Side-by-Side Comparison of Five Chelators for 89Zr-Labeling of Biomolecules: Investigation of Chemical/Radiochemical Properties and Complex Stability

Simple Summary The positron emitter 89Zr4+ is an important radionuclide for the preparation of radiolabeled antibodies, being applied in highly specific and sensitive positron emission tomography (PET) imaging of malignancies. The introduction of 89Zr4+ into biomolecules is performed using chelating agents, wrapping up the radiometal and preventing its release from the antibody by forming so-called complexes. Desferrioxamine B (DFO) is the clinical gold standard chelator for the preparation of 89Zr antibodies despite its known inability to stably encapsulate the radiometal, resulting in 89Zr release and associated challenges such as decreased image quality and radiation dose to healthy tissues. Therefore, several research groups have been working to develop new chelating agents able to stably encapsulate the 89Zr4+ ion. However, there are no data available directly comparing the stability of the formed 89Zr complexes of the most promising chelating agents developed so far. Here, we report on the comparison of five different chelators with high potential for stable complexation of 89Zr and determined two of them—DFO* and 3,4,3-(LI-1,2-HOPO)—to be highly interesting for the preparation of 89Zr-based radiolabeled agents and routine clinical application. Abstract In this work, five different chelating agents, namely DFO, CTH-36, DFO*, 3,4,3-(LI-1,2-HOPO) and DOTA-GA, were compared with regard to the relative kinetic inertness of their corresponding 89Zr complexes to evaluate their potential for in vivo application and stable 89Zr complexation. The chelators were identically functionalized with tetrazines, enabling a fully comparable, efficient, chemoselective and biorthogonal conjugation chemistry for the modification of any complementarily derivatized biomolecules of interest. A small model peptide of clinical relevance (TCO-c(RGDfK)) was derivatized via iEDDA click reaction with the developed chelating agents (TCO = trans-cyclooctene and iEDDA = inverse electron demand Diels-Alder). The bioconjugates were labeled with 89Zr4+, and their radiochemical properties (labeling conditions and efficiency), logD(7.4), as well as the relative kinetic inertness of the formed complexes, were compared. Furthermore, density functional theory (DFT) calculations were conducted to identify potential influences of chelator modification on complex formation and geometry. The results of the DFT studies showed—apart from the DOTA-GA derivative—no significant influence of chelator backbone functionalization or the conjugation of the chelator tetrazines by iEDDA. All tetrazines could be efficiently introduced into c(RGDfK), demonstrating the high suitability of the agents for efficient and chemoselective bioconjugation. The DFO-, CTH-36- and DFO*-modified c(RGDfK) peptides showed a high radiolabeling efficiency under mild reaction conditions and complete 89Zr incorporation within 1 h, yielding the 89Zr-labeled analogs as homogenous products. In contrast, 3,4,3-(LI-1,2-HOPO)-c(RGDfK) required considerably prolonged reaction times of 5 h for complete radiometal incorporation and yielded several different 89Zr-labeled species. The labeling of the DOTA-GA-modified peptide was not successful at all. Compared to [89Zr]Zr-DFO-, [89Zr]Zr-CTH-36- and [89Zr]Zr-DFO*-c(RGDfK), the corresponding [89Zr]Zr-3,4,3-(LI-1,2-HOPO) peptide showed a strongly increased lipophilicity. Finally, the relative stability of the 89Zr complexes against the EDTA challenge was investigated. The [89Zr]Zr-DFO complex showed—as expected—a low kinetic inertness. Unexpectedly, also, the [89Zr]Zr-CTH-36 complex demonstrated a high susceptibility against the challenge, limiting the usefulness of CTH-36 for stable 89Zr complexation. Only the [89Zr]Zr-DFO* and the [89Zr]Zr-3,4,3-(LI-1,2-HOPO) complexes demonstrated a high inertness, qualifying them for further comparative in vivo investigation to determine the most appropriate alternative to DFO for clinical application.


