Animal experimental procedures were performed in accordance with the National Research Council’s Guide for the Care and Use of Medical Laboratory Animals (Ministry of Health, China). All the animal experimental protocols were approved by the Institutional Animal Care and Research Ethics Committee of Shenzhen Hospital of Southern Medical University. Male NOD-SCID mice aged 6–8 weeks and healthy male SD rats aged 8–10 weeks (200 ± 10 g) were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). All mice were kept in a specific pathogen-free (SPF)-grade animal house under 12-h light/dark cycles with controlled temperature (24°C ± 2°C) and relative humidity (50%–60%) and were provided with a standard rodent chow diet and water ad libitum. When used for experiments, the average weight of the mice was 22 ± 2 g, and the age was approximately 12 weeks. Pancreatic cancer cell line PANC-1 was obtained from the American Type Culture Collection (ATCC), pentobarbital sodium was purchased from Sigma-Aldrich (Darmstadt, Germany), rabbit anti-TGFβ monoclonal antibody (#3711S) was purchased from Cell Signaling Technology (Danvers, MA, USA), and secondary antibodies and Hoechst 33258 staining kit were obtained from Beyotime Biotechnology (Shanghai, China).

TGFβ-specific blocking tetradecapeptide P144 is a fragmental analog of the extracellular domain of TGFβR3 (28), and its synthesis was entrusted to a third-party organization company (Hangzhou Chinese Peptide Company, Hangzhou, China). P144-DOTA (molecular weight, 2,114.4 Da) was obtained by coupling P144 with the biological chelator DOTA through a polyethylene glycol short chain (n = 3), and the high-performance liquid chromatography (HPLC) method was used for quality control. Then, P144-DOTA was labeled with radionuclide [⁶⁸Ga]. A radiolabeling module (Smart Module-X) was designed and developed for the synthesis of radionuclide-labeled tracers. The radiolabeling synthesis method for [⁶⁸Ga]Ga-DOTA-P144 was conducted based on a similar methodology as previous reports with modifications (29). The ⁶⁸Ge/⁶⁸Ga generator was rinsed with 0.1 M of HCl, and the pH value of the rinsing solution was adjusted to between 3.5 and 4.5 with 1.25 M of acetic acid sodium buffer. A volume of 200 μL (200 μg) of the precursor was taken after it was mixed evenly, and then it was added into a volume of 3 mL of rinsing solution (radioactivity of ⁶⁸Ga was approximately 370 MBq, that is, 10 mCi), and it was heated in a constant-temperature metal bath at 100°C for 10 min. After cooling, 20 mL of sterile water was added for dilution, and the liquid was passed through the C18 Sep-Pak light solid-phase extraction column and rinsed with 5 mL of sterile water for injection to remove free ⁶⁸Ga ions, ⁶⁸Ge ions, and water-soluble impurities. Then, it was eluted with 10 mL of 75% ethanol, and the ethanol solvents were removed from the radiolabeled compound solution using a nitrogen stream. Finally, the sterile filter membrane was placed into the sterile sealed bottle, and sterilizing filtration of [⁶⁸Ga]Ga-DOTA-P144 product solution was performed before injection.

Radio-HPLC and radio-iTLC assays were conducted for the quality control of product [⁶⁸Ga]Ga-DOTA-P144 and free [⁶⁸Ga] ions. American alltech1500 high-performance liquid chromatography and online radioactive detector were used. Chromatographic column: YMC-Pack Pro C18 RS (5 μm, 250 mm × 4.6 mm). Mobile phase: A) water + 0.1% formic acid; B) acetonitrile. Method: 0 min, 95% A and 5% B; 1 min, 95% A and 5% B; 10 min, 70% A and 30% B; 18 min, 70% A and 30% B; 25 min, 95% A and 5% B; flow speed, 1 mL/min. The quality control standard was that the chemical and radiochemical purities of the products were greater than 95%. Radio-iTLC analytical quantification methods/conditions were as follows. The stationary phase was iTLC glass fiber (1 cm × 8 cm instant silica gel strip), and the developing solvent was 2 mM of sodium citrate/citric acid buffer solution (pH 6.5). performed according to General Rule 1401 in Chinese Pharmacopoeia (2020 Edition), the method for the determination of radioactive drugs, the Rf value of the radioactive peak of the product to be tested was approximately 0.1–0.4, and there was a single main peak under normal conditions. If there were impurities, especially free [⁶⁸Ga] ions, the corresponding peaks should appear at the Rf value of approximately 0.6–0.8.

