Kinetic stability (robustness) of copper chelates

1.10.1 Kinetic properties of metal and chelator

Radiochemistry has shown that though the contents of different elements in a cell remain stable over time, the atoms of the cell are continuously exchanged with other atoms of the same element. Normally it does not matter if for instance one copper atom is exchanged with the other, but if it a radioactive isotope is

introduced it makes a huge difference. If a ⁶⁴Cu labelled tracer molecule exchange copper with an enzyme on its way to a tumour, then the scanning will not image the position of the tumour but of the enzyme. A high thermodynamic stability of a chelate only means that there is a high possibility for the chelator to be part of a complex with the specified metal ion. In technical terms: Kf is high.

In a system where C + M ⇄ CM and C is the chelator, M is the metal ion and CM is the chelate, Kf can be defined in two ways. The first is from the reactant concentrations:

] ][

[ ] [

f C M

K = CM 1·1

However, a high Kf does not guarantee a low exchange between chelated metal and free metal. This can be seen from the other definition of Kf.

off on

f k

K = k Where kon is the rate coefficient for C + M → CM, and koff for CM → C + M. 1·2

A high Kf means that kon is much larger than koff but koff can still be significant. If a chelator is very rigid, both kon and koff are small compared to a flexible chelator due to higher rearrangement energies. Therefore, to minimize koff the chelator should both have a high Kf and be rigid [151]. A chelate with a very small koff is said to be kinetically stable or robust.

Elements, for which there are no natural chelating molecules in vivo, might be recaptured if they leave the chelator and consequently a high kon is preferable. On the other hand, elements, for which there are plenty of natural chelating molecules in vivo, can be expected to be lost to competing chelating molecules if they leave the chelator and consequently a small koff is required. That is the case for copper for which competing chelators are abundant [169,224].

It should be noted that the robustness is a function of both the chelator and the metal ion. Rigid chelators are more robust than flexible chelators, and some metals such as iridium form complexes which are more robust than those form by labile metals such as copper [48,198]. The chelation speed of a metal ion can be

estimated from its exchange rate of H2O [48,56]. One might think that a labile chelator with a “robust” metal ion is just as suited for TRT as a robust chelator with a labile metal ion, but this is not the case. If the metal ion is the speed-limiting factor, then there is a high risk that other contaminating metal ions will bind before the intended, which is especially devastating if there are contaminating radiometals in the synthesis.

Therefore, a radiophysical interesting element such as iridium is ruled out due to slow complex formation.

On the other hand, if the chelator is limiting, then the radiometal can bind before any slower contaminating metals. This enables synthesis of copper chelates of high purity even in the presence of other first transition row metals.

1.10.2 Effects of metal reduction on kinetic stability

Another reason why a high Kf is not a guarantee for slow dissociation is that the Kf value applies for one oxidation state and if the metal ion is reduced, the Kf can change several orders of magnitude:

(Kf)^b (βn) [Cu(CN)4]²⁻ = 10²² M⁻⁴ (Kf)^b (βn) [Cu(CN)4]³⁻ = 10^30.5 M⁻⁴ Kf [Fe(EDTA)]^– = 10^24.2 M⁻¹ Kf [Fe(EDTA)]^2– = 10^14.3 M⁻¹

As Cu(II) favours octahedral coordination geometry and Cu(I) tetrahedral, hexadentate chelators will loose at least two bonds when Cu(II) is reduced to Cu(I), and generally the lonepairs of the electron donors will be less attracted to the a metal ion when its charge is reduced [147]. The effect of tetrahedral preference can be seen in the increase in β value from Cu(II) to Cu(I) cyanide, whereas the charge reduction dominates the drop in Kf from Fe(III) to Fe(II) EDTA, as both EDTA ferrates are 6 coordinated.

To avoid potential reductive transchelation three strategies are available: Making the rearrangement energy from the Cu(II) to the Cu(I) chelate sufficiently high to impair reduction, making the C(I) chelate sufficiently robust or avoiding/blocking the reductive mechanism. It appears odd that Woodin et al. hypothesize that it is resistance towards reduction that is important for in vivo stability (strategy 1) and then compare the

robustness of the Cu(I) chelates (stategy 2) by cyclic voltammetry [232]. However, the Cu(II/I) redox potential of a chelate is closely related to the stability of the Cu(II) complex. Therefore, from cyclic

voltammetry the redox potential gives an estimate of the robustness of the Cu(II) complex, including towards reduction and the reversibility gives an estimate of the robustness of the Cu(I) complex, depending on competing Cu(I) binding solutes [182]. A chelator that can fulfil the first strategy is so rigid that chelation of Cu²⁺ at physiological conditions becomes too slow for therapeutic isotopes. It is known that the redox potential of tumour cells is lower than average and this is connected to hypoxia and high metabolic rates [234]. This has been used to create alkylated derivates of bis(thiosemicarbazonato)copper(II) complexes, which selectively are reduced in hypoxic tumour cells. The effect is self enhancing as ionizing radiation creates free radicals capable of reducing chelated Cu(II) [194]. It has not yet been possible to create a bis(thiosemicarbazonato)copper(II) complex with a redox potential so low that even hypoxic cells are incapable of reducing it [51].

