Py-macrodipa: A Janus Chelator Capable of Binding Medicinally Relevant Rare-Earth Radiometals of Disparate Sizes

Abstract

Nuclear medicine leverages different types of radiometals for disease diagnosis and treatment, but these applications usually require them to be stably chelated. Given the often-disparate chemical properties of these radionuclides, it is challenging to find a single chelator that binds all of them effectively. Towards addressing this problem, we recently reported a macrocyclic chelator macrodipa with an unprecedented “dual-size-selectivity” pattern for lanthanide (Ln³⁺) ions, characterized by its high affinity for both the large and the small Ln³⁺ (J. Am. Chem. Soc. 2020, 142, 13500). Here, we describe a second-generation “macrodipa-type” ligand, py-macrodipa. Its coordination chemistry with Ln³⁺ was thoroughly investigated experimentally and computationally. These studies reveal that the Ln³⁺−py-macrodipa complexes exhibit enhanced thermodynamic and kinetic stabilities compared to Ln³⁺−macrodipa, while retaining the unusual dual-size selectivity. Nuclear medicine applications of py-macrodipa for chelating radiometals with disparate chemical properties were assessed using the therapeutic ¹³⁵La³⁺ and diagnostic ⁴⁴Sc³⁺, radiometals representing the two size extremes within the Ln³⁺ series. Radiolabeling and stability studies demonstrate that the rapidly formed complexes of these radionuclides with py-macrodipa are highly stable in human serum. Thus, in contrast to gold standard chelators like DOTA and macropa, py-macrodipa can be harnessed for the simultaneous, efficient binding of radiometals with disparate ionic radii like La³⁺ and Sc³⁺, signifying a substantial achievement in nuclear medicine. This concept could enable the facile incorporation of a breadth of medicinally relevant radiometals into chemically identical radiopharmaceutical agents. The fundamental coordination chemistry learned from py-macrodipa provides valuable insight for future chelator developments.

Graphical Abstract

INTRODUCTION

Nuclear medicine is a rapidly growing and critically important branch of radiology that uses radionuclides to diagnose and treat diseases.^1–4 Radionuclides that are suitable for different applications within nuclear medicine are chosen based on their radioactive emission properties and physical half-lives. Based on these criteria, nearly the entire periodic table is represented by elements with potential relevance in nuclear medicine.^5–9 As such, the chemical properties of these radionuclides can vary widely, requiring different strategies for making them into useful radiopharmaceutical agents. For metallic radionuclides, their incorporation into useful therapeutic and diagnostic moieties requires the use of a chelator. These chelators need to be designed and optimized to rapidly and efficiently bind to the desired radiometal and form stable coordination compounds that remain intact in vivo.¹⁰ The chelators that have been used in clinically approved metal-based radiopharmaceutical agents are all derivatives of two structures, 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA, Chart 1)¹¹ and diethylenetriaminepentaacetic acid (DTPA, Chart 1). Although both chelators are highly effective for many smaller ions, like Lu³⁺ and In³⁺, they are insufficient for use with larger ions. As such, new chelators are needed for medicinally valuable radionuclides with different chemical properties.

Chart 1.

Structures of Ligands Discussed in This Work.

A key challenge in this area has been to design ligands that are effective for very large metal ions. Due to their more diffuse charge density and steric requirements, chelating large metal ions are typically inefficient with the conventional chelators DOTA and DTPA, thus driving a need for more suitable analogues. In this context, the ligand macropa or bp18c6 (Scheme 1), which possesses an unusual selectivity for larger lanthanide (Ln³⁺) ions over smaller ones (Table 1),¹² has arisen as a promising candidate. This chelator has shown particular promise for its ability to chelate and stably retain large radiometal ions including ^132/135La³⁺ and ²²⁵Ac³⁺.^13–15 A potential shortcoming of macropa for its broader use in nuclear medicine, however, is its poor affinity and stability in binding smaller ions, a feature that renders it incompatible with medicinally relevant radiometals of this type. Consequently, chelators that are capable of stably holding radionuclides of disparate sizes would be of significant value for use in nuclear medicine.

Scheme 1.

Depiction of the Conformational Toggle Present in “Macrodipa-Type” Ligands when Chelating Ln³⁺ Ions with Different Sizes. Reproduced from J. Am. Chem. Soc. 2020, 142, 13500. Copyright 2020 American Chemical Society.

Table 1.

Stability Constants of the Ln³⁺ Complexes Formed with py-macrodipa, macrodipa, and macropa.

