Evaluation of Astatine-211-Labeled Fibroblast Activation Protein Inhibitor (FAPI): Comparison of Different Linkers with Polyethylene Glycol and Piperazine

Fibroblast activation proteins (FAP) are overexpressed in the tumor stroma and have received attention as target molecules for radionuclide therapy. The FAP inhibitor (FAPI) is used as a probe to deliver nuclides to cancer tissues. In this study, we designed and synthesized four novel 211At-FAPI(s) possessing polyethylene glycol (PEG) linkers between the FAP-targeting and 211At-attaching moieties. 211At-FAPI(s) and piperazine (PIP) linker FAPI exhibited distinct FAP selectivity and uptake in FAPII-overexpressing HEK293 cells and the lung cancer cell line A549. The complexity of the PEG linker did not significantly affect selectivity. The efficiencies of both linkers were almost the same. Comparing the two nuclides, 211At was superior to 131I in tumor accumulation. In the mouse model, the antitumor effects of the PEG and PIP linkers were almost the same. Most of the currently synthesized FAPI(s) contain PIP linkers; however, in our study, we found that PEG linkers exhibit equivalent performance. If the PIP linker is inconvenient, a PEG linker is expected to be an alternative.


Introduction
The most important aspect of molecular-targeted cancer therapy is the selection of excellent molecular targets. Cancer specificity must be ensured in order to prevent adverse effects. Cancer tissue is a heterogeneous population in which cancer cells and various other cell types are mixed. Therefore, especially in cancer types that possess stromal cells that play a wall-like role, it is difficult to deliver drugs to them. By targeting fibroblast-activated protein (FAP), which is also a marker for stromal cells, we may be able to efficiently target cancer tissues.
FAP inhibitor (FAPI), which has recently attracted attention as a cancer-targeting molecule, causes the accumulation of FAP, which is overexpressed in the tumor stroma and is involved in cancer growth. From a structural perspective, the presence of carbonitrile on the pyrrolidine ring of FAPI increases its affinity for FAP [1,2]. Furthermore, Jansen et al. argued that the quinoline moiety of FAPI forms a cation-π interaction with the guanidine moiety of Arg123 in FAP [3]. Lindner et al. compared the EC 50 of FAPI(s) using competition experiments and found that the pharmacokinetics changed significantly according to the structure of the linker [4]. PET imaging using 68 Ga-labeled FAPI-02, FAPI-04, and FAPI-13 revealed favorable tumor accumulation in vivo in HT-1080-FAP xenografted mice. Diagnostic positron emission tomography/computed tomography (PET/CT) scans using FAPI have been demonstrated to be useful for the development of clinical applications in patients [5,6]. Using 68 Ga-labeled FAPI-02 and FAPI-04, excellent tumor accumulation was observed in a mouse sarcoma model and in pancreatic cancer patients [4,7]. In addition, new FAPI(s) (e.g., FAPI-21, FAPI-34, FAPI-46, and FAPI-74) have been developed to further improve their beneficial effects [8][9][10][11][12][13][14]. The 68 Ga-Labeled FAPI dimer linked by polyethylene glycol (PEG) with DOTA exhibited outstanding tumor uptake and longer tumor retention than 68 Ga-FAPI-46 [15]. 18 F-labeled FAPI with a PEG linker without piperazine (PIP) has also been employed for clinical PET/CT imaging and nasopharyngeal cancers, in addition to 18 F-FDG [16]. 18 F-NOTA-FAPI has also been used for PET/CT imaging of patients [17]. Biodistribution of 68 Ga-DOTA.SA.FAPI in patients was also investigated, and clear accumulation was shown by PET/CT imaging [18]. As it was possible to clearly visualize the cancerous lesion, it is speculated that [ 68 Ga] Ga-DATA5m. SA.FAPI is a promising substrate for PET/CT imaging [19]. As the clinical application of FAPI-PET steadily progresses, an efficient and stable supply of 18 F-labeled FAPI is urgently needed. Recently, we developed an automated radiosynthesis method for [ 18 F]AIF-FAPI-74 [20].
FAPI, which was created by taking advantage of the cancer specificity of FAP, has been shown to have excellent performance as a PET/CT probe [6]. Thus, FAPI is a promising cancertreating molecule. We synthesized 177 Lu-labeled FAPI-46 ( 177 Lu-FAPI) and administered it to PANC-1 xenograft mice [21]. 177 Lu-FAPI showed a strong antitumor effect with 30 MBq of radioactivity. We also labeled FAPI-04 with 225 Ac as an α-emitter and administered it to PANC-1 xenograft mice [22]. Although a low radioactivity level of 225 Ac was used, excellent anti-tumor activity was achieved owing to the irradiation of particles, which have high linear energy transfer (LET). Furthermore, we focused on 211 At as an emitter. 211 At has a relatively short half-life of 7.2 h, and exhibits a simple decay chain without harmful daughter nuclides ( 207 Bi and 207 Pb). Therefore, if α nuclear medicine treatment becomes widely used, it is expected that the length of hospital stay would be shortened. Hospitalization may not be necessary in the future. If so, α nuclear medicine treatment will be able to further contribute to improving the quality of life of cancer patients.
Our  [23]. They also claimed that no toxicity was observed in the kidneys, liver, stomach, or thyroid tissue. We also synthesized 211 At-FAPI(s) using dihydroxyboryl astatine substitution reaction [24]. We have previously established an astatination method that does not require toxic reagents [24,25]. In this study, we investigated the usefulness of 211 At-FAPI(s) synthesized using an established safe method and a linker in the FAPI structure.

