Development of 18F-Labeled Bispyridyl Tetrazines for In Vivo Pretargeted PET Imaging

Pretargeted PET imaging is an emerging and fast-developing method to monitor immuno-oncology strategies. Currently, tetrazine ligation is considered the most promising bioorthogonal reaction for pretargeting in vivo. Recently, we have developed a method to 18F-label ultrareactive tetrazines by copper-mediated fluorinations. However, bispyridyl tetrazines—one of the most promising structures for in vivo pretargeted applications—were inaccessible using this strategy. We believed that our successful efforts to 18F-label H-tetrazines using low basic labeling conditions could also be used to label bispyridyl tetrazines via aliphatic nucleophilic substitution. Here, we report the first direct 18F-labeling of bispyridyl tetrazines, their optimization for in vivo use, as well as their successful application in pretargeted PET imaging. This strategy resulted in the design of [18F]45, which could be labeled in a satisfactorily radiochemical yield (RCY = 16%), molar activity (Am = 57 GBq/µmol), and high radiochemical purity (RCP > 98%). The [18F]45 displayed a target-to-background ratio comparable to previously successfully applied tracers for pretargeted imaging. This study showed that bispyridyl tetrazines can be developed into pretargeted imaging agents. These structures allow an easy chemical modification of 18F-labeled tetrazines, paving the road toward highly functionalized pretargeting tools. Moreover, bispyridyl tetrazines led to near-instant drug release of iTCO-tetrazine-based ‘click-to-release’ reactions. Consequently, 18F-labeled bispyridyl tetrazines bear the possibility to quantify such release in vivo in the future.


Introduction
Radioimmunoconjugates have emerged as important tools; e.g., in the diagnosis and treatment of cancer [1,2]. A subclass within these important substances are monoclonal antibodies (mAbs). They can be designed and produced with exquisite target affinity and selectivity. From a nuclear molecular imaging point of view, mAb-based agents can result in high target-to-background ratios with a low nondisplaceable binding component, which make them almost ideal tracers. Unfortunately, at the same time they possess slow pharmacokinetic properties; i.e., target accumulation and blood clearance takes days rather than hours [3,4]. Consequently, long-lived radionuclides must be used to match the pharmacokinetic profile of these vectors [3]. This results in unnecessary radiation burden for Pretargeted imaging consequently exploits the unique targeting properties of mAbs and the rapid pharmacokinetics properties of small molecules, enabling exceptional target-to-background ratios within hours [3]. Currently, the most promising reaction for pretargeted imaging is the tetrazine ligation between a tetrazine (Tz) and a trans-cyclooctene (TCO) ( Figure 1B). Several preclinical examples have been reported that demonstrated the potential of this reaction [6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22]. More importantly, in late 2020, the first clinical phase I trial based on tetrazine ligation was initiated (NCT04106492) [23]. From a clinical point of view, a fluorine-18 ( 18 F)-labeled Tz would be ideal for positron emission tomography (PET) applications, since 18 F possesses almost perfect physical characteristics for molecular imaging [24][25][26]. However, due to the intrinsic instability of the tetrazine scaffold to basic conditions-typically employed in direct 18 F-fluorination approaches-no 18 F-tetrazine was synthesized until few years ago [10,[27][28][29][30][31]. Recently, we reported the first direct 18 F-fluorination of highly reactive tetrazines, which could be labeled via copper-mediated oxidative 18 F-fluorination [32]. This methodology allowed us to develop a low-lipophilicity, fast-clearing, and highly reactive tetrazine PET tracer (1) (Figure 2). In a similar way, we were able to develop the first direct aliphatic 18 F-radiolabeled Tz (2) (Figure 2). This was possible using ultralow basic conditions [33,34]. Both compounds (1 and 2) showed high imaging contrast during in vivo experiments, and are currently subjects in further studies to evaluate their potential to be translated into the clinic. Pretargeted imaging consequently exploits the unique targeting properties of mAbs and the rapid pharmacokinetics properties of small molecules, enabling exceptional targetto-background ratios within hours [3]. Currently, the most promising reaction for pretargeted imaging is the tetrazine ligation between a tetrazine (Tz) and a trans-cyclooctene (TCO) ( Figure 1B). Several preclinical examples have been reported that demonstrated the potential of this reaction [6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22]. More importantly, in late 2020, the first clinical phase I trial based on tetrazine ligation was initiated (NCT04106492) [23]. From a clinical point of view, a fluorine-18 ( 18 F)-labeled Tz would be ideal for positron emission tomography (PET) applications, since 18 F possesses almost perfect physical characteristics for molecular imaging [24][25][26]. However, due to the intrinsic instability of the tetrazine scaffold to basic conditions-typically employed in direct 18 F-fluorination approaches-no 18 F-tetrazine was synthesized until few years ago [10,[27][28][29][30][31]. Recently, we reported the first direct 18 Ffluorination of highly reactive tetrazines, which could be labeled via copper-mediated oxidative 18 F-fluorination [32]. This methodology allowed us to develop a low-lipophilicity, fast-clearing, and highly reactive tetrazine PET tracer (1) (Figure 2). In a similar way, we were able to develop the first direct aliphatic 18 F-radiolabeled Tz (2) (Figure 2). This was possible using ultralow basic conditions [33,34]. Both compounds (1 and 2) showed high imaging contrast during in vivo experiments, and are currently subjects in further studies to evaluate their potential to be translated into the clinic.

