Benzoazacrown compound: a highly effective chelator for therapeutic bismuth radioisotopes

Ligand H₄BATA forms highly stable complex with bismuth(iii) in 1–2 min at room temperature.

Abstract

A new benzoazacrown ligand H₄BATA was synthesized and its complexation ability towards bismuth cations was evaluated. Binding of cation occurs at room temperature in a few minutes and formed complex exhibits the same level of inertness as highly stable complex with the well-known H₄DOTA in biologically relevant and challenging media under in vivo conditions.

Introduction

At the present time, targeted alpha therapy with ²²⁵Ac and ²¹³Bi has already shown promising results in the treatment of cancerous diseases such as bladder cancer, glioma, neuroendocrine tumors, etc.1,2 Chelating agents, such as H₄DOTA and H₅DTPA, are used to label target molecules.3,4 Complexes with azacrown ether H₄DOTA are highly stable under challenging conditions in vitro and in vivo but the kinetics of radionuclide binding is quite slow and only elevated temperatures up to 100 °C facilitate complexation. In contrast, H₅DTPA chelates cations instantaneously at room temperature but the formed complexes release radiometals in vitro and in vivo. In our work, we attempted to combine the macrocyclic effect of azacrown ethers with the mobility of acyclic ligands, from which the ligand H₄BATA with a large macrocyclic cavity and 4 carboxylic pendant arms was synthesized (Scheme 1). In this ligand, the benzene ring plays a dual role: it gives greater rigidity to the macrocycle and can be used to add a functional group for conjugation with a biomolecule. Moreover, it can be expected that conjugation with a biomolecule through a substituent in the benzene ring will not alter the ability of the macrocycle to coordinate cations.

Scheme 1. Synthesis of the ligand H₄BATA. Conditions: (i) BH₃·THF; (ii) tert-butyl bromoacetate, K₂CO₃, MeCN; (iii) H₂O.

Results and discussion

Synthesis of the ligand H₄BATA was carried out according to Scheme 1. Macrocyclic bisamide 1, synthesized in accordance with a previous work,5 was reduced by BH₃ in THF to give benzoazacrown 2 in high yield. Carboxylic pendant arms were introduced into the macrocycle in two steps. Compound 2 was N-alkylated with tert-butyl bromoacetate in MeCN under reflux in the presence of K₂CO₃ as the base to obtain 3, which was readily deprotected by hydrolysis in boiling water without any catalyst, in quantitative yield.

The protonation constants of H₄BATA and its stability constants with Bi³⁺ were determined at 25 °C in a 0.1 M KNO₃ aqueous solution using potentiometric titration.6,7 The calculated protonation constants and stability constants of Bi³⁺ complexes are presented in Table 1, which also lists the protonation and stability constants reported for H₄DOTA.8 It was shown9 that the protonation of linear polyaminopolyacetic ligands proceeds by the successive addition of protons on all the nitrogen atoms after which the carboxylate groups can be protonated. On the other hand, the H₄DOTA protonation scheme is different from the protonation scheme of its non-cyclic analogue, such as H₄TTHA. This difference consists in the fact that two nitrogen atoms of DOTA turn out to be less basic than carboxylate groups.8 A comparison of the protonation constants determined for H₄BATA and those reported for H₄DOTA shows that all the four amino groups of H₄BATA are more basic than the amino groups of H₄DOTA. Thus, sequential protonation of H₄BATA is closer to the protonation scheme of the non-cyclic analogue H₄TTHA than to that of H₄DOTA, more likely due to the large enough macrocyclic cavity of H₄BATA. The latter does not provide rigidity making protonation of the neighboring amines easier than in the case of H₄DOTA. Apart from H₄TTHA,9 the obtained stepwise protonation constant values of H₄BATA are close to the values of other polyaminopolyacetic ligands such as H₄TETA, H₅PEPA and H₆HEHA.10 The value of log K_H5L = 2.6 corresponds to one of the four carboxylic arms and is in line with protonation of the COOH group of EDTA, DTPA and other amines.9 Under the considered conditions, it was hard to reliably evaluate the formation of further protonated forms.

Table 1. Overall (log β_HhL, log β_MHhL, and log β_ML) and stepwise (log K_HhL) protonation constants and stability constants of H₄BATA and its complexes with Bi³⁺ in aqueous solution at 25.0 °C and 0.1 M KNO₃.