Establishment of FAPI1 to 5
PIP has been mainly employed as a linker for FAPI to improve its hydrophilicity. However, the PIP linker FAPI is degraded in the body [16]. The PEG linker is more stable than the PIP linker. Moreover, the production of byproducts during 211 At treatment can be suppressed in the absence of an amino group. To confirm the effectiveness of the PEG linker FAPI, we synthesized four PEG linker FAPI(s). Their structures are illustrated in Figure 1a. FAPI1 has a simple PEG linker. In FAPI2, to reserve the space for the FAP binding site, the length of the PEG linker was extended to increase the distance between the radionuclide-labeling site and the biologically active site. FAPI3 increases stability by adding Fluorine, and FAPI4 decreases lipid solubility and increases urinary excretion by adding glucose. We also synthesized a PIP linker, FAPI, to compare its efficacy with that of the PEG linker, FAPI(s). The FAPI(s) were labeled with 211 At or 131 I at the same site in their structures (Figure 1b,c). The radiochemical purities (RCPs) of 211 At-FAPI1, 2, 3, 4, and 211 At-FAPI5 were 100%, 100%, 96%, 78%, and 95%, and those of 131 I-FAPI1 and 131 I -FAPI5 were 91% and 99%, respectively. The labeling efficiency of 211 At-labeled FAPI with a PEG linker decreased with increasing complexity. The purity of 131 I-labeled FAPI1 was lower than FAPI5.

Establishment of FAPI1 to 5
PIP has been mainly employed as a linker for FAPI to improve its hydrophilicity. However, the PIP linker FAPI is degraded in the body [16]. The PEG linker is more stable than the PIP linker. Moreover, the production of byproducts during 211 At treatment can be suppressed in the absence of an amino group. To confirm the effectiveness of the PEG linker FAPI, we synthesized four PEG linker FAPI(s). Their structures are illustrated in Figure 1a. FAPI1 has a simple PEG linker. In FAPI2, to reserve the space for the FAP binding site, the length of the PEG linker was extended to increase the distance between the radionuclide-labeling site and the biologically active site. FAPI3 increases stability by adding Fluorine, and FAPI4 decreases lipid solubility and increases urinary excretion by adding glucose. We also synthesized a PIP linker, FAPI, to compare its efficacy with that of the PEG linker, FAPI(s). The FAPI(s) were labeled with 211 At or 131 I at the same site in their structures (Figure 1b,c). The radiochemical purities (RCPs) of 211 At-FAPI1, 2, 3, 4, and 211 At-FAPI5 were 100%, 100%, 96%, 78%, and 95%, and those of 131 I-FAPI1 and 131 I -FAPI5 were 91% and 99%, respectively. The labeling efficiency of 211 At-labeled FAPI with a PEG linker decreased with increasing complexity. The purity of 131 I-labeled FAPI1 was lower than FAPI5.
The synthesis results and labeling procedures for each FAPI are shown in the Supplementary Materials S1.   The synthesis results and labeling procedures for each FAPI are shown in the Supplementary Materials S1.

