Preclinical Evaluation of a Fluorine-18 Labeled Probe for the Detection of the Expression of PSMA Level in Cancer.

Prostate-specific membrane antigen (PSMA) is a prospect biomarker for the treatment of prostate cancer. Meanwhile, positron emission tomography (PET) is being developed as a significant imaging modality in cancer diagnosis. A new PET probe Glu-ureido-Lys-naphthylalanine-tranexamic acid-Gly(AMBF3)-triiodobenzoic acid (18F-GLNTGT) was radiosynthesized by a one-step 18F-labeled method. 18F-GLNTGT was obtained with a radioactivity yield (RCY) of 12.16 ± 6.4% and a good radiochemical purity (RCP > 96%). The cell uptakes of 18F-GLNTGT were determined to be 15.9 ± 0.43% ID and 9.47 ± 1.26% ID at 15 min in LNCaP cells and PC-3 cells, respectively. The cell internalization of 18F-GLNTGT was determined to be 12.89 ± 0.94% ID and 5.34 ± 0.15% ID at 15 min in LNCaP cells and PC-3 cells, respectively. It is suggested that the probe has good specificity targeting PSMA. From the results of 18F-GLNTGT binding affinity with PSMA, it has a higher affinity and a K i value of 0.49 nM (95% confidence interval (CI): 0.35-0.67 nM). In PET imaging, 18F-GLNTGT showed the highest tumor uptake of 3.51 ± 0.15% ID/g at 45 min and the maximum tumor/muscle (T/Mmax) ratio of 3.68 ± 0.29 at 60 min post-injection (p.i.) in LNCaP tumors. The control probe 18F-AlF-NOTA-RGD2 presented the highest tumor uptake of 4.2 ± 0.54% ID/g at 7.5 min and the T/Mmax ratio of 2.72 ± 0.63 at 45 min p.i. in LNCaP tumors. The results showed that the probe has a higher tumor/muscle ratio compared with the control probe 18F-AlF-NOTA-RGD2. Although the probe 18F-GLNTGT has some limitations for CT signal detection both in cells and in vivo, it is still a promising PET probe for targeting PSMA membrane protein.

Introduction

Prostate cancer is a normal malignant tumor all over the world, which has become the fifth primary inducements of cancer deaths among adult male.[1] Currently, the gold standard for clinical diagnosis of early prostate cancer is a combination of prostate-specific antigen (PSA) blood test and prostate biopsy. Meanwhile, a large number of imaging techniques have been applied to the diagnosis of prostate cancer, such as bone scan, computed tomography (CT) scan, magnetic resonance imaging (MRI), lymph node biopsy, position emission tomography (PET) scan, and so on.[2−4] Among all of the clinical image techniques of diagnosis malignant tumors, CT imaging is widely applied because of its higher spatial resolution and anatomical imaging ability. However, the low soft tissue resolution is the disadvantage of CT imaging.[5] PET probes are applied to image a variety of tumor biomarkers, which will further promote tumor diagnosis efficiency. Meanwhile, noninvasive in vivo PET/CT imaging is utilized to precisely ensure the location of tumor.[6] Currently, the application of molecular imaging technology for prostate cancer is usually a combination of PET with the CT imaging technology to achieve early diagnosis and staging of tumors.[7] For example, 11C and 18F-choline probes and 11C-acetic acid probes achieved accurate PET/CT imaging of prostate cancer cell membrane and fatty acid metabolism. Furthermore, the diagnosis efficacies of the above two types of probes in diagnosing prostate cancer at different stages were still low.[8,9] Recently, researchers have shown an increased interest in exploring biomarkers in connection with the presence or development of prostate cancer. Prostate-specific membrane antigen (PSMA) is a transmembrane protein, which is highly expressed in prostate cancer. It is expressed strongly in low-differentiated, metastatic, and androgen-independent prostate cancers.[10] Therefore, PSMA is a promising biomarker in prostate cancer.

