Metal Chelating Crosslinkers Form Nanogels with High Chelation Stability

Abstract

We present a series of hydrogel nanoparticles (nanogels) incorporating either acyclic or cyclic metal chelates as crosslinkers. These crosslinkers are used to formulate polyacrylamide-based nanogels (diameter 50 to 85 nm) yielding contrast agents with enhanced relaxivities (up to 6-fold greater than Dotarem®), because this nanogel structure slows the chelator's tumbling frequency and allows fast water exchange. Importantly, these nanogels also stabilize Gd³⁺ within the chelator thermodynamically and kinetically against metal displacement through transmetallation, which should reduce toxicity associated with release of free Gd³⁺. This chelation stability suggests that the chelate crosslinker strategy may prove useful for other applications of metal-chelating nanoparticles in medicine, including other imaging modalities and radiotherapy.

Introduction

Nanogels, nanoscale hydrogel particles formed by chemically or physically crosslinked polymer chains, have attracted much attention in recent years for their high drug-carrying capacity, environmental responsiveness, and stability in aqueous media. ^1-2 While they have been employed in several studies to deliver small therapeutic agents ^3-4 and biomacromolecules, ^5-6 their advantages for use as biomedical imaging agents, such as water access throughout the particle ⁷ and protection of the payload have not been thoroughly explored.

Metals have many applications in medicine, including as contrast agents in multiple biomedical imaging modalities (PET, MRI), and as radiotherapeutics. Incorporating metals into nanogels can impart properties useful in medicine. However, no versatile and reproducible chemistry yet exists to incorporate metals into nanogels in a stable manner that is of clinical relevance. Clinical development of high-contrast magnetic resonance imaging (MRI) agents is an important area of investigation that relies on Gd³⁺ chelates. Gd based MRI agents could allow tissue- or disease-specific imaging using this accurate diagnostic imaging modality capable of providing submillimeter spatial resolution and unrivalled soft tissue contrast. ^8-10 Previously employed strategies combining Gd³⁺ chelates with nanogels include the encapsulation of small contrast agent molecules inside an electrostatically crosslinked polymer ¹¹ or the post polymerization functionalization of nanogels with contrast agents, ¹² both of which enhance relaxivity by slowing the chelates’ molecular tumbling rate. ^{14, 15} These designs follow the example of numerous other groups that attach Gd³⁺ chelates to high molecular weight macromolecules, including dendrimers, ^13-16 polymers, ^17-19 micelles ^20-21 and metallostars ²² or bind them to micro- ²³ or nanoparticles ^24-25 or proteins through covalent ^26-30 or non-covalent interactions. ^31-33By using gadolinium complexes directly as crosslinkers, the whole nanoparticle incorporates contrast agents, in contrast to post-functionalized nanoparticles or gadolinium-based inorganic nanoparticles, in which only surface gadolinium can relax water protons. This strategy also allows us to keep the nanogel surface available for further functionalization, for example with targeting agents. We also hypothesized that incorporating the contrast agent into a macromolecular network as a crosslinker would further increase rigidity and thus relaxivity. Importantly, metal chelating crosslinkers can be used with a variety of metals, such as radioisotopes for positron emission tomography or cancer radiotherapy. ³⁴

The design of the target crosslinker 1 (Figure 1) was based on a diethylenetriaminepentacetatic acid (DTPA) and was first selected because of its ease of synthesis. The second crosslinker structures were selected for their potential for greater safety in vivo; it has been reported that cyclic structures are more stable than DTPA, ³⁵ better protecting and binding the central Gd³⁺. A cyclic crosslinker would thus decrease toxicity, which can increase the risk of nephrogenic systemic fibrosis (NSF) via transmetallation with endogenous ions such as zinc or copper and resulting release of free Gd³⁺. Thus we decided to design two other crosslinkers based on the 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) chelating agent. The design of these macrocyclic crosslinkers 2 and 3 (Figure 1) was based on a DOTA unit bearing a bisacrylamide moiety. The DOTA macrocycle is C-substituted in order to avoid any interference with the coordination site of the ligand. This choice is based on previous observations that chelates with a hydration number greater than one are relatively unstable; q = 1 for all commercial agents. Therefore, though other designs allow a greater hydration number, which yields higher relaxivity, ^36-39 we chose C substitution to minimize dechelation. Furthermore, incorporating the chelate into a nanogel both enhances relaxivity by limiting movement and improves Gd³⁺ chelating stability by reducing access of other metal ions. To the best of our knowledge, this work is the first example of nanogels incorporating DOTA chelating agents as crosslinkers.

