Albumin Binder–Conjugated Fibroblast Activation Protein Inhibitor Radiopharmaceuticals for Cancer Therapy

Visual Abstract

Fi broblast activation protein (FAP) is overexpressed in cancerassociated fibroblasts, which are one of the main tumor stroma components and constitute a major proportion of cells within the tumor (1,2). Though stromal cells are not malignant, the growth factor and chemokine produced by stromal cells, especially cancer-associated fibroblasts, can lead to the direct stimulation of tumor cell growth, migration, and progression (3). Considering the vital role in tumor survival and cancer growth, cancer-associated fibroblast-targeted diagnosis and therapy via the biomarker FAP have become an attractive strategy for tumor treatment (4,5). FAP-targeted radiopharmaceutical therapy might deliver therapeutic radioisotopes to cancer-associated fibroblasts. It damages the stromal cells and the neighboring tumor cells through the crossfire effect of the bor a-emitting radionuclides, potentially augmenting the therapeutic efficacy (6).
Recently, a variety of quinolone-based FAP inhibitor (FAPI) radiopharmaceuticals has been developed and demonstrated excellent uptake in different FAP-positive tumors of cancer patients (7)(8)(9). For FAP-targeted radiotherapy, an emerging strategy is to directly modify the inhibitor structure to enhance tumor uptake and retention while keeping the accumulation in nontarget tissues unchanged or decreasing it (10)(11)(12). A series of FAPI probes, including FAPI-04, FAPI-21, and FAPI-46, has been successfully developed; their improved pharmacokinetic properties make them promising candidates for therapeutic outcome improvement (13)(14)(15)(16). Though this is one of the optimal ways to develop therapeutic radiopharmaceuticals, the relatively rapid washout from the tumor is still a considerable limitation. Besides, it may be challenging to achieve notable enhancement of pharmacokinetic properties by subtle structure modification.
Several studies have shown that prolonging the blood circulation of drug molecules using albumin-binder moieties could remarkably improve the therapeutic dose (17). 4-(p-iodophenyl) butyric acid and truncated Evans blue moieties are the most widely used albumin binders. The previous studies suggest that they can enhance the tumor uptake and retention of radiopharmaceuticals, resulting in improved therapeutic efficacy (18)(19)(20)(21)(22)(23). Further clinical translation studies also validate the promise of this strategy to be a platform technology for radiopharmaceutical development (24)(25)(26). Therefore, we were curious whether attaching an albumin-binder moiety to the FAPI molecules would improve the FAP-targeted radiotherapy efficacy at the expense of increased retention in blood.
In this study, 2 albumin binder-FAPI conjugates, TEFAPI-06 and TEFAPI-07, were developed by logistic fabrication of 3 functional components: a quinoline-based FAPI originating from FAPI-04, a chelator (i.e., DOTA group) that allows radionuclide labeling for imaging ( 68 Ga or 86 Y) or therapy ( 177 Lu), and an albumin binder: 4-(p-iodophenyl) butyric acid moiety (TEFAPI-06) or truncated Evans blue moiety (TEFAPI-07). The purpose of the study was to evaluate whether the modification improves tumor retention in vivo and which albumin binder better matches FAPI molecules. A series of detailed experiments and comparisons, including cell binding assays, a PET imaging study, a biodistribution study, and a radiotherapy study, was performed. The results demonstrated these 2 albumin binder-conjugated FAPI radiotracers to have high FAP binding affinity and specificity, enhanced tumor retention, and improved radiotherapy efficacy.

Ligands and Radionuclides
The synthesis route and chemical characterization of TEFAPI-06 and TEFAPI-07 are described in Supplemental Figures 1-20 (supplemental materials are available at http://jnm.snmjournals.org). 68 Ga-Cl 3 was eluted with a solution of 0.6 M hydrochloride from a 68 Ge-68 Ga generator (iThemba LABS). 86 Y-Cl 3 was produced with a 14.6-MeV cyclotron; the target design follows our previous report (27), and the purification procedure follows the previous protocol (28). 177 Lu-Cl 3 in a solution of 0.1 M hydrochloride was purchased from ITG.

Radiolabeling and Stability In Vitro
The radiolabeling of 68 Ga, 86 Y, and 177 Lu was performed by incubation with 50 nmol of precursor at pH 4.5-5.0 at 90 C for 10 min. The product was purified by C18 column extraction, and the radiochemical purity was determined by high-performance liquid chromatography equipped with a radioactivity detector. The stability of 177 Lu-TEFAPI-06 and 177 Lu-TEFAPI-07 in saline and human serum was monitored from 2 to 168 h using radio-high-performance liquid chromatography (13). More details about the radiochemistry, the quality control testing, and the stability assay can be found in the supplemental materials (section 3).

