Development of FAPI Tetramers to Improve Tumor Uptake and Efficacy of FAPI Radioligand Therapy

Visual Abstract

the tumor stroma in many epithelial carcinomas, play a pivotal role in tumor growth, tissue remodeling, and immune evasion (1). Fibroblast activation protein (FAP), a type II transmembrane glycoprotein, is overexpressed in cancer-associated fibroblasts but expressed at low levels in normal fibroblasts (2). Therefore, FAP is considered a promising target for tumor imaging and therapy.
In our previous study, a dimeric FAPI molecule, DOTA-2P(FAPI) 2 , was designed and synthesized (12). Preclinical and clinical PET studies have demonstrated that 68 Ga-DOTA-2P(FAPI) 2 exhibits significantly higher tumor uptake and longer retention than 68 Ga-FAPI-46 (12). Similar results were obtained for other FAPI dimers, including DOTAGA, (SA.FAPi) 2 , and BiOncoFAP (13,14). Therefore, polyvalency may be an effective strategy for developing FAP-targeted radiopharmaceuticals with higher tumor uptake because of their increased FAP-recognition ability. Moreover, FAP-targeted radioligand therapy could be more effective if further improvements in tumor retention and absolute uptake are achieved.
In this paper, we report the design, synthesis, and preclinical evaluation of a tetrameric FAPI molecule based on the polyvalency principle. It was constructed on the FAPI-46 motif with 4 diethylene glycol (mini-polyethylene glycol [PEG]) spacers between the 4 FAPI motifs, denoted as 4P(FAPI) 4 . This FAPI tetramer was conjugated to the chelator DOTA or NOTA and labeled with 68 Ga or 64 Cu for PET imaging. It was also labeled with 177 Lu for radioligand therapy applications. This study aimed to investigate the tumor-targeting potential of FAPI tetramers in vitro and in vivo and whether this form is more effective than its monomeric and dimeric analogs.

Chemistry and Radiolabeling
Details of the reagents, chemicals, synthesis route, radiochemistry, and quality control of the FAPI tetramer are described in the supplemental materials (available at http://jnm.snmjournals.org) (12) 4 were diluted with 450 mL of NaAc (0.5 M) and incubated with 50 mL of 64 CuCl 2 (740 MBq in 0.01 M HCl) at 90 C for 20 min. All 3 products were purified using a C18 Plus Short Cartridge (WAT020515; Waters Corp.). Radio-high-performance liquid chromatography was used for quality control.

Preparation of Cell Line-Derived Xenograft Models
All animal experimental procedures were approved by the Animal Care and Ethics Committee of Xiamen University and performed in accordance with the Guidelines for the Care and Use of Animals of the Xiamen University Laboratory Animal Center. For in vivo experiments, 6-wk-old BALB/c nude mice (Beijing Vital River Laboratory Animal Technology Co., Ltd.) were subcutaneously inoculated with HT-1080-FAP or U87 cells (5 3 10 6 in 100 mL of phosphate-buffered saline) in the right shoulder.

Small-Animal PET and SPECT Studies
Dynamic PET, static PET (with or without competition), and SPECT scans with radiolabeled monomeric, dimeric, and tetrameric FAPIs were performed on HT-1080-FAP tumor-bearing mice for pharmacokinetic evaluation. Additionally, static PET with 68 Galabeled monomeric, dimeric, and tetrameric FAPIs was performed and compared in U87MG tumor-bearing mice.

