Chemistry Reference
In-Depth Information
(Figure 8.5) uses an entirely different ligand framework to achieve a diaqua complex, exploiting an acyclic backbone and all
oxygen donor set to achieve q = 2 in an eight-coordinate complex in which there is sufficient space for associative (and there-
fore rapid) water exchange [61-63]. This approach has been extended to q = 3 systems in gd.tacn-1,2-HOPO [64].
[gd.AAZTA]
-
(Figure 8.5) offers another approach to achieving q = 2 complexes; these are highly stable with unchanged
relaxivities over a broad range of pH (7.1 mMgd
-1
s
-1
at 20 MHz in the range pH = 2-11) and show no anion or protein
binding [65]. Furthermore, they offer high kinetic stability when challenged with endogenous ions and can be synthesised
without recourse to the use of a macrocyclic backbone. This not only reduces the cost of synthesis, but also allows for facile
variation of structure and bioconjugation [66]. As with the other systems we have discussed, there are two isomeric forms in
solution with greatly differing solvent exchange rates [67]. At the time of writing, AAZTA-derived systems appear to have
great promise.
We have yet to mention control of rotational correlation time. From the discussion above, it is clear that extending τ
R
should enhance the observed relaxivity. τ
R
for a spherical molecule is directly related to the size of the molecule. For a sphere
of radius
a
in solution in a medium with viscosity η,
a
/
τ
= 4
πη
3
kT
.
(8.22)
R
Because τ
R
will vary with the cube of the molecular radius, large effects are to be expected from changing the size of the
molecule. However, the disadvantage of this purely theoretical approach is obvious in that solvated molecules are not
spherical. What's more, they tend to become less spherical as molecular size increases. Furthermore, a molecule must be
rigid if it is to attain maximum theoretical relaxivity. If a small(ish) molecule like a dOTA-monoamide is tethered to a larger
edifice by a single tether, rotational motion around the tether is often still possible, meaning that τ
R
is often much shorter
than would be estimated from a simple equation. τ
R
can be measured in number of ways, most notably by making measure-
ments on analogous vanadyl(IV) complexes (although this method has an obvious limitation in that the coordination chem-
istry of vanadyl is very different to that of gd(III)) [68], through
17
O NMR measurements on the gd complex, or through
fluorescence polarisation measurement (although this obviously requires the presence of a fluorescent chromophore within
the molecule) [69].
Parker and co-workers have controlled the relaxivity by placing the lanthanide binding motif at the centre of a dendritic
architecture derived from [gd.gdOTA]
5-
[70]. These systems essentially place the lanthanide at the hub of a wheel in which
the rest of the dendrimer provides the spokes and hub, ensuring that all motion relating to the lanthanide centre is coupled to
the molecular tumbling and removing any potential alternative routes that would decrease the rotational correlation time. For
the best of these systems [gd.dENgdOTA]
-
(Figure 8.6), the observed relaxivity of 23.5 mmolgd
-1
s
-1
is very high indeed,
and the rotational correlation time of 390 ns, which the authors estimated by fitting the field-dependence of the observed T
1
data, is very long compared to that observed in simple complexes (for gd.dOTA, τ
R
= 56 ps) [71].
Alternatively, it is possible to exploit protein binding to change the rotational correlation time. Many lanthanide complexes
bind to serum albumin and other proteins in blood. This phenomenon has long been recognised and leads to an enhancement
of the rotational correlation time as a consequence of the increased bulk of the assembly [45]. Binding to human serum
albumin is favoured by lipophilic side chains appended to the binding site, including aryl groups and steroids. Both dOTA
and dTPA analogues have been shown to demonstrate these effects [72-75]. To deal with one example in detail, MS-325
(Figure 8.1) not only demonstrates strong affinity for serum albumin proteins [76], but also gives rise to very different vari-
ations in relaxivity with applied field between the protein bound and protein unbound forms of the complex. Because the
protein-bound form displays strong field dependence of the relaxivity around 1.5 T, while the unbound form displays very
little variation with field in this region, this can be exploited to provide direct images of albumin concentration. Use of
dreMR (a variable field approach to MRI imaging) allowed Caravan and co-workers to obtain images of protein concentration
following inflammation in the human hand (Figure 8.7) [77]. This quantitative approach to MRI (where the same probe is
used to provide both a signal and a reference channel depending upon the applied field) represents a landmark and one that
is sure to be developed further once whole body dreMR scanners become available.
Perhaps the most obvious trick of all is to incorporate more than one gadolinium centre into a molecular complex, thereby
increasing the local concentration of gadolinium. Provided such systems remain soluble, this approach has much to recom-
mend it, because there will often be complementary enhancements to the relaxivity arising from increases in the rotational
correlation time. For instance, in the m-xylyl bridged dO3A derivative (m-XyL, Figure 8.5), the observed relaxivity at
400 MHz in water is 13.8 mmolgd
-1
s
-1
(cf. 5.5 mmolgd
-1
s
-1
for gd.dO3A) [27]. In other words, each gadolinium contrib-
utes more than twice the relaxivity of an analogous monomeric system as a consequence of the increase in molecular diam-
eter. On a larger scale, Meade and co-workers have incorporated seven lanthanide complexes onto a β-cyclodextrin scaffold,
giving a system (HEPTAgAdOdEXTRIN, Figure 8.6) with q = 1 on each of the seven sites. The resulting array has
r
1
= 12.2 mmol
-1
gds
-1
(cf. 3.2 for a monomeric analogue) [78]. This approach gives rise to relatively small molecules with