Figure 2.
Compound evaluation and rationale-from previous work to this project: The 18 F-labeling of H-Tz has already been established (previous work); labeling of bispyridyl Tzs succeeded in this study. Bispyridyl Tzs have intrinsic high-reaction kinetics and allow higher chemical flexibility. Pyridyl unit A can be used to introduce the necessary polarity for in vivo pretargeting, whereas pyridyl unit B can be used for labeling purposes.
However, both labeling methods (Cu-mediated 18 F-fluorinartion and direct aliphatic 18 F-labeling) have thus far only been successful for H-Tzs. Bispyridyl scaffolds could not be labeled with Cu-mediated approaches due to the chelating effect of the pyridyl nitrogen [32,35]. Aliphatic labeling was only applied on monosubstituted Tzs [34]. Bispyridyl Tzs are, however, very appealing subjects to be used for pretargeted experiments ( Figure  2). Beside H-Tzs, these structures are the only other structural class that have been successfully applied in vivo [6,[35][36][37][38]. Bispyridyl Tzs have intrinsic high-reaction kinetics and allow for greater flexibility with respect to structural modifications compared to H-Tzs. For example, one pyridyl unit can be used for labeling purposes, whereas the other can be used to introduce polarity or other functional scaffolds as NIR turn-on dyes or additional chelators for pretargeted radiotherapy [3,17,39,40]. Recently, bispyridyl Tzs also were shown to be superior for instant and near-quantitative 'click-to-release' approaches [41]. Radiolabeled bispyridyl Tzs could be used as such to quantify drug release in vivo.
In this work, we describe the design, synthesis, and in vivo evaluation of the first 18 Flabeled bispyridyl Tz. In order to develop such a Tz, we exploited one pyridyl ring to introduce polarity. We recently showed that a clogD7.4 of at least -3 is a necessary parameter for successful in vivo pretargeting [17]. The other pyridyl ring was used for labeling purposes using a set of different linkers. The selection of the linker for the 18 F-labeling was initially performed using H-Tzs due to their more accessible chemistry. Promising structures in term of radiochemical conversions (RCC) were translated into their bispyridyl Tz counterparts. Finally, the best candidate was applied for pretargeted PET experiments.

Evaluation of Different Fluoroethyl Linkers Using H-Tz as Model Structures
We have previously reported that fluoroethyl moieties can be used to introduce 18 Ffluoride into H-Tz employing non-nucleophilic bases [33,34]. We believed that these conditions could also be suitable to label bispyridyl-based Tzs. In order to test the influence on the labeling step of different conjugation strategies between the fluoroethyl moiety and the Tz, we synthesized a set of structures possessing various linkers such as esters, amines, ethers and amides ( Figure 3). We decided to use H-Tzs for this study, as they are easier to access. Seven different molecules (3)(4)(5)(6)(7)(8)(9) and their corresponding precursors were designed.  However, both labeling methods (Cu-mediated 18 F-fluorinartion and direct aliphatic 18 F-labeling) have thus far only been successful for H-Tzs. Bispyridyl scaffolds could not be labeled with Cu-mediated approaches due to the chelating effect of the pyridyl nitrogen [32,35]. Aliphatic labeling was only applied on monosubstituted Tzs [34]. Bispyridyl Tzs are, however, very appealing subjects to be used for pretargeted experiments ( Figure 2). Beside H-Tzs, these structures are the only other structural class that have been successfully applied in vivo [6,[35][36][37][38]. Bispyridyl Tzs have intrinsic high-reaction kinetics and allow for greater flexibility with respect to structural modifications compared to H-Tzs. For example, one pyridyl unit can be used for labeling purposes, whereas the other can be used to introduce polarity or other functional scaffolds as NIR turn-on dyes or additional chelators for pretargeted radiotherapy [3,17,39,40]. Recently, bispyridyl Tzs also were shown to be superior for instant and near-quantitative 'click-to-release' approaches [41]. Radiolabeled bispyridyl Tzs could be used as such to quantify drug release in vivo.
In this work, we describe the design, synthesis, and in vivo evaluation of the first 18 F-labeled bispyridyl Tz. In order to develop such a Tz, we exploited one pyridyl ring to introduce polarity. We recently showed that a clogD 7 . 4 of at least -3 is a necessary parameter for successful in vivo pretargeting [17]. The other pyridyl ring was used for labeling purposes using a set of different linkers. The selection of the linker for the 18 F-labeling was initially performed using H-Tzs due to their more accessible chemistry. Promising structures in term of radiochemical conversions (RCC) were translated into their bispyridyl Tz counterparts. Finally, the best candidate was applied for pretargeted PET experiments.