Comparison of Cellular Uptake between FAPI1 to 5
In an uptake experiment using HEK293 cells (Figure 2), the uptake of 211 At-FAPI2 and 3 was slightly lower, whereas that of 211 At-FAP1, 4, and 5 appeared to be the same. Because FAPI1-4 are PEG linker compounds and FAPI5 is a PIP linker compound, FAPI1, the simplest PEG linker, and FAPI5, a PIP linker, were compared. The difference was not notable in A549 cells (FAPα/A549), and the uptake followed the same trend as-overexpressing HEK293 (FAPα/293) cells. No significant difference was observed between the uptake of 211 At-FAPI1 and 211 At-FAPI5 by A549 cells (Figure 3). As expected, there were significant differences in the uptake between the parental and FAPα-overexpressing cell lines for all compounds, from 211 At-FAPI1 to 211 At-FAPI5. All FAPI(s) had specificity to FAPα.
compound. (c) PIP linker compound. PIP linker compound was synthesized based on the already reported compound for comparison with the PEG linker compound.

Comparison of Cellular Uptake between FAPI1 to 5
In an uptake experiment using HEK293 cells (Figure 2), the uptake of 211 At-FAPI2 and 3 was slightly lower, whereas that of 211 At-FAP1, 4, and 5 appeared to be the same. Because FAPI1-4 are PEG linker compounds and FAPI5 is a PIP linker compound, FAPI1, the simplest PEG linker, and FAPI5, a PIP linker, were compared. The difference was not notable in A549 cells (FAPα/A549), and the uptake followed the same trend as-overexpressing HEK293 (FAPα/293) cells. No significant difference was observed between the uptake of 211 At-FAPI1 and 211 At-FAPI5 by A549 cells (Figure 3). As expected, there were significant differences in the uptake between the parental and FAPα-overexpressing cell lines for all compounds, from 211 At-FAPI1 to 211 At-FAPI5. All FAPI(s) had specificity to FAPα.

Comparison of Body Distributions of FAPI1 or FAPI5 between Two Nuclides
Iodine physiologically accumulates in the thyroid via Na + /I − symporter (NIS/SLC7A5), but 131 I-FAPI1 and 131 I-FAPI5 accumulated minimally in the thyroid ( Table  1). The expression of NIS is also known to be high in the gastrointestinal tract, and high accumulation of both 131 I-FAPI1 and 131 I-FAPI5 were observed in the gastrointestinal tract. 211 At-FAPI1 and 211 At-FAPI5 also accumulated less in the thyroid, stomach, and kidneys

Comparison of Cellular Uptake between FAPI1 to 5
In an uptake experiment using HEK293 cells (Figure 2), the uptake of 211 At-FAPI2 and 3 was slightly lower, whereas that of 211 At-FAP1, 4, and 5 appeared to be the same. Because FAPI1-4 are PEG linker compounds and FAPI5 is a PIP linker compound, FAPI1, the simplest PEG linker, and FAPI5, a PIP linker, were compared. The difference was not notable in A549 cells (FAPα/A549), and the uptake followed the same trend as-overexpressing HEK293 (FAPα/293) cells. No significant difference was observed between the uptake of 211 At-FAPI1 and 211 At-FAPI5 by A549 cells (Figure 3). As expected, there were significant differences in the uptake between the parental and FAPα-overexpressing cell lines for all compounds, from 211 At-FAPI1 to 211 At-FAPI5. All FAPI(s) had specificity to FAPα.

Comparison of Body Distributions of FAPI1 or FAPI5 between Two Nuclides
Iodine physiologically accumulates in the thyroid via Na + /I − symporter (NIS/SLC7A5), but 131 I-FAPI1 and 131 I-FAPI5 accumulated minimally in the thyroid ( Table  1). The expression of NIS is also known to be high in the gastrointestinal tract, and high accumulation of both 131 I-FAPI1 and 131 I-FAPI5 were observed in the gastrointestinal tract. 211 At-FAPI1 and 211 At-FAPI5 also accumulated less in the thyroid, stomach, and kidneys