Ln3+	py-macrodipaa	macrodipab	macropac
	log KLnL	log KLnL	log KLnL
La3+	14.31(6)	12.19	14.99
Ce3+	14.65(8)	12.50	15.11
Pr3+	14.81(6)	12.41	14.70
Nd3+	14.51(7)	12.25	14.36
Sm3+	13.66(4)	11.52	13.80
Eu3+	13.29(6)	10.93	13.01
Gd3+	12.63(5)	10.23	13.02
Tb3+	11.95(3)	9.68	11.79
Dy3+	11.47(4)	9.36	11.72
Ho3+	10.69(1)	9.36	10.59
Er3+	10.60(3)	9.71	10.10
Tm3+	10.92(2)	10.13	9.59
Yb3+	11.31(4)	10.48	8.89
Lu3+	11.54(2)	10.64	8.25
Sc3+	15.83(2)	14.37(7)

^a0.1 M KCl, this work. The values in the parentheses are one standard deviation of the last significant figure.

^b0.1 M KCl. For La³⁺−Lu³⁺, ref 16; for Sc³⁺, this work.

^c0.1 M KCl, ref 12.

We recently reported a macrocyclic chelator macrodipa (Chart 1) that shows an unprecedented selectivity pattern for the Ln³⁺ ions by exhibiting good affinities for both the largest and smallest members of this series.¹⁶ This dual-size-selectivity pattern arises from a significant conformational toggle of this chelator that enables it to accommodate Ln³⁺ ions of different sizes, as depicted in Scheme 1. For large Ln³⁺, a 10-coordinate, nearly C₂-symmetric complex is formed (Conformation A), whereas for small Ln³⁺ an 8-coordinate asymmetric complex arises (Conformation B). We believe that this concept of conformational switching upon coordinating metal ions with different sizes potentially prompts a new class of chelators for nuclear medicine that can accommodate a wide range of radionuclides with disparate chemical properties.

Despite this promising and potentially valuable selectivity pattern of macrodipa, its coordination compounds with Ln³⁺ are kinetically labile (vide infra), a property limiting its use in nuclear medicine. To overcome this shortcoming, we considered different design strategies to enhance the overall thermodynamic and kinetic stability of macrodipa, while maintaining its unusual selectivity pattern. To achieve this goal, we targeted the ligand py-macrodipa (Scheme 1), an “upgraded macrodipa” with one of the ethereal oxygen donors replaced by a pyridyl nitrogen donor. We anticipated that the pyridyl nitrogen donor would enhance both the thermodynamic and kinetic stability of macrodipa. The pyridyl nitrogen donor is a stronger donor for coordinating Ln³⁺ compared to the ethereal oxygen. For instance, in comparing the ligands Oxyaapa and pypa, which differ only by the substitution of an ethereal oxygen donor for a pyridyl donor (Chart 1), the latter forms thermodynamically more stable complexes with Ln³⁺.^17,18 Furthermore, the planar pyridyl unit is likely to rigidify the metal complex and reinforce its kinetic stability. The incorporation of rigidifying groups into ligand backbones is an established means of enhancing metal complex kinetic inertness,¹⁹ as exemplified by comparing the highly efficacious chelator CHX-A”-DTPA, which contains a rigid trans-cyclohexyl ring, to its more flexible analogue DTPA.¹⁰

Herein, we report the synthesis and characterization of the “upgraded macrodipa” ligand, py-macrodipa. Its coordination chemistry with all 14 non-radioactive Ln³⁺ ions and Sc³⁺, a class of metals ions that have similar chemical properties but different ionic radii,^20–22 was investigated, enabling us to understand how py-macrodipa interacts with ions of different sizes.²³ The candidacy of py-macrodipa for nuclear medicine applications was also evaluated in this work, demonstrated via radiolabeling studies with ¹³⁵La and ⁴⁴Sc radionuclides, radiometals with significantly different ionic radii that attain the large and small size extremes within the Ln³⁺ series. Through these experiments, we show that py-macrodipa is capable of chelating large ions like La³⁺ and small ions like Sc³⁺ with both satisfactory thermodynamic and kinetics stabilities, marking a significant improvement over macrodipa, our first-generation “macrodipa-type” ligand. This discovery indicates that py-macrodipa has great potential for use in chelating radiometals with significant size differences, enabling it to be employed with size-mismatched theragnostic pairs.

RESULTS AND DISCUSSION

Ligand Syntheses.

The synthesis of py-macrodipa (Scheme 2) involves the assembly of the previously reported macrocyclic core through a reductive animation²⁴ and the subsequent installation of the two picolinate pendent arms, with an overall yield of 23% over three steps. The identity and purity of the intermediates and final product were ascertained by NMR spectroscopy, mass spectrometry, and analytical HPLC (Figures S1−S9). The macrodipa ligand was obtained via published procedures.^16,25,26

Scheme 2.

Synthetic Route to py-macrodipa.

Ln³⁺ Complex Stability Constants.

