Macropa Scaffold Expansion for Actinium-225 Chelation: A Synthetic Strategy, Labeling Kinetics, and Theoretical Calculations

Abstract

²²⁵Ac is a key α-emitter for targeted alpha therapy. Among available chelators, Macropa currently provides some of the most stable ²²⁵Ac complexes, yet the limited stability of [²²⁵Ac]­Ac–Macropa indicates a further optimization potential. Here, we report the design and evaluation of a new Ac³⁺ chelator, coined Megapa, and assessed its suitability to produce stably labeled ²²⁵Ac-based radiopharmaceuticals using radiochemical and computational methods. Unexpectedly, Megapa showed poorer radiolabeling performance than Macropa, showing reduced ²²⁵Ac incorporation (80.2 ± 2.9% RCC vs quantitative labeling) across multiple conditions tested. In addition, [²²⁵Ac]­Ac–Megapa displayed lower kinetic inertness than [²²⁵Ac]­Ac–Macropa, with lower stability in human serum (45.8% intact after 7 days vs no detectable degradation) and substantially higher ²²⁵Ac release in La³⁺ challenge experiments (64.3% vs 0.7%). Thermodynamic stability studies supported these results, indicating a lower thermodynamic stability of La–Megapa compared to La–Macropa (log K _LaL of 10.53 vs 13.90). To rationalize these findings, quantum chemical calculations were performed on the Ac³⁺ and La³⁺ complexes of Megapa and Macropa. The computed low-energy structures were closely analogous for both chelators, indicating that the differing radiochemical behavior is unlikely to arise from intrinsic metal–ligand bonding. Instead, solvation effects and solution-phase molecular interactions are the most probable contributors to the poorer performance of Megapa.

Introduction

Targeted radionuclide therapy (TRT) is a therapeutic approach for malignant neoplasms that involves the systemic administration of radiopharmaceuticals carrying therapeutic radionuclides that emit cytotoxic radiation. The ideal endoradiotherapeutic agent selectively accumulates in the tumor tissue through specific interactions with cell surface proteins that are highly overexpressed on malignant cells, thereby maximizing tumor targeting while minimizing off-target effects. By this pathway, a high tumor uptake is achieved and the therapeutic radionuclide delivers dose, resulting in cell damage and cell death induced by different biochemical mechanisms. , An important aspect of this form of tumor therapy is its high selectivity and specificity, concentrating the cytotoxic effect on the tumor, while healthy tissues receive a comparatively low radiation dose burden.

In TRT, a range of therapeutic radionuclides can be employed, including alpha (α) and beta (β^–) emitters as well as low-energy electrons resulting from the Meitner–Auger effect. While beta-emitting radionuclides have been utilized clinically for several decades, recent years have seen a growing emphasis in both basic and translational research on alpha emitters. This shift to targeted alpha therapy (TAT) reflects the high potential of this special form of TRT for enhanced cytotoxicity, improved treatment results, and spatial precision in targeting malignant cells. This is a consequence of the much higher radiotoxicity of these nuclides, being described by the linear energy transfer, specifying how many ionizations occur per path length of the traveling particle in vivo and the radiobiological effectiveness, which indicates the relative damage caused by the respective radiation type. Moreover, α particles induce so-called “clustered” DNA damage which is more difficult and much slower to repair than isolated defects induced by β^– particles. ,− Another notable advantage of α-emitting radionuclides is their capability to exert therapeutic effects even under hypoxic tumor conditions, where resistance to β^– particle-based therapies is frequently observed. This property enhances their clinical utility in targeting aggressive and treatment-resistant malignancies. ,, Consequently, α emitter-based TRT has demonstrated to be able to achieve treatment response or remission even in such individuals resistant to β^– emitter-based TRT in the clinical context. ,−

To be suitable for TAT, the selected α-emitting radionuclide must possess physical properties, such as half-life, emission energy, and decay profile, that align with the therapeutic application. In addition, it must be capable of forming thermodynamically stable and kinetically inert complexes that remain intact under physiological conditions, ensuring an effective and safe in vivo applicability. Currently, different α emitters Astatine-211 (²¹¹At), Lead-212 (²¹²Pb), Bismuth-212 (²¹²Bi), Bismuth-213 (²¹³Bi), Radium-223 (²²³Ra) Actinium-225 (²²⁵Ac), and Thorium-227 (²²⁷Th)) are of particular clinical interest, exhibiting diverse half-lives, decay chains, and introduction chemistries. −

Among these, ²²⁵Ac is currently the clinically most important radionuclide for TAT, exhibiting an approximately 10 day half-life and emitting multiple high-energy α particles during its decay chain. Its clinical application, however, critically depends on stable introduction chemistry into the tumor-specific targeting vector. As a radiometal ion, [²²⁵Ac]­Ac³⁺ is incorporated into biological carrier molecules through chelation. A fundamental requirement for any ²²⁵Ac-based radiopharmaceutical is the use of a chelating agent capable of forming complexes with the radiometal ion that are both kinetically inert and thermodynamically stable.

Currently, the chelator DOTA (1,4,7,10-tetraazacyclododecane-N,N′,N″,N‴-tetraacetic acid) is employed for [²²⁵Ac]­Ac³⁺ complexation and thus for the preparation of ²²⁵Ac-based TAT agents in the clinical context. However, accumulating evidence suggests that the [²²⁵Ac]­Ac–DOTA complex may undergo partial radiometal dissociation under in vivo conditions, compromising therapeutic efficacy and safety. − The release of [²²⁵Ac]­Ac³⁺ can lead to its redistribution away from the intended target site, thereby reducing the radiation dose delivered to the tumor. Simultaneously, nontarget healthy tissues and organsmost notably the dose-limiting kidneys, as well as liver and spleencan be exposed to a significant radiation burden (Figure ). −

1.

Schematic depiction of the TAT principle using an ²²⁵Ac-based endoradiotherapeutic with high stability, not releasing the parent nuclide (upper row), resulting in the main dose fraction being applied to the tumor or with partial release of [²²⁵Ac]­Ac³⁺ from the agent under in vivo conditions (lower row), resulting in a considerable dose being applied to the kidneys and liver.

Therefore, new approaches for a more stable complexation of [²²⁵Ac]­Ac³⁺ have been sought after and have resulted in the development of several new chelating agents such as linear acyclic systems featuring picolinic acid arms, different aza-macrocyclic chelators, polydentate picolinates, and bispidine-based scaffolds. , Among the emerging chelators for [²²⁵Ac]­Ac³⁺, the diazacrown-based ligand Macropa (N,N′-bis­[(6-carboxy-2-pyridyl)­methyl]-4,13-diaza-18-crown-6; compound 1, Figure ) stands out as particularly promising. It enables efficient complexation with minimal ligand quantities and exhibits rapid labeling kinetics at low temperatures, a significant advantage over DOTA (Figure ), which requires elevated temperatures for labeling and cannot take up the ion that efficiently: Macropa achieves an almost quantitative uptake (98%) of [²²⁵Ac]­Ac³⁺ already at a precursor concentration of 5.6 × 10^–7 M at room temperature, whereas DOTA reaches a 95% uptake only at 85 °C and a precursor concentration of 10^–3 M. , Moreover, the [²²⁵Ac]­Ac–Macropa complex was shown to provide a higher inertness toward serum protein degradation (90% vs 85% intact complex after 7 days of incubation) and metal displacement experiments using La³⁺ as the challenging ion (91% vs 77% intact complex after 7 days) compared to [²²⁵Ac]­Ac–DOTA. ,

2.

Chemical structures of the chelators DOTA, Macropa (1), Macropid, 2, H₂BZmacropa, PYTA, and macropa-XL having been described before and Megapa (3) developed here.

Although Macropa represents the most promising chelating agent for [²²⁵Ac]­Ac³⁺ complexation to date, the results of several studies indicate room for improvement of the [²²⁵Ac]­Ac–Macropa complex. Unbound [²²⁵Ac]­Ac³⁺ exhibits pronounced accumulation in the liver, as well as uptake in the spleen and bone. However, since [²²¹Fr]­Fr⁺ and [²¹³Bi]­Bi³⁺ also show substantial uptake in the kidneys, liver, and spleen (although not in bone), discrimination of the origin of the measured activity is challenging unless an appropriate energy window is used for activity quantification by γ-counting. This is particularly relevant because release of the first daughter nuclide occurs already at the stage of the first decay in the chain to [²²¹Fr]­Fr⁺, driven by the recoil energy and the fundamentally different chemical properties of [²²¹Fr]­Fr⁺ compared to [²²⁵Ac]­Ac³⁺. If the radionuclide is not retained within the tumor cell after release, relocalization occurs, accompanied by its associated deleterious effects. In several in vivo studies evaluating ²²⁵Ac-labeled radiopharmaceuticals across different targeting vector classes, relevant levels of activity have been reported in the mentioned organs. − In most cases, the γ-counting performed did not distinguish between the parent nuclide and its daughters, such that unambiguous evidence for the release of [²²⁵Ac]­Ac³⁺ from the [²²⁵Ac]­Ac³⁺–Macropa complex cannot be conclusively established. Nevertheless, the consistently observed substantial activity levels in bone strongly suggest at least some release of [²²⁵Ac]­Ac³⁺. Of special interest is a study systematically comparing the in vivo pharmacokinetics and dose deposition of [²²⁵Ac]­Ac–Macropa–TATE vs [²²⁵Ac]­Ac–DOTA–TATE in a preclinical SSTR2-positive tumor model. Here, the authors observed a higher activity uptake for the [²²⁵Ac]­Ac–Macropa–TATE not only in the tumor but also in the liver, spleen, kidneys, and femur compared to the [²²⁵Ac]­Ac–DOTA-analog and attributed this to the metabolism or fragmentation of the compound despite its high in vitro stability. This is consistent with the relatively low thermodynamic stability of the homologous La–Macropa complex, which exhibits a log K of only 13.90. Consequently, novel 18-membered macrocyclic chelators have recently been developed; however, they did not exhibit the requisite kinetic inertness of their ²²⁵Ac-complexes required for in vivo application. −

The development of a novel chelating agent optimally suited for the thermodynamically stable and kinetically inert complexation of Ac³⁺ thus seems to remain a challenging task, primarily due to an incomplete understanding of the coordination chemistry of the actinium ion. All actinium isotopes are radioactive, with ²²⁷Ac possessing the longest half-life of 21.8 years; the remaining isotopes exhibit half-lives shorter than 10 days. This short-lived nature and the extremely limited quantities of actinium isotopes available for chemical investigation significantly complicate the elucidation of Ac³⁺ coordination chemistry, although important progress has been made within the last few years characterizing ²²⁷Ac-complexes with polydentate chelators of radiopharmaceutical interest by extended X-ray absorption fine structure (EXAFS) spectroscopy and X-ray crystallography. , As well, theoretical modeling approaches face substantial limitations due to the high electron count of actinium, pronounced relativistic effects, the extensive conformational flexibility of chelators, inadequacies in solvation models, and the absence of reliable in vivo kinetic parameters. Consequently, the rational design of suitable chelators solely on the basis of existing experimental or computational data remains a difficult task.