What have been shown in the recent studies is that PET probes targeting PSMA have been explored. Due to the low blood clearance and high background value of the first-generation PSMA targeting probe (18F-DCFBC), the second and third generations (18F-DCFPyL and 18F-PSMA-617/1007, respectively) were developed.[11,12] Studies have been devoted to developing PET imaging with targeting PSMA probes because of the higher detection effect of prostate cancer. Among these PSMA PET probes, 68Ga-labeled targeting PSMA imaging probes have been used in prostate cancer diagnosis.[13] Although the 68Ga-labeled targeting PSMA probes show good clinical imaging characteristics for prostate cancer, they have a shorter half-life and the rinsing quantities of the 68Ga generator produce less isotopes. Fluoride-18 possesses an appropriate half-life, and larger doses are easily available; it is selected as a common radionuclide for prostate cancer.[14]

This study was focused on designing a probe targeting PSMA with higher competitive binding ability and achieving potential dual-modality imaging. The probe contains not only the glutamate-ureido-lysine scaffold bearing RBF3– radioprosthetic groups for PET imaging[15] but also the triiodobenzoic acid monomer group for CT signal detection.[21] We measured the serum stability of probe and the log D7.4 value. The cell uptake, cytotoxicity, and competitive binding ability of the PSMA-targeted PET probe were evaluated in cells with different PSMA expression. Meanwhile, PET imaging of the probe in prostate cancer tumors with different PSMA expressions was performed. This further confirmed that the probe is a specific PET probe for imaging PSMA-overexpressed tumors.

Results

Synthesis of Nonradioactive Compound GLNTGT

The synthesis routes of nonradioactive compound GLNTGT are shown in Figure . We first synthesized intermediate A (yield: 460 mg, 95%) by the method reported in the literature.[10] Intermediate B (yield: 410 mg, 88.2%) was synthesized by a solid-phase peptide synthesis method. The synthesis of compound 1 (yield: 140 mg, 49%) using intermediate A and intermediate B was performed at 25 °C in a N2 atmosphere. Compound 2 (yield: 115 mg, 82.1%) could be obtained after compound 1 had been deprotected. Furthermore, the nonradioactive compound GLNTGT (yield: 20 mg, 40%) was obtained from the click reaction through compound 2. The electrospray ionization-mass spectrometry (ESI-MS), high-performance liquid chromatography (HPLC), 1H, and 13C nuclear magnetic resonance (NMR) characterizations of nonradioactive compound GLNTGT are shown in Figures S1–S11.

Figure 1

Synthesis of nonradioactive compound GLNTGT.

Radiosynthesis of 18F-GLNTGT and In Vitro Stability Study

The probe 18F-GLNTGT was obtained through a one-step 18F-labeling reaction (Figure ). The radiochemical yield of 18F-GLNTGT was 12.16 ± 6.4%. The specific activity of probe 18F-GLNTGT was 31.71 ± 7.4 GBq/μmol, and its radiochemical purity was over 96%. The probe 18F-AlF-NOTA-RGD2 was radiolabeled according to the literature.[16] The stability of 18F-GLNTGT was studied by radio-HPLC in phosphate buffered saline (PBS) and fetal bovine serum (FBS). As can be seen in Figure , 18F-GLNTGT was consistently stable in PBS and FBS during incubation for 4 h.

Figure 2

Radiochemical synthesis of probe 18F-GLNTGT.

Figure 3

Meaning of the Z-axis part in the radiolabeling and stability diagrams: Line B represents the radioactive HPLC curve of probe 18F-GLNTGT after radiolabeling; line C represents the radioactive HPLC curve of probe 18F-GLNTGT after purification; lines D, E, and F represent the radioactive HPLC curves of probe 18F-GLNTGT after 1, 2, and 4 h incubation in fetal bovine serum, respectively; line G represents the radioactive HPLC curve of the labeled substance after the probe 18F-GLNTGT was incubated for 4 h in PBS.

Lipophilicity

The log D7.4 value of probe 18F-GLNTGT was determined to be 0.39 ± 0.25 (n = 3), indicating that the probe was hydrophobic.

