Polyazamacrocycle ligands facilitate ⁸⁹Zr radiochemistry and yield ⁸⁹Zr complexes with remarkable stability

Abstract

Over the last three decades, the chemistry of zirconium has facilitated antibody development and the clinical management of disease in the precision medicine era. Scientists have harnessed its reactivity, coordination chemistry and nuclear chemistry to develop antibody-based radiopharmaceuticals incorporating zirconium-89 (⁸⁹Zr: t_½ = 78.4 h, β⁺: 22.8%, E_β+max = 901 keV; EC: 77%, E_γ = 909 keV) to improve disease detection, identify patients for individualized therapeutic interventions and monitor their response to those interventions. However, release of the ⁸⁹Zr⁴⁺ ion from the radiopharmaceutical remains a concern since it may confound the interpretation of clinical imaging data, negatively affect dosimetric calculations and hinder treatment planning. In this report we relate our novel observations involving the use of polyazamacrocycles as zirconium-89 chelators. We describe the synthesis and complete characterization of zirconium 2,2′,2”,2”’-(1,4,7,10-tetraazacyclotridecane-1,4,7,10-tetrayl)tetraacetic acid (Zr-TRITA), zirconium 3,6,9,15-Tetraazabicyclo[9.3.1] pentadeca-1(15),11,13-triene-3,6,9-triacetic acid (Zr-PCTA) and zirconium 2,2’,2”-(1,4,7-triazacyclononane-1,4,7-triyl)triacetic acid (Zr-NOTA). Additionally, we elucidate the solid-state structure of each complex using single crystal x-ray diffraction analysis. Finally, we found that [⁸⁹Zr]Zr-PCTA and [⁸⁹Zr]Zr-NOTA demonstrate excellent stability in vitro and in vivo and provide a rationale for these observations. These innovative findings have the potential to guide the development of safer and more robust immuno-PET agents to improve precision medicine applications.

Graphical Abstract

The macrocycles PCTA and NOTA are efficient zirconium-89 chelators. The resulting radiometal complexes can be formed under mild conditions and exhibit remarkable stability in vitro and in vivo.

INTRODUCTION

Zirconium-89 immuno-positron emission tomography (⁸⁹Zr-immuno-PET), which is the use of ⁸⁹Zr-radiolabelled monoclonal antibodies (mAbs) in conjunction with current whole-body PET detection, has the potential to enhance the information obtained through traditional biopsy and improve patient care.^1-3 Indeed, important clinical trials are underway to evaluate the utility of ⁸⁹Zr-immuno-PET for detecting disease, guiding biopsy, stratifying patients for therapy and monitoring therapy response by non-invasively probing receptor expression on tumors of oncology patients.^4,5 For example, Ulaner and colleagues recently sought to determine if ⁸⁹Zr-trastuzumab PET imaging could be used to detect HER2-positive metastases in patients diagnosed with HER2 negative, primary disease.^6-11 Confirming the authors’ hypothesis, ⁸⁹Zr-trastuzumab PET identified several women who although diagnosed with HER2 negative primary disease had pathologically confirmed HER2 positive distant metastases. However, the authors also observed false-positive findings; several were in osseous metastases. The authors suggested that these findings may result from the ⁸⁹Zr⁴⁺ ion being released from its chelator and non-specifically localizing in the phosphate rich hydroxylapatite where active bone remodelling is occurring due to metastatic disease. Indeed, several published reports describe the accumulation of radioactivity in bone when ⁸⁹Zr is chelated to antibodies with desferrioxamine (DFO), which is the only ⁸⁹Zr chelator currently being used in clinical studies.^12-14 These findings highlight the clinical importance of developing new ligands that can stably sequester ⁸⁹Zr during the extended circulation time that these radiopharmaceuticals encounter after patient injection.

Recently, Pandya et al. investigated the use of tetraazamacrocycles as chelators for ⁸⁹Zr.¹⁵ In this work the authors described the synthesis and characterization of several Zr-cyclen complexes containing carboxylic acid, phosphonate and amide containing pendant arms. These studies revealed all complexes to be octa-coordinate with all four ring nitrogen atoms and pendent arms participating in Zr⁴⁺ ion coordination. In vitro several of the ⁸⁹Zr-tetraazamacrocycle complexes demonstrated greater resistance to challenge by exogenous ligands, biologically relevant metal ions or serum than [⁸⁹Zr]Zr-DFO. Biodistribution and small animal PET/CT studies revealed that [⁸⁹Zr]Zr-DOTA demonstrated extraordinary in vivo stability and was far superior to the stability of [⁸⁹Zr]Zr-DFO. Although extraordinary stability was observed with the former complex, high temperature was required to synthesize it quantitatively, and this was deemed a disadvantage for the synthesis of ⁸⁹Zr-immuno-PET agents if the DOTA-mAb was going to be radiolabeled with ⁸⁹Zr directly. Although several two-step radiochemistry procedures published in the literature would alleviate this potential disadvantage,^16-20 it seemed prudent to survey similar ligands to determine if they could be quantitatively radiolabeled with ⁸⁹Zr at low temperature while providing a stability profile comparable to that of [⁸⁹Zr]Zr-DOTA. Accordingly, we synthesized Zr-PCTA, Zr-NOTA, Zr-TRITA and Zr-TETA and their ⁸⁹Zr analogs (Figure 1) to assess how the size and flexibility of the polyazamacrocycle ring influenced the radiochemistry and stability of the resulting radiometal complex in vitro and in vivo.

Figure 1.

Structures of polyazamacrocyclic ligands and their ⁸⁹Zr-complexes.

RESULTS AND DISCUSSION

Synthesis and Molecular Structure Determination.

Zirconium-89 immuno-PET has enabled significant advances in therapeutic development while its role in the clinical management of disease continues to evolve.^21,22 It has facilitated therapeutic mAb development by reducing the time needed to identify the most promising agents for clinical translation. Further, ⁸⁹Zr-immuno-PET is expected to become an increasingly powerful clinical tool in the era of personalized medicine since it has the potential to confirm malignancy, elucidate antigen density; optimize pharmacokinetics, provide dosimetry in advance of radioimmunotherapy and describe each lesion’s molecular phenotype in a comprehensive manner. However, two important criteria must be satisfied for ⁸⁹Zr-immuno-PET to be effective. First, the radiopharmaceutical must demonstrate specificity for its biological target. Secondly, once conjugated to the mAb, the ⁸⁹Zr-chelate complex must remain inert to the biological, chemical and physical forces that will be encountered while circulating within the blood pool over several days. This latter issue may confound the interpretation of clinical imaging data, negatively affect dosimetric calculations and hinder treatment planning. Solving this problem remains an active area of research.^{12-14,20,23-40}

Previously, we evaluated tetraazamacrocycles as ⁸⁹Zr chelators.¹⁵ In that report we evaluated the cyclen–based macrocycles DOTA, DOTP and DOTAM to understand the influence of various pendant arm coordinating units on radiochemistry and stability. Surprisingly, DOTA, a cyclen derivative containing four pendant arms with carboxymethylene coordinating units yielded a radiometal complex with an in vitro and in vivo stability that was superior to the other derivatives containing pendant arms with the phosphonate or amide coordinating units. Moreover, [⁸⁹Zr]Zr-DOTA also demonstrated superior in vivo behavior to [⁸⁹Zr]Zr-DFO, which is currently the radiometal complex being used in clinical ⁸⁹Zr-immuno-PET applications. This report describes the suitability of the macrocycle ligands TETA, TRITA, PCTA and NOTA as ⁸⁹Zr chelators; it posits how the size and flexibility of each macrocycle may influence the Zr chemistry of each ligand, the radiochemical synthesis of each radiometal complex, and the in vitro and in vivo stability of each ⁸⁹Zr complex.

Following synthetic procedures established for Zr-DOTA,¹⁵ we undertook the synthesis of Zr-TETA (Scheme S1), Zr-TRITA (Scheme S2), Zr-PCTA (Scheme S3) and Zr-NOTA (Scheme S4) using Zr(IV) acetylacetonate (Zr(AcAc)₄) as a zirconium source with subsequent synthetic strategies modified from the literature. Unfortunately, Zr-TETA could not be prepared even under forcing conditions using conventional thermal or microwave-assisted strategies; we were unable to characterize this non-radioactive complex further. However, the non-radioactive complexes of Zr-TRITA, Zr-PCTA and Zr-NOTA were successfully prepared in excellent yields and all were fully characterized by HPLC, NMR spectroscopy, and HR-MS analyses (Figures S1-S23). Additionally, each complex yielded single crystals suitable for x-ray diffraction analysis. The complete crystallographic parameters, data collection, refinement information, selected bond lengths, and bond angles of Zr-TRITA, Zr-PCTA and Zr-NOTA are defined in Tables S1-S5, S6-S10, and S11-S15 respectively. The molecular structures of Zr-TRITA, Zr-PCTA and Zr-NOTA are depicted in Figure 2, Figure 3, and Figure 4, respectively. In each case the ligands display a low-symmetry saddle-like conformation. This conformation is characteristic of metal complexes, which demonstrate d⁰ electron configurations and lack crystal-field stabilization. While the bond lengths and bond angles of Zr-TRITA, Zr-PCTA and Zr-NOTA are comparable to those observed in structurally characterized Zr complexes,^41-50 each complex demonstrates unique structural features, which have been summarized below.