However, several key characteristics of Ac³⁺ are established. Owing to its exceptionally large ionic radiusthe largest among all trivalent cationsAc³⁺ displays a low charge density and consequently an intermediate level of absolute chemical hardness. Furthermore, the large size of Ac³⁺ allows for a higher coordination number compared to the closely related homologue La³⁺ (e.g., up to 11 vs 9 water molecules as determined by EXAFS). Hence, an increased denticity of the chelating ligand, achieved by introducing more than just two carboxylic acids as in Macropa, might increase the stability of the formed complex by better saturating the coordination sphere of the Ac³⁺ ion. Moreover, it has been determined in a study using theoretical calculations that the ionic interactions within actinium complexes are more important than covalent contributions, suggesting that stable actinium chelation mainly relies on maximized ionic bonding interactions. However, practical studies in this field, resulting in Macropid and DOTA-derivative 2 (Figure ), found that none of these agents was able to take up [²²⁵Ac]­Ac³⁺ even at high temperatures, , suggesting that increasing the number of carboxylic acids alone is not key to achieve increased complex stability. Furthermore, increasing chelator backbone rigidityas pursued in the development of H₂BZmacropa and PYTA (3,6,10,13-tetraaza-1,8­(2,6)-dipyridinacyclo-tetradecaphane-3,6,10,13-tetraacetic acid; Figure )did not enhance the stability of the corresponding ²²⁵Ac-complexes. , Consequently, alternative strategies beyond those investigated thus far must be identified.

In the pursuit of developing an improved chelator for the stable complexation of [²²⁵Ac]­Ac³⁺, Macropa provided a promising structural scaffold. Building on prior studies and our own design rationale, we aimed to enhance complexation by expanding the macrocyclic backbone, thereby increasing the chelator’s denticity and potentially improving its coordination efficiency.

A similar strategy has very recently been explored to achieve a more stable complexation of Ra²⁺. Like [²²⁵Ac]­Ac³⁺, [²²³Ra]­Ra²⁺ is a highly relevant radioisotope for TAT, yet its use in target-specific treatments remains precluded because no chelators are currently known to form sufficiently stable complexes with this nuclide. To date, Macropa is the only chelator demonstrating some in vitro stability of the resulting complexes, although failing to maintain stability in vivo when conjugated to tumor-targeting vectors. Building on its molecular scaffold, an asymmetric Macropa-based ligand with a 1,13-diaza-[21]­crown-7 rather than a 1,10-diaza-[18]­crown-6 backbone, named Macropa-XL (Figure ), was investigated to assess whether it could more effectively saturate the large coordination sphere of Ra²⁺ and thus enhance complex stability. While this approach did not succeedresulting in lower kinetic stability in human serum compared to the corresponding Macropa complexthe strategy nevertheless remains fundamentally promising, particularly in the context of trivalent actinides: Ra²⁺ complexes are inherently more labile than their trivalent counterparts due to the lower charge, the resulting charge-to-radius ratio (Ra²⁺ exhibits an eight-coordinate ionic radius of 1.48 Å), and the distinct coordination chemistry of divalent ions.

In the context of [²²⁵Ac]­Ac³⁺ complexation, the use of a larger metal ion cavity could thus potentially be better capable of accommodating the exceptionally large +3 actinium ion while providing more donor atoms for coordination. Such an expanded framework may facilitate a more complete envelopment of the metal center, potentially improving complex stability and reducing the likelihood of radiometal release under physiological conditions. As a high symmetry of chelating agents is commonly preferable for achieving high complex stability, chelating agent 3, coined Megapa (Figure ), was thus chosen as the most promising structure for further studies.

To this end, the chelator Megapa was synthesized, thoroughly characterized, and subsequently evaluated using radiochemical methods to assess its suitability for the formation of stable and kinetically inert ²²⁵Ac-bearing radiopharmaceuticals. In addition, the thermodynamic stability of the homologous La–Megapa complex was also examined. Theoretical quantum-chemical investigations were conducted to complement and validate the experimental findings, providing deeper insights into the structural and bonding characteristics of Megapa complexes in comparison to those formed by Macropa, with both Ac³⁺ and La³⁺.

Results

Chemical Syntheses of Macropa (1) and Megapa (3)

The first step in the synthesis of the new chelator Megapa (3) was the preparation of crown ether 1,13-diaza-24-crown-8 (9) (Scheme ). The synthesis of the ring can in principle be performed via different routes. In general, the preparation is associated with two steps, first generating lactam 8, which is subsequently reduced to intended ring 9. This crown ether, analogously to commercially available Kryptofix 2.2 (1,10-diaza-[18]­crown-6) being a precursor of Macropa (Scheme ), can then be further reacted with 6-(mesyl-oxymethyl)­picolinic acid (12) to give tert-butyl-protected chelators 10 and 11.

1. Schematic Depiction of the Syntheses of Lactam 8, Picolinic Acid Derivative 12, Crown Ether 9, and Chelating Agents Macropa (1) and Megapa (3).

For the preparation of lactam 8, different reaction pathways were tested, all of which relied on the formation of the acid amides in a highly dilute solution to generate mainly intended species 8 instead of the larger rings which are formed in solutions of higher concentrations. For this purpose, bis-amine 4 was reacted with (i) bis-chloride 5, being generated in situ from bis-acid 6 using SOCl₂ in DMF under reflux for 16 h, followed by the slow addition of 4 together with DMAP (4-(dimethylamino)­pyridine) and pyridine as the base within 40 h; (ii) bis-acid 6 using EDC*HCl (N-ethyl-N′-(3-(dimethylamino)­propyl)-carbodiimide HCl salt) in CH₂Cl₂ as the solvent and DIPEA (N,N-diisopropylethylamine) as the base at ambient temperature over 1 week; or (iii) bis-methyl ester 7, having been prepared from bis-acid 6 by acidic esterification, in methanol at ambient temperature within 20 days.

Among the synthetic pathways evaluated for the preparation of 8, the reaction between bis-amine 4 and bis-methyl ester 7 yielded the cleanest conversion, producing minimal byproducts in both number and quantity. In contrast, the reaction of 4 with bis-chloride 5 proceeded significantly faster but was less selective, generating the highest levels of byproducts, which proved challenging to separate. Consequently, despite the comparatively long reaction time, spanning several weeks, the synthesis of lactam 8 was conducted via the bis-methyl ester route.

The reduction of lactam 8 to the corresponding 1,13-diaza-[24]­crown-8 crown ether 9 was evaluated using multiple different approaches: (i) BH₃*THF in THF at ambient temperature or reflux, (ii) Tf₂O, followed by NaBH₄ in THF at ambient temperature or reflux, and (iii) LiAlH₄ in THF at ambient temperature or reflux. For this reduction step, different results were obtained depending on the reaction conditions applied. The slowest and most incomplete reduction of the lactam was observed in the case of the borane tetrahydrofuran complex. Analogously, reduction via the triflate/sodium tetrahydroborate pathway proceeded incompletely, yielding a mixture of mono- and bis-reduced lactam species. In contrast, the reduction using lithium aluminum hydride gave the completely reduced product 9.

After the crown ether was obtained, 9 and Kryptofix 2.2 were efficiently reacted with 12 to give tert-butyl-protected chelators 10 and 11. 12 was obtained from first bis-tert-butyl-protection of pyridine-2,6-dicarboxylic acid, yielding 13, followed by the one-sided reduction to the corresponding alcohol 14 which was further converted into mesylate 12 to enable an efficient reaction with crown ethers. 10 and 11 were finally converted into the target chelators Macropa and Megapa by incubation with a cleavage cocktail consisting of TFA (trifluoroacetic acid), TIS (triisopropylsilane), and water (95:2.5:2.5, v/v/v) for 24 h.

Overall, both chelators were synthesized via this route in quantities of 31–74 mg, affording acceptable overall isolated yields of 16% for Macropa and 12% for Megapa in excellent purities of >97%. The identities of the compounds were confirmed by ¹H and ¹³C NMR spectroscopy as well as high-resolution mass spectrometry (see Supporting Information for details).

Comparative Radiolabeling Experiments of Macropa and Megapa with [²²⁵Ac]­Ac³⁺

In order to determine the general suitability of the newly developed chelator Megapa for ²²⁵Ac-labeling purposes, the comparative labeling efficiency and precursor quantities of Megapa and Macropa required to achieve complete complexation were investigated first.

For the typical radiolabeling experiment, 44–52 kBq [²²⁵Ac]­AcCl₃ in 0.1 M HCl was added to a solution of 10 nmol of precursor Macropa or Megapa in TRIS buffer (2 M, pH 7.0, 55 μL, TRIS: tris­(hydroxymethyl)­aminomethane) before the solution was diluted with water to a final volume of 100 μL, exhibiting a pH of 7.6–8.4. For radiolabeling reactions with [²²⁵Ac]­Ac³⁺, pH values of 5.5–7.0 were described, , but reactions are also possible at a pH of up to 9.4 due to the high pK _1H of Ac³⁺. As we found the radiolabeling reaction of Macropa to be more efficient at pH 7.6 compared with pH 6.0, we used the described parameter set as standard conditions for the following experiments. After incubation of the mixtures for 15 min at 37 °C, 10 μL of the respective solution was analyzed by analytical radio-HPLC and radio-iTLC. During analytical radio-HPLC, fractionated collection from the HPLC runs (five fractions per minute) and subsequent gamma counting using gamma energy windows of 70–270 keV for ²²¹Fr (E _γ: 218 keV) and 362–462 keV for ²¹³Bi (E _γ: 440 keV) were performed. The gamma counting measurements of the radio-HPLC fractions as well as the analyses of developed iTLCs were performed after 6 h to allow most of ²¹³Bi to have decayed.