FAP-Targeted Radioligand Therapy
When the tumor volume reached approximately 100 mm 3 , the mice were randomized into 4 groups for radioligand therapy with 177 Lu-labeled monomeric, dimeric, and tetrameric FAPIs (6/group): group A, saline; group B, 29.6 MBq of 177 Lu-FAPI-46; group C, 29.6 MBq of 177 Lu-DOTA-2P(FAPI) 2 ; and group D, 29.6 MBq of 177 Lu-DOTA-4P(FAPI) 4 . The frequency of administering 177 Lu radiopharmaceuticals to U87MG mice was based on the administration frequency used in our previous study on hepatocellular carcinoma patient-derived xenograft tumor models, which showed a significant reduction in tumor uptake after 72 h after injection (15). HT-1080-FAP, a FAP-transfected tumor xenograft with much higher levels of FAP expression than U87MG, was also used in this study. Therefore, the frequency of administration was higher in U87MG tumor-bearing mice (every 72 h, 3 doses in total) than in the HT-1080-FAP models (a single dose). Weight and tumor volume were monitored every 2 d, and the mice were euthanized when the average tumor volume exceeded 1,500 mm 3 , when the tumor was ulcerated, or when weight loss was more than 20%. To further assess radioligand therapy-related toxicity effects, the main organs were collected from the 177 Lu-DOTA-4P(FAPI) 4 group on day 22 after hematoxylin and eosin staining (16).

Statistics
Quantitative data are expressed as mean 6 SD. Statistical analyses were performed using SPSS Statistics for Microsoft Windows, version 22.0 (IBM Corp.). The Student t test was used to determine differences between 2 groups, and 1-way ANOVA was used to compare differences among multiple groups. Statistical significance was set at a P value of less than 0.05.

Synthesis and Radiolabeling
Tetrameric FAPIs containing 4 PEG 3 groups and the chelator DOTA or NOTA were synthesized ( Fig. 1 4 ) after incubation in phosphatebuffered saline and fetal bovine serum via radio-high-performance liquid chromatography analysis, demonstrating the high stability of the products (Supplemental Fig. 3).

Small-Animal PET Imaging of HT-1080-FAP Tumors
To comprehensively evaluate the in vivo pharmacokinetics of 68 Ga-DOTA-4P(FAPI) 4 , a 60-min dynamic PET scan was performed on HT-1080-FAP tumor-bearing mice. As illustrated in Figure 3A, 68 Ga-DOTA-4P(FAPI) 4 was rapidly taken up by the tumor, and the uptake increased from 10 to 60 min after injection. In contrast, the radiotracer uptake rapidly declined over the same period in the heart, kidneys, and liver. Additional late-time-point static scans performed on tumor-bearing mice revealed that tumor uptake remained constant up to 4 h after injection (Fig. 3B). Moreover, 68 Ga-DOTA-4P(FAPI) 4 was eliminated predominantly through the kidneys and bladder, resulting in low background activity and favorable tumor-to-background ratios, especially at later time points. Similar tumor uptake and retention were observed for 68 Ga-DOTA-2P(FAPI) 2 (Fig. 3C); however, a significant decrease in tumor uptake over time was observed on 68 Ga-FAPI-46 PET (Fig. 3D).
Target specificity was evaluated using an in vivo blocking assay. Coinjection with an excess of unlabeled FAPI-46 successfully blocked tumor uptake at 1 h after injection (SUV mean without blocking, 1.87 6 0.08, vs. SUV mean with blocking, 0.16 6 0.03; 92% reduction in tumor uptake), demonstrating that the uptake of the major fraction of 68 Ga-DOTA-4P(FAPI) 4 in tumors was FAP-mediated (Supplemental Fig. 6).
To observe the entire process of tracer accumulation and washout from the tumor tissue, a radionuclide with a longer halflife (12.7 h, 64 Cu) was used to label the   Whole-body SPECT imaging and biodistribution studies were performed to further explore the in vivo characteristics of the 177 Lu-labeled FAPI tetramer. Representative SPECT images of the FAPI tetramer, dimer, and monomer are presented in Figure 6 (3/group), and the ex vivo biodistribution data of the 3 tracers are presented in Supplemental Figure 8 (3 per group). Similar to the observation with 64 Culabeled analogs, HT-1080-FAP tumors clearly contained 177 Lu-labeled dimer and tetramer at all time points examined (Fig. 6). The uptake of 177 Lu-DOTA-4P(FAPI) 4 reached 21.4 6 1.7 %ID/g 24 h after injection, with relatively slow tumor clearance (19.2 6 0.6 %ID/g, 18.8 6 2.1 %ID/g, and 14.8 6 0.9 %ID/g at 48, 72, and 96 h, respectively). The tumor uptake of 177 Lu-DOTA-2P (FAPI) 2 was 17.1 6 3.9 %ID/g 24 h after injection, which was slightly lower than that of 177 Lu-DOTA-4P(FAPI) 4 . Tumor washout of the FAPI dimer was faster than that of the tetramer, with uptake values of 18.8 6 4.1 %ID/g, 13.8 6 2.6 %ID/g, and 13.1 6 0.7 %ID/g at 48, 72, and 96 h, respectively. Unsurprisingly, the tumor uptake of 177 Lu-FAPI-46 was significantly lower than that of 177 Lu-DOTA-4P(FAPI) 4 24 h after injection (3.4 6 0.7 %ID/g, P , 0.001). Because 177 Lu-FAPI-46 was rapidly cleared from the blood and exhibited extremely low accumulation in the tumor 48 h after injection (2.0 6 0.4 %ID/g), no further scans were performed for this radiotracer.

