Cross-bridged Macrocyclic Chelators for Stable Complexation of Copper Radionuclides for PET Imaging

Abstract

Copper-64 (t_1/2 = 12.7 h, β⁺: 17.4%, E_β+max = 656 keV; β⁻: 39%, E_β-max = 573 keV) has emerged as an important non-standard positron-emitting radionuclide for PET imaging of diseased tissues. A significant challenge of working with copper radionuclides is that they must be delivered to the living system as a stable complex that is attached to a biological targeting molecule for effective imaging and therapy. Significant research has been devoted to the development of ligands that can stably chelate ⁶⁴Cu, in particular, the cross-bridged macrocyclic chelators. This review describes the coordination chemistry and biological behavior of ⁶⁴Cu-labeled cross-bridged complexes.

Introduction

Non-traditional positron-emitting radionuclides, particularly those of the transition metals, have gained considerable interest for imaging with positron emission tomography (PET) because of increased production and availability. Significant research effort has been devoted to the copper radionuclides because they offer a varying range of half-lives and positron energies as depicted in Table 1[1]. In addition, the well-established coordination chemistry of copper allows for its reaction with a wide variety of chelator systems that can potentially be linked to antibodies, proteins, peptides and other biologically relevant small molecules.

Table 1.

Decay Characteristics of Copper Radionuclides

Isotope	t1/2	β− MeV (%)	β+ MeV (%)	EC (%)	γ MeV (%)
60Cu	23.4 min	---	2.00 (69%) 3.00 (18%) 3.92 (6%)	7.0%	0.511 (186%) 0.85 (15%) 1.33 (80%) 1.76 (52%) 2.13 (6%)
61Cu	3.32 h	---	1.22 (60%)	40%	0.284 (12%) 0.38 (3%) 0.511 (120%)
62Cu	9.76 min	---	2.91 (97%)	2%	0.511 (194%)
64Cu	12.7 h	0.573 (38.4%)	0.655 (17.8%)	43.8%	0.511 (35.6%) 1.35 (0.6%)
67Cu	62.0 h	0.395 (45%) 0.484 (35%) 0.577 (20%)	---	---	0.184 (40%)

Coordination Chemistry of Copper

The aqueous solution coordination chemistry of copper is limited to its three accessible oxidation states (I-III) [2-4]. The lowest oxidation state, Cu(I) has a diamagnetic d¹⁰ configuration and forms complexes without any crystal-field stabilization energy. Complexes of this type are readily prepared using relatively soft polarizable ligands like thioethers, phosphines, nitriles, isonitriles, iodide, cyanide and thiolates; however, Cu(I) complexes are typically not used for in vivo imaging due to their lability. Copper (II) is a d⁹ metal of borderline softness which favors amines, imines, and bidentate ligands like bipyridine to form square planar, distorted square planar, trigonal pyramidal, square pyramidal, as well as distorted octahedral geometries. Jahn-Teller distortions in six-coordinate Cu(II) complexes are often observed as an axial elongation or a tetragonal compression. Due to the presence of some crystal-field stabilization energy, Cu(II) is generally less labile toward ligand exchange and is the best candidate for incorporation into radiopharmaceuticals. A third oxidation state Cu(III) is relatively rare and difficult to attain without the use of strong π-donating ligands.

Production of ⁶⁴Cu

Copper-64 can be effectively produced by both reactor-based and accelerator-based methods. One method of ⁶⁴Cu production is the ⁶⁴Zn(n,p)⁶⁴Cu reaction in a nuclear reactor [5]. Most reactor-produced radionuclides are produced using thermal neutron reactions, or (n,γ) reactions, where the thermal neutron is of relatively low energy, and the target material is of the same element as the product radionuclide. For producing high-specific activity ⁶⁴Cu, fast neutrons are used to bombard the target in a (n,p) reaction. This method enabled the production of high specific activity ⁶⁴Cu at the Missouri University Research Reactor (MURR) in amounts averaging 250 mCi [5].

The production of no-carrier-added ⁶⁴Cu via the ⁶⁴Ni(p,n)⁶⁴Cu reaction on a biomedical cyclotron was proposed by Szelecsenyi et al. In this study small irradiations were performed demonstrating the feasibility of ⁶⁴Cu production by this method [6]. Subsequent studies by McCarthy et al. were performed, and this method is now used to provide ⁶⁴Cu to researchers throughout the United States [7]. Recently, Obata et al. reported the production of ⁶⁴Cu on a 12 MeV cyclotron, which is more representative of the modern cyclotrons currently in operation [8]. They utilized very similar methods to those previously published [7]. A remote system was described for separation of the ⁶⁴Cu from the ⁶⁴Ni target.