These initial radiolabeling experiments revealed an unexpected chelating behavior for Megapa. While Macropa quantitatively incorporated [²²⁵Ac]­Ac³⁺ using 10 nmol of the chelator together with radionuclide activities of up to 800 kBq, Megapa failed to achieve complete complexation under identical conditions. As determined by analytical radio-HPLC and iTLC, a substantial fraction of unbound radionuclide activity was consistently observed for Megapa, typically ranging from 16.4% to 24.4%, corresponding to a radiochemical conversion (RCC) of 80.2 ± 2.9% (mean ± SD; Figure B). These findings were corroborated by iTLC measurements, which likewise indicated 18.7 ± 9.9% unbound activity in the case of Megapa. In principle, this apparent nonchelated activity could arise from the decay daughters of [²²⁵Ac]­Ac³⁺, namely, [²²¹Fr]­Fr⁺ and [²¹³Bi]­Bi³⁺, which might be more efficiently complexed by Macropa than by Megapa, thereby mimicking incomplete ²²⁵Ac-complexation in the Megapa system. However, this explanation can be excluded. HPLC fraction collection followed by delayed gamma counting in the 70–270 keV energy window (6 h) confirmed that the activity detected at the solvent front of the analytical radio-HPLC chromatograms originated from ²²⁵Ac/²²¹Fr, thereby demonstrating incomplete complexation of [²²⁵Ac]­Ac³⁺ by Megapa. In contrast, no such effect was observed in analogous ²²⁵Ac-labeling experiments with Macropa, which consistently yielded high RCCs of 99.2 ± 0.9% and showed no detectable activity at the solvent front under the same conditions (Figure A). In principle, one would expect that even in the case of less efficient complexation by the new chelator, quantitative complex formation could be achieved, provided that sufficiently high excesses of Megapa are employed. Contrary to this expectation, increasing the amount of the chelator did not result in complete incorporation of [²²⁵Ac]­Ac³⁺; instead, a plateau in the achievable RCC was observed with increasing chelator amount (Figure C) which might indicate an initially complete chelation of [²²⁵Ac]­Ac³⁺ but instability of one of the formed species under HPLC conditions.

3.

Typical analytical radio-HPLC chromatograms of the raw products of the radiolabeling reactions of Macropa (A) and Megapa (B) with [²²⁵Ac]­Ac³⁺ being obtained by fractionated collection of analytical radio-HPLC eluates and deferred gamma counting. (C) Correlation between the precursor amount of Macropa/Megapa and RCC with [²²⁵Ac]­Ac³⁺. (D) Influence of the pH on the [²²⁵Ac]­Ac³⁺ incorporation rate into Megapa. (E) Influence of reaction time and temperature on the [²²⁵Ac]­Ac³⁺ incorporation rate into Megapa under otherwise identical standard conditions (10 nmol precursor, TRIS buffer, pH 7.8–8.0). Shown radio-HPLC chromatograms were obtained using an RP-18e Chromolith Performance column together with a gradient of 0–100% acetonitrile in H₂O (+0.1% TFA) within 5 min at a flow rate of 4 mL/min and deferred γ-counting of the fractions at 6 h after collection.

Another unexpected observation was the considerably different shapes of the product peaks of [²²⁵Ac]­Ac–Macropa and [²²⁵Ac]­Ac–Megapa when investigated with analytical radio-HPLC. While [²²⁵Ac]­Ac–Macropa eluted from the HPLC column as a single uniform peak (Figure A), [²²⁵Ac]­Ac–Megapa showed a tailing bulge in front of the actual product peak (Figure B, R _t‑bulge: 1.1–2.3 min, R _t‑peak: 2.5 min). However, both species do not seem to represent different compounds as confirmed by collecting peak and bulge separately, followed by reinjecting into analytical radio-HPLC, showing an interconversion of both species. Changing the standard analytical system using another HPLC column did not give differing results, also showing the same inhomogeneous product peak. Taken together, this indicates that the bulge and peak both represent [²²⁵Ac]­Ac–Megapa but might consist of different conformers of the same complex.

As Megapa seemed to be less efficient in taking up the [²²⁵Ac]­Ac³⁺ activity compared to Macropa at a precursor amount of 10 nmol, it was tested if variation of the reaction conditions would result in a complete incorporation of the [²²⁵Ac]­Ac³⁺ activity into the Megapa chelator. For this purpose, the precursor amount (Figure C), the pH (Figure D), and the reaction time and temperature (Figure E) were varied. However, none of these parameters exerted a significant influence on the achievable RCCs. Notably, neither increasing the precursor quantity nor elevating the reaction temperature improved the efficiency of [²²⁵Ac]­Ac³⁺ incorporation into Megapa. To exclude thermal decomposition of the chelating agent as a potential cause of the observed incomplete complexation at elevated temperatures, the thermal stability of Megapa was also evaluated. However, no detectable decomposition was observed over a period of 3 days (Figure S25). Similarly, variation in pH within the range of 6.8–9.4 exhibited no discernible effect on the labeling outcomes. Because buffer salts can potentially affect radiolabeling results as well, TRIS was replaced with ammonium acetate, MOPS (3-(N-morpholino)­propanesulfonic acid), HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid), or sodium carbonate buffer systems to evaluate their influence on [²²⁵Ac]­Ac³⁺ incorporation into Megapa under standard reaction conditions. In these experiments, the use of sodium carbonate efficiently prevented product formation (RCC: 12.6 ± 4.5%), whereas the usually applied ammonium acetate buffer resulted in higher conversion rates to [²²⁵Ac]­Ac–Megapa (RCC: 56.7%), being however lower when compared to the standard TRIS buffer (RCC: 80.2 ± 2.9%). In contrast, MOPS (RCC: 86.7%) as well as HEPES (RCC: 70.1 ± 7.1%) gave comparable results to TRIS, thus also being well suited for [²²⁵Ac]­Ac³⁺ complexation using Macropa-derived chelators. To exclude radiolysis as a potential cause of [²²⁵Ac]­Ac³⁺ release from the [²²⁵Ac]­Ac–Megapa complex, the reaction mixture was further supplemented with ascorbic or gentisic acid to assess possible improvements in the labeling efficiency. However, none of these additives had any effect on the observed RCCs; rather, the ratio of unbound activity, bulge, and product peak remained unchanged, indicating that radiolysis is not responsible for the unbound activity detected in the reaction mixtures to [²²⁵Ac]­Ac–Megapa.

Determination of the Relative Complex Stabilities of [²²⁵Ac]­Ac–Macropa and [²²⁵Ac]­Ac–Megapa

Following the investigations on radiolabeling characteristics, the kinetic inertness of [²²⁵Ac]­Ac–Macropa and [²²⁵Ac]­Ac–Megapa against transchelation was studied in direct comparison, being an integral factor for the in vivo applicability of the systems.

For this purpose, both complexes were examined with regard to resistance against transchelation in human serum and by a complex challenge using excess La³⁺. For comparative evaluation of human serum stabilities, [²²⁵Ac]­Ac–Macropa and [²²⁵Ac]­Ac–Megapa were incubated with commercially available pooled human serum of healthy donors at 37 °C. At defined time-points of 0, 15 min, 1 h, 2 h, 3 h, 1 day, 6 days, and 7 days, aliquots of the mixtures were taken, the serum proteins were precipitated, the protein pellets and remaining solutions were measured for radioactivity, and the solutions were analyzed by analytical radio-HPLC. Also, in this case, radio-HPLC analyses were performed by fractionated collection of the HPLC runs and subsequent delayed gamma counting using the gamma energy window of 70–270 keV for ²²⁵Ac/²²¹Fr. This enabled the determination of the fraction of intact ²²⁵Ac-complexes, revealing that [²²⁵Ac]­Ac–Macropa remained largely intact over the entire 7 day period (Figure B + E), whereas [²²⁵Ac]­Ac–Megapa released substantial amounts of the radiometal within 15 min (Figure A + D).

4.

Overview of the results of the kinetic inertness studies of [²²⁵Ac]­Ac–Megapa and [²²⁵Ac]­Ac–Macropa against transchelation by human serum (A + B) and complex challenge of both complexes using an excess of La³⁺ (C). Typical analytical radio-HPLC chromatograms of serum stability assays for [²²⁵Ac]­Ac–Megapa and [²²⁵Ac]­Ac–Macropa over time are depicted in (D, E), respectively. Typical analytical radio-HPLC chromatograms of complex challenge experiments with La³⁺ for [²²⁵Ac]­Ac–Megapa and [²²⁵Ac]­Ac–Macropa over time are shown in (F, G), respectively.

Likewise, the kinetic inertness of [²²⁵Ac]­Ac–Macropa and [²²⁵Ac]­Ac–Megapa was determined in direct comparison by La³⁺ complex challenge experiments, using a 4000-fold excess of lanthanum­(III)-trifluoromethanesulfonate as the challenging agent. This assay allows assessment of the relative kinetic inertness of the complexes, mimicking the challenge of a radiometal complex by endogenous substances being present in high excess under in vivo conditions. For this purpose, product solutions of [²²⁵Ac]­Ac–Macropa and [²²⁵Ac]­Ac–Megapa were incubated over a period of 7 days with lanthanum triflate, and the extent of transchelation was determined by analytical radio-HPLC for both complexes. The results of these evaluations (Figure C) demonstrated a high kinetic inertness of the [²²⁵Ac]­Ac–Macropa complex against metal ion challenge (Figure G), whereas [²²⁵Ac]­Ac–Megapa showed a fast and considerable displacement of [²²⁵Ac]­Ac³⁺ by La³⁺ already within 15 min (Figure F). This confirms the results found in the serum stability assay, indicating a considerably lower kinetic inertness of the [²²⁵Ac]­Ac–Megapa complex compared to that of [²²⁵Ac]­Ac–Macropa.

This lower kinetic inertness of the [²²⁵Ac]­Ac–Megapa complex compared to [²²⁵Ac]­Ac–Macropa is also consistent with the labeling experiments, demonstrating a significantly higher efficiency of [²²⁵Ac]­Ac³⁺ incorporation into Macropa compared to Megapa. These findings match those from other studies, indicating that comparatively lower radiolabeling efficiencies tend to be accompanied by the reduced complex stability of the resulting metal complexes. For example, a recent highly systematic investigation of the radiolabeling of various chelators with [²²⁵Ac]­Ac³⁺ and the subsequent evaluation of the inertness of the formed ²²⁵Ac-complexes demonstrated that chelators exhibiting poor uptake of the radionuclide generally also formed less stable complexes. Similar trends have been reported for other radionuclides, suggesting that a low radiolabeling efficiency may be indicative of unfavorable complex geometries or suboptimal electronic properties of the resulting complexes which, in turn, manifest as comparatively lower kinetic inertness. These observations are consistent with the results presented herein.

This assumptionthat the Megapa chelator exhibits overall less favorable complexation properties for Ac³⁺is supported by the characterization of its nonradioactive La complex homologue (full characterization data, including analytical HPLC, HR-MS, and ¹H- and ¹³⁹La NMR, are provided in the Supporting Information). For example, analytical HPLC measurements revealed a pronounced acid lability of La–Megapa: the complex underwent complete decomposition in the presence of commonly used HPLC mobile phase additives (0.1% trifluoroacetic acid or 0.1% formic acid), preventing its detection under these standard acidic conditions. Detection by HPLC was possible only after switching to a neutral mobile phase system consisting of 20 mM NH₄HCO₃ and MeCN. This experimentally observed low kinetic inertness of La–Megapa was also corroborated by solution thermodynamic studies conducted on the Megapa ligand and La–Megapa.