In Vitro Uptake Study and Cytotoxicity Determination

The PSMA-mediated uptake of probe 18F-GLNTGT was evaluated with LNCaP cells and PC-3 cells. As shown in Figure a, the cell uptake of probe 18F-GLNTGT quickly increased to 15.90 ± 0.43% ID at 15 min and then slowly increased up to 17.61 ± 1.71% ID at 120 min in LNCaP cells. The cell uptake of probe 18F-GLNTGT also gradually increased from 9.47 ± 1.26% ID at 15 min to 13.02 ± 0.65% ID at 120 min in PC-3 cells. The cytotoxicity of nonradioactive compound GLNTGT was determined in the prostate cancer cells with different PSMA expressions. As seen in Figure b, the cell viability of nonradioactive compound GLNTGT decreased slightly from 0.98 ± 0.02 at 6.25 μM to 0.92 ± 0.01 at 100 μM in LNCaP cells. The cell viability of nonradioactive compound GLNTGT also decreased slightly from 0.99 ± 0.01 at 6.25 μM to 0.94 ± 0.01 at 100 μM in PC-3 cells.

Figure 4

(a) Cell uptake assessment of probe 18F-GLNTGT in LNCaP and PC-3 cells at 15, 30, 60, and 120 min (*P < 0.05, **P < 0.01, ***P < 0.001). (b) Cytotoxicity assessment of nonradioactive compound GLNTGT in LNCaP and PC-3 cells at concentrations of 6.25, 12.5, 25, 50, and 100 μM.

Cell Internalization

The cell internalization was performed to verify the specific targeting of the probe 18F-GLNTGT in different PSMA expression cell lines. As shown in Figure , the internalization degree of probe 18F-GLNTGT in LNCaP cells gradually increased from 12.89 ± 0.94% ID at 15 min to 15.96 ± 0.93% ID at 120 min. The internalization degree of the probe 18F-GLNTGT in PC-3 cells increased slightly from 5.34 ± 0.15% ID at 15 min to 6.01 ± 0.10% ID at 120 min. The results indicated that the probe has specific targeting to cells with different PSMA expressions.

Figure 5

Cell internalization assessment of probe 18F-GLNTGT in LNCaP and PC-3 cells at 15, 30, 60, and 120 min (*P < 0.05, **P < 0.01, ***P < 0.001).

Binding Assay

The inhibition constants (Ki) of probe 18F-GLNTGT for PSMA were determined by the competitive binding method using LNCaP cells and the nonradioactive compound GLNTGT. As shown in Figure , as the concentration of the nonradioactive compound GLNTGT gradually increased, the competitive binding ability of the probe 18F-GLNTGT to PSMA gradually decreased. Compared with the Ki value of 18F-DCFPyL (Ki = 2.0±0.8),[15] the Ki value of 18F-GLNTGT (0.49 nM (95% confidence interval (CI): 0.35–0.67 nM)) was lower. This indicates that the probe 18F-GLNTGT has better binding affinity to PSMA.

Figure 6

Competitive inhibition curve of probe 18F-GLNTGT in LNCaP cells.

In Vitro CT Experiment

The CT values increased linearly with the concentration of the nonradioactive compound GLNTGT, and the CT images of the nonradioactive compound GLNTGT showed an obvious concentration-dependent brightening effect (Figure S13).

In Vivo PET Imaging

The high stability and specificity of 18F-GLNTGT showed that this probe was fit for further in vivo experiments. As can be seen in the PET images in Figure a, uptake of the probe 18F-GLNTGT was clearly observed in the LNCaP tumors from 15 to 60 min, and there was almost no uptake in the PC-3 tumors. The uptake curves of probe 18F-GLNTGT in LNCaP and PC-3 tumors are shown in Figure b. The highest uptake of probe 18F-GLNTGT in LNCaP tumor (3.51 ± 0.15% ID/g at 45 min p.i.) showed a higher tumor uptake than the highest uptake of probe 18F-GLNTGT in PC-3 tumors (1.19 ± 0.31% ID/g at 15 min p.i.). The T/M ratios of probe 18F-GLNTGT in LNCaP and PC-3 tumors are shown in Figure c. The T/M ratio of probe 18F-GLNTGT in LNCaP tumors gradually increased (T/Mmax 3.68 ± 0.29 at 60 min p.i.) during the PET dynamic scanning time. The probe 18F-GLNTGT reached the T/Mmax ratio at 1.46 ± 0.39 at 30 min p.i., which then gradually decreased in PC-3 tumors.