Chelating Ligands for ⁶⁴Cu

Utilizing ⁶⁴Cu of high specific activity with a chelator that forms a stable complex in vivo is critical in achieving high uptake of the copper radionuclide in the tissue or organ of interest while minimizing the non-selective binding or incorporation into non-target organs or tissues. Ligands that can form radio-copper complexes with high kinetic inertness to Cu(II) decomplexation (proton-assisted as well as transchelation or transmetallation) are ideal, since this is more significant than thermodynamic stability after the radio-copper complex is injected into a living organism [9, 10]. Reduction of Cu(II) to Cu(I) and subsequent Cu(I) loss may also be a pathway for loss of radio-copper, and resistance of the radio-copper complex to Cu(II)/Cu(I) reduction as well as reversibility can also be important [11]. Rapid complexation kinetics are also essential to allow for the facile formation of the radio-copper complex. Finally, chelators also must be designed with available functional groups that allow them to be covalently linked to targeting peptides, proteins, and antibodies.

Chelators Based on Cyclam and Cyclen Backbones

The most widely used chelators for attaching ⁶⁴Cu to biological molecules are tetraazamacrocyclic ligands with pendant arms that utilize both the macrocyclic and chelate effects to enhance stability. By far the most extensively used class of chelators for ⁶⁴Cu has been the macrocyclic polyaminocarboxylates shown in Fig. (1). These systems have been thoroughly investigated, and in vitro and in vivo testing have shown them to be superior to acyclic chelating agents for ⁶⁴Cu [12]. This enhanced stability is most likely due to the greater geometrical constraint incorporated into the macrocyclic ligand that enhances the kinetic inertness and thermodynamic stability of their ⁶⁴Cu complexes [13-15]. Two of the most important chelators studied were DOTA (1) and TETA (2). While DOTA has been used as a BFC for ⁶⁴Cu, its ability to bind many different metal ions, and its decreased stability compared to TETA make it less than ideal [16-21]. The tetraazamacrocyclic ligand TETA therefore, has been extensively used as a chelator for ⁶⁴Cu, and successful derivatization of this ligand has allowed researchers to conjugate it to antibodies, proteins, and peptides [22-32].

Figure 1.

Cyclic polyaminocarboxylates and their derivatives.

Although ⁶⁴Cu-TETA complexes are more stable than ⁶⁴Cu-DOTA and ⁶⁴Cu-labeled complexes of acyclic ligands, their instability in vivo has been well documented. Bass et al. demonstrated that when ⁶⁴Cu-TETA-octreotide was injected into normal Sprague-Dawley rats, nearly 70% of the ⁶⁴Cu from ⁶⁴Cu-TETA-octreotide was transchelated to a 35 kDa species believed to be superoxide dismutase (SOD) in the liver 20 h post-injection [33]. These results are supported by the observations of Boswell et al. [34].

Despite the considerable efforts made by researchers to use tetraaza-tetracarboxylate macrocyclic ligands as effective BFCs for ⁶⁴Cu, it is evident that the in vivo instability of these ⁶⁴Cu complexes emphasizes the need for more inert ⁶⁴Cu chelators. With this in mind, new ligand systems, including those based upon the cross-bridged (CB) tetraazamacrocycles have been developed to complex ⁶⁴Cu more stably.

The Cross-bridged Tetraamine Ligands (Figure 2)

Figure 2.

Cross-bridged macrocyclic complexes and analogs compared to the sarcophogene chelator, diamsar.

This class of chelators was first conceived of and synthesized by Weisman and Wong and coworkers in the 1990's [35, 36], and they were originally designed to complex metal cations like Li⁺, Cu²⁺, and Zn²⁺ within their clamshell-like clefts. Numerous copper complexes of these and related ligands have since been prepared and studied by Wong, Weisman and coworkers as well as other research groups [37-42]. With available structural data, the expected cis-folded coordination geometry of these chelators has been confirmed in all cases. Attachment of two carboxymethyl pendant arms to CB-cyclam (8) to give CB-TE2A (10) further ensures complete envelopment of a six-coordinate Cu(II) as shown in Fig. (3).

Figure 3.

A comparison of the structures of Cu(II) complexes of the CB ligands CB-TE2A and CB-DO2A with those of TETA and DOTA based on published crystallographic data from references [14, 15, 34, 42].