Evaluation of Different Fluoroethyl Linkers Using H-Tz as Model Structures
We have previously reported that fluoroethyl moieties can be used to introduce 18 Ffluoride into H-Tz employing non-nucleophilic bases [33,34]. We believed that these conditions could also be suitable to label bispyridyl-based Tzs. In order to test the influence on the labeling step of different conjugation strategies between the fluoroethyl moiety and the Tz, we synthesized a set of structures possessing various linkers such as esters, amines, ethers and amides ( Figure 3). We decided to use H-Tzs for this study, as they are easier to access. Seven different molecules (3)(4)(5)(6)(7)(8)(9) and their corresponding precursors were designed.

Synthesis of H-Tz Derivatives
Compound 8 and its precursor 8a were synthesized as previously reported [33]. Compounds 3 and 4 and their corresponding precursor were obtained as reported in Scheme 1.

Synthesis of H-Tz Derivatives
Compound 8 and its precursor 8a were synthesized as previously reported [33]. Compounds 3 and 4 and their corresponding precursor were obtained as reported in Scheme 1.

Synthesis of H-Tz Derivatives
Compound 8 and its precursor 8a were synthesized as previously reported [33]. Compounds 3 and 4 and their corresponding precursor were obtained as reported in Scheme 1. Briefly, Tz 10 was synthesized from 4-cyanobenzoic acid using a Pinner-like, sulfurmediated procedure in modest yields [42]. Subsequent alkylation with 1-fluoro-2-iodoethane or 2-bromoethanol respectively gave compounds 3 and 11. The latter was then reacted with nosyl chloride to yield precursor 3a. Coupling of 4-cyanobenzoic acid with 2-fluoroethylamine or ethanolamine afforded 12 and 13; these were then converted into the corresponding Tz derivatives 4 and 14. These intermediates were further nosylated under basic conditions. However, we were not able to isolate 4a, since an intramolecular substitution occurred that could not be prevented. Tzs 5 and 6 and their respective precursors were obtained in a similar manner as reported in Scheme 2. The yields obtained were comparable to those of 3 and 4. However, the amide precursor 6a could be isolated in a yield of 83%. This most likely was possible due the lower reactivity of the tetrazine core compared to that of 4a. Briefly, Tz 10 was synthesized from 4-cyanobenzoic acid using a Pinner-like, sulfurmediated procedure in modest yields [42]. Subsequent alkylation with 1-fluoro-2-iodoethane or 2-bromoethanol respectively gave compounds 3 and 11. The latter was then reacted with nosyl chloride to yield precursor 3a. Coupling of 4-cyanobenzoic acid with 2-fluoroethylamine or ethanolamine afforded 12 and 13; these were then converted into the corresponding Tz derivatives 4 and 14. These intermediates were further nosylated under basic conditions. However, we were not able to isolate 4a, since an intramolecular substitution occurred that could not be prevented. Tzs 5 and 6 and their respective precursors were obtained in a similar manner as reported in Scheme 2. The yields obtained were comparable to those of 3 and 4. However, the amide precursor 6a could be isolated in a yield of 83%. This most likely was possible due the lower reactivity of the tetrazine core compared to that of 4a. Compounds 7 and 9 were obtained using a different synthesis strategy (Scheme 3). 4-Hydroxybenzonitrile was reacted with 1-fluoro-2-iodoethane or 2-bromoethanol to respectively give 20 and 21 in 71% and 92% yields. These were converted into the corresponding Tzs 7 and 22 in modest yields (37% and 54%, respectively). The latter was further modified to give the nosylate precursor 7a. Intermediates 17 and 18 were obtained from 4-(bromomethyl)benzonitrile and the corresponding amines. N-Boc protection and reaction with hydrazine gave Tzs 27 and 28. Deprotection of 27 afforded compound 9 in an almost quantitative yield. Reaction of nosyl chloride with 28 resulted in an intramolecular reaction, and 9a could not be obtained. For this reason, a different protective group was selected. Then, 28 was deprotected and reacted with trityl chloride to give 31. In order to have the corresponding reference compound, 9 was converted to 30 as well. Finally, reaction with nosyl chloride gave 9b in a yield of 56%. Compounds 7 and 9 were obtained using a different synthesis strategy (Scheme 3). 4-Hydroxybenzonitrile was reacted with 1-fluoro-2-iodoethane or 2-bromoethanol to respectively give 20 and 21 in 71% and 92% yields. These were converted into the corresponding Tzs 7 and 22 in modest yields (37% and 54%, respectively). The latter was further modified to give the nosylate precursor 7a. Intermediates 17 and 18 were obtained from 4-(bromomethyl)benzonitrile and the corresponding amines. N-Boc protection and reaction with hydrazine gave Tzs 27 and 28. Deprotection of 27 afforded compound 9 in an almost quantitative yield. Reaction of nosyl chloride with 28 resulted in an intramolecular reaction, and 9a could not be obtained. For this reason, a different protective group was selected. Then, 28 was deprotected and reacted with trityl chloride to give 31. In order to have the corresponding reference compound, 9 was converted to 30 as well. Finally, reaction with nosyl chloride gave 9b in a yield of 56%.