Figure 7

PET dynamic imaging in 60 min. The tumor nude mice were fixed and anesthetized (2% isoflurane mixed with oxygen), and then 18F-GLNTGT (5.4–5.7 MBq) was injected through the tail vein. (a) PET dynamic imaging in 60 min of mice with the injection of 18F-GLNTGT. (b) Average tumor uptake of the probe 18F-GLNTGT in the whole tumor region of interest of LNCaP tumors and PC-3 tumors; (c) T/M ratios of 18F-GLNTGT in LNCaP tumors and PC-3 tumors. Data were presented as mean ± SD of n = 3 mice (*P < 0.05, **P < 0.01, ***P < 0.001).

As can be seen in the PET images in Figure S14a, the uptake of probe 18F-AlF-NOTA-RGD2 was clearly observed in both the LNCaP and PC3 tumors from 15 to 30 min p.i. The tumor uptake curves of probe 18F-AlF-NOTA-RGD2 in LNCaP and PC-3 tumors are shown in Figure S14b. The highest uptakes of probe 18F-AlF-NOTA-RGD2 in LNCaP and PC-3 tumors were 4.20 ± 0.54% ID/g and 5.70 ± 0.79% ID/g at 7.5 min p.i., respectively. The uptake of probe 18F-A1F-NOTA-RGD2 was significantly absorbed in LNCaP and PC-3 tumors during the PET dynamic scanning time. The T/M ratios of probe 18F-A1F-NOTA-RGD2 in LNCaP and PC-3 tumors are shown in Figure S14c. The T/Mmax ratios of probe 18F-A1F-NOTA-RGD2 were 2.72 ± 0.63 at 45 min p.i. and 2.08 ± 0.79 at 15 min p.i. within LNCaP and PC-3 tumors, respectively.

Discussion

The PET probe targeting PSMA with glutamate-ureido-lysine-naphthylalanine-tranexamic acid as the main structure has been extensively studied over the past years. Currently, the purpose of modification of glutamate-ureido-lysine-naphthylalanine-tranexamic acid analogues is to increase tumor accumulation and expand the application range of probes.[15,18−20,22] Meanwhile, since the triiodobenzoic acid monomer group is a common CT signal group,[21] it is introduced into the glutamate-ureido-lysine-naphthylalanine-tranexamic acid structure to achieve PET/CT dual-modal imaging.

We measured some characterizations of probe 18F-GLNTGT in vitro and in vivo. It is worth noting that the one-step 18F-labeled method was easy to operate and shortened the production time. Therefore, nonradioactive compound GLNTGT was elected with this method to radiolabel. Meanwhile, the results indicated that 18F-GLNTGT had good stability and could be further studied in vitro and in vivo experiments. The octanol–water partition coefficient was very important in the physicochemical properties of the probe, and it is reported to affect the pharmacokinetic parameters of the probe.[23] The octanol–water partition coefficient of probe 18F-GLNTGT verified its hydrophobicity. Therefore, the above characteristics of probes contributed to the further study in vitro experiments.

The probe 18F-GLNTGT has a very significant difference in cell uptake between LNCAP cells and PC-3 cells (P < 0.01), which indicated that the probe showed specificity toward prostate cancer cells with different PSMA expressions. The result of cell internalization can be further verified by the specific targeting of the probe. From the results of the cytotoxicity assay, nonradioactive compound GLNTGT at different concentrations had high cell viability in both LNCAP and PC-3 cells. The result indicated that the nonradioactive compound GLNTGT had better biocompatibility in cells with different PSMA expressions and was suitable for further in vivo experiments.[24] The binding affinity assay of probe 18F-GLNTGT showed that the binding affinity of the probe to PSMA also reached to subnanomolar level after being modified with an iodine-containing group. This result verifies that probe 18F-GLNTGT has achieved the design intention of improving the competitive binding ability with PSMA.[15] Furthermore, considering that the structure of 18F-GLNTGT contained a triiodobenzoic acid monomer group,[21] the nonradioactive compound GLNTGT might have potential CT signals. From the results of in vitro CT imaging, the nonradioactive compound GLNTGT has CT signal only at very high concentrations. Therefore, it is not possible to inject a large dose of nonradioactive compound GLNTGT through the tail vein to achieve CT signal detection in vivo because of the hydrophobicity of the probe.