Cell Culture and Assay
The human fibrosarcoma cell line (HT-1080), and the HT-1080 cell line transfected with human FAP gene (HT-1080-FAP, from WuXi AppTec), were cultivated in Eagle minimum essential medium containing 10% fetal bovine serum, 1% antibiotic-antimycotic, and a 4 mg/mL concentration of blasticidin S at 37 C under conditions of 5% carbon dioxide. For competition assays, HT-1080-FAP cells were seeded in 6-well plates and cultivated until they reached about 1.2 3 10 6 cells per well. The cells were incubated simultaneously with unlabeled FAPI-04, TEFAPI-06, or TEFAPI-07 (10 25 -10 29 M) with 68 Ga-FAPI-04 in 1 mL of fresh medium without fetal bovine serum for 1 h. The medium was removed, and the cells were washed twice with phosphate-buffered saline (PBS). Subsequently, the cells were lysed with 0.5 mL of 1 M NaOH and washed with 0.5 mL of PBS twice, and the NaOH (0.5 mL) and PBS (0.5 mL 3 2) were collected to determine the uptake counts. For saturation binding assays, HT-1080-FAP and HT-1080 cells were seeded in 24-well plates and cultivated until they reached about 2 3 10 5 cells per well. 68 Ga-FAPI-04, 68 Ga-TEFAPI-06, or 68 Ga-TEFAPI-07 was diluted to a concentration 0.01-200 nM in fresh medium without fetal bovine serum. The cells were incubated in the above solution for 1 h and then washed twice with PBS. The lysed cells and the PBS for washing were collected to determine the counts.

Tumor-Bearing Animal Models
All animal care and experimental procedures were performed by following the animal protocols (CCME-LiuZB-2) approved by the ethics committee of Peking University. The mice were from the Beijing Vital River Laboratory Animal Technology Co., Ltd. For cell-line-derived xenograft models, 5 3 10 6 HT-1080-FAP or HT-1080 cells were subcutaneously inoculated into the right shoulder of 6-wk-old female nu/nu-mice. To establish the patient-derived xenograft (PDX) model, tumor specimens were obtained from patients who underwent presurgical 68 Ga-FAPI-04 PET/CT imaging to confirm that the tumor was FAP-positive. After surgical resection, the tumor specimens were immediately placed in ready-touse fresh tissue preservation solution (TM2701-100) and transported under refrigerated conditions within 2 h. The research protocol was approved by the Institutional Ethics Committee of Peking Union Medical College Hospital (JS-2628). Six-week-old female nonobese diabetic/severe combined immunodeficiency (NOD/SCID) mice were used to establish the PDX models. After being removed from the preservation solution, the tumor specimens were immediately immersed in sterile PBS solution and minced with scissors, and the fragments were then implanted subcutaneously into the left and right shoulders of the mice, which were anesthetized with isoflurane in advance. Engraftment efficiency was determined by 68 Ga-FAPI-04 PET/CT imaging (Supplemental Fig. 21A). Immunohistochemical staining (Supplemental Fig. 21B) demonstrated that the pancreatic cancer PDX model used in this study was indeed FAP-overexpressed.

Small-Animal PET Imaging
All PET scans were performed on a Mediso nanoScan PET 122S small-animal PET/CT imaging system. For the 60-min dynamic PET

Biodistribution Study
PDX-bearing mice were injected with 925.0 kBq of 177 Lu-TEFAPI-06 or 177 Lu-TEFAPI-07 for an ex vivo biodistribution study. The mice were killed at 24 h and 96 h after injection, the counts of the different organs were measured with a g-counter, and the data were normalized to percentage injected dose (%ID)/g using 1% of total counts. Histopathologic staining was performed with an antihuman FAP monoclonal antibody (ab207178; Abcam), and hematoxylin and eosin staining was performed as previously described (29).

Radiochemistry and Stability In Vitro
The radiolabeling yield of TEFAPI-06 and TEFAPI-07 (Fig. 1A) was over 90%, and the radiochemical purity was over 99% (n . Lu-TEFAPI-07 in saline and human serum was analyzed using radio-high-performance liquid chromatography, as shown in Supplemental Figure 22.
The radiochemistry purity of both 177 Lu-TEFAPI-06 and 177 Lu-TEFAPI-07 was still over 90% after incubation in saline and human serum for 7 d.