Species	H 4 BATA			H 4 DOTA a
	log βHhL	log βHhLM	log KHhL	log KHhL8	log βLM11
HL	11.9(1)		11.9	11.08
H2L	21.9(1)		10.0	9.23
H3L	30.3(1)		8.4	4.24
H4L	35.5(1)		5.2	4.18
H5L	38.1(1)		2.6	1.88
H2LBi		35.9(4)			—
HLBi		34.5(1)			—
LBi		30.3(1)			30.3
LBiH–1		20.9(1)			—

^aLiterature data for H₄DOTA8 are provided for comparison.

Potentiometric titration of H₄BATA has been carried out in the presence of equimolar Bi³⁺ ions to determine the stability constants of the corresponding metal complexes. Protonated forms of the complexes H₂BATABi⁺ and HBATABi and complexes BATABi^– and BATABi(OH)^2– have been detected over the studied pH range. To calculate the stability constants, the possibility of the formation of hydroxy and chloride complexes of bismuth(iii) was taken into account. The following values of the formation constants were used: log βBi(OH)^2– = 12.4, log βBi(OH)^2– = 23.48, log βBi(OH)₃ = 31.9, log βBi(OH)₄⁺ = 33.5 (ref. 12–15), log βBiCl²⁺ = 3.6, log βBiCl₂⁺ = 5.6, log βBiCl₃ = 7.0, and log βBiCl₄^– = 8.0.13

The species distribution diagram calculated for the complexes of H₄BATA with Bi³⁺ shows the presence of BATABi^– as the main component in the range of pH 5–9. HBATABi is mainly formed at lower pH while BATABi(OH)^2– prevails in an alkaline medium (Fig. S2†).

It should be noted that the stability constant obtained for the BATABi^– complex in a 0.1 M KNO₃ aqueous solution (log K_BATABi = 30.3) is equal to that reported for DOTABi^– (log K_DOTABi = 30.3).11 The pM values (pM = –log[Metal]_free), calculated for a specific pH value taking into account the known values of the stability constants, are often considered as a better estimate of the complex stability. The pBi values obtained for the complexes of Bi³⁺ with H₄BATA and H₄DOTA in aqueous solution at pH = 7.4, [Bi³⁺] = 1 μM and [Ligand] = 9 μM are equal to 23.06 and 25.69 for H₄BATA and H₄DOTA respectively. This result shows that bismuth(iii) binds to H₄BATA a little worse than to H₄DOTA in aqueous solution at pH = 7.4 but still with high affinity.

The fast formation of the complex is essential for radiopharmaceutical and, in particular, for therapeutic radioisotopes of ^212,213Bi which have fairly short half-lives (61 and 46 minutes respectively). Therefore, the chelator must be able to rapidly coordinate with Bi³⁺ under mild conditions, must be suitable for use in combination with biological vectors, and must provide a high yield of labeling. Thus, the labeling efficacy was determined at room temperature and at 80 °C after 1 hour of incubation. The H₄BATA labeling by ²⁰⁷Bi was quantified using thin layer chromatography (TLC) with further measurements of radioactivity on each half of a plate using a γ-scintillation counter. For this purpose, into a solution of H₄BATA (1.0 μM to 770 μM) buffered at pH 6.1 or 8.0, a solution of [²⁰⁷Bi]BiCl₃ was added. The radiolabeled ligand has an R_f = 0.9 as determined by TLC under the experimental conditions and the free Bi³⁺ stays at the start line with an R_f = 0.

The radiolabeling yield immediately after preparation (that is, after 2–3 minutes) and during 1 hour of equilibration did not change regardless of temperature: 44% at pH 6.1 and 1 μM H₄BATA concentration at 80 °C versus 49% at room temperature (Table 2). According to the literature,16 the slow kinetics of complexation by H₄DOTA is due to the slow introduction of a cation into the cage formed by the macrocycle and carboxylic arms, with simultaneous deprotonation of the macrocyclic N-atoms. In the case of H₄BATA, fast complexation can be provided by a much less rigid azacrown-ether cavity due to the larger sized 18-crown-6 of H₄BATA compared to the 12-crown-4 of H₄DOTA. This is the reason why all the labeling experiments were performed at room temperature without additional equilibration time.