Figure 2.

The crystal structure of [Zr(C₁₇H₂₆N₄O₈)] – 3.33 H₂O (Zr-TRITA). The asymmetric unit cell contains three crystallographically-independent molecules. The introduction of the additional methylene unit into the TRITA backbone increases the cis-N-Zr-N bite angle of the cyclam portion of the ligand by 15° when compared to the cis-N-Zr-N bite angle exhibited by the cyclen portion of the ligand. Only a partial numbering scheme is provided. Solvent molecules and hydrogen atoms are omitted for clarity. Thermal ellipsoids are drawn at the 50% probability level.

Figure 3.

Crystal structure of [Zr(C₁₇H₂₁N₄O₆)]₂O - 8 H₂O (Zr-PCTA). (Left) The seven coordinate monomers are bridged by one oxygen atom, which lies on the crystallographic 2-fold axis at ½, y, ¼ in the unit cell, to fulfill Zr⁴⁺ ion’s requirement of an octa-coordinate ligand sphere. (Right) The monomer is depicted so that the coordination sphere around the Zr⁴⁺ ion is visualized more easily. Nitrogen one (N1) represents the nitrogen atom of the pyridyl ring. Only a partial numbering scheme is provided. Solvent molecules and hydrogen atoms are omitted for clarity. Thermal ellipsoids are drawn at the 50% probability level. The molecules are labeled to facilitate the identification of the symmetry related atoms only.

Figure 4.

Crystal structure of [Zr(C₁₂H₁₈N₃O₆)(OH)]₂ - 6 H₂O (Zr-NOTA). Zr-NOTA crystalizes as a binuclear complex. The six coordinate monomers, which are defined by the Zr-NOTA subunits are bridged by two OH groups (O7A and O7B) and are related by the crystallographic inversion center at ½, ½, 0 in the unit cell. Only a partial numbering scheme is provided. Solvent molecules and hydrogen atoms are omitted for clarity. Thermal ellipsoids are drawn at the 50% probability level. The molecules are labeled to facilitate the identification of the symmetry related atoms only.

X-ray diffraction of Zr-TRITA crystals (Figures 2, S24-S30) revealed that the asymmetric unit cell contains three crystallographically-independent molecules, which co-crystallize with three water molecules in the monoclinic Cc space group. The molecular geometry can best be described as a square antiprism. Similar to the Zr-DOTA structure, the Zr⁴⁺ ion lies closer in proximity to the carboxylate containing pendant arms than to the macrocycle. Analysis of the three crystallographically distinct molecules reveals that the average perpendicular distance from the metal center to the plane described by the four acetate-containing pendant arms of the macrocycle is 1.030 ± 0.002Å while the average perpendicular distance from the metal center to the plane described by the four nitrogen atoms of the macrocycle is 1.280 ± 0.002Å; both distances do not vary significantly from those observed in the Zr-DOTA solid state structure. Interestingly, the introduction of the additional methylene unit into the TRITA backbone increases the cis-N-Zr-N bite angle of the cyclam portion of the ligand by 15° when compared to the cis-N-Zr-N bite angle exhibited by the cyclen portion of the ligand (85.5(1)° vs. 71.0(1)°). This bite angle is also 13° larger than that observed in the analogous Zr-DOTA complex.¹⁵ The expanded bite angle of the former complex may be a contributing factor to the difficult radiochemistry and poor stability exhibited by [⁸⁹Zr]Zr-TRITA (vide infra).

X-ray diffraction analysis revealed that Zr-PCTA (Figures 3, S31-S34) crystalizes as a binuclear complex in the monoclinic C2/c space group along with four water molecules and its molecular geometry is best described as a distorted square antiprism. The seven coordinate monomers are bridged by one oxygen atom, which lies on the crystallographic 2-fold axis at ½, y, ¼ in the unit cell, to fulfil Zr⁴⁺ ion’s requirement of an octa-coordinate ligand sphere. Interestingly, the pyridyl rings of the PCTA ligands are related by the 2-fold axis and are oriented in a nearly parallel configuration; the angle between them deviates by 7.1°. The perpendicular distance from the metal center to the plane described by the three acetate-containing pendant arms of the macrocycle is 0.9813(4)Å, while the perpendicular distance from the metal center to the plane described by the four nitrogen atoms of the macrocycle is 1.5091(6)Å. This suggests that the Zr⁴⁺ ion is more closely associated with the pendant arms than the macrocycle. The distances between the Zr⁴⁺ ions of each monomer subunit and the bridging oxygen atom were observed to be 1.9809(3)Å, while the Zr-O-Zr angle was observed to be 167.32(7)°. These values are similar with published literature describing other Zr-μ-oxo complexes and several observations may explain these results.^51-54 The pyridyl ring of the PCTA ligand most likely increases the rigidity and preorganization of the resulting Zr-PCTA complex. Additionally, numerous inter-and intramolecular hydrogen bonding motifs are observed in the solid state. Both factors probably minimize repulsion energies through enhanced crystal packing and give rise to the shortened bond lengths and pseudo-linear μ-oxo bridge. Finally, the intramolecular distance between the centroids of the pyridyl rings of the PCTA ligands is 6.399Å. The large separation between the rings and their unfavorable orientation about the 2-fold axis disfavors π-π stacking, which is a stabilizing phenomenon that has been observed to occur in other transition metal complexes containing similar ligand motifs.^55,56

X-ray diffraction analysis revealed that Zr-NOTA (Figures 4, S35-S41) also crystalizes as a binuclear complex in the monoclinic P2₁/n space group with six water molecules per dimer. Unlike the other two structures, its geometry is best described as a triangular dodecahedron. The six coordinate monomers, which are defined by the Zr-NOTA subunits are bridged by two OH groups and are related by the crystallographic inversion center at ½, ½, 0 in the unit cell.

The arrangement of the three methylene groups around N3 of the structure demonstrate a slight rotational disorder (93%/7%) around the Zr-N3 bond. The average perpendicular distance from the metal center to the plane described by the three acetate-containing pendant arms of the macrocycle is 0.3820(7) Å, while the perpendicular distance from the metal center to the plane described by the three nitrogen atoms of the macrocycle is 1.8253(5) Å. Similar to the Zr-PCTA structure, this suggests that the Zr⁴⁺ ion is more closely associated with the pendant arms than the macrocycle. The distances between the Zr⁴⁺ ions of each monomer subunit and the bridging hydroxide molecules were observed to be unequal with an average distance of 2.1553(7)Å; the Zr-OH-Zr angles were observed to be approximately 111.70(3)°. These observations are consistent with μ-hydroxo and μ-oxo binuclear In³⁺ complexes but dissimilar to binuclear Fe³⁺ complexes reported in the literature.^57,58

Radiochemical Synthesis.