Labeling of H-Tz Derivatives
As we recently reported, nosyl leaving groups are optimally suited to aliphatically 18 F-label high-reactive Tzs. A mesylate or tosylate base precursor resulted in low radiochemical yields (RCY) [33,43]. Radiolabeling only succeeded using low basicity conditions, since H-Tzs are too sensitive for standard-rather, basic-aliphatic 18 F-fluorination approaches [29,33]. Within this study, we applied the same labeling conditions. Furthermore, we also investigated to what extent the precursor concentration influenced the RCC. Two amounts (3.1 and 9.3 nmol) were selected in this respect. With the exception of compound 3, both concentrations resulted in the same RCC range. Labeling attempts revealed further that the linkers influenced RCCs strongly. Only tetrazines 3, 7, and 8 could be synthesized. RCCs were in the range of 23-53% (based on radio-HPLC, n = 3). No or only minimal product formation could be observed with a different substitution profile. In light of that, bispyridyl Tzs 33, 38, and 41 were designed, synthesized, and then radiolabeled.

Synthesis of the Bispyridyl Tzs
Bispyridyl analogues of 3, 7, and 8 were synthesized as described in Scheme 4. 6-Cyanonicotinic acid was converted to the corresponding bispyridyl Tz 32 following a previously reported procedure [44]. The latter was alkylated with 1-fluoro-2-iodoethane or 2bromoethanol to respectively give compounds 33 and 34. The alcohol derivative was then transformed to the nosylate precursor 33a. Differently, 5-methylpicolinonitrile was brominated and reacted with ethylene glycol to afford 36. Formation of the Tz ring and conversion to the fluorine analogue gave 38 [45]. Nosylation of the same alcohol intermediate

Labeling of H-Tz Derivatives
As we recently reported, nosyl leaving groups are optimally suited to aliphatically 18 F-label high-reactive Tzs. A mesylate or tosylate base precursor resulted in low radiochemical yields (RCY) [33,43]. Radiolabeling only succeeded using low basicity conditions, since H-Tzs are too sensitive for standard-rather, basic-aliphatic 18 F-fluorination approaches [29,33]. Within this study, we applied the same labeling conditions. Furthermore, we also investigated to what extent the precursor concentration influenced the RCC. Two amounts (3.1 and 9.3 nmol) were selected in this respect. With the exception of compound 3, both concentrations resulted in the same RCC range. Labeling attempts revealed further that the linkers influenced RCCs strongly. Only tetrazines 3, 7, and 8 could be synthesized. RCCs were in the range of 23-53% (based on radio-HPLC, n = 3). No or only minimal product formation could be observed with a different substitution profile. In light of that, bispyridyl Tzs 33, 38, and 41 were designed, synthesized, and then radiolabeled.