FAP-Targeted Radioligand Therapy with 177 Lu-FAPI Tetramer
In HT-1080-FAP tumor-bearing mice, rapid tumor growth was observed in groups A (control) and B (29.  4 ), significant inhibition of tumor growth was observed, and most tumors started to shrink from day 6 and remained small until days 12-14, after which tumor volumes increased (Fig. 7B). No systemic toxicity due to radioligand therapy, determined by monitoring the body weight of the mice, was observed in any of the 4 groups. To further evaluate the toxic effects, hematoxylin and eosin staining of the selected nontarget organs was performed, which revealed no differences between the control and radioligand therapy groups (Supplemental Fig. 9). In U87MG tumor-bearing mice, tumors in the control and 177 Lu-FAPI-46 therapy groups both demonstrated fast growth, and  all mice (6/6) in the control group and half the mice (3/6) in the 177 Lu-FAPI-46 therapy group were euthanized by day 14 because of excessive tumor volumes. Although a better antitumor efficacy was observed in the 177 Lu-FAPI dimer group (median survival time not reached) than in the control group (median survival, 12 d) and the 177 Lu-FAPI-46 group (median survival, 14 d), the 177 Lu-FAPI tetramer (median survival time not reached) yielded the greatest inhibition of tumor growth among all 4 groups (Fig. 7). In brief, the tumor volume in 177 Lu-FAPI tetramer group was significantly less than in the FAPI dimer, FAPI-46, and control groups at day 14 after treatment (140. 28