Metal complex stability constants provide a quantitative measure of the thermodynamic affinity of ligands for metal ions. The magnitudes of these stability constants for various ligands are valuable for assessing their use in different metal chelation applications.^27,28 In order to study the influence of the newly-introduced pyridyl group in the macrocycle of py-macrodipa, we determined the protonation constants (K_i, Table S1) of py-macrodipa, as well as the stability constants (K_LnL, Table 1) of its Ln³⁺ complexes, via either potentiometric titrations or UV−Vis spectrophotometric titrations.^29–31 These values for macrodipa, which we previously reported¹⁶ except for the newly measured stability constant for Sc³⁺, are also compared in Tables 1 and S1. These quantities are defined in eqs 1–2, where the concentration terms represent those at chemical equilibrium, and L signifies the uncomplexed ligand in its fully deprotonated state.

$$K_i=\left[{\text{H}}_{\mathrm{i}}\text{L}\right]/\left[{\text{H}}^{+}\right]\left[{\text{H}}_{\text{i-1}}\text{L}\right]$$ (1)

$$K_{\text{LnL}}=\left[\text{LnL}\right]/\left[{\text{Ln}}^{3+}\right]\left[\text{L}\right]$$ (2)

Figure 1 plots log K_LnL values against the Ln³⁺ 6-coordinate ionic radius²² for py-macrodipa and macrodipa. Similar to macrodipa, py-macrodipa shows a dual-size selectivity across the Ln³⁺ series, as reflected by its greater affinity for the large and small ions over the intermediately sized ions. This result demonstrates that the incorporation of the pyridine within the macrocycle does not alter the unique size-selectivity profile and further suggests that py-macrodipa is similarly undergoing a conformational toggle upon traversing the Ln³⁺ series. Remarkably, py-macrodipa shows an overall enhancement in its thermodynamic affinity for the entirety of the Ln³⁺ series compared to macrodipa. For the large Ln³⁺ ions, py-macrodipa exhibits an approximate 2-log-unit increase in K_LnL over macrodipa, whereas a 1-log-unit enhancement in stability is observed for the small Ln³⁺.

Figure 1.

Stability constants of Ln³⁺ complexes formed with py-macrodipa and macrodipa plotted versus ionic radii.

The presence of the pyridine donor in py-macrodipa thus improves the overall Ln³⁺-binding affinity, while maintaining the novel dual-size-selectivity pattern. Although the K_LnL values of py-macrodipa are several orders of magnitude smaller than those of the conventional chelating agent DOTA,³² for the large Ln³⁺ ions they compare favorably to those of macropa (Figure 1), which has been successfully applied in nuclear medicine applications for radiometals of this type.^13,14 A key advantage of py-macrodipa over macropa, however, is that it retains high affinity for the small Ln³⁺ ions for which macropa is an ineffective chelator.

Complex Structures—NMR Spectroscopy.

To investigate the inferred switch between Conformations A and B in py-macrodipa across the Ln³⁺ series, we used NMR spectroscopy to characterize the solution structures of its complexes with the diamagnetic La³⁺, Lu³⁺, Sc³⁺, and Y³⁺ ions, with ionic radii spanning from 103.2 pm to 74.5 pm. The 90-pm ionic radius²² of Y³⁺ aligns well with the minimum affinity among all Ln³⁺ complexes (Figure 1). Thus, an investigation of the coordination chemistry of Y³⁺−py-macrodipa provides insight on the conformational change that is likely occurring as this minimum of stability is traversed.

The ¹H and ¹³C{¹H} NMR spectra of these four complexes were acquired in D₂O at pD = 7 (Figures 2 and S28−S35). A gradual change of the complex conformation is observed as the Ln³⁺ gets smaller. For La³⁺−py-macrodipa, the complex is present as the symmetric Conformation A, whereas the complexes of smaller Lu³⁺ and Sc³⁺ exist solely as the asymmetric Conformation B. The complex of the intermediately sized Y³⁺ ion shows a mixture of both Conformations A and B in a molar ratio of 3.6:1. These conformation are identified and assigned based on our previous detailed multinuclear NMR studies on the La³⁺, Y³⁺, and Lu³⁺ complexes of macrodipa.¹⁶ Considering the relatively large difference in ionic radius between La³⁺ and Y³⁺, we obtained the ¹H NMR spectrum of the py-macrodipa complex formed with paramagnetic Eu³⁺ (ionic radius = 94.7 pm).²² which clearly represents a symmetric Conformation A. The clearly resolved paramagnetic ¹H NMR spectrum of this complex (Figure S36) shows that it adopts the symmetric Conformation A. In this study, we also acquired the ¹H and ¹³C{¹H} spectra of the Sc³⁺−macrodipa complex (Figures S37−S38), which reveal it to attain Conformation B, as expected based on the small ionic radius of Sc³⁺. Collectively, these NMR spectroscopic results show that the inclusion of a pyridyl moiety within py-macrodipa does not alter the conformational toggle responsible for accommodating both large and small Ln³⁺ ions that is observed for macrodipa. Consistent with our expectations and prior studies with macrodipa, large Ln³⁺ ions form symmetric complexes that attain Conformation A, whereas the asymmetric Conformation B is preferred for small Ln³⁺ ions.