Since 18F-AlF-NOTA-RGD2 targeted the integrin αvβ3, it normally has high expression on the neovascularization of malignant tumors.[16] Meanwhile, PSMA is also expressed in the neovascularization of malignant solid tumors but not in normal vasculature.[25,26] Therefore, 18F-AlF-NOTA-RGD2 was chosen as a control probe to conduct a comparative study of PET imaging with probe 18F-GLNTGT. To evaluate the diagnostic effect of the probes 18F-GLNTGT and 18F-AlF-NOTA-RGD2 in the prostate cancer tumors with different PSMA expressions, PET imaging was performed in mice having prostate cancer tumors with different PSMA expressions. According to the results of micro-PET image, the probe 18F-GLNTGT and the control probe 18F-AlF-NOTA-RGD2 were both uptaken in LNCaP tumors. However, the tumor uptakes of probe 18F-GLNTGT and control probe 18F-AlF-NOTA-RGD2 in PC-3 tumors were completely different. It is suggested that the uptake of probe 18F-GLNTGT could be distinguished accurately in prostate cancer tumors with different PSMA expressions compared to probe 18F-AlF-NOTA-RGD2 (Figures a and S14a). As shown in Figures b and S14b, the uptake of probe 18F-GLNTGT was significantly different (P < 0.001) in prostate cancer tumors with different PSMA expressions. Although the control probe 18F-AlF-NOTA-RGD2 had higher tumor uptake than 18F-GLNTGT, there was no extremely significant difference in prostate cancer tumors with different PSMA expressions (P > 0.001). Figures c and S14c show that the T/M ratios of the probe 18F-GLNTGT were significantly increased in prostate cancer tumors with different PSMA expressions, while the T/M ratio of the control probe 18F-AlF-NOTA-RGD2 kept decreasing slowly. 18F-GLNTGT was a hydrophobic probe, which undergoes defluorination of AMBF3 labeling method, thus resulting in the probe 18F-GLNTGT with high uptake in nonspecific organs (mainly in metabolic organs such as liver, kidney, and intestine). However, these problems will not affect the uptake of probe 18F-GLNTGT in tumors and the accurate distinction between tumors and normal tissues. The probe could be used for the noninvasive detection of PSMA level in vivo, which is considered a promising probe for the diagnose of prostate cancer.

Conclusions

Using a one-step 18F-labeled method, a probe targeting PSMA with higher competitive binding capacity was designed and synthesized by click reaction between AMBF3 and the corresponding alkyne precursors. Since the probe had good stability in vitro, it is appropriate for subsequent in vitro and in vivo experiments. The nonradioactive compound GLNTGT showed good biocompatibility in prostate cancer cells with different PSMA expressions. The cell internalization of 18F-GLNTGT determined that the probe 18F-GLNTGT has higher tumor cell uptake. Compared to probe 18F-DCFPyL,[14] the probe 18F-GLNTGT has higher binding affinity to PSMA. The results of PET imaging indicated that the mice of the 18F-GLNTGT group showed a higher T/M ratio than those of the control probe 18F-AlF-NOTA-RGD2. The probe holds promising candidates for PET imaging of prostate cancer.

Experimental Section

Materials

The reagents and solvents used in this study were purchased from the Bide Pharmatech Co., Ltd., SAAN Chemical Technology Co., Ltd., Sangon Biotech Shanghai Co., Ltd., Sinopharm Chemical Reagent Co., Ltd., and Sigma-Aldrich Co., Ltd. 1H NMR spectra were recorded on a Bruker Ultrashield 400 MHz spectrometer. Mass spectrometric information for the synthetic intermediates of the nonradioactive compound GLNTGT was obtained by electrospray ionization mass spectrometry system (ESI-MS, Waters). The analytical high-performance liquid chromatography (1525, HPLC, Waters) and semipreparative HPLC (2545, Waters) of the nonradioactive compound GLNTGT were performed by the Breeze2 system. The analytical HPLC and semipreparative HPLC conditions are listed in Tables S1 and S2. Radioactivity was measured in a dose calibrator (CRC-15R, Capintec). The CT imaging of the nonradioactive compound GLNTGT was performed using a Micro-CT (MCT-1113, Suchow Hejun Technology Development Co., Ltd., China).