Increasing the stability of the Cu(I) chelate can be done in several ways. By increasing the size of a ring chelator it can easier adapt a tetrahedral formation. However, the increased flexibility not only potentially increases the Kf of the Cu(I) chelate but also the koff. Therefore, one option is to vary the ring size while measuring the kinetic stability of the Cu(I) chelate. Another option is to replace N-donors with S-donors as Kaden did to make a chelator for Ag⁺ [108]. Unfortunately the only robust complex is 6-coordinated, which does not resolve the problem with Cu(I). The lack of success with N/S-exchange could be explained by the observation that the soft Ag⁺ ion prefers the tertiary amine donors to the soft S-donor atoms [81].

A solution could be to look at copper binding enzymes as most of them are redox-active and thus able to bind both Cu(I) and Cu(II) stably. Moreover, they manage to do it with the copper ion open to interaction with the substrate. However, thermodynamic, not kinetic stability is of importance for enzymes as chelation must be swift and if the metal is lost, a replacement can be acquired from the copper chaperones.

Nonetheless, wild type SOD and many mutants are not only thermodynamically stable but also kinetically stable in the metalated (holo) form, which probably protects SOD from aggregation [135]. It should be noted though that the measured kinetic stability of SOD is against unfolding, not against metal transchelation.

An important question to answer before spending years and millions in search of kinetically stable Cu(I) chelators is: Is it an advantage to loose copper as Cu(I) in the nuclei of target cells? In a comparison of ⁶⁴

Cu-labelled Bombesin using either CBTE2A or DOTA as chelator, the CBTE2A tracer molecule showed markedly better non-target tissue clearance, but also markedly lower tumour retention compared to the DOTA tracer molecule explained by resistance to transchelation of CBTE2A [71]. If a BFC, which is completely (by measuring terms) resistant to transchelation is found, them a trapping mechanism to keep the chelate inside the target tissue (or even better inside target nuclei) is required to fully benefit from it.

1.10.3 Measurements of kinetic stability

The kinetic stability of Cu(II) complexes are normally compared by acid-decomplexation [222,232]. This is a stepwise process where the nitrogen copper coordination bonds one by one are broken and replaced with a protonation of the nitrogen atoms. Woodin [232] refer to Kotek [117] claiming, “Aqueous acid-assisted decomplexation of polyazamacrocyclic copper complexes is a convenient and useful indicator of their kinetic inertness”. However, reading the cited articles by Kotek reveals that the usefulness of the method solely relies on the fact that he uses it, which seems to be a rather thin justification. Beside the fact that the method is easy and well used, it is hard to see why the half-life of copper complexes in 5 M HCl should be indicative of their half-life in vivo, when the in vivo [H⁺] is only about 10^-7,4 M making acid-decomplexation unlikely [12]. Some groups use HClO4 for acid-decomplexation, thereby removing the factor that Cl^- forms

complexes with Cu²⁺. This increases the decomplexation time; in one example by 10 fold [206]. Heroux et al. moderates the description of the method to: “convenient and popular” [88], and after describing some of the weaknesses they conclude that “acid decomplexation nonetheless appears to serve as a useful first indicator of its likely in vivo integrity towards metal loss due to protons, competing biometals or biological ligands”. 4 reservations (underlined) and reduction, probably the most important factor for copper loss in vivo, is not mentioned.

Another method for measuring the kinetic stability is incubation in plasma or serum at 37 °C but incubation in excess of a strong competing chelator such as EDTA or diethylenediaminetetraacetic acid (DTPA) could give an even better indication of the in vivo stability [123], although neither serum nor ligand challenge assays take reduction into account.

Despite that it seems likely that reduction of the chelated Cu(II) to Cu(I) has a major impact on kinetic stability in vivo, there are no papers of the Kf of the common BFC’s measured in reducing environments. A reductive challenge assay recently published is likely to change that [12]. Cyclic voltammetry has revealed that some BFC’s rapidly loose copper when reduced, but these experiments have not been quantitative [71].

A dibenzyl derivate of cross bridged cyclam has been crystallized both as Cu(I) and Cu(II) complex but the redox potential is unknown [98].