322
V. Maheshwari et al. / Inorganica Chimica Acta 393 (2012) 318–323
In marked contrast to 64Cu-DOTA, there is minimal loss of 64Cu
from 64Cu-TETA with >90% retention of 64Cu at pH 7.5 in the pres-
ence of 5
M Cu2+ (Fig. 4), despite the overall similarity of the li-
however, there are no studies than compare the kinetic stability
of these complexes to DOTA, TETA, or NOTA. The method described
above will allow us, in future studies, to measure the rate constants
for dissociation of 64Cu from these complexes, providing quantita-
tive measures of their relative kinetic stability. It will also allow the
quantitative evaluation of the effects of non-radioactive Cu2+ and
pH on the kinetic stability of these complexes under physiologi-
cally relevant conditions.
A second avenue is the more detailed evaluation of the pro-
posed mechanism for the loss of 64Cu from Cu-DOTA and Cu-TETA.
The aforementioned TE2A and DO2A ligands provide the opportu-
nity to test this hypothesis. If the proposed mechanism is valid, the
rate of loss of 64Cu from both of these ligands should be signifi-
cantly slower than from the tetraacetic acid analogs.
A third avenue that might be explored using this method is the
evaluation of the possible role of reduction of Cu2+ to Cu+ in the
in vivo loss of 64Cu from various bifunctional chelators. Although
this has been suggested as a possible mechanism by which 64Cu
may be lost from bifunctional chelates with high thermodynamic
stability [18], to date there have been no studies in which potential
in vivo reductants (e.g., glutathione) have been evaluated for their
effect on the kinetic stability of these complexes.
It is also worth noting that these results apply only to the macro-
cyclic ligands themselves, not necessarily to the protein conjugates.
It is well known that addition of substituents to the backbone of a
macrocyclic chelator significantly decreases the rate of loss of the
metal from the complex (cf. [33]). Accordingly, it is reasonable to ex-
pect that addition of a p-aminobenzyl moiety, such as is present in
many bifunctional chelators, will have a significant effect on the rate
of demetallation of the 64Cu complexes. Using the system described
in the present study, it is possible to compare the rate of loss of 64Cu
from the bare chelator, a simple alkylated derivative (e.g., methyl),
the p-nitrobenzyl derivative (rather than the p-aminobenzyl, which
would introduce an additional donor atom), and the chelator conju-
gated to proteins of various sizes, from peptides to antibodies.
l
gands. One obvious difference between the two complexes is
that, in contrast to 64Cu-DOTA, in 64Cu-TETA the metal fits within
the macrocyclic core (Fig. 6b) and is, therefore, much less exposed.
In keeping with the hypothesis proposed above for the mechanism
of loss of 64Cu from Cu-DOTA, for 64Cu-TETA the rate of loss is min-
imally dependent on pH and increases slightly with increasing
[Cu2+], perhaps due to a combination of factors, the extra carboxyl-
ate arms increasing the susceptibility to associative loss and the in-
creased core size decreasing the susceptibility to associative loss of
64Cu because the metal atom now fits in the core of the macrocycle.
An interesting aspect of the crystal structure of the unsubstitut-
ed Cu-TETA complex is that it exists in two isomers, one in which
the carboxylate moiety is located along the Z-axis (and thus subject
to bond lengthening due to Jahn–Teller distortion) and one in
which the N atoms of the macrocycle are located along the Z axis
(with elongated Cu–N bonds compared to Cu–Naxial) [31]. In an ear-
lier study, Kukis et al. compared the serum stability of 64Cu-labeled
antibody conjugates of two TETA derivatives, BAT-2 and BAT-6,
which differ only in the location of the bromoacetamidobenzyl
conjugation moiety on the backbone of the core cyclam macrocy-
cle, and found that the BAT-2 derivative was much less stable in
serum than the BAT-6 derivative [14]. The crystal structure of
Cu-BAT-2 has not been determined, but the crystal structure of
Cu-BAT-6 shows that two of the N atoms of the macrocycle are lo-
cated along the Z-axis of the octahedral structure resulting in sig-
nificant elongation of the Cu–N bonds compared to the Cu–N
equatorial bonds (2.4 Å vs. 2.0 Å), a result similar to that observed
by Silversides et al. [31]. It is possible that the different substitu-
tion isomers lead to the formation of different coordination iso-
mers, one of which is more kinetically stable. This possibility is
supported by the results of the current study in which, in contrast
to the results for 64Cu-DOTA, a plateau is observed for the rate of
loss of 64Cu from 64Cu-TETA. This may reflect the presence of the
two isomers in the product mixture, one of which is significantly
more kinetically stable than the other.