To this end, ligand protonation and metal–ligand stability constants were determined by potentiometric titrations in 0.1 M NaCl (Figures S22 and S23). Comparing the protonation steps of Megapa, Macropa, and Macropa-XL (Figure and Table ), it appears that one more protonation step was accessible for Megapa by our experimental setup, corresponding to the deprotonation of one of the amines (log K ₁ ^H of 9.69(2)). The next four protonation constants were similar to those determined for Macropa and Macropa-XL and can be assigned to the second amine (log K ₂ ^H of 7.62(1)) and the pyridyl groups (log K ₃ ^H of 7.41(1) and log K ₄ ^H of 4.02(1)), , whereas the last protonation (log K ₅ ^H = 2.26(1)) corresponds to one of the carboxylates of the picolinic acid moieties. Protonation of the second carboxylate was not accessible, as the starting pH was chosen to be 2.4 for practical reasons.

1. Protonation and Stability Constants of Megapa^2– (I = 0.1 M NaCl, 25 °C) Compared to the Values Reported in the Literature for Macropa^2– and Macropa-XL^2– ,

	Megapa	Macropa	Macropa-XL
log K 1 H	9.69(2)	7.73	7.92
log K 2 H	7.62(1)	6.80	6.99
log K 3 H	7.41(1)	3.13	3.27
log K 4 H	4.02(1)	2.40	2.44
log K 5 H	2.26(1)	-	-
log K 1–5 H	31.00	20.06	20.62
log K LaL	10.53(2)	13.9	-
log K 1 LaL H	8.83(3)	-	-
log K 2 LaL H	7.34(3)
log K 3 LaL H	6.79(2)
log K 4 LaL H	2.97(3)
pLa	10.31	15.6	-

^aData in 0.1 M NaCl, 25 °C.

^bData in 0.1 M KCl, 25 °C.

^cCalculated for [L] = 10 [La³⁺] = 10 μM; pM = −log­([La³⁺]_free) at pH = 7.4 and 25 °C.

Further, the La–Megapa complex was studied. All data sets were modeled using a 1:1 (M/L) stoichiometry, and inclusion of additional protonation equilibria for the complex led to a marked improvement in the fits, consistent with the comparatively low stability of the La–Megapa complex (log K _LaL = 10.53(2)). The presence of multiple complex protonation steps is indicative of an out-of-cavity coordination mode, leaving one amine donor and both pyridine nitrogen atoms available for protonation. Accordingly, the stability constant is substantially lower than that reported for [La­(Macropa)]⁺ (log K _LaL = 13.90). From the comparison of the titration curves of Megapa^2– and Megapa^2– in the presence of La­(III), a decrease of the measured pH values is evident (Figure S23), which suggests the binding of the metal ion by Megapa^2–. Further, the comparison shows that one protonation constant is absent in the presence of La­(III).

Speciation diagrams calculated from the equilibrium constants (Figure ) show that deprotonation of the Megapa chelator occurs only at relatively high pH. At physiological pH (7.4), the mono- and diprotonated 1:1 complexes are present in approximately equal proportions, further underscoring the limited stability of the La–Megapa complex. To assess thermodynamic stability under physiologically relevant conditions, pLa values were calculated, where pLa is defined as −log­([La³⁺]_free) for [L] = 10­[La³⁺] = 10 μM at pH 7.4 and 25 °C. Consistent with the potentiometric data, Megapa^2– exhibits a significantly lower pLa value than Macropa, indicating a poorer interaction of La³⁺ with the Megapa chelator compared to Macropa.

5.

Speciation diagrams of Megapa^2– (left) calculated for c _L = 10^–3 M and Megapa^2– in the presence of 1 equiv of La³⁺ (right) calculated for c _La3+ = c _L = 10^–3 M (I = 0.1 M NaCl, 25 °C).

Theoretical Assessment of the La/Ac–Megapa and La/Ac–Macropa Complexes

In order to understand the unexpected results obtained, the structural and bonding properties of the complexes were assessed by quantum chemical calculations. In fact, they provided the only clue to the probable structure of the Megapa complexes because no suitable crystals of La–Megapa could be obtained for X-ray crystallographic structure elucidation. The large conformational space of Megapa required a thorough conformational search, which was performed at the molecular mechanics level in an aqueous solution. The final geometries and energetics were evaluated by using advanced density functional theory (DFT) calculations.

The reliability of our computational procedure was verified by the results on the La–Macropa complex. In the conformational search, all the 8 significant conformers of La–Macropa , were found, and the known preference of the δλδ,δλδ conformer (Figure S28) was also reproduced by the present DFT calculations (Table S2).

The computed La–Megapa and Ac–Megapa complexes were inspected, and the low-energy structures for the two metals were found to be analogous, though with somewhat different relative energies. The relative energies of the seven lowest-energy conformers are compiled in Table S3, where the Cartesian coordinates of the selected optimized structures are available too. The found lowest-energy conformer with picolinate groups interacting from opposite sides of the macrocycle (added as conformer VIII for both La–3 and Ac–3 in Table S3) was computed to lie by ca. 30 kJ/mol higher in energy than the global minimum. The two most significant conformers I and II of Ac–3 are depicted in Figure .

6.

Comparison of the most stable Ac–Megapa calculated complex structures (Ac: cyan; C: gray; N: blue; O: red) with atoms in the back gradually faded. For facilitated comprehension of the calculated structures, they are also shown schematically below. The staggered lines indicate the coordination to Ac. Conformer I is shown on the left and conformer II on the right.

Whereas the relative stabilities of the lowest-energy La–Macropa and Ac–Macropa complexes were found to be similar, this is not fully true for the Megapa complexes. For six of the seven most stable Megapa complexes of La and Ac, the energy ordering is the same, and conformer I is preferred for both f elements (Table S3). The striking exception is conformer II, which has a stability identical to that of I in the case of Ac–Megapa while having significantly lower (by 14 kJ/mol in terms of ΔG _SMD) stability in the case of La–Megapa. This indicates that the cavity of II is significantly more suitable for the larger Ac³⁺ ion than that for La³⁺. This feature is analyzed further on the basis of the M–ligand distances given in Table .

2. Selected Average Computed Metal–Ligand Distances (≤3 Å) of the Two Most Stable M–Megapa and M–Macropa Conformers (M = La, Ac) .

	Megapa				Macropa
	I		II		δλδ,δλδ
distance	Ac	La	Ac	La	Ac	La
M···Opic	2.512 (2)	2.434 (2)	2.521 (2)	2.424 (2)	2.515 (2)	2.427 (2)
M···Ocyc	2.800 (4)	2.765 (4)	2.900 (5)	2.914 (5)	2.786 (4)	2.724 (4)
M···Npic	2.756 (2)	2.669 (2)	2.788 (2)	2.697 (2)	2.731 (2)	2.663 (2)
M···Ncyc	2.916 (2)	2.855 (2)	2.955 (2)	2.895 (2)	2.944 (2)	2.916 (2)

^aNumber of contacts in parentheses.

Comparison of the two M–Megapa conformers (M = Ac, La) revealed 10 metal–ligand coordinations in I and 11 ones in II (cf. Figure ). Analysis of the distances revealed the major role of the M···O_pic interactions in the formation of the complexes, having significantly shorter distances than the other ones (M···O_cyc, M···N_pic, and M···N_cyc). Three of these four distance types reflect reasonably the difference between the ionic radii of Ac³⁺ and La³⁺. Remarkable differences appear, however, in the M···O_cyc average distances of II: in spite of the smaller ionic radius of La³⁺ (by 0.058 Å), the La···O_cyc distances are longer than the Ac···O_cyc ones. This proves that the La³⁺ ion cannot fit so effectively into the cavity of II as the Ac³⁺ ion does, which is a likely explanation for the decreased stability of this La–Megapa conformer.

The related data of the decacoordinate Macropa complexes given in Table facilitate a comparison of complex formation by the two chelators. Most of the M–ligand coordination distances are slightly shorter in the M–Macropa complexes as compared to those in M–Megapa. These slight differences may imply slightly stronger M–ligand interactions in the Macropa complexes. This feature is evaluated in more detail in the following energy analysis.

Quantitative bonding information, i.e., energetics of the electrostatic, orbital, and repulsive steric interactions, can be estimated by energy decomposition analysis (EDA) on the isolated molecule using M³⁺ and chelator^2– fragments. In this model, the theoretical interaction energy between the fragments, ΔE _int, is defined as

$${\mathrm{\Delta }E}_{\mathrm{int}}={\mathrm{\Delta }E}_{\mathrm{e}\mathrm{l}\mathrm{s}\mathrm{t}}+{\mathrm{\Delta }E}_{\mathrm{P}\mathrm{a}\mathrm{u}\mathrm{l}\mathrm{i}}+{\mathrm{\Delta }E}_{\mathrm{o}\mathrm{i}}$$ 1

where ΔE _elst corresponds to the classical electrostatic interaction between the charge distributions of the two fragments after being brought together in the complex, ΔE _Pauli is the repulsion between occupied orbitals (practically the steric repulsion), and ΔE _oi is the orbital interaction energy between the fragments in the complex, accounting for electron pair bonding, charge transfer, and polarization. The EDA results are compiled in Table .

3. Main Results of the EDA (kJ/mol) of Selected M–Megapa and M–Macropa Complexes.

	Megapa				Macropa
	I		II		δλδ,δλδ
	Ac	La	Ac	La	Ac	La
ΔE int	–4140.0	–4322.0	–4187.0	–4361.6	–4149.7	–4337.2
ΔE elst	–3245.5	–3283.2	–3239.4	–3282.8	–3303.9	–3345.3
ΔE oi	–1619.7	–1744.6	–1594.4	–1714.6	–1585.4	–1709.7
ΔE Pauli	763.6	744.8	685.7	675.3	773.9	752.7

As the EDA analysis could be performed only under isolated conditions, principally, it lacks the polarizing effects of the solvent. However, some effect of water solvent is incorporated in the geometries, which were optimized under aqueous conditions. The main observations in the data of Table are

• 1.Conformer I of Ac–Megapa exhibits a somewhat lower stability in terms of ΔE _int than conformer II. Their above-reported identical stability in terms of ΔG _SMD may be achieved by larger solvent stabilization of the former conformer.
• 2.The La complexes of Megapa are significantly more stable compared to the Ac ones in terms of ΔE _int, similarly to the reported case of La–Macropa­(H₂O)⁺ and Ac–Macropa­(H₂O)⁺. This higher stability arises from the more favorable character of all the three interactions (in eq ) between La³⁺ and both ligands. Particularly interesting is the higher ΔE _oi term in spite of the known higher covalent capability of actinides vs lanthanides due to the more diffuse valence f orbitals of the former f elements. Behind the stronger attraction interactions stand the generally shorter La–ligand distances compared to the Ac–ligand ones (cf. Table ), while behind the smaller Pauli repulsion stands a likely more favorable steric arrangement of La³⁺ in the ligand cavities.