Synthesis of the Bispyridyl Tzs
Bispyridyl analogues of 3, 7, and 8 were synthesized as described in Scheme 4. 6-Cyanonicotinic acid was converted to the corresponding bispyridyl Tz 32 following a previously reported procedure [44]. The latter was alkylated with 1-fluoro-2-iodoethane or 2-bromoethanol to respectively give compounds 33 and 34. The alcohol derivative was then transformed to the nosylate precursor 33a. Differently, 5-methylpicolinonitrile was brominated and reacted with ethylene glycol to afford 36. Formation of the Tz ring and conversion to the fluorine analogue gave 38 [45]. Nosylation of the same alcohol intermediate yielded 38a. The last compounds, 41 and 41a, were obtained starting with 5-hydroxypicolinonitrile and using the same procedure employed for Tz 7.  41 were labeled from their corresponding precursors 33a, 38a, and 41a following the same procedure reported in Section 2.3. The results are reported in Table 1. As expected from previous results, radiolabeling was successful in all cases. RCCs were in the range of 20-30%. However, only compounds [ 18 F]33 and [ 18 F]38 could be radiolabeled in a similar conversion when lower precursor concentrations were used. Due to that and its easier synthetic access compared to 33, we decided to use 38 as a template for our next steps, aiming to develop a pretargeted imaging agent.
a Precursor was not obtained. b No reaction observed. c Precursor protected with trityl. d Precursor decompose during reaction. n.d. = not determined. * Radiochemical conversions (RCC) were calculated for each compound by radio-HPLC and radio-TLCs as recently reported [43]. All results were based on n = 3 experiments.

Synthesis and Evaluation of 45 by Ex Vivo Blocking Assay
Compound 45 was designed based on a fluorothexy moiety identified to yield in the highest RCCs of previous results in this study. To decrease lipophilicity, we further conjugated a diacetic acid moiety to the structure. This combination should have resulted in an easy-to-label compound, as well as an imaging agent with the necessary polarity to be applied for in vivo pretargeted imaging. The pathway to synthesize 45 is depicted in Scheme 5. Alkylation of di-tert-butyl iminodiacetate by compound 35 under basic conditions yielded 43. The latter was reacted with an excess of 39 and hydrazine hydrate to give 44. Deprotection and purification by preparative-HPLC afforded 45 as a TFA salt.
a Precursor was not obtained. b No reaction observed. c Precursor protected with trityl. d Precursor decompose during reaction. n.d. = not determined. * Radiochemical conversions (RCC) were calculated for each compound by radio-HPLC and radio-TLCs as recently reported [43]. All results were based on n = 3 experiments.

Synthesis and Evaluation of 45 by Ex Vivo Blocking Assay
Compound 45 was designed based on a fluorothexy moiety identified to yield in the highest RCCs of previous results in this study. To decrease lipophilicity, we further conjugated a diacetic acid moiety to the structure. This combination should have resulted in an easy-to-label compound, as well as an imaging agent with the necessary polarity to be applied for in vivo pretargeted imaging. The pathway to synthesize 45 is depicted in Scheme 5. Alkylation of di-tert-butyl iminodiacetate by compound 35 under basic conditions yielded 43. The latter was reacted with an excess of 39 and hydrazine hydrate to give 44. Deprotection and purification by preparative-HPLC afforded 45 as a TFA salt.
Next, 45 was evaluated for its abilities in pretargeting. We recently developed a blocking assay that allowed us to assess the in vivo ligation performance of unlabeled tetrazine derivatives, omitting the time-consuming development of radiolabeled tetrazines for each tested ligand. It was based on the ability of Tzs to block the binding of the literature-known pretargeted imaging agent [ 111 In]47 to the pretargeting vector CC49-TCO (administered 72 h prior) in tumor-bearing mice [17,32,34]. In this setup, 45 showed almost complete blocking. This blocking effect was as good as the ones observed with currently successfully applied 'state-of-the-art' pretargeted imaging agents (Table 2). This result allowed us to assume that 45 could indeed be a suitable candidate for pretargeted imaging. Next, 45 was evaluated for its abilities in pretargeting. We recently developed a blocking assay that allowed us to assess the in vivo ligation performance of unlabeled tetrazine derivatives, omitting the time-consuming development of radiolabeled tetrazines for each tested ligand. It was based on the ability of Tzs to block the binding of the literature-known pretargeted imaging agent [ 111 In]47 to the pretargeting vector CC49-TCO (administered 72 h prior) in tumor-bearing mice [17,32,34]. In this setup, 45 showed almost complete blocking. This blocking effect was as good as the ones observed with currently successfully applied 'state-of-the-art' pretargeted imaging agents (Table 2). This result allowed us to assume that 45 could indeed be a suitable candidate for pretargeted imaging.   , and the uptake was normalized to a group of animals in which no blocking was performed (control). Data represent mean from n = 3 mice/group; detailed information can be found in the Materials and Methods section. g Blocking data from [26]. h Blocking data from [27]. * Tz 41 was tested to demonstrate that the blocking effect was related to clogD7.4 and not the bispyridyl scaffold. TPSA = topological polar surface area.