Introduction
In medical applications, whole-body imaging of malignant tissue represents a standard procedure for the diagnosis of cancer. For this purpose, different methods such as magnetic resonance imaging (MRI), computed tomography (CT) and positron emission tomography (PET) are used, each with its own specific advantages. PET, for example, offers the unique advantage of requiring such small amounts of radiotracers that target structure-specific imaging of malignant tissues becomes possible with high sensitivity. Thus, PET allows not only delineation but also the functional characterization of tumors. Furthermore, PET enables the determination of an appropriate dose of a potential endoradiotherapeutic agent and the assessment of a therapeutic response to a therapeutic agent, as well as therapy monitoring.
The radiotracers used can be based on known drugs, comprising compounds of low molecular weight, peptides and artificial functionalized nanocarriers, as well as antibodies. Antibodies have the advantage to bind to their target structure with very high specificity and binding strength, leading to an improved accumulation at the target site. Hence, radiolabeled antibodies can be used to obtain high-contrast diagnostic images that delineate the tumor with high specificity and sensitivity. Due to the relatively slow in vivo pharmacokinetics and long blood pool residence time, the physical half-life of the radionuclide used for the labeling of antibodies has to match its biological half-life. This is the reason why 89 Zr 4+ is mostly used for antibody labeling, as it has a physical half-life of 3.3 days, perfectly matching the slow pharmacokinetics of antibodies. The increasing number of clinical studies performed with 89 Zr-labeled antibodies also reflects the clinical relevance of the compound class [1].
In clinical studies, 89 Zr 4+ is usually introduced into the biomolecule carrier using desferrioxamine B (DFO; Figure 1A), a natural siderophore. However, there is strong evidence that the kinetic inertness of the formed 89 Zr-DFO complex is limited [2][3][4], resulting in release of the 89 Zr cation. Free 89 Zr 4+ attaches to the hydroxyapatite of bones, resulting in significant uptake into bones and joints. This is of course problematic, as it can result in a relevant dose to hematopoietic bone marrow and, furthermore, reduces the quality of the obtained images. On the one hand, this is caused by the higher background accumulation of freely circulating 89 Zr 4+ , and on the other hand, the bone uptake compromises the visualization of bone metastases. For this reason, several new chelating agents have been developed over the last years, some of which have been shown to be significantly more suitable than DFO to form kinetically inert complexes with 89 Zr [3,5]. All of these attempts to develop stable 8 9 Zr complexes rely on the complete saturation of the coordination sphere of 89 Zr 4+ . In the 89 Zr-DFO complex, the three hydroxamates occupy only six of the preferred eight coordination sites of the Zr 4+ ion, leaving a gap in the ligand sphere where other ions and molecules can interact, destabilize or break the complex. A complete saturation of the coordination sphere, together with a complete spatial embedment of the central ion, are thus equally important for the formation of stable 89 Zr complexes.
There has been no comparison of these most promising representatives of this class of new chelating agents for 89 Zr regarding their kinetic inertness. It is thus still not clear which chelator is the most suitable for clinical translation to replace the commonly used gold standard DFO in clinical applications.
Very recently, a highly interesting study reported on the prediction of the thermodynamic stability of different 89 Zr-based radiotracers [6]. Here, the absolute and relative formation constants of 23 different zirconium complexes were determined by means of density functional theory (DFT) calculations. In this study, which differentiated between DFO chelator analogs and alternative chelating agents, it was shown that some of the complexes investigated exhibited very promising thermodynamic stabilities.
In the group of alternative chelating agents, the [ 89 Zr]Zr-CTH36 complex ( Figure 1B) deserves special consideration, as it exhibited a particularly high complexation constant of β = 52.84. This is the result of a nearly optimal complex geometry that is close to the lowest energy structure of Zr(MeAHA)4, being formed by Zr 4+ and bidentate MeAHA (Nmethyl-acetohydroxamic acid), and the macrocyclic structure of the chelator. This macrocyclic structure results in a preorganization of the hydroxamates and in a reduced entropic penalty during complex formation compared to acyclic chelates (such as DFO). The high flexibility, due to the eight atom chains between each set of carbonyl and N−O donor groups, further reduces steric strain and allows the donor atoms to adopt the preferred geometry.
Within the group of DFO-based chelators, [ 89 Zr]Zr-DFO* ( Figure 1B) also showed a very high formation constant of β = 51.56, which can be explained by the complete saturation of the coordination sphere of the Zr 4+ ion.
These theoretical considerations are supported by experimental studies demonstrating that the chelators CTH36 and DFO* form complexes of significantly increased stability compared to DFO in in silico complex challenges and/or in vivo imaging studies [7][8][9][10]. Therefore, these two chelating agents are of high interest with regard to further comparative investigation and also potential clinical application.
Other chelating agents that showed a significantly higher stability of the formed 89 Zr complexes were 3,4,3-(LI-1,2-HOPO) [11,12] and DOTA (1,4,7,10-tetraazacyclododecane- For this reason, several new chelating agents have been developed over the last years, some of which have been shown to be significantly more suitable than DFO to form kinetically inert complexes with 89 Zr [3,5]. All of these attempts to develop stable 89 Zr complexes rely on the complete saturation of the coordination sphere of 89 Zr 4+ . In the 89 Zr-DFO complex, the three hydroxamates occupy only six of the preferred eight coordination sites of the Zr 4+ ion, leaving a gap in the ligand sphere where other ions and molecules can interact, destabilize or break the complex. A complete saturation of the coordination sphere, together with a complete spatial embedment of the central ion, are thus equally important for the formation of stable 8 9 Zr complexes.
There has been no comparison of these most promising representatives of this class of new chelating agents for 89 Zr regarding their kinetic inertness. It is thus still not clear which chelator is the most suitable for clinical translation to replace the commonly used gold standard DFO in clinical applications.
Very recently, a highly interesting study reported on the prediction of the thermodynamic stability of different 89 Zr-based radiotracers [6]. Here, the absolute and relative formation constants of 23 different zirconium complexes were determined by means of density functional theory (DFT) calculations. In this study, which differentiated between DFO chelator analogs and alternative chelating agents, it was shown that some of the complexes investigated exhibited very promising thermodynamic stabilities.
In the group of alternative chelating agents, the [ 89 Zr]Zr-CTH36 complex ( Figure 1B) deserves special consideration, as it exhibited a particularly high complexation constant of β = 52.84. This is the result of a nearly optimal complex geometry that is close to the lowest energy structure of Zr(MeAHA) 4 , being formed by Zr 4+ and bidentate MeAHA (N-methyl-acetohydroxamic acid), and the macrocyclic structure of the chelator. This macrocyclic structure results in a preorganization of the hydroxamates and in a reduced entropic penalty during complex formation compared to acyclic chelates (such as DFO). The high flexibility, due to the eight atom chains between each set of carbonyl and N−O donor groups, further reduces steric strain and allows the donor atoms to adopt the preferred geometry.
Within the group of DFO-based chelators, [ 89 Zr]Zr-DFO* ( Figure 1B) also showed a very high formation constant of β = 51.56, which can be explained by the complete saturation of the coordination sphere of the Zr 4+ ion.
These theoretical considerations are supported by experimental studies demonstrating that the chelators CTH36 and DFO* form complexes of significantly increased stability compared to DFO in in silico complex challenges and/or in vivo imaging studies [7][8][9][10]. Therefore, these two chelating agents are of high interest with regard to further comparative investigation and also potential clinical application.
Other chelating agents that showed a significantly higher stability of the formed 89 Zr complexes were 3,4,3-(LI-1,2-HOPO) [11,12] and DOTA (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetrayl)tetraacetic acid) [13] (Figure 1B), which are thus additional promising candidates for further comparative investigation of complex stability and clinical applicability. The aim of the current study was therefore to directly compare the mentioned four chelating agents, as well as the commonly used DFO with regard to the relative kinetic inertness of the 89 Zr complexes formed under identical conditions for direct comparability of the obtained results, and to be able to identify the most useful chelating agent for stable 89 Zr complexation.
For this, analogs of these chelating agents were to be developed enabling an efficient introduction into biomolecules by a chemoselective and biorthogonal conjugation reaction to facilitate a high-yield derivatization of even sensitive biomolecules such as antibodies. For this purpose, a necessary functional group for bioconjugation had to be introduced in a position of the molecular structure of the chelators not interfering with 89 Zr complex formation. This requires a backbone functionalization of the respective chelators, leaving the hydroxamate or carboxylate functional groups needed for 89 Zr complexation uncompromised. Furthermore, the same biorthogonal and chemoselective conjugation reaction should find application in all cases, thus excluding the possibility that the bioconjugation chemistry itself influences 89 Zr complex formation or kinetic inertness.
A popular and customizable click chemistry reaction is the inverse electron demand Diels-Alder (iEDDA) conjugation reaction between tetrazines and TCOs (TCO = transcyclooctene), which has already found widespread application in radiochemistry [14][15][16]. For this reason, we decided (i) to synthesize backbone tetrazine-modified analogs of DFO, CTH-36, DFO*, 3,4,3-(LI-1,2-HOPO) and DOTA, leaving the coordination sphere of the respective agents unaltered to achieve a high kinetic inertness of the resulting 89 Zr complexes, and (ii) to introduce them into c(RGDfK); (iii) to radiolabel the resulting bioconjugates with 89 Zr and (iv) to determine the inertness of the resulting equally modified and conjugated 89 Zr complexes by challenge experiments under identical conditions to be able to directly compare and evaluate their relative stability, with the aim to identify the most promising candidate for stable 89 Zr complexation.
Analytical and semipreparative high-performance liquid chromatography (HPLC) analyses and purifications were carried out on Dionex UltiMate 3000 systems (Thermo Fisher) equipped with Chromolith Performance (RP-18e, 100-4.6 mm, Merck, Darmstadt, Germany) or Chromolith SemiPrep (RP-18e, 100-10 mm, Merck) columns, respectively. A flow rate of 4 mL/min and the eluents H 2 O and acetonitrile (MeCN) containing either 0.1% trifluoroacetic acid (TFA) or 0.1% formic acid (FA) were used. Nuclear magnetic resonance (NMR) spectroscopy was carried out on a 500-MHz Varian NMR System spectrometer, a 700-MHz Bruker Avance III HD NMR spectrometer and a 300-MHz MERCURYplus NMR spectrometer, respectively. The signals of the deuterated solvents were used as references. All chemical shifts (δ) are reported in ppm and the coupling constants (J) in Hz. The matrix-assisted laser desorption/ionization mass spectroscopy (MALDI-MS) was carried out with on a Bruker Daltonics Microflex spectrometer (Bremen, Germany) and for the highresolution electrospray ionization mass spectroscopy (HR-ESI-MS), a Thermo Finnigan LTQ FT Ultra Fourier Transform Ion Cyclotron Resonance (Dreieich, Germany) mass spectrometer was used. Radioactivity was measured using an ISOMED 2010 (Kappeln, Germany) activimeter. Analytical radio-HPLC chromatography was performed on a Dionex UltiMate 3000 system (Thermo Fischer, Dreieich, Germany) equipped with a radio detector GabiStar (Raytest) and a Gemini column (C18, 5 µm, 250-4.6 mm, Phenomenex) at a flow rate of 2 mL/min using the eluents H 2 O and MeCN containing 0.1% TFA. As the gamma counter, the 2480 Wizard system (PerkinElmer) was used. Radio-iTLC (instant thin-layer chromatography) analyses were carried out using ITLC-SG strips (Agilent Technologies) together with citrate buffer as the eluent (0.1M, pH 5), which were analyzed using a Scan-RAM radio-TLC scanner (LabLogic) using LAURA software (Jahnsdorf, Germany, for the analyses of radio-HPLC, TLC and GC chromatography, version: 4.1.12.89). (4-(1,2,4,5-Tetrazin-3-yl)phenyl)methanamine hydrochloride 6 (15 mg, 67.1 µmol) was added to a solution of succinic anhydride (8 mg, 80.5 µmol) in DMF (1 mL). After the addition of triethylamine (9.3 µL, 67.1 µmol), the mixture was stirred under exclusion of light for 4 h at ambient temperature. Then, the solvent was removed under reduced pressure, and the crude product was purified by semipreparative HPLC using a gradient of 0-40% MeCN + 0.1% TFA in 8 min (R t = 5.51 min). Finally, the product was isolated as pink solid in a yield of 86% (17 mg, 57.8 µmol). 1  trifluoroacetic acid (TFA) or 0.1% formic acid (FA) were used. Nuclear magne (NMR) spectroscopy was carried out on a 500-MHz Varian NMR System sp 700-MHz Bruker Avance III HD NMR spectrometer and a 300-MHz MERCU spectrometer, respectively. The signals of the deuterated solvents were used All chemical shifts (δ) are reported in ppm and the coupling constants (J) in trix-assisted laser desorption/ionization mass spectroscopy (MALDI-MS) w with on a Bruker Daltonics Microflex spectrometer (Bremen, Germany) and resolution electrospray ionization mass spectroscopy (HR-ESI-MS), a Ther LTQ FT Ultra Fourier Transform Ion Cyclotron Resonance (Dreieich, Ge spectrometer was used. Radioactivity was measured using an ISOMED 20 Germany) activimeter. Analytical radio-HPLC chromatography was perfor onex UltiMate 3000 system (Thermo Fischer, Dreieich, Germany) equipped detector GabiStar (Raytest) and a Gemini column (C18, 5 µm, 250-4.6 mm, at a flow rate of 2 mL/min using the eluents H2O and MeCN containing 0.1% gamma counter, the 2480 Wizard system (PerkinElmer) was used. Radiothin-layer chromatography) analyses were carried out using ITLC-SG st Technologies) together with citrate buffer as the eluent (0.1M, pH 5), which w using a Scan-RAM radio-TLC scanner (LabLogic) using LAURA software (Ja many, for the analyses of radio-HPLC, TLC and GC chromatography, versi