Figure 2.

¹H NMR spectra of La³⁺−, Y³⁺−, Lu³⁺−, and Sc³⁺−py-macrodipa complexes (600 MHz, D₂O, pD = 7, 25 °C).

Complex Structures—X-ray Crystallography.

To gain further insight on the conformational toggle of py-macrodipa across the Ln³⁺ series, we characterized the La³⁺, Lu³⁺, and Sc³⁺ complexes by X-ray crystallography. Crystal structures of these three complexes are shown in Figure 3. These crystal structures are consistent with the NMR spectroscopic data, which show the presence of different conformations for Ln³⁺ ions of different sizes. As previously observed for macrodipa, the La³⁺ center in the py-macrodipa complex attains a 10-coordinate geometry, forming a compound with the Conformation A geometry and a slightly distorted C₂ symmetry. For this complex, the central La³⁺ ion is fully encapsulated in the macrocycle, interacting with all of these donors. The two pendent picolinate arms coordinate La³⁺ from above and below the plane of the macrocycle. The complexes of the smaller ions, Lu³⁺ and Sc³⁺, attain the 8-coordinate Conformation B, which is of low symmetry. In Conformation B, in addition to the two picolinate donor arms, only three of the macrocyclic donors, N1–N3, appreciably interact with the metal centers. Furthermore, the Lu³⁺ and Sc³⁺ complexes are highly comparable except for presence of an inner-sphere water molecule in the Sc³⁺ complex that is absent in the Lu³⁺ complex. With this inner-sphere water molecule, the Sc³⁺–py-macrodipa complex is consistent with what was previously observed for the Lu³⁺–macrodipa complex,¹⁶ with both structures showing hydrogen-bonding interactions between the inner-sphere water and the macrocycle ethers. By contrast, the coordination sphere of the Lu³⁺ center in the py-macrodipa complex is completed by one of the ethereal oxygens on the macrocycle in place of the inner-sphere water. Although we consider both Sc³⁺ and Lu³⁺ complexes to attain Conformation B, we distinguish them further by referring to the anhydrous complex as “Conformation B0” and the complex with the inner-sphere water molecule to be “Conformation B1”.

Figure 3.

Crystal structures of (a) [La(py-macrodipa)]⁺, (b) [Lu(py-macrodipa)]⁺, and (c) [Sc(py-macrodipa)(OH₂)]⁺ complexes. Thermal ellipsoids are drawn at the 50% probability level. Solvent, counterions, and nonacidic hydrogen atoms are omitted for clarity. Only one of the two [La(py-macrodipa)]⁺ or [Sc(py-macrodipa)(OH₂)]⁺ complexes in the asymmetric unit is shown.

The presence of the inner-sphere water molecule may be a consequence of the conditions under which these crystals were obtained. Crystals of Lu³⁺–py-macrodipa were grown from the vapor diffusion of Et₂O into a MeOH solution containing the complex, reflecting a situation with a limited H₂O content. By contrast, crystals of Sc³⁺–py-macrodipa were obtained directly from an aqueous solution, a condition that would most likely favor H₂O coordination. Moreover, the previously reported Lu³⁺−macrodipa structure, which also attains Conformation B1, was afforded via the analysis of crystals grown from an aqueous solution as well.¹⁶ A recent survey on interatomic distances (r) found within the crystal structures in the Inorganic Crystal Structure Database (ICSD) of Ln³⁺ complexes with direct oxygen atom coordination revealed an average value of r(Lu−O) = 2.342 ± 0.103 Å for 8-coordinate Lu³⁺ complexes.³³ The distance between the Lu³⁺ center and ethereal oxygen donor O1 within [Lu(py-macrodipa)]⁺, r(Lu−O) = 2.509 Å, is substantially longer than the majority of Lu³⁺−O interactions in the ICSD, suggesting this interaction to be weak. Furthermore, for [Lu(macrodipa)(OH₂)]⁺, the distance between Lu³⁺ and the bound OH₂ is r(Lu−O) = 2.281 Å, indicating a typical coordination interaction between these two species. Based on these findings, we anticipate that in aqueous solution, which is of the most significant relevance to the potential application of this chelator for nuclear medicine, the hydrated Conformation B1 is the dominant species.

DFT Calculations.