Binding Assay
As shown in Figure 1B and Supplemental Figure 23A, cellular uptake of 68 Ga-FAPI-04 can be significantly inhibited by treatment with cold TEFAPI-06 and TEFAPI-07. The ligand concentrations required for 50% inhibition (half-maximal inhibitory concentration) of TEFAPI-06 and TEFAPI-07 are 12.  2.56 nM and 7.81 6 2.28 nM (Fig. 1C), respectively, which are comparable to that of 68 Ga-FAPI-04 (1.91 6 0.62 nM, Supplemental Fig.  23B). As shown in Figure 1D, both 68 Ga-TEFAPI-06 and 68 Ga-TEFAPI-07 exhibited almost negligible uptake in HT-1080 cells but had significant uptake in HT-1080-FAP cells. We also performed the binding assays in 0.05% human serum albumin (20), with the following results. The half-maximal inhibitory concentrations of TEFAPI-06 and TEFAPI-07 were 11.39 6 1.15 nM and 27.68 6 5.00 nM, respectively, in the presence of albumin. The dissociation constants of TEFAPI-06 and TEFAPI-07 were 4.37 6 0.81 nM and 19.12 6 5.54, respectively, in the absence of albumin. The half-maximal inhibitory concentration and dissociation constant of TEFAPI-07 were slightly impacted by the presence of albumin, which may be the reason why blood clearance was faster than for TEFAPI-06.

Small-Animal PET Imaging
To evaluate the in vivo pharmacokinetics of these 2 radiotracers, dynamic PET imaging of 68 Ga-TEFAPI-06 and 68 Ga-TEFAPI-07 was performed on healthy NOD-SCID mice. The signal in heart peaked rapidly at about 2 min after injection and then declined gradually. For 68 Ga-TEFAPI-06, the signal decreased by 35.70% 6 4.74% from 10 to 60 min after injection, a decrease that was greater than that of 68 Ga-TEFAPI-07 (23.15% 6 2.16%), whereas from 60 to 240 min after injection, the signal decreased by 31.80% 6 1.15% and 40.56% 6 5.25% for 68 Ga-TEFAPI-06 and 68 Ga-TEFAPI-07, respectively, resulting in a similar proportion in the decrease of these 2 radiotracers from 10 to 240 min, at 56.18% 6 2.50% and 54.28% 6 4.98% for 68 Ga-TEFAPI-06 and 68 Ga-TEFAPI-07, respectively. As shown in Figure 2, for both 68 Ga-TEFAPI-06 and 68 Ga-TEFAPI-07, most of the radioactivity was retained in the blood circulation during the monitoring period, and the uptake in other organs, such as the liver, spleen, and kidney, was lower than in the heart or main blood vessels.
To identify the tumor-targeting ability and monitor the in vivo pharmacokinetics quantitatively over a longer period, TEFAPI-06, TEFAPI-07, and FAPI-04 were labeled with the radionuclide 86 Y, which has a half-life of 14.7 h, and the PET imaging was performed using pancreatic cancer PDX-bearing mice. As shown in Figure 3 and Supplemental Figure 24, for both 86 Y-TEFAPI-06 and 86 Y-TEFAPI-07, tumor was completely visible at 2 h after injection. The tumor SUV mean of 86 Y-TEFAPI-06 peaked at 0.73 at 18 h after injection, and that of 86 Y-TEFAPI-07 peaked at 0.81 at 8 h after injection. Then, the tumor SUV mean decreased slowly but still remained high until 36 h after injection, with a value of 0.602 and 0.606 for 86 Y-TEFAPI-06 and 86 Y-TEFAPI-07, respectively. However, the tumor SUV mean of 86 Y-FAPI-04 peaked at 0.35 at 0.2 h after injection and then decreased rapidly, and the areas under the curve for TEFAPI-07 and TEFAPI-06 were 35.5-fold and 37.9-fold that for FAPI-04.
To further confirm the FAP specificity in vivo of these 2 radiotracers, PET imaging of HT-1080-FAP and HT-1080 tumor-bearing mice was performed. As shown in Figure 4, Supplemental Figure 25, and Supplemental Figure 26, the uptake of 86 Y-TEFAPI-06 and 86 Y -TEFAPI-07 in HT-1080-FAP tumors was consistently 2-to 6-fold higher than that in HT-1080 tumors. A blocking study was also performed, as shown in Supplemental Figure 27; tumor uptake decreased at 12 h and 24 h after treatment with cold TEFAPI-06 and TEFAPI-07.

Small-Animal SPECT Imaging
To further characterize these 2 molecules, SPECT imaging was conducted on PDX tumor models for a longer time. As shown in Supplemental Figure 28, high tumor-to-nontargeted-tissue signal ratios were  observed for both 177 Lu-TEFAPI-06 and 177 Lu-TEFAPI-07 until 144 h after injection. The blood circulation properties of these 2 molecules were similar to those found in the previous PET study.