Synthesis and Labeling of [ 18 F]45
The precursor (45a) was synthesized over 4 synthesis steps (Scheme 5). Briefly, 40 was reacted with an excess of 43 and hydrazine hydrate to give, after oxidation, 46. This was then converted to the corresponding nosylated derivative 45a. Radiolabeling of [ 18 F]45 was carried out in a one-pot, two-step reaction sequence ( Figure 4A,B). The conditions were similar to those we recently reported and applied for all former labeling attempts in this study [29,32,33,46]. A protection/deprotection strategy was needed to suc- , and the uptake was normalized to a group of animals in which no blocking was performed (control). Data represent mean from n = 3 mice/group; detailed information can be found in the Materials and Methods section. g Blocking data from [26]. h Blocking data from [27]. * Tz 41 was tested to demonstrate that the blocking effect was related to clogD 7 . 4 and not the bispyridyl scaffold. TPSA = topological polar surface area.

Synthesis and Labeling of [ 18 F]45
The precursor (45a) was synthesized over 4 synthesis steps (Scheme 5). Briefly, 40 was reacted with an excess of 43 and hydrazine hydrate to give, after oxidation, 46. This was then converted to the corresponding nosylated derivative 45a. Radiolabeling of [ 18 F]45 was carried out in a one-pot, two-step reaction sequence ( Figure 4A,B). The conditions were similar to those we recently reported and applied for all former labeling attempts in this study [29,32,33,46]. A protection/deprotection strategy was needed to successfully label [ 18 F]45 in a radiochemical yield (RCY) of 16 ± 4% (n = 4), a radiochemical purity (RCP) of >98%, and a molar activity (A M ) of 57 ± 15 GBq/µmol. The total synthesis was approximately 90 min, including labeling, purification, and formulation of the final product. The maximum isolated amount was 1.2-2 GBq (EOS) starting from~17 GBq of fluoride-18. The [ 18 F]45 rapidly reacted with TCO-PNP carbonate, and was stable in PBS at room temperature for minimum of 4 h, as confirmed by radio-HPLC ( Figure 4C).

In Vivo Pretargeting PET Imaging with [ 18 F]45
The ability of [ 18 F]45 to act as a pretargeting imaging agent was evaluated in LS174T tumor-bearing mice. Mice were grouped and administered with either monoclonal anti-TAG72 antibody, CC49, or TCO-conjugated CC49 (CC49-TCO) 72 h prior to injection of [ 18 F]45. PET/CT scans were acquired 1 h later, as well as image-derived uptake values from tumor, heart, and muscle tissue extracted by manually created regions of interests ( Figure  5).