To further understand the conformational toggle of py-macrodipa across the Ln³⁺ series, as well as the origins of the enhanced thermodynamic stability of Ln³⁺−py-macrodipa complexes compared to those of Ln³⁺−macrodipa, we carried out DFT calculations with Gaussian 09.³⁴ For comparative purposes, the same theoretical methods that we had previously implemented for Ln³⁺−macrodipa (Ln = La−Lu)¹⁶ were applied to study Ln³⁺−py-macrodipa. Specifically, the dispersion-corrected hybrid functional ωB97XD was employed.^35,36 A large-core relativistic effective core potential (LCRECP) and the associated basis set was assigned to the Ln³⁺ ions (Ln = La−Lu),³⁷ whereas the 6–31G(d,p) basis set was applied to all other lighter atoms.^38,39 Solvation effects in water were accounted for using the SMD solvation model.^40,41 Furthermore, we also probed Sc³⁺−macrodipa in this work to compare it directly to its py-macrodipa analogue. A detailed set of equations describing our computational thought flow is given in the Supporting Information.

To investigate the thermodynamic preferences for Conformations A and B for all Ln³⁺−py-macrodipa complexes, these structures were optimized, and their standard free energies (G°) were calculated. Due to our observations of both the non-hydrated and hydrated Conformations B0 and B1 by X-ray crystallography, both structures were calculated. Balanced chemical reactions for the conformational switch between Conformations A and B (B0 or B1) in aqueous solution can be written as either eq 3 or 4, where the standard free energy change of these two reactions are annotated as ΔG°(A,B0) and ΔG°(A,B1), respectively.

$${\left[\text{Ln}\left(\text{py-macrodipa}\right)\right]}^{+}\left(\text{Conformation A},aq\right)\rightleftharpoons {\left[\text{Ln}\left(\text{py-macrodipa}\right)\right]}^{+}\left(\text{Conformation B}0,aq\right)$$ (3)

$${\left[\text{Ln}\left(\text{py-macrodipa}\right)\right]}^{+}\left(\text{Conformation A},aq\right)+{\text{ H}}_2\text{O }\left(l\right)\rightleftharpoons {\left[\text{Ln}\left(\text{py-macrodipa}\right)\right]}^{+}\left(\text{Conformation B}1,aq\right)$$ (4)

Figure 4a summarizes computed ΔG°(A,B0) and ΔG°(A,B1) values across all Ln³⁺. As expected from the NMR and X-ray crystallographic data, these ΔG° values are positive for early Ln³⁺ ions, but negative for late Ln³⁺ ions, signifying a change in thermodynamic preference from Conformation A to B as the series is traversed. Additionally, the ΔG°(A,B1) values are systematically more negative than those for ΔG°(A,B0) by ~30 kJ·mol⁻¹, suggesting that Conformation B1 is considerably more stable than B0 in aqueous solution. Thus, regarding Conformation B, our subsequent calculations focused primarily on Conformation B1.

Figure 4.

(a) DFT-computed standard free energy differences between Conformations A and B for Ln³⁺−py-macrodipa complexes, ΔG°(A,B0) and ΔG°(A,B1), as defined in eqs 3–4. DFT-computed (b) relative binding energies (ΔΔG_B°, eq 6) and (c) relative strain energies (ΔΔG_S°, eq 7), comparing the Ln³⁺−py-macrodipa and Ln³⁺−macrodipa complexes.

Another feature of the Ln³⁺−py-macrodipa complexes that warrants computational investigation is their enhanced thermodynamic stabilities over the Ln³⁺−macrodipa complexes. To understand this property, we analyzed the free energy of the formation of these metal-ligand coordination complexes by breaking this process up into two distinct steps. In the first step, the free ligand adjusts its structure to a conformation that is appropriate for metal binding (eq 5), the free energy change of which is defined as the strain energy (ΔG_S°). Once the ligand has adopted the proper conformation, Ln³⁺ binding occurs (eq 6), resulting in a free energy change defined as the binding energy (ΔG_B°).

$${\text{L}}^{2-}\left(\text{free},aq\right)\rightleftharpoons {\text{L}}^{2-}\left(\text{Conformation A or B}1,aq\right)$$ (5)

$${\text{L}}^{2-}\left(\text{Conformation A or B}1,aq\right)+{\text{ Ln}}^{3+}\rightleftharpoons {\left[\text{LnL}\right]}^{+}\left(\text{Conformation A or B}1,aq\right)$$ (6)

These free energy terms, ΔG_S° and ΔG_B°, were investigated for both ligand systems in both possible conformations (A and B1). These values are plotted in a relative manner (ΔΔG_S° and ΔΔG_B°, defined in eqs 7–8) for a straightforward comparison between the two ligand systems, as shown in Figures 4b and 4c.

$$\Delta \Delta G_{\text{S}}°\left(\text{A or B}1\right)=\Delta G_{\text{S}}°\left(\text{py-macrodipa, A or B}1\right)-\Delta G_{\text{S}}°\left(\text{macrodipa, A or B}1\right)$$ (7)

$$\Delta \Delta G_{\text{B}}°\left(\text{A or B}1\right)=\Delta G_{\text{B}}°\left(\text{py-macrodipa, A or B}1\right)–\Delta G_{\text{B}}°\left(\text{macrodipa, A or B}1\right)$$ (8)