DISCUSSION
In the past 3 y, many clinical studies have explored the potential of FAP-targeted radioligand therapy with 177 Lu-or 90 Y-labeled FAPIs (8,9). However, most have revealed unsatisfactory therapeutic responses, mainly because of fast blood clearance accompanied by relatively short tumor retention. Therefore, various strategies have been developed to prolong the in vivo half-life of radiolabeled FAPIs to improve tumor uptake and retention.
An important strategy to enhance tumor uptake and retention is to harness the polyvalency effect of multimerization, which has been used in the development of arginylglycylaspartic acid peptides to improve their pharmacokinetics (17). Recently, we applied the multivalency concept to develop a dimeric FAPI molecule, DOTA-2P(FAPI) 2 (12), which demonstrated enhanced tumor uptake and retention properties for dimers compared with monomers in patient-derived xenografts and patients with cancer. On the basis of those results, we synthesized tetrameric FAPI molecules with 4 repeating FAPI-46 units connected by 4 mini-PEG spacers. We hypothesized that multimerization to tetrameric FAPIs would further improve their tumor accumulation and retention because of adequate contact with the FAP-binding pocket located in the extracellular segment of cancer-associated fibroblasts.
The high labeling yield, radiochemical purity, and stability of the FAPI tetramer indicate that it is a convenient precursor for radiolabeling and application. Subsequently, radioligand-binding assays were used to examine the FAP-binding affinity of FAPI tetramers, dimers, and monomers.  However, comparable 50% inhibitory concentrations were observed for all 3 FAPI variants. Multimeric FAPI molecules are not necessarily multivalent. The key to bivalency and tetravalency is the distance between the binding motifs. In this study, a FAP-binding affinity of the FAPI tetramer and dimer comparable to that of FAPI-46 indicates that the distance between binding motifs in DOTA-4P(FAPI) 4 and DOTA-2P(FAPI) 2 may not be sufficiently long for them to achieve tetravalency or bivalency. In addition, the bivalency and tetravalency of multimeric FAPI molecules also depend on FAP density. If FAP density is low, the distance between neighboring FAP sites will be long, and it may be more difficult for multiple multimers to simultaneously bind to FAP binding sites.
The tetramer 68 Ga-DOTA-4P(FAPI) 4 exhibited prominent uptake in the FAP-transfected tumor xenograft HT-1080-FAP, and its excretion route was primarily through the kidneys. However, it exhibited a similar initial tumor uptake and slightly longer tumor retention than those of 68 Ga-DOTA-2P(FAPI) 2 and 68 Ga-FAPI-46, as may be explained by the intense FAP expression in this special tumor xenograft. In another tumor xenograft, U87MG, the tumor uptake of 68 Ga-DOTA-4P(FAPI) 4 was significantly higher than that of 68 Ga-DOTA-2P(FAPI) 2 and 68 Ga-FAPI-46. In the blocking study, the tumor uptake of 68 Ga-DOTA-4P(FAPI) 4 decreased significantly when the mice were injected with unlabeled FAPI-46 1 h after injection, suggesting that the high tumor uptake of 68 Ga-DOTA-4P(FAPI) 4 was primarily a factor of its excellent FAP-targeting ability in vivo.
However, the relatively short half-life of 68 Ga limits the observation time of tumor retention. Therefore, the FAPI tetramer and dimer were labeled with 64 Cu to further evaluate their in vivo characteristics. The tetramer 64 Cu-NOTA-4P(FAPI) 4 exhibited a slightly higher initial tumor uptake and longer retention than 64 Cu-NOTA-2P(FAPI) 2 . Compared with the molecular size of the FAPI monomer and dimer, the larger size of the FAPI tetramer may explain its longer circulation time and slower tumor washout. In contrast, as the greater number of FAP binding sites on FAPI tetramers will increase the local concentration of other FAPI motifs in the vicinity of FAP sites, the locally increased FAPI concentration may explain the higher tumor uptake of radiolabeled FAPI tetramers and dimers than of their monomeric analogs (18). The higher liver uptake of 64 Cu-labeled radiopharmaceuticals may be attributed to the dissociation of free copper ions from the radiopharmaceuticals in vivo (19,20), which was also observed in previous studies. The liver uptake of 64 Cu-NOTA-arginylglycylaspartic acid-bombesin was relatively lower than that of other 64 Cu-DOTA radiotracers but higher than that of 68 Ga-NOTA-arginylglycylaspartic acid-bombesin, possibly because of the higher chelating ability of NOTA with 68 Ga than of NOTA with 64 Cu (21). However, other factors, such as radiotracer stability and metabolism, can also contribute to the increased liver uptake. Increased liver uptake of a 64 Cu-NOTA agent was also reported in PEG2-RM26 studies, partly because of the transchelation of 64 Cu 21 to the serum components or superoxide dismutase that can accumulate in the liver tissue (22). Further studies are needed to fully elucidate the mechanisms underlying the liver uptake of 64 Cu-labeled radiotracers.
