Development of the First 18F-Labeled Radiohybrid-Based Minigastrin Derivative with High Target Affinity and Tumor Accumulation by Substitution of the Chelating Moiety

In order to optimize elevated kidney retention of previously reported minigastrin derivatives, we substituted (R)-DOTAGA by DOTA in (R)-DOTAGA-rhCCK-16/-18. CCK-2R-mediated internalization and affinity of the new compounds were determined using AR42J cells. Biodistribution and µSPECT/CT imaging studies at 1 and 24 h p.i. were carried out in AR42J tumor-bearing CB17-SCID mice. Both DOTA-containing minigastrin analogs exhibited 3- to 5-fold better IC50 values than their (R)-DOTAGA-counterparts. natLu-labeled peptides revealed higher CCK-2R affinity than their natGa-labeled analogs. In vivo, tumor uptake at 24 h p.i. of the most affine compound, [19F]F-[177Lu]Lu-DOTA-rhCCK-18, was 1.5- and 13-fold higher compared to its (R)-DOTAGA derivative and the reference compound, [177Lu]Lu-DOTA-PP-F11N, respectively. However, activity levels in the kidneys were elevated as well. At 1 h p.i., tumor and kidney accumulation of [19F]F-[177Lu]Lu-DOTA-rhCCK-18 and [18F]F-[natLu]Lu-DOTA-rhCCK-18 was high. We could demonstrate that the choice of chelators and radiometals has a significant impact on CCK-2R affinity and thus tumor uptake of minigastrin analogs. While elevated kidney retention of [19F]F-[177Lu]Lu-DOTA-rhCCK-18 has to be further addressed with regard to radioligand therapy, its radiohybrid analog, [18F]F-[natLu]Lu-DOTA-rhCCK-18, might be ideal for positron emission tomography (PET) imaging due to its high tumor accumulation at 1 h p.i. and the attractive physical properties of fluorine-18.


Introduction
In 2022, an estimated number of 43,800 new thyroid cancer cases will occur in the United States [1]. Out of these, medullary thyroid carcinoma (MTC) comprises only 2-3% and thus rarely occurs. However, due to comparably late tumor detection in advanced stages and limited treatment options of this disease, the 5-and 10-year survival of MTC (65-89% and 71-87%, respectively) is lower compared to that of the more common types of differentiated thyroid cancer [2,3]. The 10-year survival rate for patients developing metastatic MTC is only 10%, which underlines the importance of an early diagnosis as well as novel therapeutic options [4].
In contrast to conventional diagnostic methodologies, nuclear medicine provides an opportunity to exhibit biochemical information through non-invasive molecular imaging applications. Amongst other uses, this allows for the localization of tumor lesions and metastases because most malignant cells overexpress certain target structures that can be addressed by radiopharmaceutical drugs. In the case of MTC, over 90% of tumors overexpress the cholecystokinin-2 receptor (CCK-2R) in high density [5]. In spite of this characteristic, the gold standard for MTC imaging is 18 F-based positron emission tomography (PET) using

Materials and Methods
Characterization of all CCK-2R-targeted compounds is provided in the Supplementary Materials (Figures S1-S17). Electrospray ionization mass spectrometry for characterization of the substances were acquired on an expression L CMS mass spectrometer (Advion Ltd., Harlow, UK).

Materials and Methods
Characterization of all CCK-2R-targeted compounds is provided in the Supplementary Materials (Figures S1-S17). Electrospray ionization mass spectrometry for characterization of the substances were acquired on an expression L CMS mass spectrometer (Advion Ltd., Harlow, UK).