Synthesis of Nonradioactive Compound GLNTGT

Synthesis of Compound 1

Intermediates A and B were obtained according to the previous report.[8]O-Benzotriazole-N,N,N′,N′-tetramethyl-uronium-hexafluorophosphate (HBTU) (72 mg, 0.19 mmol), N,N-dimethylformamide (DMF) (5 mL), intermediate A (100 mg, 0.20 mmol), and intermediate B (150 mg, 0.16 mmol) were added to a flask (50 mL). N,N-diisopropylethylamine (DIPEA) (80 μL, 0.47 mmol) was added to the reaction solution to adjust the pH to 8–9. After stirring at 25 °C for 3 h under nitrogen, the reaction product was dried under vacuum. The product was distilled by column chromatography (silica gel, CHCl3/MeOH = 100:3). Compound 1 (140 mg, 49%) was obtained and directly used for the next reaction. ESI-MS (m/z): calculated for C57H76I3N6O11 ([M + H]+) = 1401.27, observed as 1400.93.

Synthesis of Compound 2

Compound 1 was mixed with trifluoroacetic acid (TFA) (3 mL), CH2Cl2 (3 mL), and triisopropylsilane (TIPS) (0.3 mL) in a round-bottom flask (25 mL). The mixture was reacted at room temperature for 1 h and then evaporated under vacuum to obtain compound 2 (115 mg, 82.1%). ESI-MS (m/z): calculated for C45H52I3N6O11 ([M + H]+) = 1233.08; observed as 1232.95.

Synthesis of Nonradioactive Compound GLNTGT

Compound 2 (50 mg, 0.04 mmol), ammoniomethyl-trifluoroborate (AMBF3) (31 mg, 0.16 mmol), CuSO4·5H2O (5 mg, 0.02 mmol), tris(3-hydroxypropyltriazolylmethyl)amine (THPTA) (9 mg, 0.02 mmol), and sodium l-ascorbate (8 mg, 0.04 mmol) were dissolved in a mixture solvent (DMF/H2O = 4:1). The reaction mixture was stirred at room temperature for 30 min under nitrogen. After reaction, the nonradioactive compound GLNTGT (20 mg, 40%) was obtained by semipreparative HPLC purification using water–acetonitrile with 0.1% TFA as the eluent. 1H NMR (400 MHz, dimethyl sulfoxide (DMSO)) δ 8.80 (dd, J = 23.4, 8.1 Hz, 1H), 8.24 (dd, J = 21.6, 16.4 Hz, 1H), 8.10–7.85 (m, 2H), 7.92–7.64 (m, 3H), 7.75–7.64 (m, 1H), 7.55–7.27 (m, 2H), 6.33 (t, J = 8.7 Hz, 1H), 4.84 (dd, J = 14.8, 7.6 Hz, 1H), 4.73–4.42 (m, 1H), 4.16–3.94 (m, 1H), 3.73 (dt, J = 13.5, 6.8 Hz, 1H), 3.59 (ddd, J = 35.0, 20.7, 14.2 Hz, 3H), 3.21–3.04 (m, 5H), 2.95 (dd, J = 22.9, 16.6 Hz, 6H), 2.71 (d, J = 16.8 Hz, 1H), 2.53 (d, J = 18.9 Hz, 6H), 2.46–2.19 (m, 2H), 2.07 (s, 1H), 1.69–1.54 (m, 2H), 1.49 (d, J = 6.2 Hz, 1H), 1.26 (d, J = 6.6 Hz, 13H), 1.18 (t, J = 7.3 Hz, 3H), 0.90–0.72 (m, 1H). 13C NMR (101 MHz, DMSO) δ 127.83 (d, J = 16.2 Hz), 53.91 (s), 53.56 (s), 44.08 (d, J = 13.0 Hz), 42.19 (s), 40.61 (s), 40.40 (s), 40.13 (d, J = 13.2 Hz), 39.99 (s), 39.79 (s), 39.68 (d, J = 18.9 Hz), 39.46 (d, J = 21.0 Hz), 38.79 (d, J = 15.5 Hz), 30.22–29.45 (m), 29.20 (s), 12.83 (s). ESI-MS (m/z): calculated for C50H62BF3I3N10O11 ([M – H]−) = 1427.18; observed as 1426.96.