DFO* 17
The synthesis was carried out according to the literature, with minor chan DFO*-Cbz (100 mg, 0.10 mmol) was dissolved in methanol (80 mL). A small am  was added, and the reaction mixture was stirred for 30 h. D reduced pressure, and the residue was washed three times wi times with distilled water (3 mL). After lyophilization, the pro orless solid in a yield of 67%. 1

DFO* 17
The synthesis was carried out according to the literature, DFO*-Cbz (100 mg, 0.10 mmol) was dissolved in methanol (80 of Pd/C (30 mg) was added. The reaction was stirred under (balloon) for 12 h at ambient temperature. The Pd/C was filt

DFO* 17
The synthesis was carried out according to the literature, with minor changes [7]. Bn-DFO*-Cbz (100 mg, 0.10 mmol) was dissolved in methanol (80 mL). A small amount (10%) of Pd/C (30 mg) was added. The reaction was stirred under hydrogen gas atmosphere (balloon) for 12 h at ambient temperature. The Pd/C was filtered off, and DFO* was obtained after the evaporation of methanol in a yield of 70% (54 mg, 0.07 mmol). 1

Tris-Boc-spermine-trifluoroacetamide 18 and Tris-Boc-spermine 19
The synthesis of both compounds was carried out following in principle a published procedure [12]. Spermine (526 mg, 2.60 mmol) was dissolved in methanol (40 mL), and the solution was cooled in an acetone-liquid nitrogen cooling bath to −78 • C. A solution of ethyl trifluoroacetate (0.44 mL, 2.6 mmol) in methanol (26 mL) was added dropwise over one hour under argon atmosphere. The reaction mixture was stirred another 30 min at −78 • C before the temperature was slowly increased to 0 • C. A solution of di-tertbutyldicarbonate (3.583 g, 15.60 mmol) in methanol (26 mL) was added, and the reaction was stirred overnight. The solvent was removed on a rotary evaporator, and the obtained residue was redissolved in DCM (20 mL) and washed three times with water (15 mL). The crude product 18 was purified by column chromatography using a gradient of 30-100% ethyl acetate in cyclohexane + 1% NH 3 (aq.). The product could be isolated in pure form, but residues of tetra-Boc-spermine-trifluoroacetamide were still contained. 1   The synthesis of both compounds was carried out following in principle a publis procedure [12]. Spermine (526 mg, 2.60 mmol) was dissolved in methanol (40 mL), the solution was cooled in an acetone-liquid nitrogen cooling bath to −78 °C. A solu of ethyl trifluoroacetate (0.44 mL, 2.6 mmol) in methanol (26 mL) was added dropw over one hour under argon atmosphere. The reaction mixture was stirred another 30 at −78 °C before the temperature was slowly increased to 0 °C. A solution of di-terttyldicarbonate (3.583 g, 15.60 mmol) in methanol (26 mL) was added, and the reaction stirred overnight. The solvent was removed on a rotary evaporator, and the obtained idue was redissolved in DCM (20 mL) and washed three times with water (15 mL). crude product 18 was purified by column chromatography using a gradient of 30-10 ethyl acetate in cyclohexane + 1% NH3 (aq.). The product could be isolated in pure fo but residues of tetra-Boc-spermine-trifluoroacetamide were still contained. 1  18 was used in the next step w out further purification. For this purpose, 18 was dissolved in methanol (40 mL), an 30% ammonia solution was added to adjust the pH to 11-12. The mixture was stirred four days, until no more 18 was detected. The product 19 was obtained as a mixture w tetra-Boc-spermine (overall yield: quant.). 1  2.2.8. 1-Hydroxy-6-oxo-1,6-dihydropyridine-2-carbonic Acid 21 The synthesis was carried out on the basis of the literature protocol with mi changes [20]. 6-Hydroxypicolinic acid (2.00 g, 14.38 mmol) was dissolved in acetic a (8.2 mL) and neat TFA (12.2 mL). Under argon atmosphere, a 35% solution of perac acid in acetic acid (6.9 mL, 36.82 mmol) was added, and the reaction was stirred 1 ambient temperature, followed by stirring 17 h at 80°C. Next, the mixture was cooled °C for 12 h. The formed precipitate was filtered off and washed with ice-cold metha Compound 21 was obtained as a colorless solid in a yield of 69% (1.545 g, 4.149 mm 2.2.8. 1-Hydroxy-6-oxo-1,6-dihydropyridine-2-carbonic Acid 21 The synthesis was carried out on the basis of the literature protocol with minor changes [20]. 6-Hydroxypicolinic acid (2.00 g, 14.38 mmol) was dissolved in acetic acid (8.2 mL) and neat TFA (12.2 mL). Under argon atmosphere, a 35% solution of peracetic acid in acetic acid (6.9 mL, 36.82 mmol) was added, and the reaction was stirred 1 h at ambient temperature, followed by stirring 17 h at 80 • C. Next, the mixture was cooled at 4 • C for 12 h. The formed precipitate was filtered off and washed with ice-cold methanol. Compound 21 was obtained as a colorless solid in a yield of 69% (1.545 g, 4.149 mmol). 1