Chemical Synthesis and Labeling Procedures
Synthesis of all compounds was conducted via standard Fmoc-based solid phase peptide synthesis (SPPS) on a H-Rink Amide ChemMatrix ® resin (35-100 mesh particle size, 0.4-0.6 mmol/g loading, Merck KGaA, Darmstadt, Germany) either manually or with a Liberty Blue peptide synthesizer (H-Rink Amide ProTide resin, 100-200 mesh particle size, 0.6-0.8 mmol/g loading, CEM GmbH, Stuttgart, Germany). Purification of the peptide precursors was carried out by reversed-phase high-performance liquid chromatog-  Ga]gallium was performed as previously published [20,21]. A detailed description of 18 F-labeling is provided in the Supplementary Materials. Briefly, 18 F-fluorination of [ nat Lu]Lu-DOTA-rhCCK-18 (1 nmol) was conducted via an isotopic exchange reaction at the SiFA building block at 60 • C for 5 min using previously dried [ 18 F]fluoride (approx. 400 MBq). Afterwards, the 18 F-labeled peptide was purified via an Oasis ® HLB (30 mg) Light Cartridge (Waters GmbH, Eschborn, Germany).

In Vitro Experiments
Cell-based experiments and determination of lipophilicity (logD 7.4 ) were performed as previously published [19]. A detailed description of the in vitro experiments is provided in the Supplementary Materials.
Human serum albumin (HSA) binding was determined in analogy to a previously published ultracentrifugation method [22]. Therefore, the peptides of interest were incubated in a solution of HSA (700 µM in phosphate-buffered saline) at 37 • C for 30 min (n = 6). All values were corrected for unspecific binding.
In vitro stability studies in human serum after incubation at 37 • C for 24 h were performed as described in the Supplemental Material.

In Vivo Experiments
All animal experiments were conducted in accordance with general animal welfare regulations in Germany (German animal protection act, in the edition of the announcement, dated 18 May 2006, as amended by Article 280 of 19 June 2020, approval no. ROB-55.2-1-2532.Vet_02-18-109 by the General Administration of Upper Bavaria) and the institutional guidelines for the care and use of animals. CB17-SCID mice of both genders and aged 2-12 months (Charles River Laboratories International Inc., Sulzfeld, Germany) were allowed to acclimate at the in-house animal facility for at least one week before inoculation was performed. Tumor xenografts were established as previously reported [19]. Exclusion criteria for animals from an experiment were either weight loss higher than 20%, a tumor size above 1500 mm 3 , an ulceration of the tumor, respiratory distress or a change of behavior. None of these criteria applied to any animal from the experiment. Neither randomization nor blinding was applied in the allocation of the experiments. Health status is SPF according to FELASA recommendation.
Acquired data were statistically analyzed by performing a Student's t-test via Excel (Microsoft Corporation, Redmond, WA, USA) and OriginPro software (version 9.7) from OriginLab Corporation (Northampton, MA, USA). Acquired p values of <0.05 were considered statistically significant.

Synthesis and Radiolabeling
The precursors were synthesized via standard Fmoc-based SPPS with subsequent RP-HPLC purification in yields of 5-20% (chemical purity >95%, determined by RP-HPLC at λ = 220 nm). Labeling with [ nat Lu]lutetium as well as [ nat Ga]gallium was achieved in quantitative yields using a 2.5-fold excess of LuCl 3 and Ga(NO 3 ) 3 , respectively. No purification step was performed because an excess of free ions was not shown to have any impact on overall affinity data [23]. All compounds were labeled manually with lutetium-177, resulting in quantitative radiochemical yields and purities (RCYs, RCPs) and molar activities (A m ) of 10-50 GBq/µmol. After radiolabeling, no further purification steps were conducted. 18 F-Labeling of [ nat Lu]Lu-DOTA-rhCCK-18 was performed manually at 60 • C for 5 min. After purification of the 18 F-labeled peptide via an Oasis ® HLB (30 mg) Light Cartridge, RCYs (without further optimization) of 10-30% and molar activities of A m~8 5 GBq/µmol and RCPs > 95% were achieved.

In Vitro Characterization
The affinity data of all compounds evaluated are outlined in Figure 2 and Table S1.
Pharmaceutics 2023, 15, x FOR PEER REVIEW 5 of 14 cation step was performed because an excess of free ions was not shown to have any impact on overall affinity data [23]. All compounds were labeled manually with lutetium-177, resulting in quantitative radiochemical yields and purities (RCYs, RCPs) and molar activities (Am) of 10-50 GBq/µ mol. After radiolabeling, no further purification steps were conducted. 18 F-Labeling of [ nat Lu]Lu-DOTA-rhCCK-18 was performed manually at 60°C for 5 min. After purification of the 18 F-labeled peptide via an Oasis ® HLB (30 mg) Light Cartridge, RCYs (without further optimization) of 10-30% and molar activities of Am ~85 GBq/µ mol and RCPs > 95% were achieved.

