Modulating water-exchange rates of lanthanide(III)-containing polyaminopolycarboxylate-type complexes using polyethylene glycol

Abstract

We have synthesized a series of Ln^III-containing polyethylene glycol conjugates and studied the structural and electronic properties of these complexes. These studies demonstrate that polyethylene glycol can be used to fine-tune water-exchange rates of Ln^III-containing polyaminopolycarboxylate-type complexes; this control is desirable in developing Ln^III-containing contrast agents for magnetic resonance imaging.

The efficiency of Ln^III-containing complexes as conventional (T₁-reducing) and paramagnetic chemical exchange saturation transfer (PARACEST) contrast agents for magnetic resonance imaging (MRI) is governed by the interactions among the structural and electronic properties of these complexes and by the interactions of these complexes with the environment.^1,2 The optimal value for a specific parameter is a moving target that varies as a function of magnetic field strength and the other parameters.^1–4 Of these parameters, the exchange rate between coordinated and bulk water, k_ex, plays a critical role in establishing contrast-enhancing efficiency.^1,2 Hence, the ability to precisely control the water-exchange rates of Ln^III-containing complexes is imperative to achieving optimal efficiencies of contrast agents for both conventional and PARACEST imaging.

For example, conventional, Gd^III-containing, small molecular-weight agents require a fast water-exchange rate (~10⁸ s⁻¹) to achieve optimal efficiency at the clinically relevant field strength of 1.5 T,^3,5 and macromolecular Gd^III-based agents require a faster water-exchange rate (~10¹⁰ s⁻¹).⁴ However, these target values change with changes in field strength. Different from conventional agents, Ln^III-based PARACEST agents require slow-to-intermediate water-exchange rates on the NMR time scale, and the optimal value for the water-exchange rate depends on the frequency of the pulse used to presaturate the exchangeable protons.⁶ The estimated optimal water-exchange rate for PARACEST agents is between 0.3 × 10³ and 1 × 10³ s⁻¹ at clinically relevant pulse frequencies (50–100 Hz).⁶

The water-exchange rates of clinically approved Gd^III-containing small molecular-weight contrast agents are slower (1 × 10⁶ to 4 × 10⁶ s⁻¹) than the optimal value,⁷ and the water-exchange rates of most Ln^III-based complexes developed as potential PARACEST agents are faster (0.3 × 10⁴ to 1 × 10⁴ s⁻¹) than the optimal value.⁶ Consequently, there is a need to tune the water-exchange rates of Ln^III-containing complexes to achieve optimum efficiencies for conventional and PARACEST agents. This need has been the focus of a great deal of research. Tuning water-exchange rates of Ln^III-containing complexes has been achieved by modifying (1) the charge of the Ln^III-based complex;⁸ (2) the accessibility of the metal center to bulk water;^7,9–11 (3) the mechanism of water exchange;¹² and (4) the ratio between twisted square anti-prism and square anti-prism isomers for Ln^III-based complexes with 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA)-type ligands.¹³ These studies demonstrate that small modifications in ligand structure can have large impacts on the water-exchange rates of Ln^III-containing complexes, but most of these studies are specific to the system being studied.

We envisioned that a modular and tuneable system of modifications could be incorporated into Ln^III-based DOTA-type systems to tune water-exchange rates. We hypothesized that hydrophilic oligomers of polyethylene glycol (PEG) could be used to modulate water-exchange rates by altering the accessibility of bulk water to the Ln^III-center through both steric interactions and hydrogen bonding in a systematic fashion based on oligomer length. Here, we report the influence of PEG length on the molecular parameters, including water-exchange rate, of Ln^III-containing DOTA-type complexes.

To investigate the influence of PEG oligomers on the water-exchange rates of Ln^III-containing complexes, we designed and synthesized a PEG-containing model system (1a–d and 2a–d in Fig. 1) to modulate the water-exchange rates of Ln^III-DOTA-based complexes. We expected that PEG would hydrogen bond to water through the large number of oxygen-based hydrogen-bond acceptors, thereby changing the extent of the hydrogen-bonding network as PEG length is varied. Our hypothesis is consistent with reports of changes in water-exchange rates being observed as the length of PEG was varied with PEG-conjugates of Gd^III-containing hydroxypyridonate (HOPO) complexes.¹¹ The HOPO-based systems have a water-coordination number of two and display fast water-exchange rates (~10⁸ s⁻¹). The PEG moiety in those systems led to slower water-exchange rates and a decrease in water-coordination number from two to one. We expected that PEG could be used to fine-tune the water-exchange rates of other systems including Ln^III-containing DOTA-type complexes with water-coordination numbers of one and relatively slow water-exchange rates (~10⁶ s⁻¹).

Fig. 1.

Structures of Ln^III-containing PEG conjugates 1a–d and 2a–d and Y^III-containing conjugates 3a–d.

Our model system includes Ln^III-containing complexes without PEG (1a and 2a) as well as three complexes with different length oligomers of PEG (1b–d and 2b–d). The synthetic route to complexes 1a–d and 2a–d is shown in Scheme 1. Briefly, the α-chloroamide derivative of tert-butoxycarbonyl-protected ethylenediamine, 4, was reacted with tri-tert-butyl-1,4,7,10-tetraazacyclododecane-1,4,7-triacetate, 5, to obtain protected ligand 6 that was treated with HCl to yield deprotected ligand 7. Metallation was carried out using metal chlorides to obtain Gd^III-and Eu^III-containing complexes 8 and 9, respectively. Complexes 8 and 9 were reacted with succinimidyl ester derivatives of PEG or acetate (a, b, c, or d) to yield 1a–d or 2a–d. The yttrium analogues, 3a–d, were synthesized as diamagnetic controls for the variable-temperature ¹⁷O NMR experiments following the same procedures.