In Vivo Pretargeting PET Imaging with [ 18 F]45
The ability of [ 18 F]45 to act as a pretargeting imaging agent was evaluated in LS174T tumor-bearing mice. Mice were grouped and administered with either monoclonal anti-TAG72 antibody, CC49, or TCO-conjugated CC49 (CC49-TCO) 72 h prior to injection of [ 18 F]45. PET/CT scans were acquired 1 h later, as well as image-derived uptake values from tumor, heart, and muscle tissue extracted by manually created regions of interests ( Figure 5). The tumors of the mice in the group receiving CC49-TCO had a significant higher uptake of [ 18 F]45 (1.8 ± 0.3%ID/mL, mean ± SEM, n = 5) compared to the animals in the control group (0.3 ± 0.1%ID/mL, n = 5). Uptake in muscle tissue was generally low, resulting in a tumor-to-muscle ratio of 8.3 in animals pretargeted with CC49-TCO. In contrast, the uptake in blood (heart used as a surrogate) in this group of animals was high, resulting in a tumor-to-blood (T/B) ratio of 0.8. The tumor-to-liver (T/L) ratio was 0.3. Therefore, the uptake and target-to-background contrast found for [ 18 F]45 were at comparable levels to previous findings [17,32,34]. However, there is a general need to improve the target-tobackground contrast before clinical translation. This could either be achieved by developing primary and secondary vectors with improved performance, thereby increasing the tumor uptake. Further, the use of a clearing agent could also potentially accelerate clearance of the primary vector and thus improve the T/B ratio in the future [14]. The tumors of the mice in the group receiving CC49-TCO had a significant higher uptake of [ 18 F]45 (1.8 ± 0.3%ID/mL, mean ± SEM, n = 5) compared to the animals in the control group (0.3 ± 0.1%ID/mL, n = 5). Uptake in muscle tissue was generally low, resulting in a tumor-to-muscle ratio of 8.3 in animals pretargeted with CC49-TCO. In contrast, the uptake in blood (heart used as a surrogate) in this group of animals was high, resulting in a tumor-to-blood (T/B) ratio of 0.8. The tumor-to-liver (T/L) ratio was 0.3. Therefore, the uptake and target-to-background contrast found for [ 18 F]45 were at comparable levels to previous findings [17,32,34]. However, there is a general need to improve the target-to-background contrast before clinical translation. This could either be achieved by developing primary and secondary vectors with improved performance, thereby increasing the tumor uptake. Further, the use of a clearing agent could also potentially accelerate clearance of the primary vector and thus improve the T/B ratio in the future [14].

General
All reagents and solvents were purchased from commercial suppliers and used without further purification. Anhydrous tetrahydrofuran (THF) was obtained from an SG water solvent purification system (Pure Process Technology). Anhydrous dimethyl sulfoxide (DMSO), N,N-dimethylacetamide (DMA), MeCN, and pyridine were purchased from commercially suppliers and stored under argon. Reactions requiring anhydrous conditions were carried out under an inert atmosphere (nitrogen or argon) and using oven-dried glassware (152 • C). Syringes used to transfer anhydrous solvents or reagents were purged with argon prior to use. Other solvents were analytical or HPLC grade and were used as received. NMR spectra were acquired on a 600 MHz Bruker Avance III HD ( Thin-layer chromatography (TLC) was carried out on silica gel 60 F 254 plates from Merck (Germany). Visualization was accomplished by UV lamp (254 nm). Preparative HPLC was carried out on an UltiMate HPLC system (Thermo Scientific) consisting of an LPG-3200BX pump (10 mL/min), a Rheodyne 9725i injector, a 10 mL loop, a MWD-3000SD detector (254 nm), and an AFC-3000SD automated fraction collector, using a Gemini-NX C18 column (21.2 × 250 mm, 5 µm, 110Å) (Phenomenex) equipped with a guard. Purifications were performed using linear gradients of 0.1% TFA in MilliQ-H 2 O (A) and 0.1% TFA, 10% MilliQ-H 2 O in MeCN (B). Data were acquired and processed using Chromeleon software v. 6.80. Semipreparative HPLC was performed on the same system using a Luna 5µ C18 column (250 × 10 mm) with a flow rate of 3 mL/min. Automated flash column chromatography was performed on a CombiFlash NextGen 300+ system supplied by TeleDyne ISCO, equipped with RediSep silica-packed columns. Detection of the compounds was carried out by means of a UV-vis variable wavelength detector operating at 200 to 800 nm and by an evaporative light-scattering detector (ELSD). Solvent systems for separation were particular for each compound, but consisted of various mixtures of heptane, EtOAc, CH 2 Cl 2 , and MeOH. Microwave-assisted synthesis was carried out in a Biotage Initiator apparatus operating in single mode; the microwave cavity produced controlled irradiation at 2.45 GHz (Biotage AB, Uppsala, Sweden). The reactions were run in sealed vessels. These experiments were performed by employing magnetic stirring and a fixed hold time using variable power to reach (during 1-2 min) and then maintain the desired temperature in the vessel for the programmed time period. The temperature was monitored by an IR sensor focused on a point on the reactor vial glass. The IR sensor was calibrated to the internal solution's reaction temperature by the manufacturer. Mass spectra analysis was completed using MS-Acquity-A: Waters Acquity UPLC with QDa-detector. CC49-TCO was kindly provided by Tagworks Pharmaceuticals, and it was obtained as previously described [14].