While the measurement of stability constants of Cu-CB complexes have been limited by the proton-sponge nature of these chelators, available data for Cu-CB-cyclam (log K_f = 27.1) revealed very similar values to non-bridged Cu-cyclam (log K_f = 27.2) and related complexes [43]. On the other hand, their kinetic inertness, especially in aqueous solution, has been shown to be truly exceptional [11, 44]. Proton-assisted decomplexation is a convenient indicator of solution inertness. Under pseudo-first order conditions of high acid concentration (e.g. 5 M HCl), decomplexation half-lives can provide a comparative gauge. Remarkable resistance of Cu-CB complexes toward such processes has recently been demonstrated [11, 44]. As shown in Table 2, Cu-CB-cyclam is almost an order of magnitude more inert than Cu-cyclam in 5 M HCl at 90°C, while Cu-CB-TE2A is 4 orders of magnitude more inert. Impressively, the latter complex resists acid decomplexation even better than the fully-encapsulated sarcophagine complex Cu(II)-diamsar. It was confirmed that both the cross-bridged cyclam backbone as well as presence of two enveloping carboxymethyl arms are required for this unusual kinetic inertness.

Table 2.

Pseudo-first order half-lives for acid-decomplexation and reduction potentials of Cu(II) complexes (all values are from Heroux et al. [44] unless otherwise noted).

Chelator	5 M HCl 90°C	12 M HCl 90°C	Ered (V) vs Ag/AgCl
DOTA	<3 min	<3 min	−0.94 (irrev.)a
cyclam	<3 min	<3 min	−0.68 (quasi-rev)
TETA	4.5(5) min	<3 min	−1.18 (irrev.)
CB-cyclam	11.7(1) min	<3 min	−0.52 (quasi-rev)
CB-TE2A	154(6) h	1.6(2) h	−1.08 (quasi-rev.)
CB-DO2A	< 3 min	n.d.	−0.92 (irrev.)
CB-TE2LA	100 h	39(5) min	−0.68 (quasi-rev.)
CB-DO2LA	< 3 min	n.d.	−0.78 (irrev.)
CB-TEAMA	n.d.	n.d.	−0.96 (quasi-rev.)
diamsar	40(1) h	< 3 min	−1.1 (irrev)

^a[34]

The effect of pendant arm length on acid stability was also investigated. Heroux et al. compared the properties of Cu(II) complexes of cross-bridged cyclam and cyclen having N-carboxyethyl pendant arms (CB-DO2LA (14) and CB-TE2LA (15)) to the corresponding CBTE2A and CB-DO2A complexes [44]. The inertness of Cu-CB-TE2LA in 5 N HCl at 90°C of 100 h was very high compared to other chelators, though not quite as good as 154 h for Cu-CB-TE2A (Table 2).

With respect to ease of Cu(II)/Cu(I) reduction, cyclic voltammetric studies of Cu(II) complexes of a variety of tetraazamacrocyclic complexes revealed that Cu-CB-TE2A (16) is not reduced in 0.1 N sodium acetate until a relatively negative potential of −1.07 V (vs. Ag/AgCl) [11]. Further, unlike the Cu-DOTA (19), Cu-TETA (18) and Cu-diamsar cyclic voltammograms, this reduction is quasi-reversible, suggesting the innate ability of the cross-bridged cyclam ligand to adapt to a geometry suitable for Cu(I) coordination [11] (Gustafson L, Wong Edward H. Unpublished results. 2006). The reduction of the Cu-CB-TE2LA was also quasi-reversable, but was 400 mV more easily reduced than Cu-CB-TE2A [44]. Both resistance to proton-assisted decomplexation and the Cu(II)/Cu(I) reduction of Cu-CB-TE2A suggest that the ligand may be an especially promising chelator candidate for ^64/67Cu(II).

Biological stability of several of the ⁶⁴Cu complexes of the CB-macrocycle ligands shown in Fig. (2) has been investigated [34, 43, 48]. The ⁶⁴Cu complex of CB-TE2A was formed under carrier-added and no-carrier-added conditions in less than 2 h at 55°C [43]. Serum stability experiments indicated that these complexes are stable in rat serum out to 24 h. Results of biodistribution studies of these ⁶⁴Cu complexes in female Sprague Dawley rats were highly dependent upon the chelator. The complex ⁶⁴Cu-CB-TE2A was determined to be the most stable and was cleared most rapidly from the blood, liver, and kidney.