After completing synthesis and characterization of the reference complexes, we attempted to radiolabel each ligand using [⁸⁹Zr]Zr(ox)₂ and procedures established for preparation of [⁸⁹Zr]Zr-DFO. However, radiochemical yields were poor (Scheme S5, Table S16). Thus, we used [⁸⁹Zr]ZrCl₄ (Figure S42) as a radioactive precursor and observed that PCTA and NOTA were quantitatively radiolabeled within 60 minutes at 37 °C (Scheme S6, Tables S17, S18, and Figures S43, S46), while TRITA was radiolabeled 80.1 ± 1.2 % within 120 minutes at 99 °C (Table S19, Figure S49). In an attempt to improve the reaction kinetics, a microwave-assisted synthesis using forcing conditions was required to quantitatively prepare [⁸⁹Zr]Zr-TRITA (Table S20, Figure S50). Despite repeated attempts to prepare [⁸⁹Zr]Zr-TETA using conventional and microwave-assisted strategies, radiochemical yields were poor even under the most forceful conditions (Table S21, Figure S53). The radiochemical yield and purity of [⁸⁹Zr]Zr-PCTA, [⁸⁹Zr]Zr-NOTA, and [⁸⁹Zr]Zr-TRITA were confirmed by radio-TLC (Figures S44, S47, and S51) and radio-HPLC (Figures S45, S48, and S52). Optimized radiochemical synthesis conditions are presented in Table 1. The molar activity (A_m) for each radiometal complex is in good agreement with the A_m of other ⁸⁹Zr complexes reported in the literature.¹⁵ Based upon our attempts to prepare [⁸⁹Zr]Zr-TETA, [⁸⁹Zr]Zr-TRITA and our previous work with [⁸⁹Zr]Zr-DOTA, the ease of radiochemical synthesis can be described as [⁸⁹Zr]Zr-DOTA > [⁸⁹Zr]Zr-TRITA >>> [⁸⁹Zr]Zr-TETA. While increasing the cavity size would have been thought to facilitate ⁸⁹Zr radiochemistry, perhaps the introduction of additional methylene units increases the cavity size and decreases the molecular orbital overlap between the bonding orbitals of the ⁸⁹Zr⁴⁺ ion and the bonding orbitals of the macrocycles TRITA and TETA. Although we were unable to determine the molecular structure of Zr-TETA, we were fortunate enough to elucidate the crystal structures of Zr-TRITA and Zr-DOTA. In these structures we did observe that the cis-N-Zr-N bite angles were different and may be enough to disrupt the molecular orbital interactions necessary for metal-ligand bonding. Similar observations were reported to influence the preparation and stability of Ln³⁺ tetraazamacrocycle complexes. For example, ease of synthesis and kinetic inertness of Ln³⁺ tetraazamacrocycle complexes decreased by several orders of magnitude when considering the reactivity of the 12-membered DOTA, 13-membered TRITA and then the 14 membered TETA macrocycles.^59,60 Additionally, the degree of preorganization of each ligand may also influence the radiochemistry.^59,61-63 The non-complexed DOTA⁴⁻ crystal structure reveals that the four carboxymethyl pendant arms are in the syn configuration with respect to the macrocycle plane. On complexation with ⁸⁹Zr⁴⁺ ion, the ligand would have to undergo minimal reorganization to accommodate the radioactive ion. Contrarily, the TETA⁴⁻ structure reveals that the adjacent carboxymethyl pendant arms are in an anti-configuration with respect to the macrocycle plane. As a result, significant ligand rearrangement would be required to achieve the syn configuration necessary to form the [⁸⁹Zr]Zr-TETA complex. While the crystal structure of TRITA⁴⁻ is not available, it is conceivable that this ligand has at least one carboxymethylene containing pendant arm in the anti configuration, and this may explain why forcing conditions are necessary to prepare [⁸⁹Zr]Zr-TRITA. Furthemore, additional literature suggests that enthalpic effects that arise from the inability of larger macrocycles to effectively chelate large metal ions also plays a role in the lack of complex formation and instability.⁶⁴

Table 1.

Summary of optimized radiochemistry conditions to prepare [⁸⁹Zr]Zr-complexes with [⁸⁹Zr]ZrCl₄

Radiochemistry conditions	Ligands (n = 30)
	PCTA	NOTA	TRITA	TETA
Quantity (μg)	15-20	15-17	18-24	10-50
[89Zr]ZrCl4 added (MBq)	41.6-55.5	41.7-55.7	40.7-51.8	16.5-20.5
Reaction Buffer	0.5 M HEPES	0.5 M HEPES	0.5 M HEPES	0.5 M HEPES
Final Reaction pH	6.9-7.2	6.9-7.2	3.9-4.2	6.9-7.3
Reaction Temperature (°C)	37 a	37 a	180b	99 a
Reaction time (min)	60	60	20	120
Radiochemical yield (%)	100	100	100	0
Molar activity (Am; MBq μmol−1)	1035 ± 6	995 ± 7	890 ± 9	-

^aConventional heating.

^bMicrowave heating.

Lipophilicity (LogP) Determination.

Lipophilicity (LogP) represents a fundamental physiochemical property that influences the adsorption, distribution, metabolism and elimination of ⁸⁹Zr-complexes in vivo.⁶⁵ Based upon the water/octanol partition method all complexes demonstrate hydrophilic character, which is attributed to the numerous hydrogen bonds these complexes experience in solution and is supported by the multiple water molecules engaged in hydrogen bonding with each Zr complex as revealed by single crystal analysis. All values (Table S22) suggest that renal excretion would be a preferred route of elimination after in vivo injection. Interestingly, [⁸⁹Zr]Zr-NOTA has a LogP value (−2.51 ± 0.01) that is more positive than the LogP value of [⁸⁹Zr]Zr-DFO (−2.83 ± 0.03).¹⁵ Despite the pyridyl ring of the PCTA macrocycle, its LogP value (−3.09 ± 0.03) was more negative than the LogP of [⁸⁹Zr]Zr-DFO and similar to that of [⁸⁹Zr]Zr-TRITA (−3.13 ± 0.02). However, all were more positive than the LogP value reported for [⁸⁹Zr]Zr-DOTA (−3.80 ± 0.04).¹⁵

In Vitro Stability Studies.

Similar to our previous work evaluating ⁸⁹Zr-tetraazamacrocycles,¹⁵ the in vitro stability of [⁸⁹Zr]Zr-NOTA, [⁸⁹Zr]Zr-TRITA and [⁸⁹Zr]Zr-PCTA were evaluated by challenging them with excess EDTA, high concentrations of biologically relevant metal ions, or human serum proteins. Over seven days, these experiments revealed that [⁸⁹Zr]Zr-PCTA was inert to transchelation by a 1000-fold excess of EDTA at pH 5 or pH 7. After seven days in the presence of a 1000-fold excess of EDTA at pH 5, approximately 35% of the activity initially associated with [⁸⁹Zr]Zr-NOTA was transchelated by EDTA, while approximately 90% of the initial activity associated with the TRITA ligand was transchelated to EDTA under the same conditions. Based upon our EDTA challenge studies, the order of ⁸⁹Zr complex stability can be described as [⁸⁹Zr]Zr-PCTA > [⁸⁹Zr]Zr-NOTA >> [⁸⁹Zr]Zr-TRITA (Table 2).

Table 2.

Chronological stability study of [⁸⁹Zr]Zr-complexes in EDTA (pH 5 and pH 7) challenge study maintained at 37°C for 7 days and analyzed by Radio-ITLC (n = 3 for each [⁸⁹Zr]Zr-complex at each pH and time point)

Complex	EDTA	pH	% Intact of [89Zr]Zr-complexes (n = 3)
			0 h	1 d	3 d	5 d	7 d
[89Zr]Zr-PCTA	100-fold	7.0	100	100	100	100	100
		5.0	100	100	100	100	100
	500-fold	7.0	100	100	100	100	100
		5.0	100	100	100	100	100
	1000-fold	7.0	100	100	100	100	100
		5.0	100	100	100	100	100
[89Zr]Zr-NOTA	100-fold	7.0	100	94.5 ± 0.5	85.3 ± 0.8	82.9 ± 0.7	80.8 ± 0.4
		5.0	100	93.0 ± 0.3	82.7 ± 0.7	80.3 ± 0.2	78.4 ± 0.6
	500-fold	7.0	100	93.1 ± 0.1	83.8 ± 0.7	81.2 ± 0.3	73.3 ± 1.4
		5.0	100	83.7 ± 1.3	75.8 ± 1.8	72.9 ± 1.2	71.2 ± 1.7
	1000-fold	7.0	100	93.1 ± 0.4	81.8 ± 1.6	76.7 ± 0.8	70.7 ± 1.0
		5.0	100	78.5 ± 0.7	71.8 ± 0.8	66.4 ± 1.0	65.3 ± 1.2
[89Zr]Zr-TRITA	100-fold	7.0	99.4 ± 0.3	44.7 ± 0.8	20.5 ± 1.1	1.1 ± 0.5	0
		5.0	99.7 ± 0.2	85.7 ± 1.6	64.0 ± 1.4	40.5 ± 0.2	36.3 ± 0.4
	500-fold	7.0	99.4 ± 0.3	23.1 ± 1.2	3.8 ± 0.9	0	0
		5.0	99.7 ± 0.2	75.6 ± 1.0	45.5 ± 2.5	19.6 ± 0.6	16.7 ± 0.5
	1000-fold	7.0	99.4 ± 0.3	13.1 ± 0.9	1.9 ± 0.7	0	0
		5.0	99.7 ± 0.2	69.6 ± 0.7	35.8 ± 1.7	13.8 ± 3.5	9.6 ± 0.1

Polyazamacrocycles can chelate numerous, biologically relevant metal ions potentially creating a second mechanism for ⁸⁹Zr⁴⁺ ion dissociation from its chelator in vivo.²⁵ To assess the ability of the ⁸⁹Zr complexes to resist demetallation, we mixed each radiometal complex with an excess concentration of metal salts in aqueous buffer. No demetallation of [⁸⁹Zr]Zr-PCTA was observed over the 7-day experiment (Table 3). In contrast, [⁸⁹Zr]Zr-NOTA remained only 88% and 90% intact when challenged with Cu²⁺ ions or Ga³⁺ ions, respectively during the same study period. Finally, [⁸⁹Zr]Zr-TRITA remained only 60% and 62% intact when challenged with Cu²⁺ ions or Ga³⁺ ions, respectively by the end of the study. The overall order of ⁸⁹Zr complex stability mirrored our results in the EDTA challenge experiments.

Table 3.