Comparison of Body Distributions of FAPI1 or FAPI5 between Two Nuclides
Iodine physiologically accumulates in the thyroid via Na + /I − symporter (NIS/SLC7A5), but 131 I-FAPI1 and 131 I-FAPI5 accumulated minimally in the thyroid (Table 1). The expression of NIS is also known to be high in the gastrointestinal tract, and high accumulation of both 131 I-FAPI1 and 131 I-FAPI5 were observed in the gastrointestinal tract. 211 At-FAPI1 and 211 At-FAPI5 also accumulated less in the thyroid, stomach, and kidneys compared with Na 211 At [26], whereas astatine labeled compounds accumulated at higher levels in the thyroid gland 3 h after injection (Table 2). High accumulation of both 211 At-FAPI1 and 211 At-FAPI5 was observed in the gastrointestinal tract. The accumulation in tumors of 211 At-labeled FAPI1 and FAPI5 were more than twice that of 131 I-labeled FAPI(s). For both nuclides, FAPI compounds were highly excreted in the small intestine and colon. When iodine was used, comparing 1 h (a) and 3 h (b), the amount of excretion increased, but its accumulation in the tumors was low. On the other hand, at 1 h after injection, the urine excretion of 131 I-FAPI1 was 30.79 ± 15.80% ID, and that of 131 I-FAPI5 was 9.21 ± 3.52% ID. At 3 h after injection, the urine excretion of 131 I-FAPI1 was 14.46 ± 4.04% ID, and that of 131 I-FAPI5 was 3.99 ± 0.91% ID. These values were calculated using the urine retained in the bladder, indicating that 131 I-FAPI1 was excreted faster. At 1 h after 211 At-FAPI(s) injection, the urine excretion of 211 At-FAPI1 was 4.92 ± 1.74% ID, and that of 211 At-FAPI5 was 0.26 ± 0.19% ID. The feces excretion of 211 At-FAPI1 was 2.68 ± 1.13% ID, and that of 211 At-FAPI5 was 5.19 ± 1.62% ID. At 3 h after injection, the urine excretion of 211 At-FAPI1 was 0.03 ± 0.01% ID, and that of 211 At-FAPI5 was 0.63 ± 0.55% ID. The feces excretion of 211 At-FAPI1 was 16.44 ± 4.71% ID, and that of 211 At-FAPI5 was 7.65 ± 2.11% ID. The urine excretion values were calculated using the urine retained in the bladder, which indicated that 211 At-FAPI1 was excreted faster. After measuring the concentration in each organ, we also measured the rest of the carcass, feces, and floor paper by γ-counter to estimate all distribution of injected dose. While measuring the floor paper, urination was observed, but not much urination was observed after administration. These results showed faster excretion of FAPI1 than FAPI5 and iodine compared with astatine. The difference between FAPI1 and FAPI5 was not marked, and it can be said that they were equivalent. The iodine and astatine levels were clearly different. Particularly in tumors, it was clear that 211 At-FAPII(s) had superior terms of tumor retention compared with 131 I-FAPI(s) (Tables 1 and 2).

Comparison of Anti-Tumor Effects due to Differences in Linker
In terms of therapeutic effects, 211 At-FAPI1 showed better therapeutic effects. No obvious toxicity was observed for 211 At-FAPI1 or 211 At-FAPI5 (Figure 4a). For both compounds, a single administration did not result in significant tumor regression; however, a clear growth inhibitory effect was observed compared to the control (Figure 4b). A significant difference in tumor weight was observed between the groups (control vs. 211 At-FAPI1, p < 0.001; control vs. 211 At-FAPI5, p < 0.05; 211 At-FAPI1 vs. 211 At-FAPI5, p < 0.05). 211 At-FAPI1 was found to have a greater tumor growth inhibitory effect than 211 At-FAPI5 (Figure 4c). At the time of dissection, organs other than the tumor were observed; however, no metastatic tissues were confirmed, and no obvious abnormalities were observed visually. The TLC analysis results of the administered 211 At-FAPI(s) for the therapeutic experiment are shown in Figure 5. RCPs of 90% or more were ensured in both cases.