t Butyl-(4-(2-((3-((4-((3-aminopropyl)amino)butyl)amino)propyl)amino)ethyl)phenyl)-carbamate 25
The synthesis was carried out on the basis of the literature protocol [20]. Spermine (300 mg, 1.48 mmol) was dissolved in acetonitrile (45 mL), and potassium carbonate (410 mg, 2.97 mmol) was added. A solution of 4-((tert-butoxy-carbonyl)amino)-phenethyl-4-methylbenzenesulfonate 24 (296 mg, 0.76 mmol) in acetonitrile (45 mL) was prepared and added dropwise while cooling with ice. After the reaction was refluxed for 16 h, the potassium carbonate was filtered off, and the solvent was removed on a rotary evaporator. The residue was redissolved in acetonitrile/water and purified via semipreparative HPLC using a gradient 0-35% MeCN + 0.1% FA in 8 min (R t = 6.35 min). After lyophilization, the product was isolated as a colorless solid in a yield of 63% (202 mg, 48 mmol). 1 The synthesis was carried out on the basis of the literature protocol [20]. Spermine (300 mg, 1.48 mmol) was dissolved in acetonitrile (45 mL), and potassium carbonate (410 mg, 2.97 mmol) was added. A solution of 4-((tert-butoxy-carbonyl)amino)-phenethyl-4methylbenzenesulfonate 24 (296 mg, 0.76 mmol) in acetonitrile (45 mL) was prepared and added dropwise while cooling with ice. After the reaction was refluxed for 16 h, the potassium carbonate was filtered off, and the solvent was removed on a rotary evaporator. The residue was redissolved in acetonitrile/water and purified via semipreparative HPLC using a gradient 0-35% MeCN + 0.1% FA in 8 min (Rt = 6.35 min). After lyophilization, the product was isolated as a colorless solid in a yield of 63% (202 mg, 48 mmol). 1    1-(Benzyloxy)-6-oxo-1,6-dihydropyridin-2-carbonic acid 22 (0.393 g, 1.60 mmol) was dissolved in benzene (5.2 mL). Oxalyl chloride (0.25 mL, 2.90 mmol) was added dropwise under a flow of nitrogen. Every two hours, two drops of DMF were added until no more gas formation was observed (6 h). Benzene and the excess of oxalyl chloride were removed by evaporation. The obtained 1,2-HOPO acid chloride was used in the next step without further purification and was dissolved in DCM (4.7 mL). This solution was added drop- 1-(Benzyloxy)-6-oxo-1,6-dihydropyridin-2-carbonic acid 22 (0.393 g, 1.60 mmol) was dissolved in benzene (5.2 mL). Oxalyl chloride (0.25 mL, 2.90 mmol) was added dropwise under a flow of nitrogen. Every two hours, two drops of DMF were added until no more gas formation was observed (6 h). Benzene and the excess of oxalyl chloride were removed by evaporation. The obtained 1,2-HOPO acid chloride was used in the next step without further purification and was dissolved in DCM (4.7 mL). This solution was added dropwise to a mixture of 25 (0.130 g, 0.31 mmol) in DCM (3.5 mL) and a potassium carbonate solution (40%, 0.70 mL) over a period of 30 min. The mixture was stirred for 48 h at ambient temperature. In the following, the water phase was extracted with DCM (3 × 10 mL). After evaporation, the crude product was purified by semipreparative HPLC using a gradient of 10-100% MeCN + 0.1% TFA in 12 min (R t = 6.35 min). The product was isolated as a colorless solid in a 17% yield (70.3 mg, 52.8 µmol). 1

3,4,3-(LI-1,2-HOPOBn)-NH3Cl 27
The synthesis was established on the basis of a pub HOPOBn)-Ph-NH-Boc, 26 (60.0 mg, 0.045 mmol), was diss mL) under Ar atmosphere. A solution of BCl3 in p-xylene added, and the reaction was stirred for 16 h at ambient tem precipitate was centrifuged off and washed with acetone × 20 mL). The product was obtained as a colorless solid mmol). 1

Determination of logD (7.4)
To a mixture of 1-octanol (800 µL) and 0.05-M phosphate buffer (775 µL, pH 7.4), a solution of the respective 89 Zr-labeled peptide conjugate (5µL of the before-obtained product solution, 0.8-1.2 MBq, 0.37-0.45 nmol) was added. The mixture was vigorously shaken for two minutes. Afterwards, the phases were separated by centrifugation. Two hundred microliters of each, the organic and the aqueous, phase were measured in a gamma counter. Each experiment was performed at least thrice, each in duplicate.

EDTA-Based 89 Zr Complex Challenge Experiments
For each challenge experiment, three separate solutions of EDTA (14.61 mg, 50 µmol EDTA) in HEPES buffer (0.25 M, pH = 7.0, 380 µL) were prepared. The pH of the solutions was adjusted to pH 7.0 by the addition of a NaOH solution (30%, 12.5 µL), and Tracepur water (7.5 µL) was added to give a final volume of 400 µL of the EDTA solutions. To these solutions was added a solution of the respective 89 Zr-labeled peptide conjugate (5 nmol, 9.7-13.9 MBq), and the solutions were kept at 25 • C over the course of the experiment. At predefined time points, each solution was analyzed by analytical radio-HPLC for determination of the amount of 89 Zr transchelation. Each experiment was performed at least twice, each in triplicate.

DFT Calculations
The DFT calculations were all conducted as implemented in Spartan'20 (1.0.0) [21] using B3LYP [22][23][24] exchange correlation functionals, and 6-31G * polarization basis sets were assigned for all elements, at which LANL2DZ [25,26] with an effective core potential was employed for the 4-d transition metal zirconium. Solvated-phase calculations (C-PCM dielectric constant = 78.30) were used as implemented using the PCMRAD keyword (PCMRAD = ZR~2.68). The characterization of each optimized structure as the local minimum on the potential energy surface was carried out by a harmonic frequency analysis based on the second derivative.
Start geometries for the structure optimization were taken from the literature [6] and extended with respect to the new conjugation in Spartan. Only that part of the TCO reactant essential for the conjugation was added, and the rest of the biomolecule was omitted. The structure was subsequently optimized in Spartan.