Discussion

In this study, we pursued the development of a Macropa-derived chelator based on the expanded crown ether 1,13-diaza-[24]­crown-8 scaffold and evaluated both its ability to coordinate [²²⁵Ac]­Ac³⁺ and the kinetic inertness of the resulting complex. For the synthesis of the chelator scaffold, several reaction pathways were explored, ultimately affording the target chelator Megapa. Megapa, together with its smaller analogue Macropa, was subsequently subjected to comparative evaluation with respect to [²²⁵Ac]­Ac³⁺ radiolabeling efficiency and the relative kinetic inertness of the resulting complexes [²²⁵Ac]­Ac–Megapa and [²²⁵Ac]­Ac–Macropa.

Megapa exhibited a markedly lower [²²⁵Ac]­Ac³⁺ incorporation efficiency than Macropa and a reduced stability of the [²²⁵Ac]­Ac–Megapa complex against transchelation in complex-challenge experiments using human serum and excess La³⁺. Further, thermodynamic stability studies on the [²²⁵Ac]­Ac–Megapa complex homologue La–Megapa supported these results, indicating a lower thermodynamic stability of La–Megapa compared to La–Macropa.

As suitable crystals of the La–Megapa complex could not be obtained to elucidate the structure of [²²⁵Ac]­Ac–Megapa, complementary insights into the structure and bonding properties of the La³⁺ and Ac³⁺ complexes of Macropa and Megapa were derived from quantum chemical calculations.

These calculations gave two preferred conformers with comparable stability for Ac–Megapa (Figure ), whereas only one preferred conformer was obtained for La–Macropa, La–Megapa, and Ac–Macropa. This is in agreement with the experimental analytical HPLC results obtained for [²²⁵Ac]­Ac–Megapa and La–Megapa, indicating two different species formed in the case of [²²⁵Ac]­Ac–Megapa, whereas only one species could be detected in the case of its lanthanum analog (Figures B vs S26). The results obtained for the Macropa complexes of both metals are in good agreement with the structure data reported in the literature, demonstrating the reliability of the approach. The existence of two different comparably stable conformers in the case of Ac–Megapa could explain the unexpected observations during the [²²⁵Ac]­Ac³⁺ radiolabeling reactions of the ligand, showing the formation of two different species being in equilibrium (Figure B). Further, the calculations suggest comparable bond lengths of Ac–Macropa and Ac–Megapa as well as La–Macropa and La–Megapa complexes (Table ). Alike, the energy decomposition analyses did not show significant differences between Ac–Macropa and Ac–Megapa as well as La–Macropa and La–Megapa (Table ). Thus, the observed differences in the kinetic inertness of [²²⁵Ac]­Ac–Macropa and [²²⁵Ac]­Ac–Megapa are likely attributable to factors beyond structural considerations, specifically molecular interactions occurring in solution.

This conclusion is corroborated by the experimental observation that both, La–Megapa and [²²⁵Ac]­Ac–Megapa, complexes exhibit a considerable instability in solution. A possible explanation for this behavior, despite the binding interactions of the isolated molecules being comparable to those of the corresponding Macropa complexes as indicated by EDA analysis, is the larger ring size of the Megapa chelator. Computational analysis revealed that this larger ring leads to an increased distance between the central metal ion and the ring oxygen atoms. This spatial arrangement may enhance the likelihood of interactions between the ring oxygens or the central metal ion with protons, other ions, or biomolecules (as in the case of the serum stability experiments), thereby destabilizing the complex and promoting metal ion exchange or decomplexation. Previous studies have established that a mismatch between the cavity size of a chelator and its central metal ion adversely affects complex stability. , Interestingly, in this case, the observed destabilization is pronounced, despite conformer I showing donor–metal distances comparable to those of the corresponding Macropa complex. This suggests that additional factors may contribute to the observed complex instability. Notably, it has been reported that providing more donor atoms than required for complex formation can also destabilize a complex, , likely due to dynamic donor exchange. In the present system, one to two ring oxygen atoms of Megapa were found not to participate in La³⁺/Ac³⁺-coordination, which could further contribute to the observed reduced stability of the resulting Megapa complexes as the noncoordinating O donors are located at the periphery of the complex and in this way are more accessible to strong interactions with surrounding ions and molecules. These interactions could further destabilize the complex in solution.

Despite the rationale presented in the introduction suggesting that larger ring sizes of Macropa-based chelators could confer higher complex stability (by potentially better fitting the large ionic radius of the metal ion and a more appropriate saturation of the coordination sphere of the Ac³⁺ ion) and despite the calculated structural properties of the Ac–Megapa complex, our results demonstrate that the enlarged cavity of Megapa does not produce the anticipated effect. Instead, it renders the complex more susceptible to challenge, resulting in a decreased kinetic inertness. This observation is in line with the results of the radium study discussed before. Even though the complexation chemistry of Ra²⁺ radioisotopes is considerably differing from that of Ac³⁺, our results in summary imply that increasing the ring size of Macropa-based chelators is not a viable strategy to increase the kinetic inertness of the corresponding complexes with large-sized metal ion radioisotopes.

A reason for this might be the high flexibility of the ring system, which is inherent to its large size and the fact that not all heteroring atoms participate in radiometal coordination, as revealed by our theoretical calculations. This high flexibility of the system may, therefore, enable greater interaction with destabilizing species present in solution, thereby favoring dissociation of the metal complex.

Megapa thus contributes to the expanding body of developments in recent years. , Although none have yet surpassed the exceptional stability and inertness of the [²²⁵Ac]­Ac–Macropa complex, each effortincluding Megapaprovides valuable insights that advance our understanding of [²²⁵Ac]­Ac³⁺ chelation chemistry and guide future design strategies.

Experimental Methods

Chemical Syntheses, Radiolabeling, and Determination of Complex Stability/Inertness

General: Chemicals, Radiochemicals, and Chemicals for Radiolabeling Purposes

All chemicals were purchased from commercial sources in at least analytical grade quality and used without further purification unless otherwise stated. Methanol, ethanol, cyclohexane, sodium borohydride, 32% sodium hydroxide solution, thionyl chloride, N,N-dimethylaminopyridine (DMAP), trifluoroacetic acid (TFA), and TRIS were purchased from Carl Roth (Karlsruhe, Germany); dichloromethane (CH₂Cl₂), acetonitrile, dimethylformamide (DMF), and ethyl acetate (EtOAc) from Häberle Labortechnik (Lonsee-Ettlenschieß, Germany),; dry tert-butanol, methylsulfonic acid anhydride, Kryptofix 2.2, (+)-sodium ascorbate, N,N-diisopropylethylamine (DIPEA), triisopropylsilane (TIS), l-ascorbic acid, LaCl₃ × 7H₂O (99.999% trace metals basis), and human serum (pooled serum from male donors, AB, clotted whole blood) from Merck KGaA (Darmstadt, Germany); 2,2′-((oxybis­(ethane-2,1-diyl))­bis­(oxy))­diacetic acid, 2,2′-((oxybis­(ethane-2,1-diyl))­bis­(oxy))­bis­(ethane-1-amine), 3,6,9,12-tetraoxatetradecandiacid, 3,6,9,12-tetraoxatetradecane-1,14-diamine, and 2,6-pyridindicarboxylic acid from BLD Pharmtech GmbH (Reinbek, Germany); lithium aluminum hydride from TCI Deutschland GmbH (Eschborn, Germany); and 2,5-dihydroxybenzoic acid, lanthanum triflate, dry pyridine, and dry tetrahydrofurane (THF) from Thermo Fischer Scientific (Life Technologies GmbH, Darmstadt, Germany). Deionized water was produced by the water purifier Aquinity² P10 from membraPure GmbH (Berlin, Deutschland). [²²⁵Ac]­AcCl₃ in 0.1 M HCl for ²²⁵Ac-radiolabeling reactions was obtained from Van Overeem Nuclear BV (Breda, The Netherlands) or Eckert & Ziegler Nuclitec (Braunschweig, Germany). H₂O (Tracepur quality), hydrochloric acid (30%, Suprapur quality), and sodium hydroxide (30%, Suprapur quality) for radiolabeling reactions were purchased from Merck (Darmstadt, Germany).

Instrumentation

Analytical HPLC-ESI-MS and HPLC: For analytical HPLC-MS, a Waters ARC system with an Acquity QDa mass detector was used, equipped with an XBridge C18 column (3.5 μm, 4.6 mm × 50 mm), Waters, Eschborn, Germany). For analytical HPLC chromatography, a Chromolith Performance (RP-18e, 100 mm × 4.6 mm, Merck, Darmstadt, Germany); for semipreparative analyses and purifications, a Chromolith (RP-18e, 100 mm × 10 mm, Merck, Darmstadt, Germany); and for preparative purifications, an XBridge Peptide BEH C18 (130 Å, 5 μm, 10 mm × 250 mm) column (Waters, Eschborn, Germany) were used together with a Dionex UltiMate 3000 system (Dionex, Dreieich, Germany). For radioanalytical chromatography, a Dionex UltiMate 3000 system (Dionex, Dreieich, Germany) equipped with a Raytest GABI Star radioactivity detector (Straubenhardt, Germany) was used together with a Chromolith Performance (RP-18e, 100-4.6 mm, Merck, Darmstadt, Germany) column. All analytical and semipreparative operations were performed with a flow rate of 4 mL/min using H₂O (supplemented with 0.1% TFA) and acetonitrile (also supplemented with 0.1% TFA) as solvents. Preparative HPLC was performed with a flow rate of 6 mL/min. Reverse phase flash column chromatography was performed on an automated Puriflash XS530 system (Interchim, Montluçon, France) using C-18 columns of appropriate size with MeCN/H₂O + 0.1% formic acid as the eluent using a gradient from 0 to 100% within 50–70 column volumes. Nuclear magnetic resonance (NMR) spectroscopy was carried out on a 500 MHz Varian NMR System or a 300 MHz MERCURYplus NMR spectrometer (Agilent Technologies, Waldbronn, Germany). MALDI-TOF-MS: Matrix-assisted laser desorption/ionization (MALDI) time-of-flight mass spectra were obtained utilizing a Bruker Daltonics Microflex spectrometer (Bruker, Bremen, Germany), linear acquisition mode, and positive ion source. HR-MALDI-TOF-MS was performed on a Bruker timsTOF fleX IM-Q-TOF spectrometer (Bremen, Germany). HR-ESI-MS: For high-resolution electrospray ionization mass spectroscopy (HR-ESI-MS), a Thermo Finnigan LTQ FT Ultra Fourier Transform Ion Cyclotron Resonance (Dreieich, Germany) mass spectrometer or a Bruker microTOF-Q II ESI spectrometer (Bremen, Germany) was used. Activities were measured using an ISOMED 2010 activimeter (Nuvia Instruments, Dresden, Germany). iTLC chromatography was performed by using iTLC-SG strips (Agilent Technologies, Waldbronn, Germany) with aqueous citrate buffer (0.05 M, pH 5) as the mobile phase. The analysis of the iTLC strips was typically performed 6 h after development on a Scan-RAM iTLC scanner from LabLogic (Koblenz, Germany). γ-Counting measurements were performed using a 2480 Wizard² gamma counter system from PerkinElmer (Rodgau, Germany) at energy windows of 70–270 keV for ²²¹Fr, 362–462 keV for ²¹³Bi, or the whole energy spectrum.