In Vitro Characterization
The affinity data of all compounds evaluated are outlined in Figure 2 and Table S1. Lipophilicity (logD7.4) and human serum albumin (HSA) binding data are summarized in Table 1. Lipophilicity (logD 7.4 ) and human serum albumin (HSA) binding data are summarized in Table 1.
In general, all compounds containing a (R)-DOTAGA chelator revealed a significantly higher lipophilicity than their DOTA-comprising counterparts (p < 0.002). Furthermore, all rhCCK derivatives that comprise a SiFA moiety displayed a distinctly higher lipophilicity (logD 7 Table S2).  [19]. These data have been determined in our lab under identical conditions, n.d.: not determined.
In general, all compounds containing a (R)-DOTAGA chelator revealed a significantly higher lipophilicity than their DOTA-comprising counterparts (p < 0.002). Furthermore, all rhCCK derivatives that comprise a SiFA moiety displayed a distinctly higher lipophilicity (logD7.  Table S2).   Table S3). While the latter signal displays the amount of the respective intact compound (21-44%), the former was attributed to their analogs comprising a hydrolyzed SiFA moiety (SiOH-containing analog, 54-69%). The number of metabolites was <7% for all four derivatives.

In Vivo Characterization
Due to its favorable in vitro data (highest CCK-2R affinity and internalization, excellent lipophilicity, preferable HSA binding), [ 19 F]F-[ 177 Lu]Lu-DOTA-rhCCK-18 was selected for further in vivo studies at 1 and 24 h p.i. (Figure 4, Table S4).   35). Blood, heart and lung levels were slightly elevated at 1 h p.i. (0.9-2.6 %ID/g), while low levels (<0.2 %ID/g) were observed at 24 h p.i. in these organs (p < 0.01). Moreover, increased activity levels were found in the stomach at 1 and 24 h p.i., which was expected due to the endogenous CCK-2R expression in this organ. High kidney uptake was observed at 1 h p.i. for [ 177

Discussion
Due to its effective trapping via decarboxylation by the aromatic L-amino acid decarboxylase (AADC), [ 18 F]F-DOPA is a clinically established neuroimaging agent, but can also be used for the detection of neuroendocrine tumors such as medullary thyroid carcinoma (MTC). Although high sensitivities are only observed for the detection of primary MTC lesions, [ 18 F]F-DOPA is still considered the gold standard for MTC imaging in clinical practice [10,24], most likely as a consequence of the favorable properties of 18 F-based positron emission tomography (PET) and the lack of reliable alternatives. Notwithstanding that the majority of MTC cells overexpress the cholecystokinin-2 receptor (CCK-2R) in high density, there is currently no CCK-2R-targeted compound available that shows promising pharmacokinetics and bears the possibility of 18 F-labeling.
In recent studies, we thus introduced a silicon-based fluoride acceptor (SiFA) moiety into the D-glutamate chain of DOTA-PP-F11N. The resulting radiohybrid (rh)-based compounds enable labeling with both fluorine-18 and radiometals such as lutetium-177 due to the presence of a chelator and a SiFA moiety. We could show that these rhCCK ligands, for example [ nat/177 Lu]Lu-(R)-DOTAGA-rhCCK-16, revealed up to 8-fold increased activity levels in the tumor but also approximately 30-fold higher levels in the kidney when compared to [ nat/177 Lu]Lu-DOTA-PP-F11N, despite its distinctly lower CCK-2R affinity [19]. While we assume that the elevated tumor uptake and retention is due to a decelerated clearance of the compound, the charge distribution within the linker section and thus in proximity to the SiFA moiety likely causes the increased kidney values. Hence, the aim of this study was to retain favorable tumor values while reducing the activity levels in the