The difference in binding energies, ΔΔG_B°, demonstrate a clear contribution to the enhanced stability of Ln³⁺−py-macrodipa complexes. Regardless of the complex conformation, both ΔΔG_B°(A) and ΔΔG_B°(B1) are negative across the whole Ln³⁺ series, indicating that Ln³⁺-binding is more favorable for py-macrodipa. This result can be attributed to the presence of the pyridyl group within the macrocycle of this ligand, which is a stronger donor than the analogous ethereal oxygen donor in macrodipa. The ΔΔG_S°(A) values are less than zero for all Ln³⁺, indicating that the rigid planar pyridine group in the macrocycle of py-macrodipa favors the complex formation by preorganizing this ligand for metal binding in Conformation A. By contrast, ΔΔG_S°(B1) fluctuates about zero across the Ln³⁺ series, suggesting that the rigidity introduced by the pyridyl moiety in py-macrodipa does not significantly influence the complex formation in Conformation B1.

Photophysical Studies

Within the Ln³⁺ series, Eu³⁺ and Tb³⁺ are known for their rich photoluminescent properties.^42–45 For a better understanding and more complete characterization on Ln³⁺−py-macrodipa complexes, the photophysical properties its Eu³⁺ and Tb³⁺ complexes were investigated. These results are summarized in Table 2.

Table 2.

Photophycial Properties of Eu³⁺− and Tb³⁺−py-macrodipa Complexes.^a

	λmax/nm	Quantum Yield/%	τ(H2O)/ms	τ(D2O)/ms	q
Eu3+−py-macrodipa	273	0.21(1)	1.12(1)	1.43(1)	−0.1
Tb3+−py-macrodipa	273	0.84(4)	3.14(3)	3.14(2)	−0.3

^apH 7.4, 22 °C. The values in the parentheses are one standard deviation of the last significant figure.

The UV−Vis absorption and emission spectra of these four complexes were acquired at pH 7.4 (Figure S42). Both complexes give rise to an absorption band centered at 273 nm, which is attributed to the π → π* transition of the pyridyl units.⁴⁶ Upon this pyridyl-based excitation, both complexes show characteristic Ln³⁺-based f-f emissions, through the well-established antenna effect.^47,48 Additionally, we determined their photoluminescence quantum yields by a relative method,⁴⁹ using [Ru(bpy)₃]Cl₂ (quantum yield = 4.0% in air-saturated water)⁵⁰ as the reference compound.

A comparison between luminescence lifetimes (τ) of Eu³⁺ and Tb³⁺ complexes acquired in H₂O and D₂O can be used to determine the number of inner-sphere water molecules (q).⁵¹ The q values of Eu³⁺ and Tb³⁺ complexes are calculated by empirically derived expressions, eqs 9–10,^52,53 where in both equations Δk_obs = 1/τ(H₂O) − 1/τ(D₂O) in ms⁻¹.

$$q_{\text{Eu}}=1.11\times \left(\Delta k_{\text{obs}}–0.31\right)$$ (9)

$$q_{\text{Tb}}=5.0\times \left(\Delta k_{\text{obs}}–0.06\right)$$ (10)

As shown in Table 2, the measured τ values in H₂O and D₂O indicate both complexes do not possess any inner-sphere water molecule (q = 0). This observation is consistent to our other experimental (Figures 1 and S36) and computational (Figure 4a) results that suggest these complexes adopt Conformation A, where an inner-sphere water is absent.

DTPA Transchelation Challenge Assays.

For nuclear medicine applications, the kinetic stabilities of radiometal complexes are of critical importance because release of free metal ions is expected to give rise to toxic side effects.¹ To assess the suitability of py-macrodipa for use in nuclear medicine, we evaluated the kinetic stabilities of the Ln³⁺−py-macrodipa complexes and compared them to the corresponding Ln³⁺−macrodipa complexes. Specifically, we monitored the transchelation reaction of these Ln³⁺ complexes by UV−Vis spectroscopy in the presence of 100-fold excess DTPA at pH 7.4 and RT (22 °C), conditions that thermodynamically favor loss of the Ln³⁺ ion from both py-macrodipa and macrodipa.^54–56 With the large excess of DTPA used, this reaction follows pseudo-first-order kinetics, enabling us to use readily determined half-lives (t_1/2) as a comparative measure of the complex kinetic stability. These t_1/2 values are listed in Table 2.