Scheme 1.

Synthetic route to Ln^III-containing PEG conjugates 1a–d and 2a–d and Y^III-containing conjugates 3a–d.

The water-coordination number, q, was obtained for Eu^III-containing complexes 2a–d using luminescence-decay measurements (Table 1).¹⁴ Values of q were close to 1, and these values agree with what is expected for Ln^III-complexes coordinated to octadentate ligands. Water-coordination number data for 2a–d suggest that steric blocking has little if any influence on the properties of the complexes.

Table 1.

Molecular parameters of complexes 1a–d.

Parameter	1a	1b	1c	1d
qa,bc	0.90 ±0.01	0.89 ± 0.01	0.96 ±0.02	0.76 ± 0.02
τm298 × 10−6 (s)a	0.37 ± 0.01	0.66 ± 0.01	1.2 ± 0.1	1.5 ± 0.07
kex298 ×106 (s−1)d	2.7 ±0.03	1.5 ±0.02	0.83 ±0.08	0.67 ± 0.05
T1e298 ×10−9 (s)a	440 ± 30	61 ± 0.7	12 ± 0.3	4.0 ± 0.04
ΔH (kJ mol−1)a	36 ± 0.8	50. ± 0.6	71 ± 4	90. ± 1.5
r1 (mM−1 s−1)a	2.5 ±0.1	4.0 ±0.1	5.3 ±0.1	6.3 ±0.1
MW (Da)	641	817	1463	5222
gL	1.98	1.98	1.99	1.98
1/T2e × 108 (s−1)	17.8	8.11	6.51	5.03
τR × 10−12 (s)	46	79	110	220

^areported as mean ± standard error,

^bfrom complexes 2a–d,

^cthe error associated with q determination is ±0.1 water molecules,¹⁴

^derror represents relative uncertainty

To investigate the influence of PEG on tuning water-exchange rates, we performed variable-temperature ¹⁷O NMR experiments for Gd^III-containing conjugates 1a–d, and the fitted molecular parameters are summarized in Table 1: the residence lifetime of coordinated water, τ_m²⁹⁸ (water-exchange rate, k_ex²⁹⁸, is 1/τ_m²⁹⁸; the longitudinal electronic relaxation time, T_1e²⁹⁸; and the enthalpy change for the water-exchange process, ΔH. Based on the results of the ¹⁷O NMR experiments, a gradual decrease in water-exchange rates was observed with increasing length of PEG from 1a (control without PEG) to 1d (long PEG). Moreover, 1.8-, 3.3-, and 4.0-fold decreases in water-exchange rates were observed from 1a to 1b, 1c, and 1d, respectively. The observed decrease in water-exchange rates with increasing length of PEG was supported by an increase in ΔH from 1a to 1d. This increase in enthalpy suggests that the exchange between coordinated and bulk water molecules becomes difficult as PEG length is increased, possibly due to a larger hydrogen-bond network. The increase in ΔH and decrease in water-exchange rates from 1a to 1d support our hypothesis that PEG is able to modify the water-exchange properties of Gd^III-containing DOTA-type complexes.

To investigate the influence of PEG on the efficiency of complexes 1a–d as conventional contrast agents, we carried out relaxivity, r₁, measurements of 1a–d at 37 °C and 1.4 T in phosphate-buffered saline (pH = 7.4). The r₁ values of 1a–d were obtained from the slopes of the plots of 1/T₁ versus Gd concentration (Table 1). A 2.5-fold increase in relaxivity was observed as the length of PEG was increased from 1a to 1d. This increase in relaxivity corresponds to the increase in molecular weight (MW, Table 1) from 1a to 1d. As the molecular weight increases, complexes tend to tumble more slowly in solution leading to higher relaxivity values.¹⁵

To more rigorously unite our ¹⁷O NMR and relaxivity data, we performed electron paramagnetic resonance spectroscopy measurements to obtain the electronic Landé g factor, g_L, and the transverse electronic relaxation rate, 1/T_2e, of 1a–d. These parameters together with r₁, k_ex, T_1e, and q were used with the Solomon–Bloembergen–Morgan equations, which describe the factors affecting the efficiency of contrast agents for MRI,³ to obtain estimated rotational correlation times, τ_R, for 1a–d (Table 1 and calculations in the ESI). The values of τ_R seem lower than expected likely due to the internal motion of the PEG moieties as suggested for the HOPO-based system,¹¹ due to the ineffective coupling between the motion of Gd-water vector and rotational motion of the entire molecule,¹⁶ or a combination of the two. An increase in the τ_R values was observed from 1a to 1d, which is consistent with the increase in molecular weight as the length of PEG is increased. This increase in τ_R is likely the cause of the observed increase in relaxivity from 1a to 1d. However, the increase in relaxivity observed from 1a to 1d is lower-than-expected based solely on the variation in τ_R from 1a to 1d (see ESI for calculations), and this observation is a likely result of the decrease in water-exchange rates as PEG length increases from 1a to 1d.

We have studied the influence of PEG oligomer length on the water-exchange rates and other properties that contribute to the efficiency of Ln^III-containing DOTA-type contrast agents. Based on our findings, PEG can be used to fine-tune, toward slower rates, the water-exchange rates of Ln^III-containing DOTA-type complexes. Our results demonstrate a similar magnitude of slowing of water-exchange rates as was reported with HOPO-based systems,¹¹ but without the associated change in water coordination number. Consequently, PEG is able to slow the water-exchange rates of Ln^III-containing complexes regardless of water-coordination number, demonstrating that conjugation of PEG represents a modular and tuneable strategy for slowing water-exchange rates that is general for Ln^III-containing complexes. We expect that these findings will be useful in the design of Ln^III-based contrast agents that require slow water-exchange rates on the NMR time scale.