Chelator Syntheses

Macropa (1)

11 (51.4 mg, 79.7 μmol) was dissolved in a mixture of TFA/TIS/H₂O (95:2.5:2.5, v/v/v, 2 mL) and stirred for 24 h. The volatile components were removed under vacuum, the residue was dissolved in H₂O + 0.1% TFA (6 mL), and the product purified by preparative HPLC using a gradient of 10–0% acetonitrile (0–8 min), 0–100% acetonitrile (8.0–9.0 min), and 100–0% acetonitrile (9.0–9.1 min) (R _t : 2.9 min). The product was obtained as a white solid in 74% yield (31.3 mg, 58.7 μmol). ¹H NMR (500 MHz, DMSO-d ₆): δ 8.12 (t, 2H, ³ J = 7.6 Hz, 4-Py-H), 8.09 (dd, 2H, ³ J = 7.8 Hz, ⁴ J = 1.4 Hz, 3-Py-H), 7.80 (dd, 2H, ³ J = 7.4 Hz, ³ J = 1.4 Hz, 5-Py-H), 4.72 (s, 4H, Py-CH ₂–N), 3.86 (t, 8H, ³ J = 5.1 Hz, O–CH ₂–CH₂–N), 3.56 (m, 16H, O–C₂ H ₄–O–CH₂–CH ₂–N). ¹³C NMR (125 MHz, DMSO-d ₆): δ 165.4 (Py-CO₂H), 151.0 (6-Py-C), 147.7 (2-Py-C), 139.2 (4-Py-C), 127.8 (5-Py-C), 124.4 (3-Py-C), 69.4 (O–C ₂H₄–O), 64.5 (N–CH₂ CH₂–O), 56.8 (Py-CH₂–N), 53.4 (N–CH₂CH₂–O). HR-ESI-MS (m/z): for [M + H]⁺ (calcd), 533.2628 (533.2606).

Megapa (3)

10 (165 mg, 226 μmol) was dissolved in a mixture of TFA/TIS/H₂O (95:2.5:2.5, v/v/v, 4 mL) and stirred for 16 h at ambient temperature. Diethyl ether (10 mL) was added, and the volatile components were removed under vacuum. The residue was dissolved in H₂O + 0.1% formic acid (7 mL), and the product purified by flash chromatography (C-18 column, flow: 15 mL/min, gradient: 0.0–1.25 min: 5% MeCN, 1.25–58.1 min: 5–35% MeCN, R _t = 8 min). 1 was obtained as a white solid after lyophilization in 42% yield (59.3 mg, 95.5 μmol). ¹H NMR (500 MHz, acetonitrile-d ₃): δ 8.31 (m, 2H, 3-Py-H), 8.23 (m, 2H, 4-Py-H), 7.87 (m, 2H, 5-Py-H), 4.81 (m, 2H, Py-CH ₂N), 3.95 (m, 8H, N–CH₂–CH ₂–O), 3.63–3.77 (m, 24H, N–CH ₂–CH₂–O–CH ₂–CH ₂–O); ¹³C NMR (125 MHz, acetonitrile-d ₃): δ 167.8 (Py-CO₂H), 150.9 (6-Py-C), 147.9 (2-Py-C), 140.9 (4-Py-C), 129.1 (5-Py-C), 126.3 (3-Py-C), 70.6 and 70.5 (O–CH₂–CH₂–O), 65.0 (N–CH₂–CH₂–O), 58.4 (Py-CH₂–O), 55.1 (N–CH₂–CH₂–O); HR-ESI-MS (m/z): for [M + H]⁺ (calcd), 621.3146 (621.3130), (m/z): for [M + Na]⁺ (calcd): 643.2954 (643.2950).

7

The synthesis was adapted from ref. To a solution of 6 (5.00 g, 22.5 mmol) in methanol (200 mL) was added sulfuric acid (3 mL), and the solution was heated to 70 °C for 16 h. After colling, a saturated NaHCO₃ solution (100 mL) was added slowly, and the solvent was removed under reduced pressure. The residue was taken up in CH₂Cl₂ (50 mL) and the organic phased washed with water (20 mL) and saturated NaCl solution (30 mL) before being dried over NaSO₄. The solvent was removed, and the product was obtained as a colorless oil in 42% yield (2.37 g, 9.45 mmol). ¹H NMR (300 MHz, CDCl₃): δ 4.17 (s, 4H, O₂C–CH ₂–O), 3.75 (s, 6H, H ₃ C–O₂C), 3.71 (m, 8H, O–C₂ H ₄–O).

8

A solution of 7 (2.60 g, 10.4 mmol) and 4 (1.99 g, 10.4 mmol) in methanol (200 mL) was stirred at ambient temperature for 20 days. The solvent was removed under reduced pressure, and the residue was dissolved in CH₂Cl₂ (100 mL). The organic phase was washed thrice with NaOH solution (1 M, 50 mL), and the combined aqueous phases were washed with CH₂Cl₂ (50 mL). The combined organic phases were washed three times with HCl solution (1 M, 50 mL) and again washed three times with CH₂Cl₂ (50 mL). The solvent of the combined organic phases was removed under reduced pressure, giving the product in form of a colorless oil which was not further purified in a yield of 82% (1.33 g, 3.52 mmol). ¹H NMR (500 MHz, CDCl₃): δ 7.75 (s, 2H, CONH–CH₂), 4.08 (s, 4H, O–CH ₂CONH), 3.71 (m, 8H, C₂ H ₄–O–CH₂CONH), 3.64 (s, 8H, C₂ H ₄–O–C₂H₄–NH), 3.61 (t, 4H, ³ J = 4.8 Hz, NH–CH₂CH ₂), 3.52 (q, 4H, ³ J = 5.3 Hz, NH–CH ₂CH₂); ¹³C NMR (125 MHz, CDCl₃): δ 171.6 (CH₂ CONH), 71.0 (CH₂ CH₂–O–CH₂CONH), 70.4 (CH₂CH₂–O–CH₂CH₂–NH), 70.3 (CH₂ CH₂–O–CH₂CH₂–NH), 70.1 (CH₂CH₂–O–CH₂CONH), 69.6 (O–CH₂CH₂–NH), 39.2 (O–CH₂ CH₂–NH); HR-MALDI-MS (m/z): for [M + H]⁺ (calcd), 379.2076 (379.2075).

9

LiAlH₄ (0.761 g, 20.0 mmol) was suspended in wet ice-cooled THF (30 mL) before a solution of 8 (1.32 g, 3.76 mmol) in THF (20 mL) was added for 10 min. The mixture was refluxed for 16 h. Afterward, the solution was cooled using wet ice before water (5 mL) was carefully added, followed by NaOH solution (15%, 10 mL), water (10 mL), and finally NaOH (32%, 10 mL). The suspension was extracted four times with CH₂Cl₂ (50 mL), the organic phases dried over NaSO₄, and the solvent was removed under reduced pressure to give 9 as a brown oil in 43% yield (0.564 g, 1.49 mmol). ¹H NMR (500 MHz, CDCl₃): δ 3.65 (m, 8H, NH–C₂H₄–O–CH₂CH ₂–O), 3.61 (m, 16H, NH–CH₂CH ₂–O–CH ₂CH₂–O), 2.79 (t, 8H, ³ J = 5.0 Hz, NH–CH ₂CH₂–O); ¹³C NMR (125 MHz, CDCl₃): δ 70.8 (NH–C₂H₄–O–CH₂ CH₂–O), 70.6 (NH–C₂H₄–O–CH₂CH₂–O or NH–CH₂ CH₂–O), 70.4 (NH–C₂H₄–O–CH₂CH₂–O or NH–CH₂ CH₂–O), 49.1 (NH–CH₂CH₂–O); HR-ESI-MS (m/z): for [M + H]⁺ (calcd): 351.2493 (351.2490).

10

9 (101 mg, 288 μmol) and K₂CO₃ (125 mg, 902 μmol, 3.1 equiv) were suspended in acetonitrile (5 mL). To this mixture was added 12 (174 mg, 606 μmol, 2.1 equiv) before it was allowed to react for 16 h at ambient temperature. The suspension was filtered, and the product in solution was purified by preparative HPLC (180 μL per injection, gradient: 10–100% MeCN in 8 min, R _t = 5.4 min). 10 was obtained as a colorless oil in 78% yield (165 mg, 226 μmol). ¹H NMR (500 MHz, DMSO-d ₆): δ 8.09 (t, 2H, ³ J = 7.7 Hz, 4-Py-H), 8.02 (d, 2H, ³ J = 7.7 Hz, 3-Py-H), 7.78 (d, 2H, ³ J = 7.7 Hz, 5-Py-H), 4.68 (s, 4H, Py-CH ₂–N), 3.85 (t, 8H, ³ J = 5.0 Hz, N–CH₂–CH ₂–O), 3.57 (m, 16H, O–C₂ H ₄–O), 3.53 (t, 8H, ³ J = 4.6 Hz, N–CH ₂–CH₂–O), 1.56 (s, 18H, C­(CH ₃)₃); ¹³C NMR (125 MHz, DMSO-d ₆): δ 163.0 (O₂ C-Py), 158.3 (q, ² J = 35.8 Hz, F₃C–CO₂H), 151.2 (6-Py-C), 148.4 (2-Py-C), 138.9 (4-Py-C), 127.9 (5-Py-C), 124.5 (3-Py-C), 115.7 (q, ¹ J = 292.2 Hz, F₃ C–CO₂H), 81.8 (C­(CH₃)₃), 69.7 and 69.6 (O–CH₂–CH₂–O), 64.7 (N–CH₂–CH₂–O), 56.6 (Py-CH₂–O), 53.2 (N–CH₂–CH₂–O), 27.7 (C­(CH₃)₃); HR-ESI-MS (m/z): for [M + H]⁺ (calcd): 733.4390 (733.4382).