Comparison of Anti-Tumor Effects due to Differences in Linker
In terms of therapeutic effects, 211 At-FAPI1 showed better therapeutic effects. No obvious toxicity was observed for 211 At-FAPI1 or 211 At-FAPI5 (Figure 4a). For both compounds, a single administration did not result in significant tumor regression; however, a clear growth inhibitory effect was observed compared to the control (Figure 4b). A significant difference in tumor weight was observed between the groups (control vs. 211 At-FAPI1, p < 0.001; control vs. 211 At-FAPI5, p < 0.05; 211 At-FAPI1 vs. 211 At-FAPI5, p < 0.05). 211 At-FAPI1 was found to have a greater tumor growth inhibitory effect than 211 At-FAPI5 ( Figure  4c). At the time of dissection, organs other than the tumor were observed; however, no metastatic tissues were confirmed, and no obvious abnormalities were observed visually. The TLC analysis results of the administered 211 At-FAPI(s) for the therapeutic experiment are shown in Figure 5. RCPs of 90% or more were ensured in both cases.

Discussion
During the preparation of 211 At-FAPI with PEG linker, the radiochemical yields (RCYs) of 211 At-FAPI1 and 2 were high (100% and 99%, respectively). When producing 211 At-FAPI5, a byproduct derived from the amino group of PIP was observed (~10%). However, PEG-linked FAPI suppressed the byproduct production. Therefore, non-PIP linkers are suitable for drug manufacturing applications. The RCYs of 211 At-FAPI3 and 4 were not very high (45% and 15%, respectively), because astatination was carried out at 50 °C to prevent production of a byproduct caused by fluorides (Figure 1a). Additionally, the presence of glucosamine units decreased the astatination reactivity. Based on these results, we presume that 211 At-FAPI1 is suitable for a stable drug supply.
Cellular uptake and FAPα selectivity were evaluated using an FAPα-overexpressing strain derived from HEK293 cells (parental; HEK293, FAPα overexpression; FAPα/293). We hypothesized that polarity is important for FAPI internalization. Based on 211 At-FAPI1, we also designed 211 At-FAPI2, which has a long PEG linker to evaluate difference of FAPα and internalization into cells (Figure 1a). However, FAPα selectivity and cellular uptake did not increase. According to Lindner et al., the presence of fluorine atoms on the pyrrole ring of the FAPI moiety increased its affinity for FAP [4]. Therefore, we compared the FAP α -binding abilities of 211 At-FAPI1 and 3. Although the RCY of 211 At-FAPI3 was not

Discussion
During the preparation of 211 At-FAPI with PEG linker, the radiochemical yields (RCYs) of 211 At-FAPI1 and 2 were high (100% and 99%, respectively). When producing 211 At-FAPI5, a byproduct derived from the amino group of PIP was observed (~10%). However, PEG-linked FAPI suppressed the byproduct production. Therefore, non-PIP linkers are suitable for drug manufacturing applications. The RCYs of 211 At-FAPI3 and 4 were not very high (45% and 15%, respectively), because astatination was carried out at 50 °C to prevent production of a byproduct caused by fluorides (Figure 1a). Additionally, the presence of glucosamine units decreased the astatination reactivity. Based on these results, we presume that 211 At-FAPI1 is suitable for a stable drug supply.
Cellular uptake and FAPα selectivity were evaluated using an FAPα-overexpressing strain derived from HEK293 cells (parental; HEK293, FAPα overexpression; FAPα/293). We hypothesized that polarity is important for FAPI internalization. Based on 211 At-FAPI1, we also designed 211 At-FAPI2, which has a long PEG linker to evaluate difference of FAPα and internalization into cells (Figure 1a). However, FAPα selectivity and cellular uptake did not increase. According to Lindner et al., the presence of fluorine atoms on the pyrrole ring of the FAPI moiety increased its affinity for FAP [4]. Therefore, we compared the FAP α -binding abilities of 211 At-FAPI1 and 3. Although the RCY of 211 At-FAPI3 was not