Chronological stability study of [⁸⁹Zr]Zr-complexes with various metals in competition study (PBS, pH 7.4) maintained at 37°C for 7 days and analyzed by Radio-ITLC (n = 3 for each [⁸⁹Zr]Zr-complex with each metal and time point)

Time Point	Complex	% Intact of [89Zr]Zr-complexes (n = 3)
		Fe3+	Zn2+	Co2+	Cu2+	Mg2+	Gd3+	Ga3+
0 h	[89Zr]Zr-PCTA	100	100	100	100	100	100	100
	[89Zr]Zr-NOTA	100	100	100	100	100	100	100
	[89Zr]Zr-TRITA	99.7 ± 0.2	99.6 ± 0.2	99.5 ± 0.2	99.7 ± 0.2	99.6 ± 0.2	99.5 ± 0.2	99.7 ± 0.2
1 d	[89Zr]Zr-PCTA	100	100	100	100	100	100	100
	[89Zr]Zr-NOTA	99.0 ± 0.1	96.1 ± 0.1	97.7 ± 0.4	96.1 ± 0.2	95.5 ± 0.3	98.3 ± 0.1	97.9 ± 0.1
	[89Zr]Zr-TRITA	91.3 ± 0.3	29.2 ± 1.2	35.6 ± 2.3	94.5 ± 0.9	90.4 ± 1.0	88.7 ± 0.8	94.6 ± 0.8
3 d	[89Zr]Zr-PCTA	100	100	100	100	100	100	100
	[89Zr]Zr-NOTA	98.3 ± 0.2	95.8 ± 0.2	95.7 ± 0.3	92.2 ± 0.3	94.3 ± 0.3	97.2 ± 0.5	95.2 ± 0.6
	[89Zr]Zr-TRITA	85.8 ± 0.9	21.1 ± 0.9	18.4 ± 0.5	76.2 ± 1.0	87.9 ± 0.3	67.7 ± 0.9	70.4 ± 0.9
5 d	[89Zr]Zr-PCTA	100	100	100	100	100	100	100
	[89Zr]Zr-NOTA	97.4 ± 0.3	94.4 ± 0.5	94.8 ± 0.6	90.2 ± 0.3	91.8 ± 0.6	97.0 ± 0.3	93.4 ± 0.1
	[89Zr]Zr-TRITA	77.8 ± 1.1	16.5 ± 2.7	15.4 ± 0.8	68.1 ± 1.5	85.5 ± 1.1	59.2 ± 0.7	67.6 ± 0.5
7 d	[89Zr]Zr-PCTA	100	100	100	100	100	100	100
	[89Zr]Zr-NOTA	96.2 ± 0.2	90.4 ± 0.4	92.9 ± 0.7	87.8 ± 0.9	90.6 ± 0.4	96.3 ± 0.4	89.5 ± 1.5
	[89Zr]Zr-TRITA	70.4 ± 0.8	9.4 ± 0.7	7.3 ± 0.6	59.7 ± 0.5	83.9 ± 0.5	57.2 ± 0.6	61.8 ± 1.2

Incubation in human serum (Figures S54, S55, Table 4) revealed [⁸⁹Zr]Zr-PCTA was the most resistant to serum protein challenge, while [⁸⁹Zr]Zr-NOTA was slightly less inert. After three days, 96% of the original activity remained chelated to the latter polyazamacrocycle. Finally, only 80% and 32% of the original radioactivity associated with the TRITA macrocycle remained chelated to it after one and seven days, respectively making [⁸⁹Zr]Zr-TRITA the most vulnerable to transchelation by serum proteins.

Table 4.

Chronological in vitro serum stability study of [⁸⁹Zr]Zr-complexes in human serum maintained at 37°C for 7 days as determined by size exclusion chromatography (n = 3 for each [⁸⁹Zr]Zr-complex at each time point)

Time point	% Intact of [89Zr]Zr-complexes (n = 3)
	[89Zr]Zr-PCTA	[89Zr]Zr-NOTA	[89Zr]Zr-TRITA
0 h	100	100	100
1 d	99.9 ± 0.1	97.7 ± 0.2	80.8 ± 1.1
3 d	99.8 ± 0.1	95.5 ± 0.4	59.1 ± 1.5
5 d	99.6 ± 0.2	91.8 ± 0.3	45.2 ± 2.4
7 d	99.4 ± 0.2	89.2 ± 0.5	32.4 ± 0.6

Biodistribution Studies in Mice.

We then evaluated the in vivo behavior of [⁸⁹Zr]Zr-TRITA, [⁸⁹Zr]Zr-PCTA or [⁸⁹Zr]Zr-NOTA in acute biodistribution studies involving normal mice. Results are shown in Tables S23-S25 and Figures 5, S56-S58. At 2 h post-injection (p.i.), animals injected with [⁸⁹Zr]Zr-TRITA demonstrated elevated levels of blood-associated radioactivity and these levels remained elevated at all time points of the study. Conversely, [⁸⁹Zr]Zr-PCTA and [⁸⁹Zr]Zr-NOTA were cleared from the blood pool much more efficiently and by 72 h p.i., the amount of radioactivity associated in the blood of animals receiving [⁸⁹Zr]Zr-TRITA was still significantly greater than observed for those animals receiving either [⁸⁹Zr]Zr-PCTA or [⁸⁹Zr]Zr-NOTA ([⁸⁹Zr]Zr-TRITA vs. [⁸⁹Zr]Zr-PCTA vs. [⁸⁹Zr]Zr-NOTA [blood (72 h p.i., %ID/g ± SD); one-way ANOVA]: 0.004 ± 0.008 vs. 0.000 ± 0.001 vs. 0.001 ± 0.0002; [F(2,13) = 29.37, p < 0.0001]. At 72 h p.i., animals injected with [⁸⁹Zr]Zr-TRITA also had 10- and 5-fold more radioactivity associated with liver tissue when compared to animals receiving either [⁸⁹Zr]Zr-PCTA or [⁸⁹Zr]Zr-NOTA, respectively ([⁸⁹Zr]Zr-TRITA vs. [⁸⁹Zr]Zr-PCTA vs. [⁸⁹Zr]Zr-NOTA [liver (72 h p.i., %ID/g ± SD); one-way ANOVA]: 0.304 ± 0.022 vs. 0.030 ± 0.003 vs. 0.069 ± 0.007; [F(2,15) = 731, p < 0.0001]. A similar trend was observed when comparing radioactivity associated with kidney tissues. Animals receiving [⁸⁹Zr]Zr-TRITA retained the most radioactivity in the kidney with only 54% of the 2 h activity being excreted at 72 h p.i. In contrast, 79% and 71% of the 2 h activity in the kidney tissue of animals receiving either [⁸⁹Zr]Zr-PCTA or [⁸⁹Zr]Zr-NOTA had been excreted at the same time point ([⁸⁹Zr]Zr-TRITA vs. [⁸⁹Zr]Zr-PCTA vs. [⁸⁹Zr]Zr-NOTA [kidney (72 h p.i. %ID/g ± SD); one-way ANOVA]: 0.59 ± 0.04 vs. 0.15 ± 0.01 vs. 0.30 ± 0.02; [F(2,15) = 460, p < 0.0001]. Finally, radioactivity in the bone tissue of animals receiving [⁸⁹Zr]Zr-TRITA was elevated (2.06 ± 0.5 %ID/g) at 2 h p.i. and increased by nearly 400% at the conclusion of the study. In contrast radioactivity levels in the bones of animals receiving either [⁸⁹Zr]Zr-PCTA or [⁸⁹Zr]Zr-NOTA were observed to be much lower at 2 h p.i. and further decreased by 10% and 50%, respectively by 72 h p.i. ([⁸⁹Zr]Zr-TRITA vs. [⁸⁹Zr]Zr-PCTA vs. [⁸⁹Zr]Zr-NOTA [bone (72 h p.i., %ID/g ± SD); one-way ANOVA]: 8.04 ± 1.5 vs. 0.094 ± 0.006 vs. 0.046 ± 0.006; [F(2,15) = 176, p < 0.0001].

Figure 5.

Biodistribution data summary of [⁸⁹Zr]Zr-PCTA, [⁸⁹Zr]Zr-NOTA, and [⁸⁹Zr]Zr-TRITA in selected tissues: (a) blood, (b) liver, (c) kidney, and (d) bone. Based upon these studies [⁸⁹Zr]Zr-TRITA demonstrated the worst in vivo behavior of the three radiometal complexes. [⁸⁹Zr]Zr-PCTA and [⁸⁹Zr]Zr-NOTA demonstrated more efficient clearance profiles from tissue suggesting that their in vivo behavior is improved compared to that of [⁸⁹Zr]Zr-TRITA.

Small Animal PET Imaging Studies.

Although [⁸⁹Zr]Zr-TRITA was not considered further due to its extremely poor biodistribution profile, [⁸⁹Zr]Zr-PCTA and [⁸⁹Zr]Zr-NOTA were evaluated in normal mice using small animal PET (Figures 6, S59). Analysis of the dynamic PET data revealed that [⁸⁹Zr]Zr-PCTA and [⁸⁹Zr]Zr-NOTA underwent rapid perfusion and clearance from the blood pool and liver tissue during the first 60 minutes of the study. Both radiometal complexes also exhibited renal excretion. However, the time activity curves were dissimilar with [⁸⁹Zr]Zr-NOTA exhibiting-more efficient renal clearance. The dissimilar urine excretion profiles may be caused by the normal intra-variability that can be associated within individual mice.^66,67 Another explanation may be that the ligand associated with the radiometal complex influences this physiological parameter. A similar phenomenon was observed when the renal clearance of [⁶⁸Ga]Ga-PCTA and [⁶⁸Ga]Ga-NOTA derivatives were evaluated in normal mice with the [⁶⁸Ga]Ga-NOTA derivatives being excreted more efficiently over time.⁶⁸ Interestingly, renal and hepatobiliary excretion of both radiometal complexes were observed as the experimental time course progressed from 1 h p.i. to 24 h p.i. While the cause is uncertain, a similar bimodal excretion pattern was observed when [⁸⁹Zr]Zr-LI-1,2-HOPO was evaluated in acute biodistribution and small animal PET/CT studies.³¹

Figure 6.

Representative static PET images of [⁸⁹Zr]Zr-PCTA and [⁸⁹Zr]Zr-NOTA over 24 h. Panels a and c represent the maximum intensity projections for each radiometal complex, respectively. Panels b and d represent the coronal (left), sagittal (top right), and transverse (bottom right) views for each radiometal complex, respectively. Rapid clearance is observed at 2 h through bladder and kidneys, with increased accumulation in the kidneys over time and minimal uptake in non-clearance tissues at 24 h.

Dosimetric Calculations.

Based upon these results and our previously published data describing [⁸⁹Zr]Zr-DOTA and [⁸⁹Zr]Zr-DFO, we calculated the dosimetric profiles in tissues for each ⁸⁹Zr-complex. While we acknowledge that the dosimetric profiles of the radiometal complexes will be completely different from the calculated tissue dosimetry obtained using a radiolabeled mAb, we thought it prudent to evaluate this criterion since dissociation of the radiometal or intact radiometal complex from the circulating mAb may influence the radiation absorbed by critical tissues such as the liver, kidney and bone marrow. Except for [⁸⁹Zr]Zr-TRITA, all complexes cleared from the whole body at exceptionally fast rates; less than 1% of the administered activity remained in the body at 24 hours post-injection. Dosimetry results (Table 5) demonstrate minimal variability in whole animal mean doses, ranging from 2.3 – 2.7 mGy/MBq. Conversely, [⁸⁹Zr]Zr-TRITA demonstrated a mean total body dose of 65 mGy/MBq. The 25-fold higher total body mean dose and 85-fold higher dose to bone with [⁸⁹Zr]Zr-TRITA can be attributed to the substantial accumulation of radioactivity in the bones, increasing from 7% IA at 2 h p.i. to 16% IA at 48 h p.i. Although only estimates from a sub-sample of each mouse skeleton, it is clear that complex instability leads to significantly increased bone accumulation, and associated whole body dose per administered activity. Among other complexes, dosimetric differences were minimal.

Table 5.

Dosimetry comparison of [⁸⁹Zr]Zr-PCTA, [⁸⁹Zr]Zr-TRITA and [⁸⁹Zr]Zr-NOTA

Tissue/Organ	Dose per Administered Activity (mGy/MBq)
	[89Zr]Zr-DFO	[89Zr]Zr-DOTA	[89Zr]Zr-NOTA	[89Zr]Zr-PCTA	[89Zr]Zr-TRITA
Brain	2.0	1.8	2.2	2.1	54.6
Large Intestine	3.1	21.0	17.1	8.5	28.8
Small Intestine	2.0	5.0	4.7	2.8	16.8
Stomach Wall	15.6	19.1	10.7	41.1	55.6
Heart	2.1	2.0	2.4	2.2	53.3
Kidneys	2.0	2.7	2.7	2.1	16.2
Liver	2.1	2.2	2.4	2.3	33.4
Lung	3.1	2.3	3.0	3.4	151.0
Pancreas	4.3	3.3	3.2	2.7	16.6
Skeleton	8.5	4.2	6.9	10.2	679.0
Spleen	8.4	3.0	3.3	3.3	17.9
Testes	2.0	2.2	2.4	2.0	13.9
Thyroid	2.4	2.1	2.6	2.5	76.0
Urinary Bladder	29.9	30.3	30.4	30.0	41.6
Total Body	2.3	2.4	2.7	2.5	65.0

Retrospective Analysis.

Similar to our previous work, we attempted to place our results in context by comparing them with published studies that describe the in vitro and in vivo behavior of other ⁸⁹Zr chelators, (Tables S26-S28).¹⁵ Variability in study design or unreported data prevented direct comparisons among all ligands, but comparisons were made when possible. Furthermore, we have excluded [⁸⁹Zr]Zr-TRITA from this comparison due to its exceedingly poor performance in vitro and in vivo.

In vitro data suggest that when challenged with 1000-fold excess EDTA (pH 7) for seven days, [⁸⁹Zr]Zr-PCTA was as resistant to transchelation as several [⁸⁹Zr]Zr-TAM and [⁸⁹Zr]Zr-HOPO complexes.^34,38 [⁸⁹Zr]Zr-NOTA was less resistant to transchelation being only 71% intact under the same conditions. Metal ion competition studies revealed that [⁸⁹Zr]Zr-PCTA and [⁸⁹Zr]Zr-NOTA were as resistant to transmetallation as [⁸⁹Zr]Zr-DOTA when exposed to Fe³⁺ ions.¹⁵ [⁸⁹Zr]Zr-PCTA also demonstrated similar resistance to transmetallation as [⁸⁹Zr]Zr-HOPO and [⁸⁹Zr]Zr-DOTA when challenged with Zn²⁺, Co²⁺, Mg²⁺, Gd³⁺, Cu²⁺ and Ga³⁺ ions.¹⁵ However, [⁸⁹Zr]Zr-NOTA was more susceptible to transmetallation when challenged with Cu²⁺ and Ga³⁺ ions.¹⁵ This latter result is not surprising since NOTA and its analogs have gained popularity in the radiopharmaceutical community as useful chelators for ⁶⁴Cu²⁺ and ⁶⁸Ga³⁺ ions.^14,25,69,70 Acute biodistribution studies revealed that [⁸⁹Zr]Zr-PCTA and [⁸⁹Zr]Zr-NOTA demonstrated in vivo behavior that was comparable to [⁸⁹Zr]Zr-LI-1,2-HOPO.³¹

However, both displayed better in vivo behavior than several ligands containing terephthalamide, 3-hydroxypyridin-2-one or hydroxyisophthalamide coordinating units.^29,34,38 For example, [⁸⁹Zr]Zr-PCTA retained 22-and 110-fold less activity in liver and kidney tissue, respectively when compared to [⁸⁹Zr]Zr-TAM-2.³⁴ Similarly, [⁸⁹Zr]Zr-NOTA retained 8-and 99-fold less activity in the same tissues after 24 h.³⁴ Finally, while neither [⁸⁹Zr]Zr-PCTA nor [⁸⁹Zr]Zr-NOTA demonstrated in vivo behavior that was superior to [⁸⁹Zr]Zr-DOTA, their in vivo behavior was similar or superior to [⁸⁹Zr]Zr-DFO and makes them worthy of further study (Figure S56-S58).¹⁵

The data presented here augments what is known about the ⁸⁹Zr radiochemistry of polyazamacrocycles and reveal that PCTA and NOTA should be considered along with DOTA as promising ⁸⁹Zr chelators. However, we acknowledge that their utility will only be recognized if they can be incorporated into a clinically relevant chelator-mAb conjugate, the conjugate can be radiolabeled with ⁸⁹Zr in high specific activity and the resulting ⁸⁹Zr-immuno-PET agent can resist the biological, chemical and physical forces that will be encountered in the systemic circulation over several days. These studies are currently underway in our laboratory.

CONCLUSION

Utilizing the reactivity, coordination chemistry and nuclear chemistry of zirconium we describe the preparation and characterization of Zr-TETA, Zr-TRITA, Zr-PCTA and Zr-NOTA complexes. Similar to Ln³⁺ tetraazamacrocycles complexes, complexes containing macrocycles that were smaller and more preorganized could be prepared under milder conditions. Additionally, the solid-state structures of the Zr complexes derived from the TRITA, PCTA and NOTA ligands further reinforce the idea that the Zr⁴⁺ ion prefers an octa-coordinate ligand environment and will form binuclear complexes to saturate its ligand sphere when the monomeric ligand is incapable of providing eight coordinating units. The ⁸⁹Zr radiochemistry paralleled our non-radioactive synthesis results. The preparation of [⁸⁹Zr]Zr-TETA could not be accomplished while the preparation of [⁸⁹Zr]Zr-TRITA could be achieved only under forcing conditions. Conversely, the quantitative radiolabeling of PCTA and NOTA with ⁸⁹Zr occurred under mild conditions. Additionally, while [⁸⁹Zr]Zr-TRITA demonstrated poor stability, [⁸⁹Zr]Zr-PCTA and [⁸⁹Zr]Zr-NOTA demonstrated robust in vitro stability and in vivo behavior. Our novel observations demonstrate that polyazamacrocycles are efficient ⁸⁹Zr chelators. Considering polyazamacrocycles can chelate a variety of radioisotopes for imaging and therapy their use in immuno-PET applications may create new strategies for theranostic development and improve precision medicine applications.

MATERIALS AND METHODS

Abbreviations

DFO: Deferoxamine; PCTA: 3,6,9,15-Tetraazabicyclo[9.3.1]pentadeca-1(15),11,13-triene-3,6,9-triacetic acid; NOTA: 2,2’,2”-(1,4,7-triazacyclononane-1,4,7-triyl)triacetic acid; TRITA: 2,2’,2”,2”’-(1,4,7,10-tetraazacyclo tridecane-1,4,7,10-tetrayl)tetraacetic acid; TETA: 2,2’,2”,2”’-(1,4,8,11-tetraazacyclotetradecane-1,4,8,11-tetrayl)tetraacetic acid; EDTA: Ethylenediaminetetraacetic acid; Zr(AcAc)₄: Zirconium(IV) acetylacetonate; Zr(ox)₂: Zirconium(IV) Oxalate; TFA: Trifluoroacetic acid; MeOH: Methanol; NH₄Cl: Ammonium chloride; NH₄OAc: Ammonium acetate; NaOAc: Sodium acetate; TMAA: Tetramethylammonium acetate; MES: 2-(N-morpholino)ethanesulfonic acid; HEPES: 4-(2-Hydroxyethyl) piperazine-1-ethanesulfonic acid; TRIS: Tris(hydroxymethyl)aminomethane; NMR: Nuclear Magnetic Resonance; HMBC: Heteroatom Multiple-Bond Correlation; HMQC: Heteronuclear Multiple Quantum Coherence; UV-HPLC: Ultra-Violate High performance liquid chromatography; RT: Retention Time; R_f: Retention Factor; ESI-HR-MS: Electrospray Ionization High Resolution Mass Spectrometry; HS: Human Serum; ITLC: Instant Thin Layer Chromatography; kBq: Kilobecquerel; MBq: Megabecquerel; %ID/g: Percent Injected Dose Per Gram; P.I.: Post-injection; ROI: Region of Interest; CPM: Counts per Minute; TLC: Thin Layer Chromatography; PET: Positron Emission Tomography; CT: Computed Tomography

Experimental Methods

Zirconium-89 (⁸⁹Zr: (t_½ = 78.4 h, β⁺: 22.8 %, E_β+max = 901 keV; EC: 77%, E_γ = 909 keV) was purchased from Sophie, Inc. (Dulles, VA). Unless otherwise noted, all other chemicals were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO, USA), and solutions were prepared using ultrapure water (18 MΩ-cm resistivity). Macrocyclic ligands (PCTA, NOTA, TRITA and TETA) were purchased from CheMatech, Inc. (Dijon, France). High-resolution mass spectrometry data was acquired using a Thermo LTQ-FT (7 Tesla) system at the University of North Carolina Chapel Hill Mass Spectrometry Core Laboratory. NMR spectra were obtained using a Bruker DRX 500 MHz spectrometer equipped with a 5mm tbi z-axis gradient probe. All data were collected and processed with Topspin 1.3 using standard Bruker processing parameters. 2D ¹H-¹H gsCOSY, gsHMQC and gsHMBC were collected as 2K by 256 point data sets at 25 °C and processed to 1K x 512 blocks. ¹H (500 MHz) chemical shifts are reported in parts per million (ppm) relative to the solvent resonances, taken as δ 4.79 for D₂O. Solvent suppression (1D ¹H NMR) was carried out with pre-saturation of the HOD signal. ¹³C NMR (126 MHz) chemical shifts were referenced externally to TSP (0.00 ppm). The NMR spectra of Zr-TRITA complex were obtained at 80 °C.

Single crystal x-ray crystallography analysis of Zr-TRITA, Zr-PCTA and Zr-NOTA was performed on a Bruker APEX CCD system equipped with a graphite monochromator and a Mo Kα sealed x-ray tube (λ = 0.71073 Å). X-rays were provided by a fine-focus sealed x-ray tube operated at 50kV and 30mA. Data was refined using commercially available software packages. the following software packages.^71-76 Additional details specific to each complex can be found in the supporting information.

Radiochemistry reaction progress and purity were analyzed by using an analytical HPLC system (Waters, Milford, MA), which runs Empower software and is configured with a 1525 binary pump, 2707 autosampler, 2998 photodiode array detector, 2475 multichannel fluorescence detector, 1500 column heater, fraction collector, HYPERCARB C18 column (5 μ, 4.6 × 100 mm, Thermo Scientific) and a Carrol Ramsey 105-s radioactivity detector (Berkeley, CA). All ligands (PCTA, NOTA, TRITA and TETA) and associated ^NatZr-complexes were monitored at 201 nm using a mobile phase consisting of 0.1% TFA/H₂O (solvent A) and 0.1% TFA/acetonitrile (solvent B), and a gradient consisting of 0% B to 70% B in 20 min at a flow rate of 1 mL/min. In addition, radio-TLC was conducted on a Bioscan AR 2000 radio-TLC scanner equipped with a 10% methane : argon gas supply and a PC interface running Winscan v.3 analysis software (Eckert & Ziegler, Berlin, DE). Microwave heating was performed using Biotage® Initiator (Biotage AB, Uppsala, Sweden) microwave synthesizer. Varian ITLC-SA strips (Agilent Technologies, Santa Clara, CA) and Merck C-18 TLC plates were employed using a 50 mM M EDTA (pH 5) and 1:1 MeOH:10% NH₄Cl solution as eluents respectively, and ⁸⁹Zr-Oxalate or ⁸⁹ZrCl₄ as a standard control. Radioactive samples were counted using either a CRC-25R radioisotope calibrator (Capintec, Inc.) or a Perkin Elmer 2480 Wizard^® gamma counter (Waltham, MA) with an energy window of 500-1500 keV. A microPET Focus 120 scanner (Concorde Microsystems, Knoxville, Tennessee) was used to obtain static scans at the desired time points with at least 10 million coincident events and ASIPro VM software was used to analyze acquired images.

For statistical analysis, all plots were generated using GraphPad Prism 5.0 software (San Diego, CA). Student’s t tests (two-tailed, unpaired) or one-way analysis of variance (ANOVA) with Tukey’s multiple comparison post-test were performed. p<0.05 was considered statistically significant. Within the text, Student’s t test data is presented as mean ± SD or mean (95% confidence intervals), while Anova data is presented as F(degrees of freedom between groups (k-1), degrees of freedom within groups (N-k)) = F ratio, p value).

Attempted Synthesis of Zr-TETA.

ZrAcAc (54 mg, 0.111 mmol) was added to a solution of TETA (40 mg, 0.092 mmol) in 20 mL of methanol. The resulting solution was refluxed for 3 h. No product formation was observed.

Synthesis of Zr-TRITA.

ZrAcAc (390 mg, 0.80 mmol) was added to a solution of TRITA (280 mg, 0.67 mmol) in 40 mL of methanol. The resulting solution was refluxed for 2 h. As the reaction proceeded, a white precipitate formed. It was filtered, washed with MeOH (3 X 10 mL), and dried in an oven (304 mg, 90% yield). Formation was confirmed by HPLC (Fig. S1), NMR spectroscopy (Figs. S2-S6), and HRMS analysis (Fig. S7). ¹H NMR (400 MHz, D₂O): δ 4.56 – 3.98 (m, 6H), 3.92 – 3.58 (m, 12H), 3.55 – 3.36 (m, 6H), 2.49 – 2.45 (m, 2H); ¹³C NMR (101 MHz, D₂O) Major and Minor: δ 21.75, 52.06, 53.14, 53.32, 53.84, 54.63, 56.34, 57.30, 57.70, 58.39, 58.83, 60.75, 60.96, 62.89, 63.39, 65.65, 66.70, 68.30, 68.62, 72.77, 177.70, 178.81, 179.17, 179.45; HRMS (ESI FT-ICR): Calculated for C₁₇H₂₇N₄O₈Zr, 505.0860 [(M+H)⁺] Found: 505.0853 [(M+H)⁺]. Crystals suitable for X-ray structure determination were grown by dissolving Zr-TRITA in water at 100 °C and allowing the solvent to slowly evaporate at room temperature.

Synthesis of Zr-PCTA.

ZrAcAc (400 mg, 0.82 mmol) was added to a solution of PCTA (259 mg, 0.68 mmol) in 30 mL of methanol. The resulting solution was refluxed for 2 h. As the reaction proceeded, a white precipitate formed. It was filtered, washed with MeOH (3 X 10 mL), and dried in an oven (294 mg, 92% yield). Formation was confirmed by HPLC (Fig. S8), NMR spectroscopy (Figs. S9-S13), and HRMS analysis (Fig. S14). ¹H NMR (400 MHz, D₂O) Major: δ 8.09 – 8.02 (m, 1H), 7.51 (d, J = 7.8 Hz, 1H), 4.74 (d, J = 17.1 Hz, 2H), 4.47 (d, J = 17.1 Hz, 1H), 4.05 (d, J = 16.9 Hz, 1H), 3.77 (d, J = 9.0 Hz, 1H), 3.63 (s, 1H), 3.08 – 2.84 (m, 2H), 2.70 – 2.52 (m, 2H). Minor: δ 8.17 – 8.10 (m, 1H), 7.58 (d, J = 7.9 Hz, 1H), 4.57 (d, J = 16.3 Hz, 1H), 4.30 (d, J = 16.4 Hz, 1H), 3.80 (d, J = 7.7 Hz, 1H), 3.50 (d, J = 16.9 Hz, 1H), 3.63 (s, 1H), 3.19 (m, 2H), 2.84 – 2.74 (m, 1H), 2.42 (m, 1H); ¹³C NMR (101 MHz, D₂O) Major and Minor: δ 57.45, 58.00, 58.20, 59.03, 63.70, 63.95, 65.30, 66.05, 122.12, 122.78, 142.90, 143.74, 158.15, 158.21, 178.60, 179.14, 179.61, 179.79·; HRMS (ESI FT-ICR): Calculated for C₁₇H₂₁N₄O₆Zr, 467.0499 [(M)⁺] Found: 467.0490 [(M)⁺]. Crystals suitable for X-ray structure determination were grown by dissolving Zr-PCTA in water at 80 °C and allowing the solvent to slowly evaporate at room temperature.

Synthesis of Zr-NOTA.

ZrAcAc (619 mg, 1.27 mmol) was added to a solution of NOTA (321 mg, 1.06 mmol) in 30 mL of methanol. The resulting solution was refluxed for 2 h. As the reaction proceeded, a white precipitate formed. It was filtered, washed with MeOH (3 X 10 mL), and dried in an oven (493 mg, 95% yield). Formation of Zr-NOTA-AcAc complex was confirmed by HPLC (Fig. S15), NMR spectroscopy (Figs. S16-S22), and HRMS analysis (Fig. S23). ¹H NMR (400 MHz, D₂O): δ 5.95 (s, 1H), 3.99 (s, 2H), 3.87 – 3.60 (m, 6H), 3.50 – 3.11 (m, 10H), 2.13 (s, 3H), 2.05 (s, 3H); ¹³C NMR (101 MHz, D₂O) δ 24.53, 27.19, 55.50, 56.54, 57.08, 65.84, 66.60, 105.96, 177.76, 178.62, 184.85, 197.38, (minor spices 24.58, 27.24, 30.46, 179.13, 184.90, 197.43); HRMS (ESI FT-ICR): Calculated for C₁₇H₂₆N₃O₈Zr, 490.0761 [(M+AcAc+H)⁺] Found: 490.0767 [(M+AcAc+H)⁺]. Crystals suitable for X-ray structure determination were grown by dissolving Zr-NOTA-AcAc in water at 90 °C and allowing the solvent to slowly evaporate at room temperature.

Radiolabeling of Polyazamacrocyclic Ligands (PCTA, NOTA, TRITA, and TETA) with [⁸⁹Zr]Zr(ox)₂.

The complexation of ⁸⁹Zr with tetraazamacrocyclic ligands (PCTA, NOTA, TRITA, and TETA) was achieved by reacting 10-50 μg (1-5 μL, 10 mg/mL in water) of each ligand with an aliquot of [⁸⁹Zr]Zr(ox)₂ (0.4-0.5 mCi, 16.5-20.5 MBq) diluted in 100 μL of water and pH adjusted to 6.9-7.3 using 1 M Na₂CO₃ or 0.5 M HEPES (pH 7.1-7.3). Reactions were incubated at 99ºC for 2 h in a thermomixer (700 rpm). Formation of [⁸⁹Zr]Zr-PCTA, [⁸⁹Zr]Zr-NOTA, [⁸⁹Zr]Zr-TRITA and [⁸⁹Zr]Zr-TETA complexes was monitored by radio-TLC using Varian ITLC-SA strips and 50 mM EDTA (pH 5) as the mobile phase.

Preparation of [⁸⁹Zr]Zr-chloride.

[⁸⁹Zr]ZrCl₄ was produced using a procedure modified from the literature.²⁴ Briefly, a [⁸⁹Zr]Zr-oxalate solution in 1.0 M oxalic acid was loaded onto an activated Waters Sep-pak Light accell plus QMA strong anion exchange cartridge (300 Å pore size, 37–55 μm particle size, 230 μeq/gram ion exchange capacity), pre-washed with 6 ml MeCN, 10 ml 0.9% saline and 10 ml water. The cartridge was then washed with water (>50 ml) to remove oxalic acid and the activity eluted with 100% recovery of ⁸⁹Zr by chloride ion exchange with 400–500 μl of 1.0 M HCl(aq.).

Radiochemical Synthesis of [⁸⁹Zr]Zr-PCTA with [⁸⁹Zr]ZrCl₄:

Complexing ⁸⁹Zr with the tetraazamacrocyclic ligand PCTA was achieved by reacting 15-20 μg (15-20 μL, 1.0 mg/mL in water) of ligand PCTA with an aliquot of [⁸⁹Zr]ZrCl₄ (1.1-1.5 mCi, 41.6-55.5 MBq) diluted in 200 μL of 0.5 M HEPES (pH 7.2) followed by 60 min incubation at 37 ºC in a thermomixer (550 rpm). The results of radiochemistry studies with different reaction buffers, time, and temperature are summarized in Table S17. Formation of [⁸⁹Zr]Zr-PCTA complex was monitored by radio-TLC using a mobile phase consisting of 50 mM EDTA (pH 5) on Varian ITLC-SA strips. In this ITLC-SA system, free ⁸⁹Zr formed a complex with EDTA and eluted with the solvent front (R_f ~ 1), while [⁸⁹Zr]Zr-PCTA complex remained at the origin (R_f ~ 0) (Figs. S43, S44). The identity of the radioactive complex [⁸⁹Zr]Zr-PCTA was further confirmed by comparing its radio-HPLC elution profile to the UV-HPLC spectrum of nonradioactive ^NatZr-PCTA (Fig. S45).

Radiochemical Synthesis of [⁸⁹Zr]Zr-NOTA with [⁸⁹Zr]ZrCl₄:

Complexing ⁸⁹Zr with the triazamacrocyclic lig- and NOTA was achieved by reacting 15-17 μg (15-17 μL, 1.0 mg/mL in water) of ligand NOTA with an aliquot of [⁸⁹Zr]ZrCl₄ (1.1-1.5 mCi, 41.7-55.7 MBq) diluted in 200 μL of 0.5 M HEPES (pH 7.2) followed by 60 min incubation at 37 ºC in a thermomixer (550 rpm). The results of radiochemistry studies with different reaction buffers, time, and temperature are summarized in Table S18. Formation of [⁸⁹Zr]Zr-NOTA complex was monitored by radio-TLC using a mobile phase consisting of 50 mM EDTA (pH 5) on Varian ITLC-SA strips. In this ITLC-SA system, free ⁸⁹Zr formed a complex with EDTA and eluted with the solvent front (R_f ~ 1), while [⁸⁹Zr]Zr-NOTA complex remained at the origin (R_f ~ 0) (Figs. S46, S47). The identity of the radioactive complex [⁸⁹Zr]Zr-NOTA was further confirmed by comparing its radio-HPLC elution profile to the UV-HPLC spectrum of nonradioactive ^NatZr-NOTA (Fig. S48).

Radiochemical Synthesis of [⁸⁹Zr]Zr-TRITA with [⁸⁹Zr]ZrCl₄ (Conventional Heating Strategy):

Complexing ⁸⁹Zr with the tetraazamacrocyclic ligand TRITA was achieved by reacting 10-15 μg (30-35 μL, 1.0 mg/mL in water) of ligand TRITA with an aliquot of [⁸⁹Zr]ZrCl₄ (0.46-0.49 mCi, 17.1-18.1 MBq) diluted in 200 μL of 0.5 M HEPES (pH 7.2) followed by 120 min incubation at 99 ºC in a thermomixer (550 rpm). The results of radiochemistry studies with different buffers are summarized in Table S19. Formation of [⁸⁹Zr]Zr-TRITA complex was monitored by radio-TLC using a mobile phase consisting of 50 mM EDTA (pH 5) on Varian ITLC-SA strips. In this ITLC-SA system, free ⁸⁹Zr formed a complex with EDTA and eluted with the solvent front (R_f ~ 1), while [⁸⁹Zr]Zr-TRITA complex moved from origin (R_f = 0.20-0.25) (Fig. S49).

(Microwave Assisted Synthesis):

Complexing ⁸⁹Zr with the tetraazamacrocyclic ligand TRITA was achieved by reacting 18-24 μg (18-24 μL, 1.0 mg/mL in water) of ligand TRITA with an aliquot of [⁸⁹Zr]ZrCl₄ (1.1-1.4 mCi, 40.7-51.8 MBq) diluted in 250 μL of water and pH adjusted to 3.9-4.2 using 0.5 M HEPES (50-70 μL, pH 7.1 -7.3). The sealed vial containing the reaction mixture was placed in the microwave reactor, stirred for 2 min, and then heated at 180 ºC for 20 min. The results of radiochemistry studies with different reaction pH and temperature are summarized in Table S20. Formation of [⁸⁹Zr]Zr-TRITA complex was monitored by radio-TLC using a mobile phase consisting of 1:1 MeOH: 10% NH₄Cl on C-18 plates and 50 mM EDTA (pH 5) on Varian ITLC-SA strips. In the C-18 system, un-chelated ⁸⁹Zr remained at the origin (R_f ~ 0), while [⁸⁹Zr]Zr-TRITA complex moved near the solvent front (R_f = 0.85-0.90). In the ITLC-SA system, free ⁸⁹Zr formed a complex with EDTA and eluted with the solvent front (R_f ~ 1), while [⁸⁹Zr]Zr-TRITA complex moved from origin (R_f = 0.20-0.25) (Figs. S50, S51). The identity of the radioactive complex [⁸⁹Zr]Zr-TRITA was further confirmed by comparing its radio-HPLC elution profile to the UV-HPLC spectrum of nonradioactive ^NatZr-TRITA (Fig. S52).

Radiochemical Synthesis of [⁸⁹Zr]Zr-TETA with [⁸⁹Zr]ZrCl₄:

Complexing ⁸⁹Zr with the tetraazamacrocyclic ligand TETA was achieved by reacting 10-50 μg (10-50 μL, 1.0 mg/mL in water) of ligand TETA with an aliquot of [⁸⁹Zr]ZrCl₄ (0.4 - 0.5 mCi, 16.6 - 20.5 MBq) diluted in 200 μL of 0.5 M HEPES (pH 7.1-7.3) followed by 120 min incubation at 99 ºC in a thermomixer (550 rpm). The results of radiochemistry studies with different reaction buffers are summarized in Table S21. Formation of [⁸⁹Zr]Zr-TETA complex was monitored by radio-TLC using a mobile phase consisting of 50 mM EDTA (pH 5) on Varian ITLC-SA strips. In this ITLC-SA system, free ⁸⁹Zr formed a complex with EDTA and eluted with the solvent front (R_f ~ 1), while [⁸⁹Zr]Zr-TETA complex remained at the origin (R_f ~ 0) (Figs. S53).

Determination of Partition Coefficients (LogP).

The partition coefficient (logP) for each complex was determined by adding 5 μL of each [⁸⁹Zr]Zr-labeled complex (approx. 5 μCi; 0.19 MBq) to a mixture of 500 μL of octanol and 500 μL of water (pH 5.5).⁶⁵ The resulting solutions (n = 5) were vigorously vortexed for 5 min at room temperature, then centrifuged for 5 min to ensure complete separation of layers. From each of the five sets, 50 μL aliquot was removed from each phase into screw tubes and counted separately in a gamma counter. Each organic phase was washed with water to remove any radioactivity remaining in the organic phase before gamma counting. The partition coefficient was calculated as a ratio of counts in the octanol fraction to counts in the water fraction. The logP values were reported in an average of five measurements.

In vitro EDTA Challenge Study.

In vitro EDTA challenge study was carried out by adding 20 μL of each [⁸⁹Zr]Zr-labeled complex (110-120 μCi, 4.07-4.44 MBq) to 500 μL of EDTA (10 mM, 50 mM and 100 mM: pH 5 and pH 7) with a 1:100, 1:500, and 1:1000 ratio of ligand/EDTA.¹⁵ The solutions (n=3) were incubated at 37 °C for 7 days in a thermomixer. Samples were analyzed at 0, 1, 3, 5, and 7 days post administration to EDTA by radio-TLC using Varian ITLC-SA strips and 50 mM EDTA (pH 5) as the mobile phase and gamma counting using an energy window of 500-1500 keV and standard protocols. All studies were performed in triplicate.

In vitro Metal Competition Study.

To a solution of metal cations [iron (III) chloride, cobalt (II) chloride, zinc (II) chloride, copper (II) chloride, magnesium (II) chloride, gallium (III) nitrate, gadolinium (III) chloride] (1 mM, 200 μL), was added [⁸⁹Zr]Zr-labeled complex (0.1 mM, 20 μL, 110-120 μCi, 4.07-4.44 MBq) in PBS, pH 7.4.¹⁵ The resulting solutions (n = 3) were incubated at 37ºC for 7 days in a thermomixer. Dissociation of ⁸⁹Zr from [⁸⁹Zr]Zr-complexes was monitored by radio-TLC at 0, 1, 3, 5, and 7 days by radio - TLC using Varian ITLC-SA strips and 50 mM EDTA (pH 5) as the mobile phase and gamma counting using an energy window of 500-1500 keV and standard protocols. All studies were performed in triplicate.

In Vitro Serum Stability.

In vitro serum stability was carried out by adding 20 μL of each [⁸⁹Zr]Zr-labeled complex (110-120 μCi, 4.07-4.44 MBq) to 500 μL of human serum.¹⁵ The solutions (n=3) were incubated at 37 °C for 7 days and were analyzed at 1, 3, 5, and 7 days by size exclusion chromatography (SEC) using a Superdex 200 10/300 GL™ column (GE Healthcare Life Sciences, Piscataway, NJ) and phosphate buffered saline (PBS) as eluent with a flow rate of 0.5 mL/min. Fractions (0.5 mL per tube) were collected and the activity in each fraction was measured in a gamma counter. The percent intact radiopharmaceutical was determined by subtracting the total integrated area under the product peak from the total integrated area generated for all peaks in the chromatogram and multiplying by a factor of 100%.

Biodistribution Studies.

Biodistribution studies were conducted using a modified literature procedure.¹⁵ Briefly, female NIH Swiss mice (6-8 wk old, n=6) were injected with each [⁸⁹Zr]Zr-labeled complex (0.55 MBq (15 μCi)/mouse) via the tail vein, and sacrificed at 2, 4, 24, 48, 72 h post-injection. Organs and tissues of interest were excised, weighted, and counted on a Perkin Elmer 2480 Wizard^® gamma counter (Waltham, MA). The percent injected dose per gram (%ID/g) and percent injected dose per organ (%ID/organ) were calculated by comparison to a weighed, counted standard for each group (Tables S23 - S25). The change in tissue associated activity over time was calculated using the formula: % 2h tissue-associated activity remaining at 72 h = 100 x [ % ID/g_72h /%ID/g_{2 h}].

Small Animal Models and PET Imaging.

Initial distribution and pharmacokinetics of the [⁸⁹Zr]Zr-PCTA or [⁸⁹Zr]Zr-NOTA in healthy NIH Swiss mice (n = 3/group) was recorded by dynamic PET scans (30 μg, 11.1 MBq) for 1 hour post-injection.¹⁵ Dynamic PET scans were acquired in list mode format over 60 min, starting just prior to radiotracer injection. The resulting data was sorted into 0.5-mm sinogram bins and 60 time frames for image reconstruction (1 × 60 s). A microPET Focus 120 scanner (Concorde Microsystems, Knoxville, Tennessee) was used to obtain static scans at the desired time points with at least 10 million coincident events and ASIPro VM software was used to analyze acquired images. ROIs were drawn manually over the heart, liver, and kidneys (individually) and raw counts were adjusted to %-injected dose/gram via decay-corrected doses of [⁸⁹Zr]Zr-PCTA and [⁸⁹Zr]Zr-NOTA.

Dosimetry.

Biodistribution data from ex vivo studies, along with organ mass data collected during counting, were used to compute the fraction of administered activity in each tissue structure at 2 h, 4 h, 24 h, 48 h, and 72 h post-injection. Time-activity curves were integrated using the trapezoidal method with assumptions of constant uptake between 0 hours and 2 hours, and only physical decay beyond 72 hours. The resulting time-integrated activity coefficients (TIACs) were used to calculate tissue dosimetry using the OLIDNA/EXM v2.1 software platform. Based on typical animal weights for 6-8 week old female NIH Swiss mice, the 25 gram mouse phantom model was selected for calculations. A TIAC for the total body/remainder category was calculated by mono-exponential fitting to the aggregate organ clearance data, followed by applying the obtained exponential rate constant to the otherwise un-accounted for administered activity. The urinary bladder contents residence was calculated using the same whole-body clearance rate, assuming a voiding interval of 0.1 hours. Resultant dosimetric data are presented with units of absorbed dose (mGy/MBq) rather than equivalent dose (mSv/MBq) to avoid confusion regarding the concept of effective dose or effective dose equivalent in mice. With that said, all radiation from ⁸⁹Zr is assumed to have a radiation-specific weighting factor of 1, and thus the conversion from absorbed dose to equivalent dose would also be a factor of 1.