For both ligand systems, the kinetic stability of their Ln³⁺ complexes follows the trends of their thermodynamic stability, with the early and late Ln³⁺ complexes displaying greater kinetic stability than those in the middle of the 4f series. Notably, the dissociation of Er³⁺−py-macrodipa complex follows a biexponential function, affording two pseudo-first order rate constants. We tentatively assign the two kinetic processes as arising from dissociations of the distinct Conformations A and B, which presumably possess different kinetic labilities. The ¹H NMR spectrum of Y³⁺−py-macrodipa (Figure 2) provides support for this hypothesis, as it clearly shows a mixture of both conformers. Based on the comparable ionic radii of Er³⁺ (91 pm) and Y³⁺ (90 pm),²² Er³⁺−py-macrodipa is expected to similarly exist in solution with both conformers, providing an explanation for the two observed kinetic processes.

Overall, the kinetic stabilities of Ln³⁺−macrodipa complexes are poor, a property that is reflected by the fact that none of their half-lives exceeds 30 min. By contrast, py-macrodipa forms Ln³⁺ complexes with significantly enhanced kinetic stability. In particular, complexes of both the largest La³⁺ ion (t_1/2 = 6.3 d) and smallest Sc³⁺ ion (t_1/2 = 16.6 h) show remarkable kinetic stabilities, a prerequisite for using this ligand in nuclear medicine. These complex t_1/2 values are substantially longer than the physical half-lives of their corresponding medicinally relevant radionuclides like ¹³²La (t_1/2 = 4.6 h), ¹³⁵La (t_1/2 = 18.9 h), and ⁴⁴Sc (t_1/2 = 4.0 h).^57,58 The excellent kinetic stability of La³⁺− and Sc³⁺−py-macrodipa indicates that py-macrodipa is a promising candidate for chelating radioactive La³⁺ and Sc³⁺.

Radiolabeling Studies.

Having shown that py-macrodipa forms complexes with La³⁺ and Sc³⁺ of both high thermodynamic and kinetic stabilities, we next sought to evaluate its ability to radiolabel medicinally relevant radioisotopes of these ions. The radionuclides ¹³⁵La³⁺ and ⁴⁴Sc³⁺, which have potential applications for Auger electron therapy and positron emission tomography imaging,^59,60 were produced and isolated following our established procedures.^13,61,62 For comparison, radiolabeling studies were also conducted using macrodipa and DOTA. For py-macrodipa and macrodipa, these experiments were executed at 25 °C, whereas for DOTA the elevated temperature of 80 °C was employed due to its known slower binding kinetics. These studies used a pH 5.5 buffer, and the reactions were monitored at 15, 30, and 60 min.

Table 4 lists the apparent molar activities (AMA) for the ¹³⁵La- and ⁴⁴Sc-labeled ligands obtained under these conditions. AMA, which describes the amount of radionuclide activity incorporated normalized to the amount of chelator, has been previously used to compare the radiolabeling efficacies of different ligands.^13,61–63 Among these three ligands, py-macrodipa affords significantly higher AMAs for both ⁴⁴Sc and ¹³⁵La in comparison to the other ligands. Notably, the AMA does not appreciably increase beyond 15 min, indicating that radiolabeling at this mild temperature (25 °C) for this ligand is complete on this short time scale. The AMAs obtained for macrodipa show that although it can bind ¹³⁵La³⁺, it is ineffective for ⁴⁴Sc³⁺. Conversely, even at elevated temperatures, the established chelator DOTA was only able to bind the smaller ⁴⁴Sc³⁺ radionuclide but failed in sufficiently complexing the large ¹³⁵La³⁺ radionuclide.

Table 4.

Apparent Molar Activity (AMA, Ci·μmol⁻¹) of ¹³⁵La³⁺ and ⁴⁴Sc³⁺ Radiolabeling with Ligands py-macrodipa, macrodipa, and DOTA.

	Radiolabeling Conditions		135La3+	44Sc3+
py-macrodipa	pH 5.5 25 °C	15 min	1.488	0.798
		30 min	1.363	0.888
		60 min	1.709	0.733
macrodipa	pH 5.5 25 °C	15 min	0.105	0.006
		30 min	0.088	0.004
		60 min	0.095	0.004
DOTA	pH 5.5 80 °C	15 min	0.005	0.419
		30 min	0.020	0.395
		60 min	0.030	0.464

Collectively, these data show that py-macrodipa exhibits versatile radiolabeling properties that enable it to effectively bind to both the large ¹³⁵La³⁺ and small ⁴⁴Sc³⁺ radionuclides. Its overall radiolabeling efficiency, as reflected by significantly large AMAs, is substantially better than those of our first-generation ligand macrodipa, thus validating our design strategy of incorporating the pyridine donor within the macrocycle. In comparison to the widely used “gold-standard” ligand DOTA, py-macrodipa can effectively be radiolabeled by ions of both types. To our best knowledge, py-macrodipa is the first chelator capable of effectively binding both La³⁺ and Sc³⁺ radionuclides, which possess substantially different ionic radii.