Discussion
During the preparation of 211 At-FAPI with PEG linker, the radiochemical yields (RCYs) of 211 At-FAPI1 and 2 were high (100% and 99%, respectively). When producing 211 At-FAPI5, a byproduct derived from the amino group of PIP was observed (~10%). However, PEGlinked FAPI suppressed the byproduct production. Therefore, non-PIP linkers are suitable for drug manufacturing applications. The RCYs of 211 At-FAPI3 and 4 were not very high (45% and 15%, respectively), because astatination was carried out at 50 • C to prevent production of a byproduct caused by fluorides (Figure 1a). Additionally, the presence of glucosamine units decreased the astatination reactivity. Based on these results, we presume that 211 At-FAPI1 is suitable for a stable drug supply.
Cellular uptake and FAPα selectivity were evaluated using an FAPα-overexpressing strain derived from HEK293 cells (parental; HEK293, FAPα overexpression; FAPα/293). We hypothesized that polarity is important for FAPI internalization. Based on 211 At-FAPI1, we also designed 211 At-FAPI2, which has a long PEG linker to evaluate difference of FAPα and internalization into cells (Figure 1a). However, FAPα selectivity and cellular uptake did not increase. According to Lindner et al., the presence of fluorine atoms on the pyrrole ring of the FAPI moiety increased its affinity for FAP [4]. Therefore, we compared the FAPα-binding abilities of 211 At-FAPI1 and 3. Although the RCY of 211 At-FAPI3 was not suitable, the RCP increased to 94% after purification, which precluded concerns regarding the effects of unlabeled 211 At components. Unexpectedly, increased FAPα selectivity was not observed (Figure 2). Considering the structure of 211 At-FAPI13, its hydrophobicity is relatively high compared with that of the FAPI series possessing chelates (e.g., FAPI-02, FAPI-04, FAPI-46, and FAPI-74). Thus, we designed and synthesized 211 At-FAPI 4 with a highly oxidized glucosamine unit to increase its hydrophilicity. However, the performance did not change (Figure 2).
The results of our investigation indicated that FAPα selectivity and cellular uptake were almost the same for 211 At-FAPI1 and 211 At-FAPI5. Considering the molecular size and efficiency of astatination, we employed 211 At-FAPI1 for comparison with 211 At-FAPI5 in in vitro experiments. 211 At-FAPI1 exhibited significant FAPα selectivity, indicating fewer side effects on normal tissues. In contrast, 211 At-FAPI5 exhibited high cellular uptake, indicating that a high therapeutic effect can be expected at a small dose. Next, we compared the biodistribution of 211 At-FAPI1 and 211 At-FAPI5. At 1 h after injection, tumor accumulation of 211 At-FAPI1 and 211 At-FAPI5 was 2.15 ± 0.24% ID and 1.40 ± 1.14% ID, which greatly surpassed the accumulation of 225 Ac-FAPI-04 [22]. After 3 h, the accumulation in tumor cells increased for 211 At-FAPI1. In contrast, 211 At-FAPI5 levels decreased (Table 2). Both FAPIs are excreted in the small intestine and urine, and FAPI1 is excreted more rapidly than FAPI5. These trends remained the same even when the nuclides were different. After 1 h, both nuclides and both linkers were present at high doses in the small intestine, and after 3 h, the dose in the colon was high; therefore, they were clearly presumed to be contained in the intestines. Translocation of the 211 At or 131 I label to the gastrointestinal contents might be thought to be via digestive juice. Excreted fecal doses were not very high and might have been reabsorbed in the colon. The amount of excreted urine from iodine-labeled FAPIs increased compared with those labeled by astatine; however, its accumulation in tumors was low, and iodine might be unsuitable as a therapeutic agent for labeling FAPI compounds. Compared with FAPI5, FAPI1 exhibited higher retention in tumors and higher non-specific excretion. Because 211 At could retain more compounds in tumors than 131 I, we conducted a therapeutic experiment using 211 At-FAPI1 and 5. Although the data are not shown in this article, it has been confirmed that FAPI uptake into cells is competitively inhibited when an unlabeled FAPI compound is present in excess [24]. Because labeling with 131 I was inefficient, the possibility that excess unlabeled substances were mixed with the labeled products cannot be ruled out. In other words, the lower % ID/g values for tumors compared to the 211 At-labeled products in this study may be due to the poor labeling rate. The astatine-labeled PIP linker, the FAPI compound (FAPI5), inevitably produced two spots on the TLC ( Figure 5). However, this could not be removed using Oasis HLB column purification (Waters Corporation, Milford, MA, USA). Therefore, it is necessary to investigate the properties of this spot, which is not the main spot. However, this contaminant may explain why tumor accumulation was slightly lower than that of the PEG linker (FAPI1). FAPI5 is difficult to handle because it readily adsorbs onto tubes and syringes. One reason for this is its high structural lipophilicity. This is also an important factor that makes the compound easy to handle for use as a therapeutic agent in nuclear medicine.
Outside Japan, 225 Ac has attracted attention in alpha-ray nuclear medicine therapy, and clinical trials are underway [27][28][29]. We are also investigating the use of 225 Ac, but there are still problems with its application in medicine, such as the lack of a mass production method in Japan.