Discussion
Due to its effective trapping via decarboxylation by the aromatic L-amino acid decarboxylase (AADC), [ 18 F]F-DOPA is a clinically established neuroimaging agent, but can also be used for the detection of neuroendocrine tumors such as medullary thyroid carcinoma (MTC). Although high sensitivities are only observed for the detection of primary MTC lesions, [ 18 F]F-DOPA is still considered the gold standard for MTC imaging in clinical practice [10,24], most likely as a consequence of the favorable properties of 18 F-based positron emission tomography (PET) and the lack of reliable alternatives. Notwithstanding that the majority of MTC cells overexpress the cholecystokinin-2 receptor (CCK-2R) in high density, there is currently no CCK-2R-targeted compound available that shows promising pharmacokinetics and bears the possibility of 18 F-labeling.
In recent studies, we thus introduced a silicon-based fluoride acceptor (SiFA) moiety into the D-glutamate chain of DOTA-PP-F11N. The resulting radiohybrid (rh)-based compounds enable labeling with both fluorine-18 and radiometals such as lutetium-177 due to the presence of a chelator and a SiFA moiety. We could show that these rhCCK ligands, for example [ nat/177 Lu]Lu-(R)-DOTAGA-rhCCK-16, revealed up to 8-fold increased activity levels in the tumor but also approximately 30-fold higher levels in the kidney when compared to [ nat/177 Lu]Lu-DOTA-PP-F11N, despite its distinctly lower CCK-2R affinity [19]. While we assume that the elevated tumor uptake and retention is due to a decelerated clearance of the compound, the charge distribution within the linker section and thus in proximity to the SiFA moiety likely causes the increased kidney values. Hence, the aim of this study was to retain favorable tumor values while reducing the activity levels in the kidneys. Therefore, (R)-DOTAGA was substituted by a DOTA chelator in two rhCCK derivatives to reduce the negative charges within the linker section and maintain the peptide sequence to retain high CCK-2R affinity.
Interestingly, substitution of (R)-DOTAGA by DOTA in the most affine rhCCK derivatives from previous studies, [ nat Lu]Lu-(R)-DOTAGA-rhCCK-16 and -18, resulted in 3-to 4-fold lower IC 50 values for the DOTA-comprising analogs, surpassing even the highly affine reference compound, [ nat Lu]Lu-DOTA-PP-F11N (Figure 2). It is thus anticipated that the additional free carboxylic group of the (R)-DOTAGA chelator at the respective site has a negative impact on the overall CCK-2R affinity. Similar observations were made for the nat Ga-labeled rhCCK ligands because the additional free carboxylic group at the Ga-(R)-DOTAGA chelate compared to the respective Ga-DOTA chelate as well as the additional free carboxylic group of the Ga-DOTA compared to the respective Lu-DOTA chelate [25][26][27] led to a decreased overall CCK-2R affinity (Figure 2). Stability studies in human serum did not reveal a lower stability for the [ 67 Ga]Ga-DOTA-rhCCK-16 or -18 as compared to their 177 Lu-labeled analogs, which can be thus excluded as a potential reason for the decreased CCK-2R affinity of the nat Ga-labeled compounds. Interestingly, the stability studies in human serum showed the formation of a slightly more hydrophilic analog (∆t R~2 min) for all four compounds tested ( Figure S18). This was attributed to their corresponding SiOH-containing derivatives, and we suggest that the SiFA building block is hydrolyzed over time under these conditions. In order to confirm this assumption, we performed RP-HPLC analysis of the nat Ga/ nat Lu-labeled SiOH-containing analogs, which were generated by treatment with sodium hydroxide. All four SiOH-containing analogs (peptide identity confirmed by ESI-MS) revealed the same retention time as their respective 67 Ga/ 177 Lu-labeled derivative, which was observed after incubation in human serum. Because the SiFA-and their respective SiOH-containing ligands only differ by the atom/group bound to the silicon atom, but the remaining compound is identical, we do not consider this a metabolite but rather an intact compound. Further studies have to be conducted to elucidate whether the SiFA moiety is also hydrolyzed in vivo.