Pancreatic cancer cell line PANC-1 was obtained from the ATCC, and cells were cultured in complete RPMI1640 medium supplemented with 10% fetal calf serum (Life Technologies, Carlsbad, CA, USA). One portion of the PANC-1 cells (~1 × 10⁶) was injected subcutaneously below the anterior axillary of the nude mice to induce the mouse heterotopic tumor model. The tumor cells were suspended in a mixture (v/v, 1:1) of medium and Matrigel (Corning Life Sciences, Corning, NY, USA). The tumor xenografts were generally palpable within 10 days after injection. Xenograft tumor sizes of the model mice were measured every 3 days. Short and long tumor diameters, tumor volume, and body weight were measured. When the xenograft tumor grew to a volume of 150–200 mm³, the mice were used for PET imaging study. A diagnostic tracer and drug were injected approximately 2 weeks after the establishment of the tumor model.

The method process for cell uptake assay of radiolabeled drug was as follows. First, the selected cell lines were inoculated into a 6-well plate, incubated for 48 h to achieve a cell coverage rate of approximately 80%–90% (1.2–2 × 10⁶ cells/well), and replaced with 1 mL of fresh culture medium without fetal calf serum (FCS). Then, a certain amount of radionuclide-labeled drug (200 MBq/L) was added to the culture well and incubated for different times (0.5–5 h). The culture medium containing the drug was removed, and the cells were washed twice with 1 mL of phosphate-buffered saline (PBS) (pH 7.4). Cells were incubated with 1 mL of glycine HCl solution (1 M, pH 2.2) at room temperature for 10 min to remove surface binding, and the cells were washed with 2 mL of ice-cold PBS (collect washing solution for detection of binding amount). After that, 1.4 mL of lysis buffer (0.3 M of NaOH, 0.2% sodium dodecyl sulfate (SDS)) was used to lyse the cells and collect this lysate. With a γ-counter, the radioactive activity in the collected lysate was measured, and the %AD value was calculated. One mio (1 × 10⁶) of cells was adopted for normalization, with the result representing the internalization uptake level of the tested radiopharmaceutical by this cell line. Each experiment was repeated three times independently, with three multiple wells set at each processing time point.

Tumor tissues and normal tissues were isolated and immersed in 10% neutral formalin. After dehydration in 30% sucrose, the tissue block was embedded in paraffin and cut into slices with a thickness of 5 μm. The tissue slices were blocked with 3% bovine serum albumin (BSA) in PBS and incubated with a concentration of 2 μg/mL of primary antibody (anti-TGFβ-1 mouse monoclonal antibody, TB21, Thermo Fisher Scientific, Waltham, MA, USA) at 4°C overnight. After washing, the horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG secondary antibody (Beyotime Biotechnology, Shanghai, China) (1:400 dilution in volume) was incubated for 1 h at room temperature and washed with PBS three times. After coloration and washing, slices were mounted on glass slides, sealed with 30% glycerin, and visualized under inverted microscopy (Olympus IX71, Olympus Co., Tokyo, Japan).

Peptide conjugate P144-DOTA and the radiolabeled tracer [⁶⁸Ga]Ga-P144 were both administrated through tail vein injection. The chemical dose of [⁶⁸Ga]Ga-P144 was 1 mg/kg, and the radioactive dose was approximately 3.7 MBq in 200 μL per mouse. As a competitive blocker with the same target TGFβ, the dose of DOTA-P144 was 100 mg/kg (47.3 μmol/kg, 100 times the mass dose of the radiolabeled peptide conjugate), and the administration of this unlabeled peptide conjugate was followed after 10 min by 1 mg/kg of the radiolabeled tracer [⁶⁸Ga]Ga-P144. All the mice were deprived of food and water for 1 h before drug administration. Then, the mice were anesthetized with pentobarbital sodium (45 mg/kg) and scanned with microPET. After the last scanning, the mice were anesthetized with 2% pentobarbital sodium at 65 mg/kg intraperitoneally and then executed by dislocation of their cervical vertebra, and the tissues were harvested for analysis.