Materials and Methods
The synthesis procedures for the precursors are described in the Supplementary Materials (Supplementary Materials S1). We established a synthetic methodology to secure intellectual property. Radiolabeled products were prepared using the following methodology:

Synthesis of 211 At-FAPI Compounds
211 At was acquired from the National Institute for Quantum Science and Technology (QST, Takasaki, Gunma, Japan) and RIKEN (Wako, Saitama, Japan) through a supply plat-form for short-lived radioisotopes. 211 At was produced according to the 209 Bi(α, 2n) 211 At reaction and was separated from the Bi target using the dry distillation method. The separated 211 At was then dissolved in pure water. FAPI(s) with a dihydroxyboryl group (B-FAPI) as a precursor of 211 At-FAPI compounds were prepared via chemical synthesis according to previously published schemes. 211 At was produced via a 209 Bi (α, 2n) 211 At nuclear reaction using a cyclotron. The purification of 211 At from the irradiated Bi target was performed by dry distillation. All the FAPI compounds were synthesized via a dihydroxyboryl astatine substitution reaction with the corresponding B-FAPI1, 2, 3, 4, and B-FAPI5. In total, 10 µL of aqueous B-FAPI (1 mg/mL), 10 µL of aqueous NaHCO 3 (7% w/v), 90 µL of H 2 O, 2-19 µL aqueous 211 At (0.28-1.01 MBq) and 30 µL of aqueous KI (0.1 mol/L) were added to a polypropylene (PP) tube at room temperature. After being heated to 50 • C or 80 • C for 45 min, the reaction mixture was cooled to ambient temperature. The radiochemical yields (RCYs) of 211 At-FAPI1, 2, 3, 4, and 211 At-FAPI5 were 100%, 99%, 45%, 15%, and 90%, respectively. The reaction mixture was then purified using an Oasis

Synthesis of 131 I-FAPI Compounds
Solutions of [ 131 I] NaI were purchased from the Japan Radioisotope Association (JRIA, Kawasaki, Kanagawa, Japan). B-FAPI(s), as precursors of 131 I-FAPI compounds, were prepared by chemical synthesis (Figure 1b) according to published schemes. All the FAPI compounds were synthesized via a dihydroxyboryl astatine substitution reaction with the corresponding B-FAPI1 and B-FAPI5. Overall, 1 mg of B-FAPI, 80 µL of H 2 O, 28-30 µL aqueous 131 I-NaI (6.76-6.77 MBq), and 30 µL of 0.4% aqueous N-bromosuccinimide (NBS) were added to a polypropylene (PP) tube at room temperature. After heating to 80 • C for 30 min, the reaction mixture was cooled to ambient temperature. The RCY of 131 I-FAPI1 and 131 I -FAPI5 are 13% and 51%. After adding 4% aqueous ascorbic acid (10 µL), the mixture was stirred for 15 min. Then, the reaction mixture was purified using an Oasis HLB column (1.0 mL of H 2 O, 0.5 mL of EtOH/H 2 O (2:3) followed by 0.5 mL of EtOH) to obtain 131 I-FAPI1 and 131 I-FAPI5. After purification using an HLB column, the radiochemical purities (RCPs) of 131 I-FAPI1 and 131 I-FAPI5 were 91% and 99%, respectively. EtOH was evaporated by heating and replaced with H 2 O for the in vitro and in vivo experiments. without calcium and magnesium (PBS (−)), the culture medium was replaced with HBSS (+). After treatment with 211 At-labeled FAPI series compounds, cells were washed twice with PBS (−). After washing, all cells were lysed with NaOH (0.1 N) and the radioactivity of the cells was counted using a 2480 Wizard 2 γ counter (Perkin Elmer, Waltham, MA, USA). Protein levels were measured using a plate reader (MultiScan FC; Thermo Fisher Scientific) and a BCA protein assay kit (FUJIFILM Wako Pure Chemical). Uptakes (CPM/µg protein) were compared between parental cells and cells overexpressing FAPα at 30 min after the incubation with 211 At-labeled FAPI series compounds.