Radiosynthesis of 18F-GLNTGT

No carrier added 18F-fluoride (7.4 GBq) was obtained by medical cyclotron (HM-7, Sumitomo Heavy Industries, Ltd., Japan) to bombard high-pressure 18O–water targets with 18 MeV protons and rinsed with pyridazine buffer (Vwater/Vpyridazine = 3.3:1, 300 μL) directly into a centrifuge tube containing nonradioactive compound GLNTGT in DMF solution (25 mM, 30 μL). The mixture was incubated for 30 min at 80 °C. The radioactive HPLC was used to detect the successful labeling of 18F-GLNTGT. The labeled 18F-GLNTGT reaction mixture was transferred to a 40 mL centrifuge tube with 20 mL of deionized water. The C18 light cartridge was rinsed twice (10 mL/time) with the mixture in the centrifuge tube. The residual radioactive dose of C18 light cartridge was measured after each wash. The C18 light cartridge was washed using 500 μL of ethanol in a 10 mL vial. The radiochemical purity of the probe was measured by radioactive HPLC.

Serum Stability Study

The probe 18F-GLNTGT (20 μL at 0.074 MBq) was added into different centrifuge tubes containing fetal bovine serum (90 μL) and phosphate-buffered saline (PBS, pH = 7.4) (90 μL), respectively. The centrifuge tubes were then incubated for 1, 2, and 4 h at 37 °C. Incubation samples (10 μL) were collected at each time point, and 10 μL of acetonitrile was added to each sample to precipitate the serum proteins. The incubation samples were centrifuged (12 000 rpm, 1 min) to obtain the supernatant for radio-HPLC analysis. The radio-HPLC analysis method of the supernatant is shown in Table S1.

The lipophilicity of probe in saturated octanol-deionized water (pH 7.4) was determined by the shake flask method. In short, 18F-GLNTGT was added to a centrifuge tube containing n-octanol (5 mL) and deionized water (5 mL). The mixture was shaken for 3 min and then centrifuged at 3000 rpm for 5 min. An aliquot of the mixture solution of the octanol (5 mL) and deionized water (5 mL) was counted by a γ-counter (2480, Perkin Elmer). log D7.4 was calculated by the formula: log D7.4 = log10 [(counts in n-octanol phase/counts in water phase)].

Cell Culture and Animal Models

Human prostate cancer LNCaP cell line and PC-3 cell line were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). The cells were cultured in RPMI-1640 medium added to 10% FBS, penicillin (100 U/mL), and streptomycin (100 μg/mL) as double antibody at 37 °C in a 5% CO2 incubator (3111, Thermo Scientific). Cells achieving 80–90% confluence were washed with PBS and passaged with trypsin digestion. The collected cell number was counted by a cell counting plate. The male SCID mice (4–5 weeks old, 15–20 g) were purchased from Cavens Laboratory Animal Co., Ltd. (Changzhou, China) for the animal experiment. The SCID male mice were divided into two groups (three mice each group) and maintained under a specific pathogen-free (SPF) environmental condition for 1 week. LNCaP and PC-3 cell suspensions (5 × 106 cells in 200 μL per mouse) were implanted subcutaneously in the right armpit of mice from different groups, respectively. Tumor volumes were measured every other 2 days after injection until the appearance was observed in the tumors. The protocol of animal experiments was approved by the Animal Care and Ethics Committee of Jiangsu Institute of Nuclear Medicine.

Cell Uptake Studies

Cell uptake experiments were performed on LNCaP cells and PC-3 cells. The LNCaP cells and PC-3 cells were digested with 0.25% trypsin solution. According to the experimental protocols, the LNCaP cells (1 × 106/tube) and PC-3 cells (1 × 106/tube) were divided into centrifuge tubes (1 × 106/tube). Then, 18F-GLNTGT (100 μL at 0.037 MBq) was added to the centrifuge tubes containing the fresh medium. All samples were incubated for 15, 30, 60, and 120 min at 5% CO2 and 37 °C. At each time point, the cells were washed in cold PBS (500 μL), centrifuged (12 000 rpm, 3 min), and the supernatant was discarded. The harvested cells were counted for their radioactivity by a γ- counter.