The tumor-bearing mice were anesthetized with pentobarbital sodium before microPET scanning. Each mouse was intravenously injected with approximately 3.7 MBq (90–120 μCi) of [⁶⁸Ga]Ga-P144. After administration, the mice were scanned by a small animal PET system (Inveon, Siemens, Munich, Germany). The body temperature of the mice was monitored by a rectal probe and kept at 37°C by a heated air stream. The scanning time was set as follows: static PET scanning for 10 min and scanning energy window 350–650 keV at four time points (1 h, 2 h, 3 h, and 4 h) after the injection. MicroPET image reconstruction was performed using a 3D OSEM PSF algorithm with five iterations. Images were processed and analyzed using PMOD4.2 software (PMOD Technologies Ltd., Zurich, Switzerland). Regions of interest (ROIs) of the brain, heart, liver, lungs, kidney, muscle, and tumor were delineated. Radioactivity values of the ROIs per unit volume were obtained, and the percentage of injected dose per tissue weight in gram (%ID/g) values of the different organs and tissues were calculated as follows:

For biodistribution assay ex vivo, the PANC-1 tumor model mice were euthanized by pentobarbital sodium at 1 h after tracer administration. The blood was harvested via cardiac puncture, and the different organs/tissues were isolated, weighed, and counted on a γ-counter for radioactivity. The amount of the injected radiotracer was measured and used to determine the total number of counts of nuclei decay per minute (CPM) by comparison with a standard of known activity. The data were background- and decay-corrected and expressed as percentages of the injection dose per tissue weight in grams (%ID/g).

Data analysis was performed using SPSS19.0 software (SPSS Inc., USA), and the quantitative data were presented as “mean ± SD”. Differences between groups were compared using a two-tailed independent samples t-test. A difference at a p-value below 0.05 was considered statistically significant.

Based on the extracellular domain fragment of TGFBR3, we focused on a series of candidate antagonistic peptides targeting TGFβ, including pentadecapeptide P17 (KRIWFIPRSSWYERA) and tetradecapeptide P144 (TSLDASIWAMMQNA). We finally chose P144 for specific targeting and binding to TGFβ considering its good performance in a study with human glioblastoma cell lines (28). As shown in Figure 1A , the connection of P144 to chelator DOTA was as follows: the free amino group on the threonine residue at one end of P144 and a carboxyl group of DOTA was connected by a covalent bond through a flexible short chain of ethylene glycol. Radioactive nuclide [⁶⁸Ga] was labeled to P144 peptide by the connected chelating agent DOTA ( Figure 1B ). After separation and detection by the radio-HPLC system, the retention time of the conjugated peptide DOTA-P144 (molecular weight = 2,114.4 Da) was 10.459 min, and the chemical purity represented by peak area percentage was 96.401% (>95%; Figure 1C ). As shown in Figure 1D , the radiochemical purity of the nuclide-labeled product [⁶⁸Ga]Ga-P144 compound was 98.87% after purification and placement at room temperature for 4 h, and the retention time was 14.866 min. By radio-iTLC analysis, we found that the Rf value of free ⁶⁸Ga³⁺ ions was 0.717, and the purity was 100% ( Figure 1E ). The Rf value of [⁶⁸Ga]Ga-P144 in the radio-iTLC chromatogram was 0.322, and the radiochemical purity was 100% ( Figure 1F ).

In the human PANC-1 cell line-established mouse pancreatic cancer animal model, immunohistochemical assay was performed to evaluate and confirm the expression and distribution of TGFβ in tumor model tissues. The PANC-1 tumor-bearing nude mice with a xenograft tumor volume over 1,000 mm³ were sacrificed, and the tumor tissue and normal liver tissue were isolated for immunohistochemistry (IHC) assay with an antibody against TGFβ. Hoechst 33258 stain solution was used as a staining indicator for cellular nuclei in the IHC assay. As shown in Figures 2A–F , the IHC result demonstrated the high-intensity distribution of TGFβ in tumor tissue of the mouse model of PANC-1 pancreatic cancer, while the expression of TGFβ in non-tumor tissue (normal liver tissue) was at a very low level.