Synthesis of the Backbone Tetrazine-Modified Chelator Analogs 1-5
To reach the outlined aims, the syntheses of the backbone tetrazine-modified analogs of the target chelators (1-5, Figure 2) had to be established. Of these, 2 [10] and 5 [17] were synthesized according to literature procedures. Compound 1 was prepared following a previously described synthesis route, with some modifications in the reaction order [10] by first reacting the tetrazine amine 6 with succinic anhydride and the resulting acid 7 with DFO mesylate using PyBOP (benzotriazole-1-yloxy-tris-pyrrolidino-phosphonium hexafluorophosphate) as the coupling agent (Scheme 13). This route gave the product in higher yields compared to the conventional, opposite pathway reacting DFO mesylate with succinic anhydride and then conjugated the resulting acid to the tetrazine amine. The purification of 1 was initially carried out by semipreparative HPLC; however, this method resulted in a considerable loss of material and was therefore not further pursued. Instead, the product was purified by repeated precipitation from DMSO adding a 1:1 mixture of MeCN and H2O, giving an overall product yield of 57%. For the synthesis of 3, DFO* (17, Scheme 14) was synthesized based on a published procedure with minor changes (cf. experimental part) [7]. First, O-benzylhydroxylamine hydrochloride 8 was t Bu-protected to yield 9, and 1-bromo-5-chloropentane 10 was reacted with sodium phthalimide to give 11. The products 9 and 11 were then reacted to 12. The exchange of the phthalimide against the Cbz-protecting group took place by deprotecting 12 first with hydrazine monohydrate, followed by the protection of the amino functionality with benzyl chloroformate, giving 13. In the following, the t Bu-protecting group was removed using HCl in 1,4-dioxane instead of TFA (trifluoroacetic acid), yielding 14, which was reacted with succinic anhydride to acid 15. Compound 15 was activated using the coupling agent PyBOP instead of HATU (N,N,N′,N′-tetramethyl-O-(7azabenzotriazole-1-yl)uronium hexafluorphosphate) and, in the following, reacted with DFO mesylate, giving higher conversion rates to 16 compared to HATU activation. Of these, 2 [10] and 5 [17] were synthesized according to literature procedures. Compound 1 was prepared following a previously described synthesis route, with some modifications in the reaction order [10] by first reacting the tetrazine amine 6 with succinic anhydride and the resulting acid 7 with DFO mesylate using PyBOP (benzotriazole-1-yl-oxy-trispyrrolidino-phosphonium hexafluorophosphate) as the coupling agent (Scheme 13). This route gave the product in higher yields compared to the conventional, opposite pathway reacting DFO mesylate with succinic anhydride and then conjugated the resulting acid to the tetrazine amine. The purification of 1 was initially carried out by semipreparative HPLC; however, this method resulted in a considerable loss of material and was therefore not further pursued. Instead, the product was purified by repeated precipitation from DMSO adding a 1:1 mixture of MeCN and H 2 O, giving an overall product yield of 57%. Of these, 2 [10] and 5 [17] were synthesized according to literature procedures. Compound 1 was prepared following a previously described synthesis route, with some modifications in the reaction order [10] by first reacting the tetrazine amine 6 with succinic anhydride and the resulting acid 7 with DFO mesylate using PyBOP (benzotriazole-1-yloxy-tris-pyrrolidino-phosphonium hexafluorophosphate) as the coupling agent (Scheme 13). This route gave the product in higher yields compared to the conventional, opposite pathway reacting DFO mesylate with succinic anhydride and then conjugated the resulting acid to the tetrazine amine. The purification of 1 was initially carried out by semipreparative HPLC; however, this method resulted in a considerable loss of material and was therefore not further pursued. Instead, the product was purified by repeated precipitation from DMSO adding a 1:1 mixture of MeCN and H2O, giving an overall product yield of 57%. For the synthesis of 3, DFO* (17, Scheme 14) was synthesized based on a published procedure with minor changes (cf. experimental part) [7]. First, O-benzylhydroxylamine hydrochloride 8 was t Bu-protected to yield 9, and 1-bromo-5-chloropentane 10 was reacted with sodium phthalimide to give 11. The products 9 and 11 were then reacted to 12. The exchange of the phthalimide against the Cbz-protecting group took place by deprotecting 12 first with hydrazine monohydrate, followed by the protection of the amino functionality with benzyl chloroformate, giving 13. In the following, the t Bu-protecting group was removed using HCl in 1,4-dioxane instead of TFA (trifluoroacetic acid), yielding 14, which was reacted with succinic anhydride to acid 15. Compound 15 was activated using the coupling agent PyBOP instead of HATU (N,N,N′,N′-tetramethyl-O-(7azabenzotriazole-1-yl)uronium hexafluorphosphate) and, in the following, reacted with DFO mesylate, giving higher conversion rates to 16 compared to HATU activation. For the synthesis of 3, DFO* (17, Scheme 14) was synthesized based on a published procedure with minor changes (cf. experimental part) [7]. First, O-benzylhydroxylamine hydrochloride 8 was t Bu-protected to yield 9, and 1-bromo-5-chloropentane 10 was reacted with sodium phthalimide to give 11. The products 9 and 11 were then reacted to 12. The exchange of the phthalimide against the Cbz-protecting group took place by deprotecting 12 first with hydrazine monohydrate, followed by the protection of the amino functionality with benzyl chloroformate, giving 13. In the following, the t Bu-protecting group was removed using HCl in 1,4-dioxane instead of TFA (trifluoroacetic acid), yielding 14, which was reacted with succinic anhydride to acid 15. Compound 15 was activated using the coupling agent PyBOP instead of HATU (N,N,N ,N -tetramethyl-O-(7-azabenzotriazole-1yl)uronium hexafluorphosphate) and, in the following, reacted with DFO mesylate, giving higher conversion rates to 16 compared to HATU activation. Finally, the Cbz-and the Bn-protecting groups were removed under hydrogen atmosphere, giving DFO* 17. Finally, the Cbz-and the Bn-protecting groups were removed under hydrogen atmosphere, giving DFO* 17.
The corresponding tetrazine derivative was then obtained by reacting 17 with 7 using PyBOP as the coupling agent. The product was obtained in moderate but reproducible yields of 19% after semipreparative HPLC purification due to its low solubility in the water/acetonitrile solvent system, resulting in a considerable loss of material during purification. Up to this point, the target chelator tetrazines 1-3 and 5 could be synthesized without significant difficulties, and only minor adjustments of the published reaction conditions were necessary to optimize the product yields and to obtain the backbone tetrazine-modified chelators instead of the deviant functionalized derivatives described.
In contrast, the synthesis of 3,4,3-(LI-1,2-HOPO)-tetrazine, 4, entailed considerable challenges, and several attempts were necessary to develop a successful synthesis route towards the target compound. Initially, to obtain 20a,b, we followed a synthesis route towards the HOPO derivatives disclosed by Deri et al. [12] (Scheme 15A). Although the synthesis of 19 worked according to the published procedure, the purification of the product by column chromatography proved to be difficult. Besides the three-fold Boc- The corresponding tetrazine derivative was then obtained by reacting 17 with 7 using PyBOP as the coupling agent. The product was obtained in moderate but reproducible yields of 19% after semipreparative HPLC purification due to its low solubility in the water/acetonitrile solvent system, resulting in a considerable loss of material during purification.
Up to this point, the target chelator tetrazines 1-3 and 5 could be synthesized without significant difficulties, and only minor adjustments of the published reaction conditions were necessary to optimize the product yields and to obtain the backbone tetrazinemodified chelators instead of the deviant functionalized derivatives described.
In contrast, the synthesis of 3,4,3-(LI-1,2-HOPO)-tetrazine, 4, entailed considerable challenges, and several attempts were necessary to develop a successful synthesis route towards the target compound. Initially, to obtain 20a,b, we followed a synthesis route towards the HOPO derivatives disclosed by Deri et al. [12] (Scheme 15A). Although the synthesis of 19 worked according to the published procedure, the purification of the product by column chromatography proved to be difficult. Besides the three-fold Bocprotected product, the four-fold protected analog was also formed, which could not be completely removed. The following reaction step applying ethyl-3-bromopropanoate, benzyl-3-bromopropanoate or benzyl acrylate to introduce the protected acid functionality into the system (20) also proved to be intricate, as not only the intended reaction products 20a,b were formed but, also, several side products. Thus, highly complex reaction mixtures were obtained that prevented the isolation of the target substances.
Cancers 2021, 13, x FOR PEER REVIEW 17 of 28 protected product, the four-fold protected analog was also formed, which could not be completely removed. The following reaction step applying ethyl-3-bromopropanoate, benzyl-3-bromopropanoate or benzyl acrylate to introduce the protected acid functionality into the system (20) also proved to be intricate, as not only the intended reaction products 20a,b were formed but, also, several side products. Thus, highly complex reaction mixtures were obtained that prevented the isolation of the target substances.  In a different approach, we aimed to first introduce the tetrazine functionality into spermine by reacting spermine with N- (4-(1,2,4,5-tetrazin-3-yl)benzyl)-3-bromopropanamide to circumvent the problem of the following heterogeneous conversion of the amino groups, which led to inseparable product mixtures before. However, also this approach to first introduce the tetrazine into the spermine system failed due to the high basicity of the system, resulting in an instant decomposition of the tetrazine group.
Finally, another attempt was made that was based on a very recent work reported by  In a different approach, we aimed to first introduce the tetrazine functionality into spermine by reacting spermine with N-(4-(1,2,4,5-tetrazin-3-yl)benzyl)-3-bromopropanamide to circumvent the problem of the following heterogeneous conversion of the amino groups, which led to inseparable product mixtures before. However, also this approach to first introduce the tetrazine into the spermine system failed due to the high basicity of the system, resulting in an instant decomposition of the tetrazine group.
Finally, another attempt was made that was based on a very recent work reported by Bhupathiraju et al., where the authors disclosed an improved synthesis route towards a 3,4,3-(LI-1,2-HOPO) isothiocyanate derivative [20]. For our purpose, the published protocol had to be adapted to obtain the target tetrazine instead of the isothiocyanate (Scheme 15B). First, 6-benzyl-6-hydroxypicolinic acid (22) and 4-(tBu-amino)phenethyl-4-methylbenzenesulfonate (24) were synthesized according to the literature, giving the products in high yields between 69 and 97%. Next, spermine was reacted with 24, resulting in a mixture of multiple products, necessitating a laborious semipreparative HPLC purification that nevertheless gave 25 in good yields of 63%. Following the published reaction pathway utilizing TEA and DMAP in DCM for the next reaction step of 25 with 22 to yield 26, no product formation could be detected. Instead, 26 could be obtained using a small amount of aqueous K 2 CO 3 solution in DCM; however, this also necessitated a purification by semipreparative HPLC due to the large amount of side products formed (mainly three-fold-reacted intermediates instead of four-fold-reacted products and others), limiting the product yield to only 17%. The benzyl-and t Bu-protecting groups were subsequently removed by treatment with BCl 3 in p-xylene, and the obtained intermediate 27 was reacted directly without further purification with 4-(1,2,4,5-tetrazin-3-yl)benzoic acid, having been activated using PyBOP to give the target tetrazine-modified chelator 4 in moderate yields of 16%.