Cytotoxicity Determination

The prostate cancer cells with different PSMA expressions were digested with a 0.25% trypsin solution and 1640 medium containing 10% fetal bovine serum. The digestion cells were seeded in a 96-well plate. Each well was added with 100 μL of medium in the prostate cancer cells with different PSMA expressions (8 × 103/well). The 96-well plate was incubated for 24 h at 37 °C in a 5% CO2 incubator. According to the experimental protocols, the nonradioactive compound GLNTGT was set up in five concentration gradients (6.25, 12.5, 25, 50, 100 μM). Each concentration gradient was set up with six parallel groups. Nonradioactive compounds GLNTGT with different concentrations in the 96-well plate were placed for 24 h at 5% CO2 in a 37 °C incubator. After incubated, the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) solution (5 mg/mL, 20 μL) was added to each well. Then, the 96-well plate was incubated for another 4 h at 37 °C in an incubator and the supernatant was discarded. Each well was added with DMSO (150 μL) and shaken for 15 min. Finally, the 96-well plate was measured by a microplate reader (BioTek, Inc., Vermont).

LNCaP cells (1 × 106/tube) and PC-3 cells (1 × 106/tube) that have been digested with 0.25% trypsin solution were added to the centrifuge tube, and then 18F-GLNTGT (100 μL, 0.037 MBq) were added to centrifuge tubes containing fresh medium. The samples were placed at 37 °C in a 5% CO2 incubator and incubated for 15, 30, 60, and 120 min. At each time point, cold PBS (4 °C, 500 μL) was added to the centrifuge tubes at different time points to stop the incubation, and then the used IP cell lysate was fully lysed, centrifuged at 12 000 rpm for 5–8 min, and the supernatant was obtained. The supernatant was counted externally, and the two types of cell pellets were internalized.[27]

Competition Binding Assays

LNCaP cells were seeded on six-well plates and incubated for 48 h (4 × 105/well). After incubating, the N-(2-hydroxyethyl)piperazine-N′-ethanesulfonic acid (HEPES) buffer (50 mM, pH 7.4, 0.9% NaCl) was replaced to ensure complete removal of the medium. After 1 h, different wells (in triplicate) contained various concentrations (0.01 nM to 100 μM) of the nonradioactive compound GLNTGT solution, to which an equal concentration of 18F-GLNTGT (0.1 nM) was added. The assay mixture was incubated for 60 min at 37 °C in a 5% CO2 incubator with slight shaking, and it was washed twice with HEPES buffer. A trypsin solution was put in each well to collect cells.[28] Radioactivity was measured by a γ-counter, and then the Ki value was calculated by GraphPad Prism 8.0 software.

In Vitro CT Imaging

The nonradioactive compound GLNTGT was divided into different concentration gradients (0–350 mM) with DMF. Micro-CT scanning was performed to determine the CT contrast effect in the nonradioactive compound GLNTGT of different concentrations. Imaging parameters were set as follows: thickness, 80 nm; voltage, 40 kV; current, 200 μA.[17] The Hounsfield unit (Hu) values were determined by the Hiscan Viewer software.

PET Imaging Studies

The probe 18F-GLNTGT with a specific activity of 5.4–5.7 MBq was injected through the tail vein into the prostate cancer xenograft tumor mice with different PSMA expressions, which had been fixed and anesthetized (2% isoflurane in oxygen). Dynamic (60 min) and static PET scans (10 min) were performed by a micro-PET scanner (Inveon, Siemens, Germany). After the scanning was completed, the PET image was reconstructed using a 2D ordered subset expectation maximization (2D-OSEM) algorithm and analyzed with Siemens IRW software. The percentage of per gram of tissue uptake (% ID/g) injected per gram of tumor and tissue was used for the semiquantitative assessment of the repeated dose for each region of interest (ROI). The imaging condition of the control probe 18F-AlF-NOTA-RGD2 was consistent with that of the probe 18F-GLNTGT for PET imaging.

Statistical Analysis

All data of cell and animal experiments were expressed as mean ± standard deviation (SD). The mean ± SD values were compared by Student’s t test; only P < 0.05 was considered statistically significant.