11

Kryptofix 2.2 (91.4 mg, 351 μmol) and DIPEA (180 μL, 1.06 mmol, 3.0 equiv) were dissolved in CH₂CL₂ (30 mL). To this mixture was added 12 (201 mg, 699 μmol, 2.0 equiv) before the mixture was allowed to react for 13 days at ambient temperature. The suspension was filtered, 1 g of Celite was added, and the solvent was removed under reduced pressure. The product was purified by flash chromatography (C-18 column, flow: 15 mL/min, gradient: 0.0–1.25 min: 5% MeCN, 1.25–58.1 min: 5–35% MeCN, R _t = 30 min) and obtained as a yellowish oil in 21% yield (47.3 mg, 73.4 μmol). ¹H NMR (500 MHz, CDCl₃): δ 8.04 (d, ³ J = 7.8 Hz, 2H, 3-Py-H), 7.95 (t, ³ J = 7.8 Hz, 2H, 4-Py-H), 7.70 (d, ³ J = 7.9 Hz, 2H, 5-Py-H), 4.79 (s, 4H, Py-CH ₂–N), 3.94 (t, ³ J = 4.5 Hz, 12H, N–CH₂CH ₂–O), 3.68 (m, 24H, N–CH ₂CH₂–O–C₂ H ₄), 1.61 (s, 18H, OC­(CH ₃)₃); ¹³C NMR (125 MHz, CDCl₃): δ 163.25 (Py-CO₂), 160.71 (2-Py-C), 149.07 (6-Py-C), 139.19 (4-Py-C), 128.02 (5-Py-C), 125.26 (3-Py-C), 83.28 (C­(CH₃)₃), 70.44 (O–C ₂H₄–O), 65.13 (N–CH₂ CH₂–O), 54.89 (Py-CH₂–N), 54.61 (N–CH₂CH₂–O), 28.17 (C­(CH₃)₃); HR-ESI-MS (m/z): for [M + H]⁺ (calcd): 645.3847 (645.3858).

12

To a solution of 14 (1.37 g, 6.53 mmol) in CH₂Cl₂ (50 mL) were added DIPEA (2.27 mL, 13.1 mmol) and methanesulfonic anhydride (1.78 g, 10.2 mmol), and the mixture was stirred at ambient temperature for 16 h. The reaction mixture was washed thrice with saturated NaCl solution (50 mL) and dried over NaSO₄, and the solvent was evaporated under reduced pressure. Column chromatography (2% acetone in CH₂Cl₂, R _f = 0.8) gave the product as a yellow solid in 69% yield (1.30 g, 4.53 mmol). ¹H NMR (500 MHz, CDCl₃): δ 8.00 (d, 1H, ³ J = 7.7 Hz, 3-Py-H), 7.87 (t, 1H, ³ J = 7.8 Hz, 4-Py-H), 7.60 (d, 1H, ³ J = 7.6 Hz, 5-Py-H), 5.41 (s, 2H, Py-CH ₂OMs), 3.18 (s, 3H, OSO₂–CH ₃) 1.61 (s, 9H, C­(CH ₃)₃); ¹³C NMR (125 MHz, CDCl₃): δ 163.6 (O₂ C-Py), 154.4 (6-Py-C), 149.4 (2-Py-C), 138.2 (4-Py-C), 124.8 (5-Py-C), 124.6 (3-Py-C), 82.7 (C­(CH₃)₃), 71.4 (Py-CH₂–O), 38.4 (O₃SCH₃), 28.2 (C­(CH₃)₃); HR-ESI-MS (m/z): for [M + Na]⁺ (calcd), 310.0720 (310.0725).

13

2,6-Pyridinedicarboxylic acid (10.0 g, 59.9 mmol) was dissolved in SOCl₂ (40 mL, 551 mmol) and DMF (2 mL, 26.0 mmol) and heated to reflux for 2 days. The volatile components were removed under reduced pressure, and the residue was dissolved in CH₂Cl₂ (20 mL) and added dropwise to a wet ice-cooled solution of tert-butyl alcohol (30 mL, 320 mmol) and DMAP (1.24 g, 10.2 mmol) in pyridine (60 mL) and CH₂Cl₂ (20 mL). After overnight reaction, the volatile components were removed under reduced pressure, and the product was purified by column chromatography (EtOAc/cyclohexane, 1:9 (v/v), R _f = 0.2). The product was obtained as a colorless crystalline solid in 72% yield (12.2 g, 43.6 mmol). ¹H NMR (500 MHz, CDCl₃): δ 8.18 (d, 2H, ³ J = 7.8 Hz, 3,5-Py-H), 7.92 (t, 1H, ³ J = 7.8 Hz, 4-Py-H), 1.64 (s, 18H, O–C­(CH ₃)₃); ¹³C NMR (125 MHz, CDCl₃): δ 163.6 (Py-CO₂), 149.8 (2,6-Py-C), 138.0 (4-Py-C), 127.3 (3,5-Py-C), 82.8 (O–C­(CH₃)₃), 28.2 (O–C­(CH₃)₃); HR-ESI-MS (m/z): for [M + Na]⁺ (calcd), 302.1363 (302.1368).

14

To a wet ice-cooled solution of 13 (3.51 g, 12.6 mmol) in methanol (50 mL) was portion wise added NaBH₄ (1.91 g, 50.4 mmol; 4 equal portions in one-hour intervals). The progress of the reaction was monitored by TLC (5% methanol in CH₂Cl₂, r _f = 0.4). The reaction was quenched with water (50 mL), and the aqueous solution was extracted three times with CH₂Cl₂ (50 mL). The combined organic phases were washed with a saturated NaCl solution (50 mL) and dried over NaSO₄, and the solvent was removed under reduced pressure. Column chromatography (2% methanol in CH₂Cl₂, R _f = 0.3) gave 14 as a colorless crystalline solid in 48% yield (1.28 g, 6.11 mmol). ¹H NMR (500 MHz, CDCl₃): δ 7.88 (d, 1H, ³ J = 7.6 Hz, 3-Py-H), 7.76 (t, 1H, ³ J = 7.7 Hz, 4-Py-H), 7.46 (d,1H, ³ J = 8.0 Hz, 5-Py-H), 4.82 (s, 2H, Py-CH ₂OH), 4.32 (br s, 1H, OH), 1.58 (s, 9H, C­(CH ₃)₃); ¹³C NMR (125 MHz, CDCl₃): δ 164.0 (O₂ C-Py), 160.3 (6-Py-C), 148.3 (2-Py-C), 137.6 (4-Py-C), 123.6 (5-Py-C), 123.3 (3-Py-C), 82.4 (C­(CH₃)₃), 64.5 (CH₂OH), 28.1 (C­(CH₃)₃); HR-ESI-MS (m/z): for [M + Na]⁺ (calcd), 232.0944 (232.0950).

La–3

To a solution of 3 (18.6 mg, 30 μmol) in NaCl solution (0.1 M, 5 mL) was added a solution of LaCl₃ × 7H₂O (10.0 mg, 27 μmol) in the same solvent (1.35 mL), and the pH of the mixture was adjusted to 2.3 by adding HCl (0.1 M, 2 mL). A NaOH solution (0.1 M) was added in portions of 50 μL to a final volume of 2.55 mL, yielding a pH of 6.0. The obtained solution was used for HR mass, ¹H NMR, and ¹³⁹La NMR analyses. ¹³⁹La NMR was carried out at 70 MHz using LaCl₃ as the external standard. ¹H NMR (500 MHz, D₂O): δ 8.13 (m, 4H, 3-Py-H, 4-Py-H), 7.73 (m, 2H, 5-Py-H), 2.5–4.2 (br m, 14H, ring protons). HR-ESI-MS (m/z): for [M]⁺ (calcd), 757.1975 (757.1965).

Solution Thermodynamics

Potentiometric measurements were performed on a Metrohm 905 Titrando instrument equipped with a BlueLine 17 pH electrode (Schott Instruments) and Metrohm Dosino 800 dosing units using standardized 0.1 M NaOH and 0.1 M HCl. Titrant delivery and data acquisition were controlled automatically with tiamo 2.3 (Metrohm). The temperature was maintained at 25 °C using a water-jacketed titration vessel connected to a Lauda Ecoline E300 thermostat. To exclude CO₂, the solutions were blanketed with a gentle argon flow throughout the experiments. All solutions were prepared in standardized 0.1 M NaCl to maintain a constant ionic strength. The electrode standard potential was determined prior to each experiment by titration of a HCl solution and subsequent evaluation with GLEE. Ligand and metal-ion stock solutions were prepared at nominal concentrations of 2 and 20 mM, respectively. The ligand concentration was determined by ¹H NMR spectroscopy using an internal standard (hexafluoropropionic acid in water). Metal-ion concentrations were established by complexometric titration with 0.01 M Na₂H₂EDTA in the presence of xylenol orange. Continuous potentiometric titrations were carried out with standardized, carbonate-free 0.1 M NaOH on aqueous solutions containing 1.57 mM Megapa. For complex stability measurements, the same ligand solution was used after addition of 1.0 equiv of metal ions. Direct titrations of the free ligand and the corresponding complex were initiated after an initial equilibration period of 15–30 min; data points were collected using a drift criterion of 0.1 mV min^–1 and an equilibration time of 2–5 min between successive additions. Protonation and complex stability constants were calculated with Hyperquad2013, using fixed total concentrations and a fixed pK _W value of 13.78, without refinement during fitting. Because precipitation occurred above pH 10, data points in this range were excluded from the determination of complex stability constants for [La–Megapa]⁺. Species distribution diagrams and pM values were generated by using HySS.

²²⁵Ac-Radiolabeling of Macropa and Megapa compared to [²²⁵Ac]­Ac–Macropa and [²²⁵Ac]­Ac–Megapa

Caution! Due to the α-, β-, and γ-particle emission from ²²⁵Ac and its daughters, radioisotopes represent a serious health hazard. All studies with ²²⁵Ac were conducted in a radiation laboratory (controlled area) with appropriate shielding and protection against radionuclide release.

For radiolabeling with [²²⁵Ac]­Ac³⁺, [²²⁵Ac]­AcCl₃ in 0.1 M HCl was obtained from Van Overeem Nuclear BV (Breda, The Netherlands) or Eckert & Ziegler Nuclitec (Braunschweig, Germany).

Radiolabeling and determination of the complex formation efficiency: To a solution of Macropa or Megapa (typically: 10 nmol, tested range: 0.1–50 nmol) in H₂O (5–25 μL) were added TRIS buffer (2 M, pH 7.0, 50 μL), followed by [²²⁵Ac]­AcCl₃ in HCl (0.1 M, 5–10 μL, 44–52 kBq) and H₂O (15–40 μL) to arrive at a final overall volume of 100 μL. The pH was measured and typically found to be between 7.6 and 8.4. The solutions were warmed to 37 °C for 15 min at 500 rpm. 10 μL of these solutions were analyzed by analytical radio-HPLC and radio-iTLC.

Radiolabeling for stability/kinetic inertness testing of [²²⁵Ac]­Ac–Macropa and [²²⁵Ac]­Ac–Megapa: To a solution of Macropa or Megapa (60 nmol) in H₂O (30 μL) were added TRIS buffer (2 M, pH 7.0, 300 μL), followed by [²²⁵Ac]­AcCl₃ in HCl (0.1 M, 10–26 μL, 439–807 kBq) and H₂O (247–260 μL) to arrive at a final overall volume of 600 μL. The pH was measured and found to be between 7.3 and 7.4. The solutions were warmed to 37 °C for 15 min at 500 rpm. 10 μL of these solutions were analyzed by analytical radio-HPLC.