Investigation of the Properties of the Different 89 Zr Complexes by DFT Calculations
A very recent study reported on the thermodynamic stability of several zirconium complexes by DFT calculations, giving very encouraging results for the complex formation of CTH36 and DFO* with Zr 4+ [6]. As these calculations were conducted omitting any backbone functionalization of the chelating agents and the complex geometry was not discussed, we performed DFT calculations for the Zr complexes of the chelating agents investigated in this work and used the aforementioned data as the starting geometries. For these calculations, the chelator tetrazines 1-5 were conjugated to a simple model TCO to mimic the molecular situation in their respective 89 Zr-labeled biomolecules. In Table 1, an enlargement of the optimized geometry of the complexes and the relevant bond lengths can be found. Full structures and atom coordinates can be found in the Supplementary Materials.
Overall, the calculated values of the optimized structures are in very good agreement with the literature, and in almost all cases, the backbone modification of the chelators did not show a detectable influence on the complex geometry. The only exception was Zr-DOTA-GA, which showed elongated Zr-N bonds compared to the Zr-DOTA complex. The effect of the zirconium not being in the center of the complex cavity but considerably closer to the plane of the oxygen atoms than to that of the nitrogen atoms might be an indicator of hindered Zr-DOTA-GA complex formation and a less inert complex.
Overall, there was no evidence that the kinetic inertness of one of the studied Zr complexes might be significantly compromised by the introduction of the conducted backbone functionalization and further TCO modification.

Syntheses of Chelator Bioconjugates 29-33
As the aim of the present study was to determine the relative kinetic inertness of the 89 Zr complexes of the studied chelating agents, the chelator tetrazines 1-5 had to be introduced into a model biomolecule. The reason for this is that the kinetic inertness of the conjugated 89 Zr complexes is much more relevant than that of the tetrazines, as the same molecular situation of conjugated complexes is present in in vivo imaging applications. In addition, the bioconjugation ability of the developed chelator tetrazines had to be demonstrated as well.
The peptide c(RGDfK), which binds to integrin α v β 3 with high affinity [27] and is consequently a valuable bioactive agent accumulating in many human tumors, was chosen as clinically relevant model biomolecule for chelator introduction and 89 Zr-labeling of the resulting conjugates. c(RGDfK) possesses-especially compared to antibody moleculesthe important advantage that it exhibits a limited size and structural complexity, making the bioconjugation and the following 89 Zr-radiolabeling reaction of the conjugates, as well as the determination of the relative kinetic inertness of the formed 89 Zr complexes, easy to analyze and follow.
Thus, a complementarily functionalized c(RGDfK) peptide 28 (Scheme 16) was synthesized, carrying a trans-cyclooctene (TCO) unit for efficient click reaction with chelator tetrazines 1-5. The cyclic peptide c(RGDfK) was built on solid support by standard Fmoc-based solid-phase peptide synthesis [28,29] and modified with TCO in solution after cleavage from the resin by reaction with the corresponding p-nitrophenyl active ester, yielding c(RGDfK)-TCO, 28 [10].