In addition to a lower CCK-2R affinity, all 177 Lu-labeled (R)-DOTAGA-comprising compounds showed higher logD 7.4 values than their DOTA-containing analogs (Table 1). Because it was assumed that the negatively charged 177 Lu-(R)-DOTAGA chelates should be more hydrophilic than the neutral 177 Lu-DOTA chelates, further investigations are necessary to understand this dedicated structure-activity relationship. Furthermore, the increased CCK-2R affinity was paralleled by an improved receptor-mediated internalization because both [ 177 Lu]Lu-DOTA-rhCCK-16 and -18 exhibited the highest internalization values at all time points. While the slope of the internalization curves of most compounds decreases after the first hours (Figure 3a), the curve of [ 177 Lu]Lu-DOTA-rhCCK-18 seems to rise with an almost unaffected slope up to the end of the experiment at 6 h. Thus, although this compound initially shows a decelerated internalization rate, its continuous cellular uptake might result in a noticeably higher overall uptake at later time points when compared with the other ligands of this series. Worth mentioning, both [ 177 Lu]Lu-(R)-DOTAGA-rhCCK-16 and -18 showed significantly higher internalization values at all time points than [ 177 Lu]Lu-DOTA-PP-F11N despite their significantly lower CCK-2R affinity, which points to a beneficial impact of the SiFA moiety on internalization kinetics.
In vivo, [ nat/177 Lu]Lu-DOTA-rhCCK-18 revealed a 1.5 and 13-fold increased activity uptake in the tumor (25.4 ± 4.7 %ID/g, Figure 4) at 24 h p.i., as compared to the previously published compound, [ 177 Lu]Lu-(R)-DOTAGA-rhCCK-16 (15.7 ± 3.3 %ID/g), and the parent peptide, [ 177 Lu]Lu-DOTA-PP-F11N (1.9 ± 0.8 %ID/g), respectively, which can be attributed to its significantly higher CCK-2R affinity and internalization [19]. Consequently, tumor-to-background ratios were higher for [ 177 Lu]Lu-DOTA-rhCCK-18 compared to the previously published compounds (Table S5). Tumor specificity was demonstrated by competition studies using excess of the CCK-2R-specific compound, [ nat Lu]Lu-DOTA-MGS5 [28], which led to tumor and stomach values <2 %ID/g. Moreover, the high tumor values obtained for [ 177  Similar to previously published rhCCK derivatives [19], tumor and kidney uptake for [ 177 Lu]Lu-DOTA-rhCCK-18 was high at 1 h p.i. and remained high at 24 h p.i. Elevated tumor and stomach retention can be attributed to the decelerated clearance kinetics (higher logD 7.4 and albumin binding) of the rhCCK ligands and their prolonged bioavailability, the increased kidney retention is likely caused by a synergistic effect of the negatively charged side chains in proximity of the SiFA building block. This assumption is supported by the fact that [ 177 Lu]Lu-DOTA-PP-F11N contains a similar number of negative charges but no SiFA moiety and does not show an enhanced kidney retention. Recent reports demonstrated that SiFA-comprising PSMA inhibitors show a higher albumin binding and thus decelerated clearance kinetics, which results in increased tumor uptake, but also increased kidney retention [21,29], which correlated well with our observations because [ 177 Lu]Lu-DOTA-rhCCK-18 also exhibited an elevated albumin binding in vitro. Furthermore, it was shown that negative charges in the direct neighborhood to the SiFA moiety cause a higher albumin binding and stronger kidney retention, which could explain our results because the rhCCK derivatives contain several negative charges in proximity to the SiFA group. As the substitution of a (R)-DOTAGA by a DOTA moiety did not result in lower kidney retention, it is assumed that most of the negatively charged γ-D-glutamic acid moieties have to be removed in future studies to prevent an elevated kidney uptake and retention. In addition, it will be interesting to see whether the impact of these negative charges of radiohybrid and other CCK-2R ligands on the kidney retention will be confirmed by the first comparative studies in humans.