Radio-HPLC analyses were performed directly after radiolabeling using a gamma detector integrated over the full energy spectrum. Further, fraction collection from these analytical radio-HPLC runs (five fractions per minute), followed by gamma counting, was carried out. Gamma counting was performed using gamma energy windows of 70–270 keV for ²²¹Fr (E _γ: 218 keV) and 362–462 keV for ²¹³Bi (E _γ: 440 keV). Gamma counting of the collected radio-HPLC fractions as well as analysis of the developed iTLC strips was conducted with a delay of 6 h to allow the majority of ²²¹Fr and ²¹³Bi initially present during the labeling experiments to decay.

Determination of the Stability of [²²⁵Ac]­Ac–Macropa and [²²⁵Ac]­Ac–Megapa in Human Serum

A sample of the respective radiolabeled chelator (250 μL, 183–336 kBq) was added to human serum (500 μL) which was prewarmed and kept at 37 °C during the course of the experiments. At defined time-points (t = 15 min, 1 h, 2 h, 3 h, 1 day, 6 days, and 7 days), 40 μL of the mixture was added to 40 μL of ice-cold ethanol, further cooled on ice for 5 min, and centrifuged at 13,000 rpm for 90 s. The supernatant was collected, the activity of the supernatant and the precipitate was measured, and the supernatant was analyzed by analytical radio-HPLC. Experiments were performed thrice for both complexes.

Determination of the Inertness of [²²⁵Ac]­Ac–Macropa and [²²⁵Ac]­Ac–Megapa against Complex Challenge with La³⁺

A sample of the respective radiolabeled chelator (250 μL, 183–336 kBq) was added to a solution containing a 4000-fold excess of lanthanum triflate in H₂O (1 M, 100 μL) which was prewarmed and kept at 37 °C during the course of the experiments. At defined time-points (t = 15 min, 1 h, 2 h, 3 h, 1 day, 6 days, and 7 days), 40 μL of the mixture was analyzed by analytical radio-HPLC. Experiments were performed thrice for both complexes.

Computational Methods

The most stable conformers in aqueous solution were determined using a three-step molecular mechanics (MM)/DFT procedure: (1) An initial set of low-energy conformers was obtained by a conformational search at the cost-effective MM level. (2) The MM conformers falling into an appropriate energy window were prefiltered with simple DFT calculations. (3) The final structures and energies were evaluated at a sophisticated DFT level applied on a small set of the lowest-energy structures from step (2).

The MM conformational search was carried out with the MacroModel code , using the OPLS2005 force field as implemented in the Schrödinger suite. The generation of conformers was performed using the mixed torsional/low-mode algorithm running up to 5000 steps. Default values were used for the probability of a torsion rotation/molecule translation (0.5) as well as for the minimum and maximum distances for low-mode moves (3.0 and 6.0 Å, respectively). For energy minimizations, the Polak–Ribiere conjugate gradient (PRCG) algorithm with iteration steps up to 10,000 and a gradient convergence criterion of 0.01 kJ/(Å mol) was used. Redundant conformers were eliminated on the basis of a distance threshold of 0.5 Å between any pair of heavy atoms. Visual analysis of the obtained conformers was performed with the Maestro 14.2 module of the Schrödinger suite.

The initial Cartesian coordinates of the δλδ,δλδ La–Macropa complex were taken from a recent paper. The initial structure of the La–Megapa complex was created manually. As the available MM solvation model lacked parameters for La, conformational searches were performed on the isolated complexes. The results on the La–Macropa complex proved the applicability of this approach (vide supra). The optimized La–Macropa and La–Megapa structures served as the initial structures for the Ac complexes after replacing La with Ac.

As the MM force field does not account properly for metal–ligand coordination interactions, for the sampling of the complex conformers, the relevant La–ligand (picolinate and macrocycle donors) distances were constrained to the following ranges:

La–Macropa: 2× La···O_pic: 2.4 ± 0.2 Å; 2× La···N_pic: 2.7 ± 0.2 Å; 4× La···O_cyc: 2.9 ± 0.4 Å.

La–Megapa: 2× La···O_pic: 2.4 ± 0.2 Å; 2× La···N_pic: 2.7 ± 0.2 Å; 6× La···O_cyc: 2.9 ± 0.4 Å.

Prefiltering of the MM conformers (a several hundred within a window of 35 kJ/mol) was performed with the low-cost B97-3c composite DFT method in aqueous solution using the conductor-like polarizable continuum model (CPCM). This composite method is based on the B97 exchange–correlation functional and includes D3 dispersion correction with three-body contribution, a short-range bond length correction, and a stripped-down triple-ζ basis. The lowest-energy B97-3c structures within an energy window of about 15 kJ/mol were selected for more sophisticated DFT calculations.

Our final DFT level consisted of the PBE0 exchange correlation functional, , a high integration grid (Orca keywords: IntAccX 5,5,5; GridX 3,3,4), VeryTight optimization convergence criteria, the D4 dispersion correction, , the CPCM solvation model, and the resolution-of-the-identity (RI) approximation for the four-index electron repulsion integrals. Because of the heavy atoms in the complexes, a relativistically recontracted ZORA-def2-TZVPP basis set was utilized, where the Zero-Order Regular Approximation (ZORA) accounts for scalar relativistic effects. The respective all-electron scalar relativistic SARC basis sets were used for La and Ac. The numbers of treated structures in the three steps are provided in Table S1 of the Supporting Information.

The minimum characteristics of the optimized PBE0 structures were confirmed by analytical frequency calculations. The thermochemical data were evaluated at 298 K and 1354 atm (mimicking the condensed phase, as derived from p = ρwRT, where ρ_w = 997.02 kg/m³ is the experimental density of liquid water at 298 K) using the rigid-rotor approximation and neglecting electronic contributions.

A more sophisticated continuum solvation model based on the quantum mechanical charge density of the solute molecule interacting with a continuum description of the solvent (SMD) by Truhlar and co-workers was also considered, taking over those radii and non-electrostatic terms for the CPCM calculation. Based on the reported often failing geometry optimizations at the latter level, only single-point SMD calculations were performed on the CPCM optimized geometries. They were extended with the thermal contributions from the above-mentioned (CPCM) frequency analyses, resulting in the tabulated ΔG _SMD data. The DFT computations have been performed with the Orca 6.0 version. −

The relative energies of selected low-energy structures obtained from the different computational levels are compared in Tables S2 and S3.

The energy decomposition analyses were performed with the Amsterdam Density Functional package (ADF2020 , ) utilizing the PBE0 exchange–correlation functional , and the D4 dispersion correction. , The scalar relativistic effects were taken into account using the ZORA model. The TZ2P all-electron basis consisted of uncontracted sets of Slater-type orbitals (STOs) optimized for use with ZORA. The small-core approximation was applied freezing the inner orbitals of the non-hydrogen elements (Acup to 4f, Laup to 4d, C and O 1s). An auxiliary set of s, p, d, f, and g STOs was used to fit the molecular density and to represent the Coulomb and exchange potentials. Based on the closed-shell character of the complexes, the spin-restricted formalism was applied, and spin–orbit interactions were neglected on the basis of their reported marginal character in related complexes. The molecular geometries were optimized in aqueous solution using the COSMO solvation model. The Morokuma–Ziegler EDA , was performed on these optimized geometries, but with the absence of the solvation model due to technical reasons.

Conclusions

In this work, we describe our attempts to develop a new chelator based on Macropa for the stable complexation of Ac³⁺ in order to make TAT safer and reduce side effects. However, after the successful synthesis of the chelating agent, radiolabeling experiments with [²²⁵Ac]­Ac³⁺ showed a relatively lower complexation efficiency together with a lower kinetic inertness of the formed complex compared to its Macropa lead. In order to understand the structural properties of both complex variants despite the lack of experimental structural data, we performed MM/DFT calculations of the La and Ac complexes of both ligands.

Our combined MM/DFT computational procedure successfully reproduced the most relevant conformers of the La–Macropa complex. On this basis, reliable predictions were made on the possible solution structures of the Megapa complexes with La³⁺ and Ac³⁺. The importance of the latter results is amplified by the lack of experimental data on the structural properties of the complexes. Analyzing the computed structures, differences were pointed out between the complex formation of La³⁺ and Ac³⁺ with the Megapa chelator, which do not appear with the smaller Macropa. Both the structural properties and the energy decomposition analyses indicated stronger complex formation of Megapa with La³⁺ compared to Ac³⁺. These results agree with the found exchange of Ac³⁺ by La³⁺ in our challenge experiments. Comparing the two chelators, the computed interaction energies of the isolated Megapa complex molecules were comparable to those of the respective Macropa complexes for both La³⁺ and Ac³⁺. The disagreement with the radiolabeling and stability experiments may mainly be attributed to the lack for kinetic effects by the computational model.

These results therefore illustrate that increasing the ring size of Macropa-derived chelators is not a viable approach to improve the complex stability with Ac³⁺ isotopes and highlight the necessity for intensive future research in this area for the advancement of ²²⁵Ac-based TAT.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.inorgchem.5c05424.

• Analytica data of 7–14, 1, 3, and La–3 (HR mass analyses, ¹H/¹³C/¹³⁹La NMR spectra, HPLC analyses), depiction of the most stable DLD conformation of the La–Macropa complex, additional computational details, lowest-energy M–Megapa conformers (M = La, Ac), relevant metal–ligand distances, sample input files, and Cartesian coordinates of the two most important M–Megapa conformers (PDF)

Nils F. Baier: experimental design, chemical and radiochemical syntheses, and data analysis. Patrick Cieslik: determination of the complex formation constant of La–Megapa. Henning Rudolf: radiochemical syntheses. Marc Pretze: experimental design. Björn Wängler: manuscript preparation (revision). Ralf Schirrmacher: manuscript preparation (first draft). Lutz Greb: supervision of complex characterization experiments. Gert Fricker: manuscript preparation (revision). Zoltán Varga: data analysis and manuscript preparation (revision). Attila Kovács: conceptualized the research work, calculations, data analysis, and manuscript preparation (first draft). Carmen Wängler: conceptualized the research work, experimental design, data analysis, and manuscript preparation (first draft). All authors finalized the manuscript through proofreading. All authors approved the final version of the manuscript.

This research project is part of the Research Campus M²OLIE and was funded by the German Federal Ministry of Research, Technology and Space (BMFTR) within the Framework “Research Campuspublic–private partnership for Innovation” under the funding code 13GW0747A. Further, a New Frontiers in Research Fund (NFRF)Transformation (NFRF-T) grant to R.S. is acknowledged (funding code GR026425/GR027195).

The authors declare no competing financial interest.