Syntheses of Chelator Bioconjugates 29-33
As the aim of the present study was to determine the relative kinetic inertness of the 89 Zr complexes of the studied chelating agents, the chelator tetrazines 1-5 had to be introduced into a model biomolecule. The reason for this is that the kinetic inertness of the conjugated 89 Zr complexes is much more relevant than that of the tetrazines, as the same molecular situation of conjugated complexes is present in in vivo imaging applications. In addition, the bioconjugation ability of the developed chelator tetrazines had to be demonstrated as well.
The peptide c(RGDfK), which binds to integrin αvβ3 with high affinity [27] and is consequently a valuable bioactive agent accumulating in many human tumors, was chosen as clinically relevant model biomolecule for chelator introduction and 89 Zr-labeling of the resulting conjugates. c(RGDfK) possesses-especially compared to antibody molecules-the important advantage that it exhibits a limited size and structural complexity, making the bioconjugation and the following 89 Zr-radiolabeling reaction of the conjugates, as well as the determination of the relative kinetic inertness of the formed 89 Zr complexes, easy to analyze and follow.
Thus, a complementarily functionalized c(RGDfK) peptide 28 (Scheme 16) was synthesized, carrying a trans-cyclooctene (TCO) unit for efficient click reaction with chelator tetrazines 1-5. The cyclic peptide c(RGDfK) was built on solid support by standard Fmocbased solid-phase peptide synthesis [28,29] and modified with TCO in solution after cleavage from the resin by reaction with the corresponding p-nitrophenyl active ester, yielding c(RGDfK)-TCO, 28 [10]. Compound 28 was reacted using iEDDA click chemistry with the chelator tetrazines 1-5 to obtain the respective bioconjugates 29-33 (Scheme 17). Due to the limited solubility of 1-5, the bioconjugation reactions were carried out in aqueous DMSO (1,(3)(4)(5) or DMSO alone (2). All reactions were complete within minutes (obvious from disappearance of the pink tetrazine-associated color and nitrogen gas development) and gave the products in good yields of 59-84% after purification. Compound 28 was reacted using iEDDA click chemistry with the chelator tetrazines 1-5 to obtain the respective bioconjugates 29-33 (Scheme 17). Due to the limited solubility of 1-5, the bioconjugation reactions were carried out in aqueous DMSO (1,(3)(4)(5) or DMSO alone (2). All reactions were complete within minutes (obvious from disappearance of the pink tetrazine-associated color and nitrogen gas development) and gave the products in good yields of 59-84% after purification. It was observed during all iEDDA-based bioconjugation reactions that a side product whose amount varied among the different reactions was formed in addition to the 4,5dihydropyridazines (DHP). When possible, the respective byproducts were isolated and analyzed by mass spectrometry, indicating the oxidization of the DHPs to their respective aromatic pyridazine systems (Scheme 18). This is in accordance with literature reports where the spontaneous oxidation of iEDDA-formed DHPs to aromatic pyridazines has been described [30][31][32]. Furthermore, we calculated the bond lengths of the formed heteroatom-containing rings by the previously mentioned DFT calculations, finding bond lengths corresponding to aromatic pyridazine (Table 2), supporting the theory of spontaneous oxidation of the formed DHPs to the respective pyridazines. It was observed during all iEDDA-based bioconjugation reactions that a side product whose amount varied among the different reactions was formed in addition to the 4,5dihydropyridazines (DHP). When possible, the respective byproducts were isolated and analyzed by mass spectrometry, indicating the oxidization of the DHPs to their respective aromatic pyridazine systems (Scheme 18). This is in accordance with literature reports where the spontaneous oxidation of iEDDA-formed DHPs to aromatic pyridazines has been described [30][31][32]. It was observed during all iEDDA-based bioconjugation reactions that a side produc whose amount varied among the different reactions was formed in addition to the 4,5 dihydropyridazines (DHP). When possible, the respective byproducts were isolated and analyzed by mass spectrometry, indicating the oxidization of the DHPs to their respective aromatic pyridazine systems (Scheme 18). This is in accordance with literature reports where the spontaneous oxidation of iEDDA-formed DHPs to aromatic pyridazines has been described [30][31][32]. Furthermore, we calculated the bond lengths of the formed heteroatom-containing rings by the previously mentioned DFT calculations, finding bond lengths corresponding to aromatic pyridazine (Table 2), supporting the theory of spontaneous oxidation of the formed DHPs to the respective pyridazines. Furthermore, we calculated the bond lengths of the formed heteroatom-containing rings by the previously mentioned DFT calculations, finding bond lengths corresponding to aromatic pyridazine (Table 2), supporting the theory of spontaneous oxidation of the formed DHPs to the respective pyridazines. Table 2. Bond lengths in the formed heterocycles obtained by DFT calculations, as well as the literature data for free pyridazine [33] given for comparison. All values are given in Å.
First, attempts were made to radiolabel the bioconjugates 29-33 using these standard reaction conditions. For this purpose, an amount of 20 nmol of the respective radiolabeling precursor 29-32 was incubated with 40-55 MBq of [ 89 Zr]Zr oxalate solution at pH 7.0 in buffered solution at 37 • C, and the complexation progress was monitored by analytical radio-HPLC.
After one hour, an 89 Zr incorporation rate of ≥96% was observed for 29-31, demonstrating the fast 89 Zr complex formation of 29-31. Under the used radiolabeling conditions, the mixtures of DHPs and pyridazines, which were used in the case of 29-31 (vide supra), homogenized, giving the oxidized pyridazines. For 31, this homogenization process was complete within the first hour of radiolabeling, whereas it took an additional one or three hours for 30 and 29, respectively, to form the uniform pyridazine products. The products [ 89 Zr]Zr-29-[ 89 Zr]Zr-31 were obtained in a nonoptimized molar activity of 2-2.75 GBq/µmol in the form of a single product peak during radio-HPLC (Figure 3).  In contrast, the 89 Zr incorporation into 32 was considerably less effective, requiring prolonged reaction times of 5 h for sufficient 89 Zr complexation of ≥96% under the same conditions. Furthermore, [ 89 Zr]Zr-32 could not be obtained in form of a single product, although the pure pyridazine was applied as the precursor (Figure 3). This effect, that more than one product peak is formed during 89 Zr-radiolabeling of 3,4,3-(LI-1,2-HOPO), has been described before, and it was assumed that this could be a result of an initial formation of a kinetically favored product that is, over time, converted into a thermodynamically favored one [11]. This reasonable assumption is, however, not able to explain the observed formation of three separate peaks observed here. A possible explanation could be that different structure conformers of the same complex are formed, an effect being further enhanced by the backbone functionalization of the chelator. This effect would not be observable during antibody labeling. In a highly complex system like an antibody, a small molecular change has no influence on the retention time on conventional or size exclusion HPLC. To evaluate even small structural differences, it is an advantage to investigate a peptide of lower molecular weight, where even small changes in the structure result in different HPLC retention times. The formation of different 89 Zr complexation products is, of course, of particular interest in terms of the kinetic inertness, since different conformers might exhibit different stabilities. If this were the case, it could limit the applicability of 3,4,3-(LI-1,2-HOPO) for sTable 8 9 Zr introduction.
The 89 Zr-radiolabeling of 33 was initially also tested using the mild reaction conditions mentioned above, but no 89 Zr incorporation could be observed at 37 °C, so the temperature was increased to 99 °C for several hours. However, even under these harsh conditions, no incorporation of the radiometal could be detected, and all 89 Zr 4+ was present in the free form. Thus, different pH values (pH 4, 5 and 9) of the reaction solutions were tested, but this did not result in any radiometal incorporation into the chelator either. As Pandya 8+ , which is much more likely to occur after exchange of the stabilizing oxalate by chloride ions and is accelerated by low chloride concentrations and long standing times at high pH at elevated temperatures, limiting the apparent molar activities during 89 Zr-DOTA labeling [35]. As we, however, used relatively high chloride concentrations of 1 mol/L and directly used 89 Zr 4+ after the counter ion exchange at different pH values (among these acidic conditions), it seems to be unlikely that this is the reason for the observed missing 89 Zr incorporation, especially as we did not see any activity uptake by the chelator at all and not only to a limited extent.
An alternative to the use of DOTA-GA for biomolecule modification and 89 Zr labeling would, of course, be the use of the corresponding DOTA tetrazine without an addi-   In contrast, the 89 Zr incorporation into 32 was considerably less effective, requiring prolonged reaction times of 5 h for sufficient 89 Zr complexation of ≥96% under the same conditions. Furthermore, [ 89 Zr]Zr-32 could not be obtained in form of a single product, although the pure pyridazine was applied as the precursor (Figure 3). This effect, that more than one product peak is formed during 89 Zr-radiolabeling of 3,4,3-(LI-1,2-HOPO), has been described before, and it was assumed that this could be a result of an initial formation of a kinetically favored product that is, over time, converted into a thermodynamically favored one [11]. This reasonable assumption is, however, not able to explain the observed formation of three separate peaks observed here. A possible explanation could be that different structure conformers of the same complex are formed, an effect being further enhanced by the backbone functionalization of the chelator. This effect would not be observable during antibody labeling. In a highly complex system like an antibody, a small molecular change has no influence on the retention time on conventional or size exclusion HPLC. To evaluate even small structural differences, it is an advantage to investigate a peptide of lower molecular weight, where even small changes in the structure result in different HPLC retention times. The formation of different 89 Zr complexation products is, of course, of particular interest in terms of the kinetic inertness, since different conformers might exhibit different stabilities. If this were the case, it could limit the applicability of 3,4,3-(LI-1,2-HOPO) for stable 89Zr introduction.
The 89 Zr-radiolabeling of 33 was initially also tested using the mild reaction conditions mentioned above, but no 89 Zr incorporation could be observed at 37 • C, so the temperature was increased to 99 • C for several hours. However, even under these harsh conditions, no incorporation of the radiometal could be detected, and all 89 Zr 4+ was present in the free form. Thus, different pH values (pH 4, 5 and 9) of the reaction solutions were tested, but this did not result in any radiometal incorporation into the chelator either.  16 ] 8+ , which is much more likely to occur after exchange of the stabilizing oxalate by chloride ions and is accelerated by low chloride concentrations and long standing times at high pH at elevated temperatures, limiting the apparent molar activities during 89 Zr-DOTA labeling [35]. As we, however, used relatively high chloride concentrations of 1 mol/L and directly used 89 Zr 4+ after the counter ion exchange at different pH values (among these acidic conditions), it seems to be unlikely that this is the reason for the observed missing 89 Zr incorporation, especially as we did not see any activity uptake by the chelator at all and not only to a limited extent.
An alternative to the use of DOTA-GA for biomolecule modification and 89 Zr labeling would, of course, be the use of the corresponding DOTA tetrazine without an additional backbone functionalization and carboxylic group, but its use would raise the question of the inertness of the complex formed, as one of the carboxylates actually responsible for 89 Zr coordination was, in this case, used for conjugation, forming an acid amide. This acid amide, however, would most probably be less suitable for radiometal coordination compared to the hard oxygen atoms of the carboxylic group, resulting in a considerably reduced inertness of the resulting complex, as has already been shown in the example of Cu 2+ and its corresponding chelators [36,37].
In the following, the logD (7.4) s of the labeled agents [ 89 Zr]Zr-29-[ 89 Zr]Zr-32 were determined in order to assess the influence of the respective chelator on the overall compound lipophilicity. A high lipophilicity of the chelating agent could, e.g., result in a considerable plasma protein binding and, thus, in an unspecific background, as well as liver uptake of a correspondingly modified biomolecule, affecting the target visualization with PET [38][39][40][41].
The logD (7.4) values of the 89 Zr-labeled peptide-chelator conjugates were determined by their distribution coefficient between n-octanol and phosphate buffer at pH 7.4. The lipophilicity was determined to decrease in the order from [ 89 Zr]Zr-32 (logD (7.4) (7.4) : −2.77 ± 0.18). Compared to the three nonaromatic chelating agents exhibiting a high hydrophilicity, the poly-aromatic 3,4,3-(LI-1,2-HOPO) system shows a relatively high lipophilicity. Although this should be considerably less relevant for the modification and labeling of antibodies compared to peptides, the use of 3,4,3-(LI-1,2-HOPO) could nevertheless result in a somewhat higher background accumulation compared to the other chelating agents DFO, DFO* and CHT-36, which could then be explained, at least in part, by the higher lipophilicity of the system [41]. serving as the gold standard. Although a complex challenge experiment cannot provide information about the absolute inertness of a complex under in vivo conditions, it does allow the determination of the relative inertness of different complexes and therefore represents the standard method for the in vitro investigation of complex stability, mimicking the challenge of a radiometal complex by different endogenous substances being present in extremely high excess under in vivo imaging conditions.