Human Serum Challenge Assays.

Having demonstrated that py-macrodipa is effectively radiolabeled by ¹³⁵La³⁺ and ⁴⁴Sc³⁺, we next evaluated the kinetic stability of the resulting complexes of these radiometals in human serum at 37 °C to assess the potential value of this ligand for nuclear medicine. The stabilities of ¹³⁵La³⁺− and ⁴⁴Sc³⁺−py-macrodipa complexes were monitored by radio-HPLC over 48 h and 8 h, respectively, time periods that account for the different physical half-lives of these radionuclides. Although the signal-to-noise of the chromatograms gradually decreases due to physical decays of the radionuclides, only a single peak in the chromatograms corresponding to ¹³⁵La³⁺− and ⁴⁴Sc³⁺−py-macrodipa complexes is observed over the duration of experiment (Figures 5a–5b), indicating that both complexes exhibit superb kinetic stabilities in human serum. It is worth noting that the signal to noise of these chromatograms is fading over time. This observation is related to the physical decay of the radionuclides, which contributes to a gradually decreasing detected intensity. For comparison, the ¹³⁵La³⁺− and ⁴⁴Sc³⁺−macrodipa complexes were also tested for their stabilities in human serum (Figures 5c−5d). Although ¹³⁵La³⁺−macrodipa is sufficiently stable in human serum over 48 h, ⁴⁴Sc³⁺−macrodipa complex shows deficient kinetic stability. These results highlight that py-macrodipa is an improvement over macrodipa because of its ability to form kinetically stable complexes with both the large ¹³⁵La³⁺ and small ⁴⁴Sc³⁺. Thus, based on its promising radiolabeling and radiometal-stabilizing properties, py-macrodipa is a valuable radiopharmaceutical candidate.

Figure 5.

Human serum challenge (37 °C) on the (a) ¹³⁵La³⁺−py-macrodipa, (b) ⁴⁴Sc³⁺−py-macrodipa, (c) ¹³⁵La³⁺−macrodipa, and (d) ⁴⁴Sc³⁺−macrodipa complexes monitored by radio-HPLC.

CONCLUSION

In this work, the ligand py-macrodipa was successfully synthesized and its coordination chemistry with the Ln³⁺ ions was investigated in detail via both experimental and theoretical methods. Building upon the unusual dual-size selectivity of our first-generation ligand macrodipa, we have shown that the presence of pyridyl moiety within the macrocycle of py-macrodipa substantially improves both the thermodynamic and kinetic stability of its Ln³⁺ complexes. The enhanced Ln³⁺-binding properties of py-macrodipa manifest in efficient radiolabeling of both ¹³⁵La³⁺ and ⁴⁴Sc³⁺ and excellent stability of the resulting complexes in human serum. This ligand is the first chelator capable of binding radiometal ions with disparate sizes such as ¹³⁵La³⁺ and ⁴⁴Sc³⁺, both with satisfactory efficacy, marking a significant improvement over current state-of-the-art gold standard chelator DOTA, which can only be used for a single type of ion. We thus describe py-macrodipa as a “Janus chelator” because of its ability to effectively bind two types of metal ions by switching its conformation. The success of py-macrodipa demonstrates that coordination chemists can tailor a ligand to effectively bind metal ions with disparate properties, rather than optimizing a ligand for a single ion type. Even though this work focused exclusively on the Ln³⁺ series, we anticipate that this concept can be expanded to other metal ions on the periodic table. Another genre of “Janus chelator” was recently reported, where the chelator switches between phenolate and pyridine pendant donors to effectively bind the Mn ion in both its chemically hard and borderline +3 and +2 oxidation states.⁶⁴ Thus, our description of py-macrodipa, as a Janus chelator that bends its backbone differently to accommodate ions of different sizes, provides a complementary approach. The design of py-macrodipa is an important achievement for inspiring future ligand design strategies and represents an important step in realizing an omnipotent “Janus chelator” that can bind all metal ions of disparate properties. Ongoing work is focused on the development and implementation of a bifunctional derivative of py-macrodipa that can be used for targeted nuclear medicine.

Table 3.

Half-lives of Ln³⁺−py-macrodipa and Ln³⁺−macrodipa Complexes when Challenged with 100 Equivalents of DTPA.^a

	Ln3+–py-macrodipa	Ln3+–macrodipa
La3+	6.3 ± 0.3 d	1678 ± 33 s
Nd3+	3.9 ± 0.2 d	615 ± 14 s
Gd3+	5524 ± 107 s	54 ± 2 s
Er3+	578 ± 23 s; 61± 3 s	5 ± 1 s
Lu3+	853 ± 10 s	65 ± 1 s
Sc3+	16.6 ± 1.7 h	782 ± 18 s

^a[LnL] = 100 μM, pH 7.4, 22 °C.