Boswell et al. directly compared the in vivo stability of the ⁶⁴Cu complexes of CB-TE2A and CB-DO2A (9), which are analogues of TETA and DOTA respectively, and developed the analytical conditions to purify and isolate the kinetic ⁶⁴Cu-CB-complexes and their metabolic analytes [34, 45]. Both CB-ligands were labeled in high radiochemical purity at 95°C using ethanol and cesium carbonate, followed by the addition of ⁶⁴CuCl₂. These relatively harsh conditions were needed to ensure that the competing reactions between ⁶⁴Cu and any trace impurity ligands were suppressed. The biodistribution of the ⁶⁴Cu-CB-DO2A complex was also completed and compared to that of ⁶⁴Cu-DOTA. At 4 h p.i., ⁶⁴Cu-CB-DO2A demonstrated significantly better clearance properties than the ⁶⁴Cu-DOTA analogue. Metabolism studies in normal rats were also conducted and demonstrated that the CB-ligands are less susceptible to ⁶⁴Cu transchelation than their non cross-bridged analogues. By 4 h ⁶⁴Cu-CB-TE2A underwent significantly less transchelation in the liver than ⁶⁴Cu-TETA (13% vs. 75%), while ⁶⁴Cu-CB-DO2A underwent less transchelation than ⁶⁴Cu-DOTA (61% vs. 90%). ⁶⁴Cu-CB-TE2A was clearly the most stable of all of the ⁶⁴Cu complexes tested and this was most evident at 20 h, where only 24% of the injected ⁶⁴Cu was transchelated to proteins; in contrast, 92% of the ⁶⁴Cu associated with TETA was transchelated. In addition, a survey of biodistribution data of several ⁶⁴Cu-tetraazamacrocycles in normal rats reveals that ⁶⁴Cu-CB-TE2A has superior clearance properties as shown in Fig. (4) [34, 43, 46]. These data correspond with the in vitro data, and demonstrate the enhanced in vivo stability that the ethylene cross-bridge and carboxylate pendant arms provide to the tetraazamacrocycles.

Figure 4.

Biodistribution data of selected ⁶⁴Cu-labeled cyclam and bridged cyclam analogs at 24 h p.i. in normal rats. Adapted from references [43, 44, 46, 48].

Studies also focused upon modifying these CB chelators for conjugation to small molecules and peptides [47]. Sprague et al. confirmed that CB-TE2A will be a valuable BFC for ⁶⁴Cu. In this study, CB-TE2A was conjugated to the somatostatin analogue Y3-TATE and directly compared to the ⁶⁴Cu-TETA-Y3-TATE conjugate [47]. ⁶⁴Cu-CB-TE2A-Y3-TATE was radiolabeled in high radiochemical purity with specific activities of 1.3-5.1 mCi/µg of peptide without the need for harsh labeling conditions. Biodistribution studies using AR42J tumors implanted in male Lewis rats revealed that this complex had greater affinity for somatostatin-positive tissues compared to the TETA conjugate. Accumulation of ⁶⁴Cu-TE2A-Y3-TATE was lower at all time points in blood and liver, and less accumulation was observed in the kidney at earlier time points when compared to ⁶⁴Cu-TETA-Y3-TATE. These data suggest that the ⁶⁴Cu-CB-TE2A-Y3-TATE is more resistant to transchelation than the TETA analogue.

To further examine the stability of cross-bridged macrocycles with one carboxylate and one amide group, a series of Cu(II) complexes of cross-bridged macrocyclic chelators (CBTEAMA (11), CB-MeTEAMA (12), and PhTEAMA (13)) were synthesized and evaluated as model compounds for ⁶⁴Cu-CB-TE2A-peptide conjugates. The biological behavior of the ⁶⁴Cu complexes was evaluated in normal rats [48]. These agents showed rapid blood clearance and relatively low liver uptake at 24 h post-injection, demonstrating that one carboxylate is likely to be sufficient for in vivo stability. In addition, Cu-TEAMA showed a highly negative, quasi-reversible reduction potential (−0.96 V) (Wong et al., unpublished results), which is consistent with the in vivo results.

Conclusion

The development, production, and use of ⁶⁴Cu as a radionuclide for diagnostic imaging and therapy have greatly increased over the last decades. Because of the choice of copper isotopes with variable emission types and energies, it has become essential to develop ligand systems that can stably complex ⁶⁴Cu to form kinetically inert radiometal complexes. To accomplish this goal, the importance of improved ligand design and synthesis as well as the employment of rigorous physical and biological screening processes to evaluate their in vivo effectiveness is highly essential for the future development of ⁶⁴Cu radiopharmaceuticals as practical diagnostic imaging and/or radiotherapeutic agents.

Abbreviations

BAT

bromoacetamidobenzyl

BFC

bifunctional chelator

CB

cross-bridged

CB-DO2A

4,10-bis(carboxymethyl)-1,4,7,10-tetraazabicyclo[5.5.2]tetradecane

CB-TE2A

4,11 - bis(carboxymethyl)-1,4,8,11-tetraazabicyclo[6.6.2]hexadecane

Ci

curie

diamsar

3,6,10,13,16,19-hexaazabicyclo[6.6.6]eicosane-1,8-diamine

DOTA

1,4,7,10-tetraazacyclododecane-1,4,7, 10-tetraacetic acid

h

hour(s)

mCi

millicurie

mmole

millimole

OC

octreotide

PET

positron emission tomography

p.i.

post injection

SOD

superoxide di smutase

SUV

Standard uptake value

TETA

1,4,8,11-tetraazacyclotetradecane-1,4,8, 11-tetraacetic acid

V

volts

Y3-TATE

tyrosine-3-octreotate