Comparative
Since it was expected that all complexes to be studied (except [ 89 Zr]Zr-DFO) would exhibit a high inertness against the challenge, the experiments were performed using a very high excess of 10,000 eq. of EDTA as the challenging agent. The transchelation was monitored for up to 54 h by analytical radio-HPLC and showed significant differences between the different complexes (Figures 4 and 5).
In this context, the 89 Zr-DFO complex of [ 89 Zr]Zr-29 showed the expected limited inertness and rapid transfer of the radiometal to EDTA.
Surprisingly, the 89 Zr-CTH-36 complex of [ 89 Zr]Zr-30 showed a slightly higher but still relatively low inertness and an associated rapid transfer of the 89 Zr, while the 89 Zr-DFO* complex of [ 89 Zr]Zr-31 and the 89 Zr-3,4,3-(LI-1,2-HOPO) complex of [ 89 Zr]Zr-32 demonstrated a high resistance to the challenge. The poor performance of CTH-36 was very astonishing, since it should actually form stable complexes with 89 Zr 4+ on the basis of preliminary studies [10] and, also, theoretical considerations, which were also reflected in the very good results of recent DFT calculations on the thermodynamic stability of its 89 Zr complex [6].   However, in comparison, DFO* and 3,4,3-(LI-1,2-HOPO) exhibited a significantly better-and comparably high-kinetic stability of their respective 89 Zr complexes. This clearly demonstrates that the determination of thermodynamic stability allows, as expected, only very limited conclusions to be drawn about the actual suitability of a chelat-    However, in comparison, DFO* and 3,4,3-(LI-1,2-HOPO) exhibited a significantly better-and comparably high-kinetic stability of their respective 89 Zr complexes. This clearly demonstrates that the determination of thermodynamic stability allows, as ex pected, only very limited conclusions to be drawn about the actual suitability of a chelat ing agent for the formation of kinetically inert complexes.   However, in comparison, DFO* and 3,4,3-(LI-1,2-HOPO) exhibited a significantly better-and comparably high-kinetic stability of their respective 89 Zr complexes. This clearly demonstrates that the determination of thermodynamic stability allows, as expected, only very limited conclusions to be drawn about the actual suitability of a chelating agent for the formation of kinetically inert complexes.
Due to the results found here, the latter two compounds would be ideal candidates to study in terms of the inertness of their 89 Zr complexes in vivo, whereas CTH-36 and DOTA-GA seem to be unsuitable for 89 Zr introduction.
We are planning the in vivo evaluation of 3 and 4 using suitably modified antibodies. For this, the determination of the immunoreactivity of the obtained conjugates, 89 Zr radiolabeling and then investigating the in vivo pharmacokinetics of the radioligands in direct comparison over a timespan of several days will follow shortly. Of particular interest would be the extent to which the significantly higher lipophilicity of 4 affects the biodistribution of the respectively modified antibody and, also, whether the different complex species formed during radiolabeling using 3,4,3-(LI-1,2-HOPO) all exhibit a high kinetic inertness or if more advantageous in vivo pharmacokinetics will be found for the [ 89 Zr]Zr-DFO*-modified antibody.

Conclusions
The results shown indicate that the chelator tetrazines 1-5 developed here are very well-suited to efficiently functionalize biomolecules and that 1-4 are applicable for the radiolabeling of biologically relevant agents with 89 Zr.
By means of DFT calculations, it could be demonstrated that the backbone modifications of the developed chelates do not negatively affect the complex geometry and, thus, the radiolabeling of the chelator cores (at least apart from the DOTA-GA chelate).
The radiolabeling of the chelator-peptide bioconjugates with 89 Zr revealed some significant differences between the chelating agents: While the DFO, CTH-36 and DFO* bioconjugates exhibited very favorable 89 Zr-radiolabeling properties and advantageously high hydrophilicities of the labeled biomolecules, the 3,4,3-(LI-1,2-HOPO) peptide showed a considerably lower 89 Zr-radiolabeling efficiency, and the formation of an inhomogeneous labeling product of considerably higher lipophilicity could be validated.
The determination of the kinetic inertness of the formed 89 Zr complexes revealed a low stability of the [ 89 Zr]Zr-DFO complex but, surprisingly, also an unexpected considerable lability of the [ 89 Zr]Zr-CTH-36 complex. Only [ 89 Zr]Zr-DFO* and [ 89 Zr]Zr-3,4,3-(LI-1,2-HOPO) proved a high inertness against the competition challenge experiments, illustrating that a high thermodynamic stability of a complex is-as expected, but sometimes implied otherwise-not a good predictor of the inertness of a radiometal complex.
In subsequent studies, 3 and 4 will be investigated in direct comparison under in vivo conditions (including introduction into an IgG antibody, radiolabeling with 89 Zr and determination of the immunoreactivity of the conjugates and application in an appropriate disease model, monitoring the in vivo pharmacokinetics over several days) to finally identify the most suitable chelating agent for future clinical applications.