Compared with FAPI dimers and monomers, the FAPI tetramer exhibited significantly higher uptake in certain nontarget organs, especially the kidney and liver, as reflected by PET and SPECT imaging and biodistribution studies. The relatively high uptake of the FAPI tetramer by the kidneys may be explained by different mechanisms. First, we speculate that the increased renal uptake of the FAPI tetramer may be partially related to the 4 mini-PEG spacers. PEGylation is a strategy widely used to improve the in vivo pharmacokinetics of radiotracers, induce hydrophilicity, and increase kidney uptake (23). Additionally, the difference in charge between the 3 FAPI molecules may cause differences in tubular reabsorption, as reported in previous studies (24). Because of the presence of more guanidine groups, tetrameric FAPI is more positively charged than dimeric and monomeric FAPI. The larger molecular size of the FAPI tetramer could cause a longer circulation time and greater retention in the liver. The fact that the background of 68 Ga-labeled FAPI tetramer was higher than that of the dimer and monomer may have had unfavorable effects on diagnostic application. However, the FAPI tetramer applied in our study was designed to improve tumor uptake and retention so as to enhance the antitumor efficacy of FAP-targeted radioligand therapy. Furthermore, FAPI monomers, such as FAPI-04 and FAPI-46, are excellent PET imaging agents for detecting FAP-positive lesions because of their favorable pharmacokinetics and high binding specificity to FAP.
However, the multimerization strategy may be a double-edged sword in the development of radiopharmaceuticals. In addition to improved tumor uptake and retention, it results in higher radiotracer uptake in normal organs, particularly the kidneys and liver. The increased accumulation of radioactivity in normal organs may result in the delivery of unnecessary radiation doses, which may affect the future clinical translation of these molecules into viable treatments. Whether the positive effects of increased tumor uptake offset the potential side effects of increased liver and kidney uptake is unclear. Increased liver and kidney uptake may be undesirable; however, it may be an acceptable trade-off if the benefits of increased tumor uptake are significant. In tumors with high expression of FAP, such as HT-1080-FAP, radioligand therapy with a FAPI dimer may lead to similar antitumor efficacy but fewer side effects than for a FAPI tetramer. However, 68 Ga PET imaging and 177 Lu-radioligand therapy in U87MG tumor-bearing mice revealed that the tetramer itself acts as a double titer of the dimer, thereby improving its antitumor efficacy. Therefore, radioligand therapy with a FAPI tetramer may be more appropriate than that with a FAPI dimer in tumors with moderate or mild expression of FAP. The selection of the dimer or tetramer ultimately depends on the specific circumstances of the treatment objectives and the potential benefits and risks associated with each option. Therefore, appropriate modifications by changing the linker or chelator are needed to improve the pharmacokinetics of FAPI-based radiopharmaceuticals (25), especially to improve their FAP-targeting capabilities and reduce radiotracer accumulation in noncancerous organs.

CONCLUSION
The radiolabeled FAPI tetramer exhibited higher accumulation and longer retention in the tumor than did its dimeric and monomeric counterparts. The improved pharmacologic properties of 177 Lu-DOTA-4P(FAPI) 4 resulted in excellent antitumor ability in HT-1080-FAP and U87MG tumor-bearing mice. The information obtained here may guide the future development of FAP-targeted imaging and radioligand therapy.

KEY POINTS
QUESTION: Compared with FAPI monomers and dimers, do FAPI tetramers demonstrate enhanced tumor uptake, prolonged tumor retention, and an improved radioligand therapeutic ability?
PERTINENT FINDINGS: FAPI tetrameric radiopharmaceuticals exhibited significantly increased tumor uptake and retention compared with their monomeric and dimeric counterparts. The 177 Lu-FAPI tetramer demonstrated remarkable inhibition of tumor growth in both HT-1080-FAP and U87MG tumors, with negligible side effects.
IMPLICATIONS FOR PATIENT CARE: The formation of FAPI tetramers via multimerization is a promising strategy in the development of FAP-targeted radiopharmaceuticals.