Tetrazine Core Reactivity Test
The reaction between [ 18 F]45 and TCO-PNB was performed by mixing the formulated [ 18 F]45 (200 µL) with 5 µL of the commercially available TCO-PNB ester dissolved in DMF (5 mg/mL) in an analytical HPLC vial. The solution was gently shaken and left for 1 min before it was injected into the analytical HPLC instrument for analysis. The human colon cancer cell line LS174T (ATCC, Manassas, VA, USA) was cultured in minimum essential medium (MEM) supplemented with 10% fetal bovine serum, 1% L-glutamine, 1% sodium pyruvate, 1% nonessential amino acids, and 1% penicillinstreptomycin (all from Thermo Fisher Scientific, Waltham, MA, USA). The cells were trypsinized and harvested for inoculation when they were in their exponential growth phase.

Blocking Assay and Ex Vivo Studies
Subcutaneous tumors were established in the flank of five-week-old female nude BALB/c mice (Janvier Labs, Le Genest-Saint-Isle, France) by inoculation of~5 × 10 6 LS174T cells (in 100 µL sterile PBS).
The tumor volume was estimated from caliper measurements using the formula: volume = 1 2 (length × width 2 ).

Blocking Experiments
The blocking experiments were performed as previously described [17,32]. Briefly, tumor-bearing animals were grouped based on their tumor volume (~100-300 mm 3 , n = 3/group) and administered 100 µg/100 µL of CC49-TCO per mouse (~7 TCO/mAb). The ability of nonradioactive Tzs to block the binding between [ 111 In]47 and CC49-TCO was evaluated three days later. First, the animals were injected with the (nonradioactive) Tzs (39 nmol) that were chosen for in vivo evaluation. After a lag time of 1 h, [ 111 In]47 (~5 MBq, 3.9 nmol) was administered, and after 22 h, the mice were euthanized. Tissues were resected and weighted, and the radioactivity was measured using a gamma counter (Wizard2, Perkin Elmer). Data were corrected for decay, tissue weight, and injected amount of radioactivity. The setup also included a control group of mice receiving the precursor of [ 111 In]47 instead of a test compound 1 h prior to [ 111 In]47 (positive control), as well as a group exclusively receiving [ 111 In]47. The tumor uptake of the different evaluated Tzs was normalized to the tumor uptake of the latter to determine the blocking effect.

Pretargeted Imaging
LS174T xenografts were established in mice as previously described. When tumors reached a size of~100 mm 3 , animals were divided into two groups based on their tumor volume (n = 5/group), and injected iv with 50 µg in 100 µL of either CC49-TCO or CC49. After 72 h the animals were administered intravenously with [ 18 F]45 (4.74 ± 1.39 MBq in 100 µL of PBS) and scanned using PET/CT (Inveon, Siemens Medical Solutions, USA) 1 h later, using a PET acquisition time of 5 min, an energy window of 350-650 KeV, and a time resolution of 6 ns; followed by a continuous 360 projection/360 • CT scan, acquired with an X-ray tube voltage of 65 kV, a tube anode current of 500 µA, and an exposure time per projection of 400 ms. Mice were anesthetized by breathing 3% sevoflurane (5% for induction) mixed in 35% O 2 in N 2 , and during scans the body temperature of the mice was kept stable using a heating pad.
Signograms from PET scans were reconstructed using a three-dimensional maximum a posteriori algorithm with correction for scatter and attenuation. The mean percentage of injected dose per grams (%ID/mL) was determined by manually creating regions of interest (ROI) on coregistered PET/CT images (Inveon Research Workplace software, Siemens Medical Solutions, Malvern, PA, USA).
GraphPad Prism 9 (GraphPad Software, La Jolla, CA, USA) was used for statistical analysis, and an unpaired t-test with Welch's correction was used to compare the tumor uptake in the two groups. Results were considered significant when p < 0.05.

Conclusions
In this study, we reported the development of the first bispyridyl Tz directly labeled with fluorine-18. The [ 18 F]45 was obtained with sufficient yield, purity, and molar activity for in vivo evaluation. This imaging agent had comparable performances and target-to-background ratios compared to previously reported successful Tz tracers. Further evaluation studies are needed in order to optimize its potential for pretargeted imaging. The developed [ 18 F]45 is of special interest in pretargeted imaging, as it may be used-as a bispyridyl-based pretargeted imaging agent-to quantify drug release based on 'click-torelease' approaches. Bispyridyls have recently been shown to near-quantitatively release drugs from iTCOs within <10 min [41]. This is so far the fastest release demonstrated. The ability to quantify in vivo drug release is essential to precisely fine-tune dosing-one of the most important factors influencing the effectiveness of a treatment.