Preparation of Animals
The animals were prepared according to previously published protocol [30]. The PANC-1 cell line was obtained from the ATCC. Male BALB/c-nu-nu (nude) mice were purchased from Japan SLC Inc. (Hamamatsu, Japan). PANC-1 cells were cultured in a DMEM containing 10% fetal bovine serum and 1% penicillin-streptomycin at 37 • C in a humidified incubator with 5% CO 2 . The cultured cells were washed with phosphate-buffered saline (PBS) and harvested using trypsin. Tumor xenograft models were established by subcutaneous injection of 1 × 10 7 cells suspended in 0.2 mL culture medium and Matrigel ® (1:1; BD Biosciences, Franklin Lakes, NJ, USA) into nude mice. Animals with tumors >50 mm 3 in size were used. This study was approved by the Osaka University Graduate School of Science Animal Care and Use Committee (permission number: 2020-02-0) and was conducted according to the Osaka University Animal Experimentation Regulations.

Measurement of Biodistribution of FAP1 and FAPI5 Labeled with 131 I or 211 At
The compounds of interest were labeled with 211 At or 131 I. Mice were dissected 1 or 3 h after injection. The radioactivities were measured with a γ counter (Wizard 2 2480) and the organ weights were also measured. Using these data, the % ID/g was calculated.
4.5.1. Biodistribution of 131 I-FAPI1 and 131 I-FAPI5 PANC-1 xenograft mice were divided into groups according to the tumor size. The average body weight of the 131 I-FAPI1 group was 22.07 ± 0.19 g, and that of the 131 I-FAPI5 group was 22.25 ± 0.18 g. The mice were treated with the following conditions: 131 I-FAPI1 (0.12 ± 0.001 MBq), 131 I-FAPI5 (0.27 ± 0.01 MBq). One or three hours after injection, the mice were euthanized and dissected, radioactivity was measured with a counter, and the weights of the organs were measured using a microbalance.

Biodistribution of 211 At-FAPI1 and 211 At-FAPI5
Animals were prepared according to previously described protocols. The average body weight of the 211 At-FAPI1 group was 22.25 ± 0.18 g, and that of the 211 At-FAPI5 group was 22.07 ± 0.19 g. The animals were treated with the following conditions: 211 At-FAPI1 (0.81 ± 0.02 MBq) and 211 At-FAPI5 (0.54 ± 0.03 MBq). After the injection, the experimental mice were treated using the same procedure as for 131 I.

Examination of Therapeutic Effect
The animals were prepared according to previously published protocol [30]. PANC1 itself did not express a high level of FAPα, but it had already been confirmed that FAPα was highly expressed in xenograft tumors [22]. PANC-1 xenograft model is commonly used to evaluate the therapeutic efficacy of FAPIs. This model represents FAPα at a low level in the cancer cells themselves, but on the other hand, it well represents the state in which FAPα is highly expressed in the cancer tissue. Mice were divided as follows: control (20.88 ± 0.50 g: saline was used for control), 211 At-FAPI5 (20.95 ± 0.62 g:0.96 ± 0.06 MBq), 211 At-FAPI1 (22.25 ± 0.31 g:0.97 ± 0.08 MBq). The mice were monitored three times per week, and their body weights and tumor sizes were measured. The animals were euthanized based on humane endpoints, but can be euthanized even when they are in good health, when tumors exceed 10% of body weight, or when tumors become necrotic. Based on these criteria, the observation period was four weeks.

Statistical Analysis
The results are expressed as the means ± the standard error. Comparisons between groups were performed using the unpaired t-test and Mann-Whitney U test in Microsoft Excel for Mac (ver. 16.71). For multiple comparisons among the three groups, Bonferroni correction was performed. Differences were considered statistically significant at p < 0.05.

Conclusions
We evaluated the cellular uptake and FAPα selectivity of 211 At-FAPI1 to 4. No differences in uptake according to PEG length were observed. Based on the results of cellular uptake, nuclide labeling efficiency, and in vivo pharmacokinetics, simple PEG-linker compounds (FAPI1) exhibited the best properties. In the future, we plan to use this compound to screen more effective treatment methods and applicable cancer types.