Table 2. Labeling yields of H₄BATA with ²⁰⁷Bi.

c(H4BATA), μM	1	5	10	77	100	300	500	770
pH 6	49 ± 5%	64 ± 6%	88 ± 5%	96 ± 4%	95 ± 5%	95 ± 5%	97 ± 3%	95 ± 5%
pH 8	61 ± 6%	63 ± 5%	70 ± 5%	76 ± 4%	78 ± 6%	—	—	95 ± 5%

Since the pH value could influence complexation, we evaluated the labeling yields for several concentrations, and no significant differences were observed except for a slight decrease in yield at higher pH values. The lower labeling yield at higher pH is due to the extremely hydrolysable character of Bi³⁺. The labeling yield reached >96% at pH 6 and 77–100 μM of H₄BATA (Table 2). This large value has not changed in other buffer systems, such as PBS, bicarbonate and acetate buffers, which means that none of these buffer systems have an effect on complexation. This allows the use of any of these buffer systems, if necessary.

To further assess the in vitro stability of Bi³⁺ complexes, the following conditions were chosen: 500 μM of H₄BATA and 50 Bq of ²⁰⁷Bi in 0.01 M MES buffer (pH 6.1) at room temperature. No dissociation of the complexes in saline (0.9% NaCl) was detected after at least 2 days (Fig. S3†). In addition, the complexes were stable in the presence of a large excess of biologically significant cations: Ca²⁺ (5 mM), Mg²⁺ (5 mM), Zn²⁺ (0.1 mM) and Cu²⁺ (0.1 mM). It has been shown (Fig. S3†) that the complexes are stable in the presence of these cations for at least 1 day.

The stability of the Bi³⁺ complexes in serum was studied in a 100-fold excess (by volume) of fetal bovine serum by precipitation of the protein-bound fraction, as described elsewhere.17,18 We did not observe the release of radionuclides from the complex within the course of the experiment, which is in line with our pBi calculation (Fig. 1). In addition, for an adequate comparison with the reference complexes, we conducted experiments on serum stability under the same conditions with [²⁰⁷Bi]DOTA-Bi, which is known to be highly stable (Fig. 1a), and added similar data from the literature concerning the easily dissociated complex with the acyclic chelator [²⁰⁷Bi]EDTA-Bi.17 The supernatant after protein precipitation was analyzed by TLC, and it was proven that the radioactive species in this fraction are the same as those immediately after the radiolabeling (R_f = 0.9) (Fig. 1b). The results show the absence of cation rechelation by serum proteins from the complex with H₄BATA, and that with H₄DOTA, in contrast to the easily dissociating chelate with H₄EDTA (Fig. 1a).

Fig. 1. a) Rechelation of ²⁰⁷Bi³⁺ by serum proteins from the complexes with H₄EDTA,17H₄DOTA and H₄BATA; b) radiographic image of the TLC plate of the supernatant fraction after 2 days of incubation (95% of the complex).

Being part of the targeted radiopharmaceutical radioactive chelate itself is expected not to display any affinity to tissues. In order to estimate the biodistribution of the ligand itself at the very first level, we developed a simple deep learning model. The model was based on the Anatomical Therapeutic Chemical classification system,19 where the top-level (one-letter) groups were taken as labels for classification (details in the ESI†). The accuracy of the model predictions on the external experimental test set of 312 molecules is about 80% which can be considered as a good prediction quality (and outperforms the existing models like described in ref. 20) and allows us to use the model as a kind of fingerprint describing drug-like molecular behavior.

We applied the model to H₄BATA and to H₄DOTA as a reference point (Fig. 2). The presented distribution can be interpreted as the Anatomical Therapeutic Chemical profile of drug-like molecules (ATC-profile). According to our model, H₄BATA should behave very similarly to H₄DOTA, indicating qualitatively the same affinity to A (alimentary tract and metabolism), G (genitourinary system) and S (sensory organs) groups. Moreover H₄BATA provides even less activity concerning the above mentioned groups. Thus, we assumed a similar biodistribution for both ligands with a similar predicted ATC profile. Additionally, we could expect even less H₄BATA accumulation compared to that of H₄DOTA.

Fig. 2. ATC profiles of H₄DOTA and H₄BATA: predicted probabilities of ATC group attributions.

The fast complex formation at room temperature, high values of stability constants, stability of the complexes in vitro and theoretically estimated behavior of H₄BATAin vivo prompted us to conduct a biodistribution study on normal mice to evaluate the stability of [²⁰⁷Bi]BATA-Bi in vivo.

