22Rv1 cells were seeded at 1 × 10 6 cells/well in six-well plates and cultured in RPMI-1640 medium for 24 h. The medium was changed to 5% FBS before the cell binding study.
[ 64 Cu]3B (185 kBq/5 µL) was added to each well in a final volume of 2 mL. The plates were incubated at 37 • C for 1, 2, 6, and 24 h. The cells were washed three times with PBS and lysed using 0.1 N NaOH. The lysate was counted using a gamma counter. Protein concentrations in the cell lysates were determined using the Bradford method. For the blocking study, cells were incubated with the radioligand in the presence of 3B (20 µM) at 37 • C for 6 h and then treated as described above. All experiments were performed in triplicate. Data are expressed as mean ± SEM (n = 3).

4.7.
In Vivo Studies 4.7.1. Animals BALB/c nude mice (male) aged five weeks were used in this study. Mice were provided drinking water and a normal diet ad libitum. They were maintained under a 12 h light-dark cycle and 50% humidity condition at 24 ± 1 • C. MicroPET images were acquired using an Inveon microPET/CT scanner (Siemens Medical Solutions, Knoxville, TN, USA).

MicroPET Imaging
The tumor xenograft model was prepared by subcutaneously inoculating PC3 (7 × 10 6 ) and 22Rv1 cells (7 × 10 6 ) suspended in 100 µL of a 1:1 mixture of Matrigel and PBS into the left (PC3) and right (22Rv1) flanks of BALB/c nude mice. MicroPET images were acquired for 10 min at 1, 14, and 17 h after intravenous injection of [ 64 Cu]3B (8.14 ± 0.06 MBq) into mice (n = 3) via the tail vein when tumor volumes had reached 119.3 ± 26.5 mm 3 (PC3) and 147.0 ± 37.0 mm 3 (22Rv1). The images were reconstructed using the three-dimensional ordered subset expectation maximization and then processed using Siemens Inveon Research Workplace 4.2. ROIs were drawn over the tumors in the left and right flanks and other major tissues, and the average signal levels in the ROIs were measured. Data are expressed as mean ± SD (n = 3).

Statistical Analysis
Data were analyzed using an unpaired, two-tailed Student's t-test using GraphPad Prism version 7.0 software, and differences at the 95% confidence level (P < 0.05) were considered significant.

Conclusions
In this study, we developed a novel PET radioligand for hepsin imaging in prostate tumor. In vitro cell binding study and in vivo characterization of the radioligand in mice, implanted with two prostate cancer cell lines displaying varying levels of hepsin, showed that [ 64 Cu]3B exhibits desirable characteristics for PET imaging of hepsin. This is the first PET radioligand for hepsin imaging, and our study can be used as a foundation to further develop and refine radioligands for hepsin imaging.