5. Conclusions
The idea that the extra carboxylate moieties decrease, rather
than increase, the kinetic stability of 64Cu-TETA is supported by a
recent study by Pandya et al. who reported the synthesis and sta-
bility of 64Cu-TE2A, the two-armed analog of TETA [32]. These
investigators found that, in 5 M HCl at 50 °C, the half-life for
decomplexation of Cu-TE2A was 93 h compared with 4 h for Cu-
TETA. They also observed that the liver concentration of 64Cu-
TE2A at 24 h post-injection was approximately one-half that of
64Cu-TETA. In a separate report, these investigators described the
conjugation of TE2A to an RGD peptide and its subsequent labeling
with 64Cu, but they did not measure the biodistribution of the
64Cu-TE2A-RGD conjugate [4].
We have developed a relatively simple assay with which the
kinetic stability of 64Cu complexes of chelating agents used in
the preparation of radiolabeled proteins can be measured. This
method can potentially be used to address a fundamental question
that underlies the development of radiometal-labeled proteins –
which chelator better retains the metal under physiologically rele-
vant conditions? – without depending on indirect measurements
such as liver uptake or physiologically irrelevant conditions such
as hot solutions of 5 M HCl. The method can also be used to deter-
mine if new chelating agents, such as CB2A [10,11] and SarAr [23],
retain 64Cu better than existing chelating agents and perhaps lead
to a better understanding of the chemical parameters that are
important for improved kinetic stability. This will ultimately result
in the development of improved 64Cu-labeled radiopharmaceuti-
cals in which there is increased confidence that the observed
biodistribution reflects the distribution of the radiolabeled protein.
The 64Cu-NOTA complex is even more stable than 64Cu-TETA
with essentially 100% retention of 64Cu through 6 h at all pHs
and [Cu2+]. Consistent with the proposed mechanism for loss of
64Cu from DOTA and, to a lesser extent, from 64Cu-TETA, the crystal
structure of Cu-NOTA shows the Cu2+ atom is tightly enclosed by
the six donor atoms of the ligand with no free carboxylates
(Fig. 6c). This result is also consistent with several biological stud-
ies that suggest that loss of 64Cu from 64Cu-NOTA is significantly
less than from 64Cu-DOTA or 64Cu-TETA [15,28,29].
Together these results strongly suggest that the preferred chela-
tor for 64Cu, at least among the three that are currently most often
used, is NOTA. These results also suggest several avenues for con-
tinued investigation. These include extension of this evaluation to
other, more recently developed, bifunctional chelators such as
CB2A and SarAr that form significantly more thermodynamically
stable complexes with 64Cu than do DOTA or TETA. To date,
Acknowledgements
This project was supported by Grant #DE-FG02-08ER64704
from the US Dept. of Energy.
References
[1] Y. Wu, X. Zhang, Z. Xiong, Z. Cheng, D.R. Fisher, S. Liu, S.S. Gambhir, X. Chen, J.
Nucl. Med. 46 (2005) 1707.
[2] R.A. Dumont, I. Hildebrandt, H. Su, R. Haubner, G. Reischl, J.G. Czernin, P.S.
Mischel, W.A. Weber, Cancer Res. 69 (2009) 3173.