Nevertheless, even in the case that such behavior would be confirmed in human studies, unfavorable kidney uptake of [ 19 F]F-[ 177 Lu]Lu-DOTA-rhCCK-18 does not necessarily prevent its use for PET imaging with the corresponding 18 F-radiohybrid. When 18 F-labeled, and taking into account the short half-life of fluorine-18, an elevated kidney accumulation will not result in an unacceptable dosimetry. Similar kidney uptake is, for example, also observed for commonly applied PSMA inhibitors [30][31][32][33]. Despite this minor disadvantage, [ 18 F]F-[ nat Lu]Lu-DOTA-rhCCK-18 seems to have great potential for the detection of even small and distant metastases in MTC patients due to the unique properties of 18 F-PET and the high overexpression of the CCK-2R on these cancer cells.
In order to confirm the expected favorable pharmacokinetics of the chemically identical [ 18 F]F-[ nat Lu]Lu-DOTA-rhCCK-18, we carried out a µPET/CT image (n = 1, Figure 5), which revealed similarly high activity levels in the tumor and kidneys compared to [ 19 F]F-[ 177 Lu]Lu-DOTA-rhCCK-18. Moreover, bone uptake was observed to be low (1.69 %ID/g) for [ 18 F]F-[ nat Lu]Lu-DOTA-rhCCK-18, underlining the high metabolic stability of the Si-18 F bond. Therefore, the formation of the SiOH-containing analog observed in stability studies in human serum does not seem to occur in vivo, at least not within the first hours after injection, and should therefore not be of concern for PET imaging using this compound. However, stability of this compound (particularly of the Si-F bond) in men must be investigated to confirm this assumption. Due to these results, particularly its high tumor accumulation at 1 h p.i., [ 18 F-labeling was carried out via an isotopic exchange reaction by a novel labeling strategy, which led to molar activities of~85 GBq/µmol in a total synthesis time of~30 min. Unlike the Munich Method [36], this strategy includes the use of ammonium formiate (in anhydrous DMSO) instead of [K + ⊂2.2.2]OH − (in anhydrous MeCN) for the elution of dry [ 18 F]fluoride from a SEP-Pak ® Light (46 mg) Accell Plus QMA cartridge (Waters GmbH, Eschborn, Germany), which enables a less time-consuming preparation and more cost efficient 18 F-fluorination method. In comparison to conventional radiofluorination techniques [37], no azeotropic drying steps must be conducted. Furthermore, anhydrous DMSO is used as the reaction solvent, which is beneficial for the 18 F-labeling of CCK-2Rtargeting peptides.
In summary, we could demonstrate a significantly higher CCK-2R affinity and thus enhanced tumor accumulation by exchanging the chelator moiety in previously published rhCCK derivatives from (R)-DOTAGA to DOTA. Nevertheless, elevated kidney retention of rhCCK derivatives could not be reduced in this study, which must be addressed in future studies by extinguishing most of the negatively charged residues within the SiFA-containing linker section. Despite increased kidney uptake, [ 18 F]F-[ nat Lu]Lu-DOTA-rhCCK-18 holds great promise as an imaging agent and is expected to be highly competitive to currently applied radiotracers for PET imaging of medullary thyroid carcinoma.

Conclusions
While a simple substitution of (R)-DOTAGA by DOTA in previously reported rhCCK derivatives led to a noticeably increased CCK-2R affinity and thus high activity levels in the tumor for [ 19 F]F-[ 177 Lu]Lu-DOTA-rhCCK-18 at 24 h p.i., kidney retention was also high. Nevertheless, due to its very high tumor accumulation at already 1 h p.i., the chemically identical [ 18 F]F-[ nat Lu]Lu-DOTA-rhCCK-18 might compete with or even surpass the detection rate of currently applied imaging agents for MTC such as 68   Inlabeled CCK-2R or SSTR2-targeted compounds and [ 18 F]F-DOPA, which is why a clinical translation of this compound for MTC imaging is recommended.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/pharmaceutics15030826/s1, general information and characterization of all CCK2R-targeted compounds, detailed description of cell-based experiments. Figure S1: