Stable Chelation of Zr-89 and Ac-225 with Dual Size-Selective Macrocyclic py₂‑macrodipa for ²²⁵Ac/⁸⁹Zr Paired Theranostics

Abstract

Radiotheranostic agents combine targeted therapeutics and diagnostic imaging, offering more efficient and personalized treatment strategies. Positron emission tomography provides high sensitivity and quantitative accuracy; however, its integration with α-particle radiotherapy remains limited by the lack of chelators capable of stabilizing radionuclides with widely different ionic radii. ⁸⁹Zr and ²²⁵Ac are promising radionuclides for PET imaging and targeted α-therapy, respectively, but require robust and versatile chelation strategies. Herein, we report stable coordination using the 18-membered macrocyclic dual size-selective chelator py₂-macrodipa with [⁸⁹Zr]­Zr⁴⁺and [²²⁵Ac]­Ac³⁺ ions. The chelator demonstrated efficient radiolabeling with both ⁸⁹Zr and ²²⁵Ac, exhibiting ⁸⁹Zr labeling efficiency comparable to the state-of-the-art chelator DFO at micromolar concentrations. The [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ complex showed high stability in human serum and in vivo, with minimal uptake in nonspecific tissues such as bone and other organs. Similarly, the [²²⁵Ac]­[Ac­(py₂-macrodipa)]⁺ complex displayed robust in vitro and in vivo stability along with a biodistribution clearance profile similar to its ⁸⁹Zr analogue. In addition, a squaramide ethyl ester-conjugated bifunctional chelator (py₂-macrodipa-sq) was synthesized for antibody conjugation and assessed using the PD-1 targeting antibody Opdivo. Both ⁸⁹Zr- and ²²⁵Ac-labeled opdivo-py₂-macrodipa-sq conjugates were obtained with >98% radiochemical purity and demonstrated high stability in vitro and in vivo. Although the therapeutic construct exhibited higher activity levels in blood and other organs than the diagnostic analogue, both cleared over time with similar distribution patterns and low bone uptake. These findings establish py₂-macrodipa and its bifunctional derivative, py₂-macrodipa-sq, as potential chelation platforms for the development of matched ⁸⁹Zr/²²⁵Ac radiotheranostic agents, demonstrating stable and comparable in vivo behavior.

Introduction

Radiotheranostics are an innovative approach in the precision medicine landscape in that they fuse diagnostic and therapeutic capabilities for pretreatment planning, real-time dosimetry, and monitoring of treatment response. Together, this approach offers improved and personalized targeted treatments while minimizing off-target toxicity. , For optimal application, quantitative assessment of the distribution of therapeutic radioisotopes is important to assist in absorbed dose measures and monitoring response over cycles of treatment. However, limitations in the quantitative accuracy, sensitivity, and resolution of single-photon emission computed tomography (SPECT) imaging when using therapeutic emitters are a challenge. , Inaccuracies are exacerbated at the ultralow activity levels involved in targeted α-particle (α) emitting therapies. − Positron emission tomography (PET) offers superior sensitivity, quantitative precision, and resolution over SPECT; , however, there are no readily available positron-emitting therapeutic isotopes.

Modification between therapeutic and diagnostic analogues of the same agent ideally involves the single-atom substitution of a positron emitter for a therapeutic radionuclide. The use of radiometals requires their stable chelation with the targeting vector (antibody or peptide) that remains intact in vivo. However, a fundamental challenge is to ensure that chemical changes to the theranostics have minimal impact on pharmacokinetic behavior, and there are few stable and effective chelators that can conjugate both positron-emitting diagnostic and therapeutic isotopes to a targeting molecule without compromising function.

Instability of the metal-chelate system can lead to premature release of isotopes, reducing effectiveness and potentially increasing toxicity. This study adopts a dual approach that combines detailed coordination chemistry investigations of radiometal-chelator systems with in vivo radiochemical evaluation to assess the translational relevance of these systems under biological conditions. While coordination chemistry provides key information on metal-ligand affinity, thermodynamic stability, and coordination geometry, the chemical speciation and stability of radiometal complexes in vivo may differ substantially due to the complexity of the physiological environment. We focused on Zirconium-89 (⁸⁹Zr), a promising candidate for preclinical and clinical diagnostic PET imaging , due to its favorable half-life (t _1/2 = 78.4 h), which is well matched with the pharmacokinetics of monoclonal antibodies, , enabling high-resolution and quantitative assessment of their in vivo distribution. In addition, ⁸⁹Zr has been successfully incorporated into peptide-based radiopharmaceuticals, enabling delayed imaging. − The radionuclide pair Zr-89 and Actinium-225 (²²⁵Ac) represents a promising theranostic combination due to their compatible half-lives. Prior studies have highlighted the potential of this combination for coordinated imaging and targeted alpha therapy. − Desferrioxamine (DFO, Figure ) remains the most commonly used chelator for ⁸⁹Zr, particularly for antibody conjugation; , however, DFO is unsuitable for α-emitting radiometals with expanded ionic radii such as Actinium-225 (²²⁵Ac), Thorium-227 (²²⁷Th), and Radium-223 (²²³Ra). The development of a single chelator capable of stably coordinating both diagnostic ⁸⁹Zr and therapeutic α-emitters would therefore significantly advance radiotheranostic applications.

1.

Structures of the ligands discussed in this work.

DOTA (1,4,7,10-tetraazacyclododecane-N,N′,N″,N‴-tetraacetic acid) is the most widely used macrocyclic chelator in radiopharmaceuticals due to its ability to form stable complexes with diverse radionuclides. Recent findings have shown that tetraazamacrocyclic DOTA may be a superior chelator for ⁸⁹Zr compared to DFO, demonstrating high in vitro and in vivo stability, yet this requires heat and time to encourage thermodynamically stable complexes with low charge density and large-radius metal ions, limiting its use for protein labeling. , Recently, we identified the 18-membered macrocyclic chelator macropa, which shows exceptional affinity for metal ions across a range of ionic radii, and its potential bifunctional derivative for ²²⁵Ac has been explored. We have extended that work to ²²³Ra, the first approved α-particle emitter, demonstrating its utility for large ionic radii but presenting limited application for radionuclides with smaller ionic radii.

Dual-size selective chelators have emerged as a promising solution. These ligands undergo conformational switching upon metal binding, forming stable complexes with both metal ions of larger ionic radii (10-coordinate, C₂-symmetric conformation A) and smaller ionic radii (8-coordinate, asymmetric conformation B). Previous investigations demonstrate stable complexation of the py-macrodipa ligand with radiometal ions of ionic radii ranging from 1.03 to 0.74 Å, including Lanthanum-135 (¹³⁵La) and Bismuth-213 (²¹³Bi, the ionic radii of 6-coordinated La³⁺ and Bi³⁺ are similar, 1.03 Å), which can be used in targeted radionuclide therapy, and a positron emitter Scandium-44 (⁴⁴Sc, 6-coordinated ionic radius of Sc³⁺: 0.74 Å); − although its stability with ²²⁵Ac is limited. Multiple ionic radii for Ac³⁺ have been reported in the literature. Zachariasen and coworkers determined a value of 1.11 Å for six-coordinate Ac³⁺, − which is consistent with Shannon’s estimate of 1.12 Å. More recently, Lundberg and Persson proposed a larger ionic radius of 1.22 Å for Ac³⁺ with a coordination number of nine. The relatively large ionic size and diffuse charge of this ion contribute to weaker metal-ligand interactions. The third-generation ligand py₂-macrodipa demonstrates improved thermodynamic and kinetic stability with ²²⁵Ac. We have expanded this work towards the diagnostic SPECT imaging radionuclide Indium-111 (¹¹¹In) with a small ionic radius.

In light of these advancements, we investigated further application of py₂-macrodipa for radiotheranostics. The six-coordinate and eight-coordinate ionic radii of Zr⁴⁺ are 0.72 Å and 0.84 Å, respectively, which are smaller than those of six-coordinate and eight-coordinate In³⁺ (0.80 Å and 0.92 Å, respectively). For stable and effective chelation, the size of the metal ion should ideally match the spatial arrangement and distance between the donor atoms. Moreover, Zr­(IV), a tetravalent and strongly oxophilic metal ion with a relatively small ionic radius, exhibits a pronounced preference for hydroxo and other oxygen-donor ligands, consistent with its hard Lewis acid character and high affinity for oxygen. Consequently, chelating Zr⁴⁺ with 18-membered macrocyclic chelator py₂-macrodipa would be expected to be challenging. Yet surprisingly, we demonstrate successful coordination of this ligand with Zr⁴⁺, identify optimal labeling conditions, and assess its suitability for both ⁸⁹Zr and ²²⁵Ac through in vitro and in vivo studies. To further evaluate the in vivo relevance of this chelator, we developed a squaramide ethyl ester-conjugated chelator, py₂-macrodipa-sq. Unlike isothiocyanates, squaramide ethyl esters are highly resistant to hydrolysis and offer extended shelf stability. − We investigated its performance when py₂-macrodipa-sq was conjugated to the FDA-approved anti-PD-1 antibody Opdivo (Nivolumab). Antibody-chelator conjugates were efficiently radiolabeled with both ⁸⁹Zr and ²²⁵Ac and assessed for stability and in vivo performance. These studies establish py₂-macrodipa as a versatile platform capable of supporting matched immuno-PET imaging and targeted alpha therapy using a single antibody construct.

Results

[Zr­(py₂-macrodipa)]²⁺ Complex Characterization

To evaluate the potential of py₂-macrodipa for ⁸⁹Zr PET imaging, the [Zr­(py₂-macrodipa)]²⁺ complex was characterized by ¹H NMR spectroscopy. The NMR sample was prepared by incubating an aqueous solution of py₂-macrodipa and Zr­(acac)₄ (1:1 metal:ligand ratio). The pH of the solution was adjusted to 5.0 using 1 M NaOH, and the mixture was stirred at room temperature overnight. The resulting product was isolated by lyophilization as a white solid and used directly for subsequent ¹H NMR analysis. Previous studies have shown that ligands similar to py₂-macrodipa form complexes with distinct geometries depending on the preferred coordination environment of the metal ion, with metal ions favoring higher coordination numbers (e.g., La³⁺) adopting a symmetric Conformation A, whereas those favoring lower coordination numbers (e.g., In³⁺) adopt an asymmetric Conformation B (Figure ). The ¹H NMR spectrum of [Zr­(py₂-macrodipa)]²⁺ displayed well-resolved aromatic resonances and split aliphatic signals, including two distinct sets of picolinate resonances, indicating an asymmetric coordination environment. Comparison with La³⁺ and In³⁺ complexes (Figure ) revealed that the Zr⁴⁺ spectrum closely resembles that of In³⁺ in terms of the number of resonances and chemical shifts, strongly suggesting that Zr⁴⁺ adopts the asymmetric Conformation B.

2.

Depiction of the conformational switch present in Ln³⁺-py₂-macrodipa complex. Reproduced with permission from A. Hu, S. N. MacMillan, J. J. Wilson. Macrocyclic ligands with an unprecedented size-selectivity pattern for the lanthanide ions. J Am Chem Soc. 2020; 142:13500–13506. Copyright 2020 American Chemical Society.

3.

¹H NMR spectra of py₂-macrodipa, [La­(py₂-macrodipa)]⁺, [In­(py₂-macrodipa)]⁺, and [Zr­(py₂-macrodipa)]²⁺ (500 MHz, D₂O, pD 5, 25 °C). The Zr⁴⁺ spectrum closely resembles that of In³⁺ strongly suggesting that Zr⁴⁺ adopts the asymmetric Conformation B.

Although we were unable to obtain crystals of the py₂-macrodipa complex, we successfully grew suitable crystals of the Zr complex with the second-generation dual-size selective ligand py-macrodipa (Figure ). The ¹H NMR spectra of these crystals clearly indicate the presence of Conformation B in the chelator arrangement, as shown in Figure S1. The solid-state structure of [Zr­(py-macrodipa)­(OMe)]⁺ reveals an 8-coordinate Zr­(IV) center with an N₅O₃ donor set. The coordination sphere is composed of the pyridyl nitrogen donors, two tertiary amine donors of the macrocycle, two picolinate oxygen atoms, and an inner-sphere methoxide ligand. The Zr–O bond distances span 1.94–2.18 Å, while the Zr–N bond distances range from 2.33 to 2.51 Å, consistent with strong coordination to oxygen donors and slightly longer interactions with nitrogen donors. The coordination geometry can be described as a distorted eight-coordinate polyhedron, deviating from an ideal square-antiprismatic arrangement due to the constrained topology of the macrocycle and the asymmetric donor distribution associated with Conformation B. In contrast to the previously reported [In­(py-macrodipa)]⁺ structure, in which the eighth coordination site is occupied by the macrocyclic ether oxygen, the [Zr­(py-macrodipa)­(OMe)]⁺ structure adopts a different coordination mode in which the ether oxygen is not bound, and a methoxide ligand completes the coordination sphere. The presence of this inner-sphere methoxide results in a more expanded macrocyclic conformation. Similar coordination of solvent-derived ligands has been observed in the related [In­(py₂-macrodipa)­(OH₂)]⁺ complex, where an inner-sphere water molecule was present. These observations suggest a degree of structural flexibility, in which exchange between an inner-sphere solvent molecule and the macrocyclic ether donor may occur. Accordingly, slight variations in crystallization conditions may favor the isolation of distinct coordination modes. The identity of the complex is further supported by ESI-HRMS (Figure S2).

4.

Crystal structure of [Zr­(py-macrodipa)­OMe]­[Cl]­DMF. Thermal ellipsoids are drawn at the 50% probability level. Solvent, counterions, and nonacidic hydrogen atoms are omitted for clarity.

Radiolabeling with Zr-89

To investigate the potential for chelation with macropa and py₂-macrodipa towards [⁸⁹Zr]­Zr⁴⁺, concentration-dependent radiolabeling studies were performed and compared with the control DFO complex. Ligands were incubated with 0.2 MBq of [⁸⁹Zr]­Zr⁴⁺ (in oxalic acid) in pH 6.5 HEPES buffer, and radiochemical yields (RCYs) were assessed by radio-TLC (iTLC-SG; Figure ). Notably, py₂-macrodipa efficiently incorporates [⁸⁹Zr]­Zr⁴⁺ at a low concentration of 10^–6 M at 90 °C within 30 min, achieving an RCY of 87% (Figure , Table S1). Following C18 cartridge purification, a radiochemical purity of 97% was obtained (Figure S12). These results are comparable to DFO, which quantitatively complexes [⁸⁹Zr]­Zr⁴⁺ at micromolar concentrations with RCY > 97% at 37 °C, highlighting that both ligands can be efficiently radiolabeled at micromolar concentrations, although py₂-macrodipa requires a higher reaction temperature. Radiolabeling of py₂-macrodipa at 37 °C afforded moderate RCYs (35–40%). By contrast, quantitative radiolabeling of DOTA with [⁸⁹Zr]­Zr⁴⁺ typically requires elevated temperature conditions of ∼90 °C. Efficient radiolabeling of macropa was observed at relatively high ligand concentrations (15 mM), as confirmed by TLC analysis and concentration-dependent studies. At lower concentrations, the radiolabeling efficiency decreased significantly (Figure ). Extending the reaction time to 60 min did not improve RCYs.

5.

Thin-layer chromatograms of (a) free [⁸⁹Zr]­Zr⁴⁺, (b) [⁸⁹Zr]­[Zr­(macropa)]²⁺, (c) [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺, and (d) [⁸⁹Zr]­[Zr­(DFO)]; iTLC-SG eluted with 50 mM EDTA, pH 5.5; free [⁸⁹Zr]­Zr⁴⁺ migrated with the solvent front (R _f = 1), whereas the complex species remained at the baseline (R _f = 0); ligand concentration used: 15 × 10^–3 M; activity of ⁸⁹Zr source: 0.2 MBq; 90 °C for (a), (b), and (c); 37 °C for (d); 30 min.

6.

Radiochemical yields% (RCY%) of ⁸⁹Zr radiolabeling with (a) py₂-macrodipa, (b) DFO, and (c) macropa at different ligand concentrations; the ligand concentrations required for 50% RCY% (RCY%₅₀) for py₂-macrodipa, DFO, and macropa were found to be 3.051 × 10^–8, 1.814 × 10^–7, and 1.499 × 10^–4 M, respectively.

Radiolabeling studies were also conducted using [⁸⁹Zr]­ZrCl₄ to assess potential differences in chelation behavior. Concentration-dependent experiments (micromolar to millimolar) were performed with ∼1.85 MBq of [⁸⁹Zr]­ZrCl₄ in pH 6.5 HEPES buffer at 25 °C for 1 h. As illustrated in Figure S13a, DFO achieved nearly 100% radiochemical conversion, while DOTA, macropa, and py₂-macrodipa failed to incorporate [⁸⁹Zr]­ZrCl₄ under these conditions (Figure S14, Table S2). At 90 °C, as expected, DOTA exhibited nearly complete conversion, while DFO failed to label due to limited thermal stability (Figure S13b). Notably, macropa and py₂-macrodipa also failed to incorporate [⁸⁹Zr]­Zr⁴⁺ under these conditions (Figure S15, Table S3), in contrast to their behavior with [⁸⁹Zr]­Zr⁴⁺ in oxalic acid. This selective radiolabeling of py₂-macrodipa with oxalate-derived [⁸⁹Zr]­Zr⁴⁺ may offer advantages for the development of ⁸⁹Zr-based PET radiotracers.

Radiolabeling with Ac-225

Having established the successful radiolabeling with the [⁸⁹Zr]­Zr⁴⁺ ion (in oxalic acid), we subsequently evaluated [²²⁵Ac]­Ac³⁺ to assess in vivo stability and its potential as a candidate for targeted α therapy (TAT). Both macropa and py₂-macrodipa were efficiently radiolabeled with [²²⁵Ac]­Ac³⁺ in 0.5 M ammonium acetate buffer (pH 6.0) at room temperature, achieving high radiochemical yields (95–98%) within 5 min under mild conditions (Figure ). These results demonstrate the favorable radiolabeling properties of py₂-macrodipa, enabling effective coordination of both metal ions with higher charge density, such as [⁸⁹Zr]­Zr⁴⁺, as well as the lower charge density ion [²²⁵Ac]­Ac³⁺.

7.

Thin-layer chromatograms of (a) free [²²⁵Ac]­Ac³⁺, (b) [²²⁵Ac]­[Ac­(macropa)]⁺, and (c) [²²⁵Ac]­[Ac­(py₂-macrodipa)]⁺; iTLC-SG eluted with 50 mM EDTA, pH 5.5; free [²²⁵Ac]­Ac³⁺ migrated with the solvent front (R _f = 1), whereas the complex species remained at the baseline (R _f ∼ 0).

In Vitro Stability

Kinetic stability in human serum was evaluated to assess the stability of these complexes under physiological conditions. The ⁸⁹Zr-complexes of py₂-macrodipa, macropa, and DFO were incubated in human serum at 37 °C for up to 48 h and analyzed by radio-TLC (Figure S17). [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ showed high stability, with >96% remaining intact after 48 h, comparable to DFO (Figure , Table S4). In contrast, only 48% of [⁸⁹Zr]­[Zr­(macropa)]²⁺ remained intact after 24 h. Although the formation of the [⁸⁹Zr]­[Zr­(macropa)]²⁺ complex was confirmed by early time point radio-TLC, the complex exhibited limited in vitro stability, as revealed by its behavior in human serum during later time point analysis. The kinetic inertness to serum of [²²⁵Ac]­[Ac­(macropa)]⁺ and [²²⁵Ac]­[Ac­(py₂-macrodipa)]⁺ is consistent with prior reports (Figures S16 and S18, Table S5). Notably, the similarly high stability of the ⁸⁹Zr and ²²⁵Ac-py₂-macrodipa complexes supports py₂-macrodipa as a promising theranostic chelator for the ⁸⁹Zr/²²⁵Ac pair.

8.

Stability of [⁸⁹Zr]­[Zr­(DFO)], [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺, and [⁸⁹Zr]­[Zr­(macropa)]²⁺ in human serum over 48 h at 37 °C; as assessed by iTLC, [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ remained more than 96% intact, comparable to the DFO complex, whereas [⁸⁹Zr]­[Zr­(macropa)]²⁺ remained 46% intact after 48 h; ligand concentration used: 15 × 10^–3 M.

In Vivo Studies Positron Emission Tomography

We next investigated in vivo stability of the [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ and [⁸⁹Zr]­[Zr­(macropa)]²⁺ complexes by PET imaging in naive mice. Projection images and quantitative volumes of interest (VOI) analysis are shown in the SI (Figure S19 and Figure S20). At 1 h post-injection (p.i.), [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ showed predominant intestinal uptake with minor bladder activity (Figure ); residual gut activity was observed at 4 h and completely cleared by 24 h. Negligible bone uptake at all time points indicated no in vivo release of free [⁸⁹Zr]­Zr⁴⁺ ions. For comparison with the control group, [⁸⁹Zr]­[Zr­(DFO)] showed only renal excretion as expected. In contrast, for the [⁸⁹Zr]­[Zr­(macropa)]²⁺ complex, activity was observed in the bladder at 1–4 h p.i. and at 24 h p.i., remaining activity was accumulated in the bone, indicating in vivo instability and [⁸⁹Zr]­Zr⁴⁺ release.

9.

PET images of representative naive Swiss Webster female mice with [⁸⁹Zr]­Zr⁴⁺, [⁸⁹Zr]­[Zr­(macropa)]²⁺, [⁸⁹Zr]­[Zr­(DFO)], and [⁸⁹Zr]­[Zr­(py₂-macrodipa)]². Free [⁸⁹Zr]­Zr⁴⁺ was located in the bone at 24 h p.i., for [⁸⁹Zr]­[Zr­(macropa)]²⁺ complex, most of the activity accumulated in the bone, indicating release of free [⁸⁹Zr]­Zr⁴⁺ ion. [⁸⁹Zr]­[Zr­(DFO)] complex was rapidly excreted through the kidneys, whereas [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ complex showed gut uptake, which was cleared rapidly, and no uptake in the bone was observed. Note, the intensity ranges for these coronal sections are fixed across the imaging sessions for each ligand complex in order to enable visualization of activity distribution.

Biodistribution Studies

Ex vivo biodistribution studies at 1, 4, and 24 h p.i. corroborated the PET findings (Figure , Table S8). The biodistribution profile of free [⁸⁹Zr]­Zr⁴⁺ revealed slow blood clearance with high bone accumulation (Figure S21). [⁸⁹Zr]­[Zr­(macropa)]²⁺ displayed a similar profile, with blood (4.4 ± 0.4 %IA/g) and kidney activity (3.1 ± 1.1 %IA/g) at 1 h p.i. and increasing bone uptake at 24 h (tibia and calvaria with 3.9 ± 0.8 and 8.3 ± 3.4 %IA/g, respectively), indicating in vivo dissociation and release of ⁸⁹[Zr]­Zr⁴⁺ (Figure S24).

10.

Organ distribution of (a) [⁸⁹Zr]­[Zr­(macropa)]²⁺, (b) [⁸⁹Zr]­[Zr­(DFO)], and (c) [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ following retro-orbital injection in mice. Naive Swiss Webster females, 4 months old, were sacrificed after 1, 4, and 24 h p.i. Values for each time point are given as %IA/g; %IA/g = percent injected activity per g of tissue; SI = small intestine and LI = large intestine. [⁸⁹Zr]­[Zr­(macropa)]²⁺ showed activity localization in the bone, indicating release of free metal ion from the ligand, [⁸⁹Zr]­[Zr­(DFO)] was rapidly renally excreted, while [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ was processed through the intestine, and negligible bone uptake at all time points was observed, indicating high in vivo stability.

By contrast, the organ distribution profile of [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ differs significantly from that of free [⁸⁹Zr]­Zr⁴⁺. It clearly showed rapid intestinal and renal excretion, minimal blood and organ retention, and negligible bone uptake at all time points, consistent with high in vivo stability. The majority of the activity localized in the intestine (10.6 ± 15 %IA/g) and kidneys (1.3 ± 0.7 %IA/g) at the early time point was cleared from the body at 24 h p.i.

As expected, [⁸⁹Zr]­[Zr­(DFO)] was primarily cleared renally, with negligible bone accumulation. The biodistribution and clearance profile of [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ differs from that of [⁸⁹Zr]­[Zr­(DFO)] complex, with lower tissue retention observed with the latter ligand; however, each ligand was cleared from the organs over time with minimal bone accumulation. The in vivo stability of [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ indicates its potential as a viable candidate for radiopharmaceutical agent development.

Given its efficient radiolabeling and in vitro stability with ²²⁵Ac, the in vivo stability and biodistribution of [²²⁵Ac]­[Ac­(py₂-macrodipa)]⁺ were evaluated and compared to those of [²²⁵Ac]­[Ac­(macropa)]⁺ (Figure ). As expected, [²²⁵Ac]­[Ac­(macropa)]⁺ showed rapid clearance, with low blood activity at 1 h p.i., predominant renal excretion, and residual intestinal activity at 4 h post administration. [²²⁵Ac]­[Ac­(py₂-macrodipa)]⁺ exhibited a similar profile, with low blood retention and rapid clearance, showing kidney, liver, and intestinal activity at 4 h p.i. Importantly, negligible activity accumulation was observed in the liver, spleen, or bone, indicating high in vivo stability and no release of [²²⁵Ac]­Ac³⁺.

11.

Organ distribution of (a) [²²⁵Ac]­[Ac­(macropa)]⁺ and (b) [²²⁵Ac]­[Ac­(py₂‑macrodipa)]⁺ following retro-orbital injection in mice. Naive mice, 4 months old, were sacrificed after 1, 4, and 24 h postadministration. Values for each time point are given as %IA/g; %IA/g = percent injected activity per g of tissue; SI = small intestine and LI = large intestine. Both complexes showed rapid intestinal and renal excretion, and negligible activity in the bone and other organs over time indicates in vivo stability.

A comparison of the biodistribution of [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ and [²²⁵Ac]­[Ac­(py₂‑macrodipa)]⁺ was also conducted. Other than greater measured uptake with the therapeutic analogue at the earliest time point of ∼1 %IA/g, the salient difference is the persistent liver signal. Globally, the distribution profiles reveal intestinal disposal, low uptake in the liver and kidneys, and little residual activity in other tissues over time. The distribution and clearance profiles are similar, though some differences in activity accumulation were observed at 24 h post-injection (Figure b) in several organs including the liver and bone; for [²²⁵Ac]­[Ac­(py₂‑macrodipa)]⁺ liver uptake (1.0 ± 0.1%IA/g) was higher than [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ (0.2 ± 0.1 %IA/g), while bone signal (tibia) was lower (0.1 ± 0.1 %IA/g) than that of [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ (0.5 ± 0.04 %IA/g). Comparison of the 4 and 24 h biodistribution data further indicates substantial clearance of radioactivity over time. Together, these biodistribution studies demonstrate that [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ and [²²⁵Ac]­[Ac­(py₂‑macrodipa)]⁺ are highly stable in vivo, an essential attribute for developing ²²⁵Ac/⁸⁹Zr paired α/positron radiotheranostics.

12.

Biodistribution comparison of [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ and [²²⁵Ac]­[Ac­(py₂-macrodipa)]⁺ complexes at (a) 4 h and (b) 24 h post-injection. Values for each time point are given as %IA/g; %IA/g = percent injected activity per g of tissue; SI = small intestine and LI = large intestine. A similar activity distribution profile was observed, with activity uptake in the intestine at 4 h, rapidly excreted at 24 h p.i., and negligible activity in the liver, kidneys, and bone was observed at 24 h p.i. Note, the %IA/g scale was reduced from ∼30 at 4 h to ∼1.5 at 24 h, reflecting extensive excretion and marked reduction in whole-body radioactivity by the later time point.

Bifunctional Chelator py₂-macrodipa-sq

For antibody conjugation, nonsymmetric macrocycles with a single, well-defined functionalization site are preferred. However, the synthesis of nonsymmetric 4,13-diaza-18-crown-6 ligands is challenging because N-alkylation typically yields mixtures of mono- and disubstituted products. Here, we adapted the reported synthesis of H₂MacropaSqOEt by introducing an alkyne substituent at the 4-position of the picolinate arm to afford compound 6-((10-((6-carboxypyridin-2-yl)­methyl)-3,13-dioxa-6,10-diaza-1,8­(2,6)-dipyridinacyclotetradecaphane-6-yl)­methyl)-4-(prop-2-yn-1-yloxy)­picolinic acid (5). Subsequent copper­(I)-catalyzed azide–alkyne cycloaddition with an azide-functionalized oligo­(ethylene glycol) squaramide ester (3-((2-(2-(2-azidoethoxy)­ethoxy)­ethyl)­amino)-4-ethoxycyclobut-3-ene-1,2-dione, 6) (Scheme ) yielded the bifunctional chelator py2-macrodipa-sq (6-((10-((6-carboxypyridin-2-yl)­methyl)-3,13-dioxa-6,10-diaza-1,8­(2,6)-dipyridinacyclotetradecaphane-6-yl)­methyl)-4-((1-(2-(2-(2-((2-ethoxy-3,4-dioxocyclobut-1-en-1-yl)­amino)­ethoxy)­ethoxy)­ethyl)-1H-1,2,3-triazol-4-yl)­methoxy)­picolinic acid (7).

1. Synthesis of Bifunctional Chelator py₂-Macrodipa-sq.

Antibody Conjugation, Radiolabeling with Zr-89 and Ac-225 and In Vitro Stability

To evaluate py₂-macrodipa-sq as a bifunctional chelator, it was conjugated to the anti-PD-1 antibody Opdivo in sodium bicarbonate buffer (pH 9.0) overnight at 37 °C using 20 equiv of the chelator. The conjugate was purified from excess unreacted chelator with a spin desalting column which is based on size exclusion chromatography. The conjugate was successfully radiolabeled with ⁸⁹Zr and ²²⁵Ac isotopes, affording radiochemical yields of 80% and 90%, respectively, and >99% radiochemical purity (RCP) after purification with size-exclusion chromatography (Figure a and b). The ²²⁵Ac labeled conjugate demonstrated high in vitro stability in human serum, with >96% remaining intact over 120 h (Figures c, S22, Table S7). The ⁸⁹Zr analogue maintained ∼80% radiochemical stability in human serum over 48 h (Figures c, S23, Table S6).

13.

Thin-layer chromatograms of (a) free [⁸⁹Zr]­Zr⁴⁺ and [⁸⁹Zr]­[Zr­(opdivo-py₂-macrodipa-sq)] and (b) [²²⁵Ac]­Ac³⁺ and [²²⁵Ac]­[Ac­(opdivo-py₂-macrodipa-sq)]⁺, iTLC-SG eluted with 50 mM EDTA, pH 5.5; free metals migrated with the solvent front (R _f = 1), whereas the complex species remained at the baseline (R _f ∼ 0). (c) Human serum stability over 48 h and 120 h at 37 °C for ⁸⁹Zr- and ²²⁵Ac-conjugates respectively; as assessed by ITLC, with >96% (for ²²⁵Ac-conjugate) and ∼80% (⁸⁹Zr-conjugate) radiochemical stability maintained.

PET and Biodistribution Study

The in vivo stability of [⁸⁹Zr]­[Zr­(opdivo-py₂-macrodipa-sq)] was evaluated by PET imaging in naive mice (Figure ; Figure S25). Imaging at 1 h p.i. demonstrated predominantly cardiac localization, with moderate liver uptake observed at 72 h. The radioconjugate demonstrated efficient circulation within the blood pool immediately after administration and throughout the imaging period. Importantly, negligible bone uptake was observed at all time points, indicating no in vivo release of free [⁸⁹Zr]­Zr⁴⁺.

14.

In vivo positron emission tomography (PET) images of [⁸⁹Zr]­[Zr­(opdivo-py₂-macrodipa-sq)] at 1 and 72 h following retro-orbital injection in naive ND4 mice, coronal view of the mice. No activity accumulation in the bone was observed over time, indicating in vivo stability of the radioconjugate.

To further support the PET findings, ex vivo biodistribution studies were conducted over a 72 h period. Corroborating imaging findings, we report high early blood pool and organ activity with minimal and nonincreasing bone uptake (Figure , Table S9), confirming high in vivo stability.

15.

Organ distribution of (a) [⁸⁹Zr]­[Zr­(opdivo-py₂-macrodipa-sq)] and (b) [²²⁵Ac]­[Ac­(opdivo-py₂-macrodipa-sq)] following retro-orbital injection in naive mice. Values for each time point are given as %IA/g; %IA/g = percent injected activity per g of tissue; SI = small intestine. Both conjugates exhibited similar distribution profiles, efficient systemic clearance of activity over time, and consistently low bone uptake, indicating in vivo stability.

The radioconjugate [²²⁵Ac]­[Ac­(opdivo-py₂-macrodipa-sq)] exhibited elevated initial activity in the blood pool and major organs. Although early organ uptake was relatively high, the radiotracer was efficiently cleared from most tissues by days 4 and 6 post administration (Figure , Table S10). Notably, bone uptake remained low throughout the study, indicating the high in vivo stability of the conjugate.

Discussion

Radiopharmaceutical therapy is a rapidly growing field that offers targeted treatment options for cancer. Theranostic approaches enable personalized treatment and monitor therapeutic response in real time by linking diagnostic imaging with the therapeutic radionuclide distribution. Although this modality holds great promise, key concerns include off-target toxicity and questions regarding the quantitative accuracy of imaging techniques.

Ac-225 is an attractive alpha-emitter for targeted alpha therapy (TAT) due to its 9.92-day half-life and the emission of four high-energy α-particles during its decay chain. The half-life is particularly well suited to antibody-based radiopharmaceuticals with slower pharmacokinetics while remaining compatible with peptide-based targeting vectors, including PSMA- and somatostatin receptor-targeted agents. , DOTA is the conventional chelator for the largest trivalent metal ion, ²²⁵Ac. However, DOTA’s inverse thermodynamic stability with metal ion size precludes stable chelation of small ionic radii PET emitters with sufficient half-life for later-phase imaging, which has been a challenge. The macrocyclic chelator macropa has been developed for coordinating large ionic radius metal ions such as [²²⁵Ac]­Ac³⁺; however, the cavity size and donor arrangement of this 18-membered ring are poorly matched to high charge-density, small ionic radii radionuclides.

For PET imaging, ⁸⁹Zr is preferred for late-phase imaging of antibodies and proteins due to its favorable emission properties, availability, and long half-life. We evaluated the coordination efficiency, radiolabeling, and stability of py₂-macrodipa with [⁸⁹Zr]­Zr⁴⁺ and [²²⁵Ac]­Ac³⁺. Both radionuclides exhibited efficient complexation and high radiochemical yields. Quantitative [⁸⁹Zr]­Zr⁴⁺ labeling was performed at 90 °C (oxalic acid precursor), whereas [²²⁵Ac]­Ac³⁺ binding proceeded under mild, room-temperature conditions. Pandya et al. reported exceptional thermodynamic stability of ⁸⁹Zr-DOTA relative to DFO. However, quantitative labeling required high temperatures, limiting applicability to protein conjugates. In contrast, py₂-macrodipa achieved 40% RCY with [⁸⁹Zr]­Zr⁴⁺ at 37 °C, whereas DOTA showed no detectable ⁸⁹Zr incorporation under identical conditions, recapitulating prior findings.

Recent studies have demonstrated that effective DOTA labeling requires the ⁸⁹Zr precursor to be in the form of [⁸⁹Zr]­ZrCl₄ rather than [⁸⁹Zr]­Zr⁴⁺ in oxalic acid. ,, Quantitative radiolabeling of 9-membered NOTA and 12-membered PCTA chelators using [⁸⁹Zr]­ZrCl₄ at 37 °C was achieved, while 13-membered TRITA required microwave heating and 14-membered TETA showed poor RCY even with such assistance. Collectively, these findings indicate that increasing the coordination environment size of the ligands weakens the metal–ligand bonding interaction.

The oxalate anion forms a stable [⁸⁹Zr]­Zr-complex in aqueous media, thermodynamically preventing trans-chelation by DOTA. In contrast, [⁸⁹Zr]­ZrCl₄ rapidly hydrolyzes to form hydroxo- and oxo-bridged compounds that are susceptible to trans-chelation by tetraazamacrocycle ligand DOTA. Based on these properties, we investigated labeling the dual-size selective chelator with [⁸⁹Zr]­ZrCl₄. Surprisingly, both macropa and py₂-macrodipa failed to incorporate [⁸⁹Zr]­Zr⁴⁺ at 25 °C, nor at elevated temperatures. Notably, labeling of [⁸⁹Zr]­Zr⁴⁺ with DFO and DOTA was achieved at 25 and 90 °C, respectively, with [⁸⁹Zr]­ZrCl₄ as previously reported. Trace metal contamination may be a factor in our results, and the distinctive selectivity of py₂-macrodipa ligand for the oxalic form of the radioisotope will be investigated further for future ligand development.

The [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ complex showed a high degree of in vitro stability in human serum; greater than 95% remained intact after 48 h. This was similar to the conventional DFO complex. As expected, the kinetic inertness of [²²⁵Ac]­[Ac­(py₂-macrodipa)]⁺ in human serum was also validated. By contrast, the macrocyclic macropa chelator showed efficient in vitro stability for [²²⁵Ac]­Ac³⁺ but was not compatible with [⁸⁹Zr]­Zr⁴⁺.

We next explored the in vivo stability of the complexes using both PET imaging and acute biodistribution studies. The dual size-selective chelator py₂-macrodipa displayed robust in vivo stability for both radionuclides. Longitudinal PET acquisitions revealed alternative excretion pathways for DFO and py₂-macrodipa ligands with Zr-89. The majority of the injected activity of the py₂-macrodipa complex was observed in the intestine, which was excreted over the course of the experiment. The DFO complex showed rapid renal excretion. Notably, both complexes showed minimal bone uptake, indicating no release of free [⁸⁹Zr]­Zr⁴⁺ ion from the ligand in vivo. Despite this difference, both complexes showed low uptake in other organs, and neither complex presented substantial accumulation over time. The amount of activity in the bone for the py₂-macrodipa complex is higher than that of the DFO complex; this could be the result of the unchelated [⁸⁹Zr]­Zr⁴⁺ cation, as we injected the [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ complex with 87% purity (Figure ). Radiochemical purity should not be a concern, as the radiolabeled compound can be purified with a C18 cartridge (Figure S12). This observation can also be inferred from the significant amount of activity detected in the bone for free [⁸⁹Zr]­Zr⁴⁺ and the [⁸⁹Zr]­[Zr­(macropa)]²⁺ complex from the biodistribution study. The biodistribution results of the [²²⁵Ac]­[Ac­(py₂‑macrodipa)]⁺ complex revealed a high degree of in vivo stability. Both [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ and [²²⁵Ac]­[Ac­(py₂‑macrodipa)]⁺ complexes demonstrated broadly similar activity clearance and kinetics, with some differences in organ-specific activity accumulation noted. The 24 h biodistribution study revealed substantial clearance and a significant reduction in whole-body radioactivity for both complexes. The combination of in vitro and in vivo data addresses the stability of the py₂-macrodipa complexes. This motivates further development of a new bifunctional chelator py₂-macrodipa-sq by incorporating a squaramide ethyl ester functional group with a PEG linker for antibody conjugation. Compared with isothiocyanates, the squaramide functional group is more stable and reacts selectively with lysine residues on antibodies to form stable amide bonds. The opdivo-py₂-macrodipa-sq quantitatively incorporates both ⁸⁹Zr and ²²⁵Ac isotopes using mild reaction conditions. The resulting radioconjugates demonstrated high kinetic inertness when challenged in human serum.

Administration of [⁸⁹Zr]­[Zr­(opdivo-py₂-macrodipa-sq)] in naive mice revealed a high degree of in vivo stability, as assessed by PET imaging and biodistribution studies. PET images acquired at early time points (1 h post injection) showed high heart uptake and moderate liver uptake, with circulation persisting up to 72 h and minimal skeletal accumulation. Biodistribution data corroborated these findings, with negligible bone uptake, further supporting the in vivo stability of the complex. The radioconjugate exhibited dual clearance pathways, undergoing both hepatobiliary and renal excretion.

Although the initial uptake of [²²⁵Ac]­[Ac­(opdivo-py₂-macrodipa-sq)] in the blood pool and major organs was relatively high compared to the ⁸⁹Zr-conjugate, both radioconjugates exhibited minimal bone accumulation. The biodistribution study showed accumulation of activity in all blood-rich organs including the liver and kidneys, indicating the involvement of both hepatobiliary and renal excretion pathways. After validating the in vivo stability of the radioconjugates, the pharmacokinetics of ²²⁵Ac- and ⁸⁹Zr-opdivo-py₂-macrodipa-sq were found to be similar (Figure ). The observed differences in activity biodistribution may result from the administration of different antibody masses (8 and 22 μg of antibody for ⁸⁹Zr and ²²⁵Ac, respectively), which can influence target saturation, tissue distribution, and clearance kinetics. Importantly, despite these early differences, both radioconjugates exhibited efficient systemic clearance over time and consistently low bone uptake, indicating neither in vivo demetalation nor radiolabel instability. In contrast to free [²²⁵Ac]­Ac³⁺ ion, which predominantly accumulates in bone and liver over time, activity associated with the ²²⁵Ac-radioconjugate was largely cleared from these tissues by days 4 and 6 post administration. Similarly, the ⁸⁹Zr-radioconjugate demonstrated no evidence of long-term accumulation in the bone, which underlines its in vivo stability.

Conclusions

In this report, we demonstrate the stable coordination of metal ions with distinct coordination characteristics, [⁸⁹Zr]­Zr⁴⁺, exhibiting high charge density and a preference for lower coordination numbers, and [²²⁵Ac]­Ac³⁺, characterized by lower charge density and a tendency toward higher coordination numbers, using the dual size-selective chelator py₂-macrodipa. The coordination chemistry of py₂-macrodipa with [⁸⁹Zr]­Zr⁴⁺ was characterized by NMR spectroscopy, indicating the formation of an asymmetric conformation. The crystal structure obtained for the [Zr­(py-macrodipa)­(OMe)]⁺ complex provides additional evidence supporting the presence of Conformation B in this class of chelators. Both [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ and [²²⁵Ac]­[Ac­(py₂-macrodipa)]⁺ complexes displayed robust stability both in vitro in human serum and in vivo in mice. PET imaging and biodistribution studies showed rapid clearance of injected activity for the [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ complex without accumulation in the organs over time. Notably, its coordination with [²²⁵Ac]­Ac³⁺ also revealed good biological behavior, rapid clearance, and absence of demetalation. In addition, both complexes exhibited similar activity distribution and excretion profiles. These results support the development of py₂-macrodipa-sq as a bifunctional chelator for antibody conjugation. The squaramide modification enables stable, selective lysine conjugation and efficient incorporation of both ⁸⁹Zr and ²²⁵Ac under mild conditions. Radioconjugates showed high kinetic inertness in serum and strong resistance to demetalation in vivo, with minimal bone accumulation and favorable clearance profiles, highlighting their potential for effective ²²⁵Ac/⁸⁹Zr theranostic applications. Evaluation of py₂-macrodipa-based radioconjugates in tumor-bearing mouse models will include side-by-side comparison of ⁸⁹Zr- and ²²⁵Ac-labeled constructs under identical conditions to assess theranostic potential, as well as the identification of improved radionuclide pairings to further enhance the theranostic potential of this chelator system.

Experimental Section

Materials and Methods

The experimental data supporting this article are provided in the SI. The X-ray crystallographic data are deposited in the Cambridge Structural Database (CCDC) as crystallographic information files (cif) under deposition number 2430371.

Preparation of Complexes for NMR Spectroscopic Analysis

NMR samples of py-macrodipa and py₂-macrodipa complexes with Zr⁴⁺ were prepared in H₂O. Ligand stock solutions (2 mM) were prepared in H₂O. Metal stock solutions (2 mM) were prepared by dissolving Zr­(acac)₄ in H₂O immediately prior to use. Complex solutions were obtained by mixing 1.0 mL of ligand stock with 1.0 mL of the corresponding metal stock to afford final concentrations of 1.0 mM ligand and 1.0 mM metal (1:1 M:L ratio). The pH of the solution was adjusted to 5.0 using 1 M NaOH, and the mixture was stirred at room temperature overnight. The resulting solutions were lyophilized to dryness. The obtained solids were subsequently redissolved in D₂O for the NMR analysis.

Single Crystal Growth of [Zr­(py-macrodipa)­(OMe)]­[Cl]­DMF

The ligand py-macrodipa·3HCl·3.5H₂O (24.6 mg, 0.033 mmol) was dissolved in methanol (MeOH, 400 μL), and triethylamine (Et₃N, 22.9 μL, 0.165 mmol) was added. A solution of Zr­(acac)₄ (17.6 mg, 0.036 mmol) in MeOH (400 μL) was added dropwise to the ligand solution. This reaction mixture was then stirred for 48 h at room temperature (RT). Afterward, the reaction mixture was transferred into a scintillation vial and air-dried overnight in the fume hood, affording a white solid (44.2 mg). Slow diffusion of diethyl ether (Et₂O) into a dimethylformamide (DMF) solution of this material afforded X-ray quality crystals of [Zr­(py-macrodipa)­(OMe)]­[Cl]­DMF.

Low-temperature X-ray diffraction data were collected on a Rigaku XtaLAB Synergy diffractometer coupled to a Rigaku HyPix detector with Cu Kα radiation (λ = 1.54184 Å) from a PhotonJet microfocus X-ray source at 100 K. The diffraction images were processed and scaled using the CrysAlisPro software. The structures were solved through intrinsic phasing using SHELXT and refined against F2 on all data by full-matrix least-squares with SHELXL, following established refinement strategies. , All non-hydrogen atoms were refined anisotropically. All hydrogen atoms bound to carbon were included in the model at geometrically calculated positions and refined using a riding model. Hydrogen atoms bound to oxygen were located in the difference Fourier synthesis and subsequently refined semi-freely with the help of distance restraints. The isotropic displacement parameters of all hydrogen atoms were fixed to 1.2 times the Ueq value of the atoms they are linked to (1.5 times for methyl groups). The X-ray crystallography data are presented in Table S11.

Synthesis of Methyl 6-(3,13-Dioxa-6,10-diaza-1,8­(2,6)-dipyridinacyclotetradecaphane-6-ylmethyl)­picolinate (3). To a solution of 3,13-dioxa-6,10-diaza-1,8­(2,6)-dipyridinacyclotetradecaphane ((1),150 mg, 0.46 mmol) in anhydrous MeCN (100 mL), Na₂CO₃ (145.2 mg, 1.37 mmol) was added. The reaction mixture was heated to reflux for 2 h and then cooled to room temperature. A solution of methyl 6-(bromomethyl)­picolinate ((2), 63.1 mg, 0.274 mmol) in anhydrous MeCN (100 mL) was added dropwise over 1 h, and the resulting mixture was heated to reflux overnight. After cooling to room temperature, the mixture was filtered, and the filtrate was concentrated under reduced pressure to afford a yellow oil. The crude product was dissolved in H₂O (2 mL, 0.1% TFA) and purified by preparative HPLC (method described above) to give the desired product as a white solid (108 mg, 49.5%). ¹H NMR (500 MHz, DMSO-d ₆) δ 7.56 (d, J = 7.3, 2.3 Hz, 4H), 7.55–7.43 (m, 2H), 7.39–7.12 (m, 3H), 4.40 (s, 4H), 4.00 (s, 4H), 3.39 (s, 4H), 3.01 (s, 2H), 2.33 (s, 2H). ESI-MS m/z: 239.6; calcd for [C₂₆H₃₁N₅O₄ + 2H]²⁺: 239.6.

Synthesis of 6-((10-((6-Carboxypyridin-2-yl)­methyl)-3,13-dioxa-6,10-diaza-1,8­(2,6)-dipyridinacyclotetradecaphane-6-yl)­methyl)-4-(prop-2-yn-1-yloxy)­picolinic Acid (5). To a solution of 3 (100 mg, 0.21 mmol) in anhydrous MeCN (10 mL) was added Na₂CO₃ (66.6 mg, 0.63 mmol), and the reaction mixture was heated to reflux for 2 h. Methyl 6-(bromomethyl)-4-(prop-2-yn-1-yloxy)­picolinate ((4), 29.8 mg, 0.10 mmol) was then added, and the resulting mixture was heated to reflux overnight. After cooling to room temperature, the mixture was filtered, and the filtrate was concentrated under reduced pressure to afford a yellow oil. The crude product was dissolved in MeOH (5 mL), and LiOH (31.9 mg, 1.33 mmol) was added. The resulting solution was stirred at room temperature overnight, then concentrated under reduced pressure to afford a yellow oil. The crude material was dissolved in H₂O (2 mL, 0.1% TFA) and purified by preparative HPLC (method described above) to give the desired product as a white solid (72.4 mg, 53%). ¹H NMR (400 MHz, MeOD) δ 8.20 (d, J = 7.7 Hz, 8H), 8.12 (m, 3H), 7.81 (d, J = 2.2 Hz, 2H), 7.74 (d, J = 7.7 Hz, 2H), 7.44 (s, 2H), 4.80 (s, 2H), 4.73 (s, 2H), 4.60 (t, J = 4.9 Hz, 4H), 3.96 (d, J = 6.2 Hz, 4H), 3.93 (s, 1H), 3.85 (d, J = 4.9 Hz, 4H), 2.50 (t, J = 6.0 Hz, 4H). ESI-MS m/z: 327.9; calcd for [C₃₅H₃₆N₆O₇+ 2H]²⁺: 327.2.

Synthesis of 6-((10-((6-Carboxypyridin-2-yl)­methyl)-3,13-dioxa-6,10-diaza-1,8­(2,6)-dipyridinacyclotetradecaphane-6-yl)­methyl)-4-((1-(2-(2-(2-((2-ethoxy-3,4-dioxocyclobut-1-en-1-yl)­amino)­ethoxy)­ethoxy)­ethyl)-1H-1,2,3-triazol-4-yl)­methoxy)­picolinic Acid (7). CuSO₄ (2.2 mg, 0.014 mmol) and sodium ascorbate (6.4 mg, 0.032 mmol) were dissolved in H₂O (5 mL) to generate the Cu­(I) species in situ. TBTA (28 mg, 0.057 mmol) was added, followed by 3-((2-(2-(2-azidoethoxy)­ethoxy)­ethyl)­amino)-4-ethoxycyclobut-3-ene-1,2-dione ((6), 18.9 mg, 0.63 mmol), and the mixture was stirred for 10 min. Compound 5 (41.1 mg, 0.63 mmol) was then added, and the reaction mixture was purged with N₂ for 15 min. Reaction progress was monitored by LC–MS, confirming complete consumption of azide species. After 1 h, Na₂EDTA (26 mg, 0.07 mmol) was added to chelate residual copper. The pale green solution was concentrated to dryness under reduced pressure and dried in vacuo. The crude material was purified by preparative HPLC (method described above); the combined fractions were lyophilized to afford an off-white powder (54.2 mg, 88%). ¹H NMR (500 MHz, MeOD) δ 7.96 (dd, J = 7.7, 1.1 Hz, 2H), 7.85 (t, J = 7.8 Hz, 2H), 7.71–7.62 (m, 3H), 7.34 (d, J = 8.4 Hz, 2H), 7.21 (dd, J = 15.0, 8.1 Hz, 2H), 6.81 (s, 1H), 4.56 (d, J = 14.4 Hz, 8H), 4.18 (t, J = 6.6 Hz, 8H), 3.90 (s, 3H), 3.81 (s, 4H), 3.75 (s, 4H), 3.63 (s, 4H), 3.07–2.96 (m, 4H), 2.81 (t, J = 5.4 Hz, 4H) ESI-MS m/z: 478.2; calcd for [C₄₈H₅₄N₁₀O₁₂ + 2H]²⁺: 478.8.

Radiolabeling of DFO, macropa, and py₂-macrodipa with Zr-89 and Ac-225

DFO, macropa, and py₂-macrodipa were dissolved in Chelex water to prepare 15 mM stock solutions. From these stock solutions, serial dilutions were performed to prepare ligand solutions at varying concentrations (10^–3–10^–12 M). For the labeling studies, the pH of the [⁸⁹Zr]­Zr⁴⁺ in 1 M oxalic acid was adjusted to 6.0–6.5 with 0.1 M sodium carbonate (Na₂CO₃). A 20 μL aliquot of each ligand solution was further diluted with 50 μL of 0.5 M HEPES buffer (pH 6.5). Subsequently, 4 μL (0.2 MBq) of balanced [⁸⁹Zr]­Zr⁴⁺ was added. The mixture was vortexed and incubated at either 37 or 90 °C for 30 min. Radiochemical yield was evaluated using iTLC-SG with 50 mM EDTA, pH 5.5, running buffer. The radiolabeled complex remained at the origin line (R _f = 0), while free [⁸⁹Zr]­Zr⁴⁺ migrated with the solvent front (R _f ≈ 1). After incubation, approximately 2 μL aliquots were spotted on iTLC-SG plates and developed. Radiochemical yields (RCY) were measured using a gas proportional scanner (Bioscan AR-2000).

For the radiolabeling of [⁸⁹Zr]­Zr⁴⁺ in 1 M HCl solution, DFO, DOTA, and py₂-macrodipa were initially dissolved in 0.1 M HEPES buffer (pH 6.5) to create 100 mM stock solutions. From these stock solutions, serial dilutions were performed to prepare ligand solutions at varying concentrations. A 5 μL aliquot of each ligand solution was further diluted with 43 μL of 0.1 M HEPES buffer (pH 6.5), and 2 μL of [⁸⁹Zr]­ZrCl₄ solution in 1 M HCl (1.85 MBq) was added to the mixture. This combination was incubated at either 25 or 90 °C for 1 h. Radiochemical yields were determined using the previously described method for labeling [⁸⁹Zr]­Zr⁴⁺ in oxalic acid.

For [²²⁵Ac]­Ac³⁺ labeling with macropa and py₂-macrodipa, dry [²²⁵Ac]­Ac­(NO₃)₃ (2.96 MBq) source received was diluted with 0.05 M HCl (200 μL). The pH of [²²⁵Ac]­Ac³⁺ solution (20 μL, 0.2 MBq) was adjusted to 6.0 with 0.5 M ammonium acetate (NH₄OAc) buffer (200 μL, pH 6.0). Afterward, the macropa ligand (100 μL, 15 mM) was added to it, and the mixture was incubated at room temperature for 15 min. A similar radiolabeling protocol was followed for py₂-macrodipa ligand. The radiochemical yield (>96%) was determined with radio-TLC.

Antibody Conjugation

Conjugation of the py₂-macrodipa-sq to the Opdivo antibody was performed using a chelator to antibody ratio of 20:1. Before conjugation, the buffer was exchanged by a spin desalting column (Zeba, 40 kDa, 0.5 mL, Thermo Scientific) to sodium bicarbonate buffer (0.5 M, pH 8.5–9.0). The mixture of Opdivo and the chelator was incubated in 0.5 M sodium bicarbonate buffer (pH 8.5–9.0) at 37 °C overnight. Before radiolabeling with Ac-225 and Zr-89, the buffer was exchanged into 0.5 M ammonium acetate (NH₄OAc, pH 6.5) and 0.5 M 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES, pH 7.0), respectively, with a Zeba column (40 kDa, 0.5 mL). The purity of the antibody conjugate was measured by FPLC using an Enrich SEC 650 (10 × 300) column.

Radiolabeling with ⁸⁹Zr and ²²⁵Ac Isotopes

For the labeling studies, the pH of the [⁸⁹Zr]­Zr⁴⁺ in 1 M oxalic acid was adjusted to 7.0 with 0.5 M sodium carbonate (Na₂CO₃). Neutralized [⁸⁹Zr]­Zr⁴⁺ (33 MBq) was reacted with the antibody conjugate (200 μg) in 0.5 M HEPES buffer (pH 7) at 37 °C for 2 h. For the Ac-225 labeling study, the dry [²²⁵Ac]­Ac­(NO₃)₃ source received was diluted with 0.05 M HCl (15 μL). The pH of [²²⁵Ac]­Ac³⁺ solution (15 μL, 1.04 MBq) was adjusted to 6.0 with 0.5 M ammonium acetate (NH₄OAc) buffer (200 μL, pH 6.0). Afterward, the conjugate (100 μg) in 0.5 M ammonium acetate buffer was added to it, and the mixture was incubated at room temperature for 1 h. Ethylenediaminetetraacetic acid (EDTA, 10 μL, 50 mM, pH 6.0) was added to each reaction mixture, then the radioconjugate was purified in saline using a preconditioned spin desalting column (40 kDa, 2 mL). The radiochemical yield and purity (RCY and RCP) were verified (Figure ) by radio-TLC (Bioscan AR-2000). Free metals migrated with the solvent front (50 mM EDTA mobile phase, R _f = 1), whereas the labeled antibodies remained at the baseline (R _f ∼ 0).

Human Serum Stability

The complexes were prepared with the radiolabeling protocol as described above for the serum stability study. A starting volume of 20 μL (25.16 MBq) of balanced [⁸⁹Zr]­Zr⁴⁺ (pH 6.5–7.0 with 0.1 M Na₂CO₃) was diluted with 150 μL of 0.5 M HEPES buffer (pH 6.5), and 100 μL of 15 mM of each ligand (DFO, macropa, and py₂-macrodipa) was added. The reaction mixtures were incubated at either 37 °C (for the DFO ligand) or 90 °C for 30 min. The radiochemical yield was determined to be 99, 87, and 86%, respectively, for ⁸⁹Zr-complexes with DFO, py₂-macrodipa, and macropa. Both [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ and [⁸⁹Zr]­[Zr­(macropa)]²⁺ compounds were further purified using a C18 cartridge with a methanol/acetonitrile (and 0.1% TFA) solvent mixture. The solvent was evaporated to dryness to give a final activity of 3.7–4.4 MBq; the activity was further diluted with 100 μL of normal saline.

To assess stability with a protein challenge, 20 μL of each radiolabeled compound (0.74–1.85 MBq) was incubated in human serum (purchased from Sigma, 400 μL) at 37 °C, under gentle shaking for 48 h. At predetermined time points, the stability was measured by radio-TLC (Figure S17), and the %intact complex values are shown in Table S4. [²²⁵Ac]­[Ac­(py₂-macrodipa)]⁺ and [²²⁵Ac]­[Ac­(macropa)]⁺ complexes were synthesized as described above. To assess the stability of the actinium complexes, we used 10 μL of each compound (0.01 MBq) incubated in human serum (400 μL) at 37 °C. Serum stability was monitored (at 5 min, 1, 4, 24, 48, 72, and 144 h) with iTLC-SG using 50 mM EDTA (pH 5.5) as the mobile phase (Figures S16, S18 and Table S5).

The [²²⁵Ac]­[Ac­(opdivo-py₂-macrodipa-sq)] radioconjugate was prepared with the radiolabeling protocol described above for the serum stability study. The ²²⁵Ac-radioconstruct (15 μg/0.07 MBq) was incubated in human serum (400 μL) at 37 °C under gentle shaking for 5 days. The stability was measured by radio-TLC (Figure S22), and the %-intact complex values are shown in Table S7.

[⁸⁹Zr]­[Zr­(opdivo-py₂-macrodipa-sq)] construct was synthesized as described above (radiolabeling section) and challenged (∼30 μg/0.93 MBq) in human serum (400 μL). The radioconjugate with an RCP > 85% was incubated at 37 °C, under gentle shaking for 48 h, and the kinetic inertness was monitored by radio-TLC (Figure S23, Table S6).

Biodistribution Study of ⁸⁹Zr- And ²²⁵Ac-Labeled Complexes

For the preparation of [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺, 30 μL of (18.5 MBq) balanced [⁸⁹Zr]­Zr⁴⁺ in 0.1 M oxalic acid (pH was adjusted to 6.5–7.0 with 0.1 M Na₂CO₃) solution was added to HEPES buffer (100 μL, 0.5 M, pH 6.5) to further adjust the pH to 6.5, and then 100 μL of 15 mM py₂-macrodipa chelator was added to it. The mixture was vortexed and incubated at 90 °C for 30 min. Radiochemical yield was evaluated using iTLC-SG with a 50 mM EDTA pH 5.5 mobile phase. Radio-TLC was developed using a gas proportional scanner (Bioscan AR-2000) with a 1 min acquisition time. For subsequent in vivo evaluation, the labeled material was diluted with 1.0 mL of normal saline (0.9% NaCl in deionized H₂O; Hospira), and activities were prepared in 30G syringes for administration. [⁸⁹Zr]­[Zr­(macropa)]²⁺ and [⁸⁹Zr]­[Zr­(DFO)] were also prepared following the labeling conditions described above in the radiolabeling section. Naive Swiss Webster female mice (4 months; Charles River) were injected via the retro-orbital with 100 μL, 2.04–4.3 MBq of each respective complex ([⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺, [⁸⁹Zr]­[Zr­(macropa)]²⁺, [⁸⁹Zr]­[Zr­(DFO)], and free [⁸⁹Zr]­Zr⁴⁺). Similarly, a biodistribution study was also performed for [²²⁵Ac]­[Ac­(py_2‑macrodipa)]⁺ and [²²⁵Ac]­[Ac­(macropa)]⁺. Each mouse was injected with an activity of 0.013–0.02 MBq (100 μL).

The organ recovery was performed at three predetermined time points: 1, 4, and 24 h post-injection. Animals were euthanized by CO₂ asphyxiation prior to tissue dissection. A blood sample was acquired, and the organs of interest were collected (heart, lungs, liver, kidneys, spleen, stomach, small intestine, large intestine, skin, muscle, fat, tibia, calvaria, brain, and bone marrow), rinsed of excess blood, weighed, and measured by γ counting (Wizard2, PerkinElmer). Data were corrected for background and decay, and the activity of each tissue was normalized to an injection standard. The organ uptake values were calculated as the percent of injected activity per gram of tissue (%IA g^–1).

PET Imaging of ⁸⁹Zr-Labeled Complexes

PET imaging of naïve Swiss Webster females was performed using free [⁸⁹Zr]­Zr⁴⁺ and [⁸⁹Zr]­[Zr­(DFO)], [⁸⁹Zr]­[Zr­(macropa)]²⁺, and [⁸⁹Zr]­[Zr­(py₂-macrodipa)]²⁺ complexes at 1, 4, and 24 h post-injection (R4 microPET, Siemens). Tracer administration was performed with the animal under anesthesia by retro-orbital injection. The imaged mouse was centered in the field of view and maintained under 1–2% isoflurane anesthesia during PET imaging. The calibration factor of the PET scanner was determined with a mouse-sized phantom composed of a cylinder uniformly filled with an aqueous solution of ¹⁸F with a known activity concentration. Acquisitions were recorded using an energy window of 350–700 keV and a coincidence-timing window of 6 ns. PET image data were corrected for detector nonuniformity, dead time, random coincidences, and physical decay, and images were reconstructed by an iterative 3D maximum a priori (MAP) algorithm.

The acquired PET images were analyzed by using ASIPro software (Siemens). Volume of interest (VOI) analysis of the acquired images was performed using ASIPro software, and the observed value (%IA/mL) represents the mean radiotracer accumulation in the organs. The sequential radioactivity measurements (%IA/mL) were plotted over time postadministration.

PET Imaging and Biodistribution Study of ⁸⁹Zr and ²²⁵Ac-Antibody Constructs

PET imaging (R4 microPET, Siemens) of naive ND4 mice (Inotiv), 5–6 months old was performed using [⁸⁹Zr]­[Zr­(opdivo-py₂-macrodipa-sq)] construct (100 μL, 1.9–2.2 MBq) at 1 and 72 h after administration of activity. VOIs were defined, and %IA/mL was determined (ASIPro).

Biodistribution study with naive ND4 was performed with injecting intravenously (retro-orbital sinus) both radioconjugates. The animals received 0.02 MBq and 1.5–1.9 MBq of Ac-225 and Zr-89 labeled antibody, respectively. Each mouse was injected with 8 and 22 μg of the ⁸⁹Zr- and ²²⁵Ac-labeled constructs. Mice were euthanized via CO₂ asphyxiation, after which organs were collected and quantified using γ-counting (Wizard2, PerkinElmer) at 1 and 72 h for the Zr-89-construct and 0-, 1-, 4-, and 6-days post-injection ²²⁵Ac analogue. The organ uptake values were calculated as percent of injected activity per gram of tissue (%IA g^–1).

Statistical Analysis

Data calculated using Microsoft Excel are expressed as mean ± SD. Student’s unpaired t test (GraphPad Prism 9) was used to determine statistical significance. Differences with p values < 0.05 were considered to be statistically significant.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.inorgchem.6c02206.

• Materials and methods, ¹H NMR spectra, ESI HRMS spectrum, HPLC chromatogram, ESI-MS spectrum, thin-layer chromatogram, radiochemical conversion, radiochemical yields, representative radio-TLCs, projection images, PET imaging, biodistribution comparison, crystal data and structure refinement (PDF)

Experiments were performed by M.C., K.K.L., A.H., T.K., A.F., A.H. M.C., K.K.L., J.J.W., and D.L.J.T. wrote the manuscript, and all authors edited and contributed to the manuscript.

All radioactive material handling and animal experiments were conducted in compliance with institutional regulations and approved by Environmental Health and Safety Radioactive Materials at Washington University in St. Louis under protocol D16–00245 and Institutional Animal Care and Use Committee protocol #22–0023.

The authors declare no competing financial interest.