We examined the binding and internalization capability of [⁶⁸Ga]Ga-P144 in PANC-1 cells and tested the in vivo safety of [⁶⁸Ga]Ga-P144 in normal male SD rats. As shown in Figure 3A , cellular uptake assay results (six series of cell assays) of [⁶⁸Ga]Ga-P144 with PANC-1 and PC3 cell lines showed high and persistent cell uptake ratio of [⁶⁸Ga]Ga-P144 in PANC-1 cells and very low level of non-specific cellular uptake in PC3 cells. Figure 3B shows the body weight growth changes of six healthy rats on different days after the administration of the radiopharmaceutical [⁶⁸Ga]Ga-P144 (500 μCi per rat), and the rapid body weight growth of all tested rats that were exposed to a relatively high dose of [⁶⁸Ga]Ga-P144 for at least 14 days after administration indicated good biosafety in vivo for the development of this radiotracer and its further clinical study.

PANC-1 xenograft tumor model of pancreatic cancer in mice was built to investigate the imaging efficiency of P144-based conjugate [⁶⁸Ga]Ga-P144. The mice were injected with approximately 3.7 MBq of [⁶⁸Ga]Ga-P144 (1 mg/kg) via a caudal vein, and microPET scans were performed at 1 h, 2 h, 3 h, and 4 h after tracer injection. Representative [⁶⁸Ga]Ga-P144 microPET images ( Figure 4A , different section images with tumor location and MIP image at 3 h after injection; Figure 4B , coronal section images at four time points) indicated higher radioactive uptake in tumor areas compared with normal tissue. Radioactive uptake in main organs and tissues was quantified at different time points ( Figure 4C ). Uptakes in the heart (%ID/g-mean, 5.738 ± 1.681), liver (4.014 ± 0.975), lung (3.560 ± 0.739), kidneys (3.322 ± 0.636), and tumors (5.764 ± 1.212) were all relatively high at 1 h after injection. At 4 h post-injection, radioactive distribution in the heart (%ID/g-mean, 3.919 ± 0.891), liver (2.906 ± 0.431), lung (2.190 ± 0.206), and kidneys (2.804 ± 0.281) slightly but continuously decreased. However, uptake in tumors slightly increased (%ID/g-mean, 6.023 ± 1.370) (n = 5 mice per time point). As shown in Figure 4D , the tumor-to-muscle ratio also slightly increased from 3.442 ± 1.531 at 1 h to 3.770 ± 1.178 at 4 h. These results indicated that [⁶⁸Ga]Ga-P144 has good targeting to TGFβ-positive tumors, and it may transfer and redistribute from the heart, liver, and lung to tumor tissues in vivo and has an adequate long retention time and good targeting efficiency for PANC-1 tumor. In addition, as shown in Figure 4E , [⁶⁸Ga]Ga-P144 radioactive distribution ex vivo in different isolated organs/tissues of PANC-1 tumor model mice (n = 5) at 1 h after tracer injection displayed considerable consistence with the result of in vivo uptakes, with relatively high radioactive biodistributions in the heart (4.886 ± 0.899), liver (1.948 ± 0.717), lung (3.156 ± 0.429), kidney (2.985 ± 0.854), and tumors (4.635 ± 1.217).

MicroPET imaging in the PANC-1 tumor model for competitive TGFβ blockade study was performed with unlabeled DOTA-P144 pretreatment (100 times the mass dose of [⁶⁸Ga]Ga-P144) for 10 min before the injection of PET tracer [⁶⁸Ga]Ga-P144 to examine and verify the targeting specificity of P144 carrier against TGFβ. As shown in Figures 5A, B , tumor uptake (%ID/g-mean) of [⁶⁸Ga]-P144 in the blocked group (1 h, 2.927 ± 0.572, p = 0.004398 vs. unblocked) was significantly lower than that of the unblocked group (5.362 ± 0.937). When compared with those in the unblocked group, radioactive uptake values in the kidneys (unblocked, 3.045 ± 0.169; blocked, 2.133 ± 0.206) and lung (3.240 ± 0.214; 1.666 ± 0.496) of the blocked group were significantly decreased (p = 0.000475 and 0.001123, respectively). The radioactive uptakes in other organs were all slightly decreased but without significant differences (p > 0.05, n = 4 per group) between the two groups at the 1-h time point. As shown in Figure 5C , the tumor-to-muscle ratios of radiouptake between the unblocked and blocked groups showed no significant difference (p = 0.3867). These results confirmed the specific targeting of P144-based conjugates to TGFβ, and molecular imaging tracer [⁶⁸Ga]Ga-P144 targeting TGFβ has potential diagnostic value for pancreatic ductal adenocarcinoma.