Radiotherapy Study
To make the assessment of therapeutic efficacy more relevant to the clinical setting, pancreatic cancer PDX-bearing mice were used for the indicated radiotherapy study (Fig. 6A). In a comparison to the group treated by saline or 3.7 MBq of 177 Lu-FAPI-04, the groups treated with 1.85 MBq or 3.7 MBq of 177 Lu-TEFAPI-06 and 177 Lu-TEFAPI-07, respectively, showed remarkable suppression of tumor growth (Fig. 6B). No statistical difference in treatment efficacy was observed between 177 Lu-TEFAPI-06 and 177 Lu-TEFAPI-07. This result corroborates the PET imaging and biodistribution studies, as they showed equally high uptake in the tumors. Except for the control group Data are %ID/g. treated with only saline, transient weight loss was observed for all treatment groups, including 177 Lu-FAPI-04, but then returned to the healthy level 7 d after the initial treatment. Hematoxylin and eosin staining of the main organs revealed that side effects from 177 Lu-TEFAPI-06 and 177 Lu-TEFAPI-07 treatment were almost negligible (Supplemental Fig. 29).

DISCUSSION
The purpose of this study was to develop FAPI-based radiopharmaceuticals that are more effective than the existing candidates for FAPtargeted radiotherapy. Two different albumin binders, 4-(p-iodophenyl) butyric acid and truncated Evans blue moieties, were chosen to be attached with FAPI-04. The resulting TEFAPI-06 and TEFAPI-07 were synthesized and radiolabeled with 68 Ga, 86 Y, and 177 Lu. The radiolabeled TEFAPIs exhibited good stability in saline and human serum and high FAP binding affinity in vitro. In addition, SPECT imaging and biodistribution studies of 177 Lu-TEFAPI-06 and 177 Lu-TEFAPI-07 showed that tumor uptake was still notable even at 6 d after the injection. Meanwhile, almost no radioactive signal could be detected for 177 Lu-FAPI-04 at 24 h after injection. We also wondered whether further modifications of the structure may lengthen the blood circulation and, thus, increase the tumor accumulation. However, it can be challenging to balance treatment efficacy against potential side effects from blood circulation.
With regard to the clearance pathway, there was no significant difference in uptake between 177 Lu-TEFAPI-06 and 177 Lu-TEFAPI-07 in tumor and main organs, except for the kidney. For TEFAPI-07, both the PET and the SPECT imaging results showed significantly higher kidney uptake than that of TEFAPI-06. Of note, imaging indicated that there was no obvious clearance of TEFAPI-07 from the kidneys over time, a finding that was consistent with the results of the biodistribution study. Besides, because both TEFAPI-06 and 177 Lu-TEFAPI-07 have relatively longer blood circulation than the classic radiopharmaceuticals, the side effects may not be negligible. Therefore, a comprehensive hematoxylin-and eosin-staining study of major organs was performed, and no tissue damage was observed (Supplemental Fig. 29).
As reported in previous studies, the radiolabeled albumin binder may target the tumor because of enhanced permeability and retention of albumin (30,31). Thus, we were curious about whether the enhanced tumor uptake and retention of 177 Lu-TEFAPI-06 and 177 Lu-TEFAPI-07 are FAP-dependent. The PET imaging results of FAP-positive (HT-1080-FAP) and FAP-negative (HT-1080) tumor-bearing mice showed much higher uptake by FAP-positive tumors than by FAPnegative tumors, demonstrating that the higher tumor uptake was dependent on the FAP-targeting ability in vivo. For the blocking study, the tumor uptake of 68 Ga-FAPI-04 decreased significantly when the mice were treated with cold TEFAPI-06 and TEFAPI-07 until 24 h after injection-a finding that supported the possibility that the prolonged tumor retention of these 2 radiotracers was also dependent mainly on their excellent FAP-targeting ability in vivo.

CONCLUSION
In this study, 2 albumin binder-conjugated FAPIs, denoted as TEFAPI-06 and TEFAPI-07, were developed to optimize the pharmacokinetics of current FAPI radiopharmaceuticals for cancer radiotherapy. Compared with 177 Lu-FAPI-04, both 177 Lu-TEFAPI-06 and 177 Lu-TEFAPI-07 showed enhanced uptake and retention in tumors. The tumor accumulations were highly FAP-selective and resulted in remarkable inhibition of PDX tumor growth, with negligible side effects. Their promising pharmacokinetics warrant further investigations toward clinical translation for the treatment of FAP-positive cancers.