All in vivo experiments were carried out in accordance with the ARRIVE guideline and in accordance with the EU Directive 2010/63/EU for animal experiments. The experimental protocol was approved by the Institute Commission for Animals. A total of 12 normal male BALB/c mice were used. A 100 μL solution of the radiolabeled ligand (pH 6.1, 0.01 M MES) was injected intraperitoneally. The mice were euthanized after 1 and 6 h. The blood and major organs were harvested and wet weighed and radioactivity was measured by gamma-spectrometry (with an HPGe-detector GR3818, Canberra Ind.). The percent injected dose per gram (% ID per g) was determined for each tissue. The mean value and standard deviation for each tissue are presented in Table 3. It can be seen that a biodistribution profile and rapid elimination from the body demonstrated by complex with H₄BATA are similar to complex with H₄DOTA. Detected radioactivity in all tissues including the whole body reached a maximum of 3% ID per g (in the kidney) and this value decreased to 2% ID per g after 6 h post-injection, identifying renal clearance of the complex [²⁰⁷Bi]BATA-Bi. After 6 h, no sufficient accumulation in other tissues was observed: only 0.1% ID per g was detected in the liver, and lower doses were found in the rest of the organs. Moreover, in some cases, the radioactivity of both [²⁰⁷Bi]DOTA-Bi and [²⁰⁷Bi]BATA-Bi in organs did not exceed the minimum detectable level, therefore it was impossible to quantify accumulation. Compared to the results obtained with the [²⁰⁷Bi]EDTA-Bi complex,17 we can conclude that the radionuclide injected as the [²⁰⁷Bi]BATA-Bi complex demonstrates high inertness in vivo as a chelate compared with the “gold standard” in radiopharmacy [²⁰⁷Bi]DOTA-Bi.

Table 3. Biodistribution (% ID per g) of the Bi³⁺ complexes with H₄DOTA and H₄BATA in normal mice ^a .

Tissue	[207Bi]DOTA-Bi		[207Bi]BATA-Bi
	Time points (h)
	1	6	1	6
Kidneys	1.83 ± 0.42	1.10 ± 0.23	3.03 ± 0.52	1.97 ± 0.14
Blood	1.22 ± 0.29	b	1.14 ± 0.63	b
Femur	0.97 ± 0.60	0.14 ± 0.07	0.58 ± 0.23	0.07 ± 0.01
Liver	0.18 ± 0.02	0.07 ± 0.01	0.40 ± 0.29	0.11 ± 0.02
Heart	0.07 ± 0.01	b	0.10 ± 0.08	0.02 ± 0.01
Lungs	0.16 ± 0.03	0.03 ± 0.01	0.22 ± 0.13	0.06 ± 0.01
Spleen	0.12 ± 0.01	0.04 ± 0.02	0.19 ± 0.13	0.08 ± 0.02
Brain	0.05 ± 0.01	b	0.02 ± 0.01	b
The rest of body	0.47 ± 0.09	0.15 ± 0.01	1.1 ± 0.5	0.25 ± 0.06

^a n = 3 for each point.

^bRadioactivity is lower than the minimum detectable level.

Comparison of the experimental data with the predicted ATC-profile confirmed the initial assumption of a similar biodistribution of ligands. According to this, we can suggest that the biodistribution of both radioactive complexes reflects a biodistribution of ligands with low radioactivity release in vivo.

Unlike qualitative results, the quantitative prediction appeared to be much worse (H₄DOTA complex's accumulation was consistently lower than that of the complex with H₄BATA), and the model at the present stage can be used only for comparative evaluation in this way. Although the ATC classification was not originally designed to demonstrate therapeutic equivalence, the results showed that in some cases the ATC-profile was useful for estimating the biodistribution and the very first evaluation of drug behavior in vivo.

Conclusions

Our studies have shown that the ligand H₄BATA, having a large macrocyclic cavity and bearing 4 carboxylic pendant arms, is characterized by the fast formation of a very stable complex with bismuth(iii) already at room temperature in a wide pH range. This complex exhibits high stability in vitro in the experiments in biologically significant media comparable to the stability of complex with H₄DOTA. Furthermore this complex formed in 1–2 minutes at room temperature possesses the same inertness in vivo and high clearance rate as [²⁰⁷Bi]DOTA-Bi.

The functionalization of H₄BATA on the phenyl core may allow the preparation of conjugates with biomolecules including heat-sensitive molecules and lead to the creation of therapeutic radiopharmaceuticals containing short-lived bismuth radioisotopes.

Conflicts of interest

There are no conflicts to declare.