Figure 1.

Chemical structures of the target DTPA and DOTA-based crosslinkers 1, 2 and 3.

Herein we report the development of model polyacrylamide (PAA)-based nanogels incorporating DTPA- and DOTA-based crosslinkers. Given the size of these nanogels (<100 nm), which should allow prolonged circulation, these chelators should be useful in the development of biodegradable nanoparticles for a number of biomedical applications involving metals.

Results and discussion

Synthesis of the crosslinkers

Crosslinker 1 was obtained in a one-step synthesis by reacting commercially available DTPA-bisanhydride with N-(3-aminopropyl) methacrylamide hydrochloride in the presence of triethylamine. The adopted strategy for the synthesis of the two DOTA-based crosslinkers 2 and 3 was convergent (Scheme 1). The C-substituted protected DOTA derivatives 5 and 7 were prepared in 4 and 5 steps, respectively, according to a procedure reported in the literature. ⁴⁰ In parallel, the bisacrylamide derivatives 4 and 9 were synthesized in one or two steps from the commercial 3,5-diaminobenzoic acid and 3,5-dinitroiodobenzene, respectively. The latter was first reduced with tin chloride to give the 3,5-diaminoiodobenzene 8. The two diamino compounds were acrylated by using acryloyl chloride under Schotten-Baumann conditions (two-phase solvent system) to give 4 and 9. These bisacrylamide derivatives were then coupled with the protected DOTA derivatives 5 and 7 via coupling reactions: an ester coupling reaction to form the protected crosslinker 6, and a Sonogashira cross-coupling reaction to form the protected crosslinker 10. Finally, the two target crosslinkers 2 and 3 were obtained after a TFA deprotection of the tert-butyl ester functions of 6 and 10.

Scheme 1.

Synthesis of crosslinkers 1, 2 and 3. a) Et₃N, r.t., 24 h, 61%; b) acryloyl chloride, K₂CO₃, EtOAc/H2O, 0 °C to r.t., 30 min, 73%; c) DCC, DMAP, DCM, 0 °C to r.t., 48 h, 59%; d) TFA/DCM 1:1, r.t., 16 h, 100%; e) propargyl bromide, tetrabutylammonium iodide, CsOH, DCM, 0 °C to r.t., 1 h, 75%; f) SnCl₂.(H₂O)₂, 70 °C, 30 min, 100%; g) acryloyl chloride, EtOAc/H2O, 0 °C to r.t., 10 min, 100%; h) Pd(PPh₃)₂Cl₂, CuI, Et₃N, THF, 50 °C, 48 h, 59%; i) TFA, r.t., 4 h, 45%.

Nanogel Preparation

The crosslinkers 1, 2 and 3 were chelated using GdCl₃ and heating at 40 °C at pH 6. Excess gadolinium was removed by adding the chelating resin Chelex-100 and filtering the solution. The gadolinium complexes 1Gd(III), 2Gd(III) and 3Gd(III) were then used to prepare polyacrylamide-based nanogels PAA/1Gd(III), PAA/2Gd(III) and PAA/3Gd(III) through an inverse emulsion process using ammonium persulfate (APS) as the initiator and by adding tetramethylethylenediamine (TMEDA) to control the radical polymerization rate (Figure 2).

Figure 2.

Preparation of nanogels through inverse emulsion process.

Nanogel characterization

The nanogels obtained were characterized by dynamic light scattering (DLS) in order to evaluate their hydrodynamic volume and size distribution (Figure 3). PAA/1Gd(III) had an average size of 65 nm with a polydispersity index (PDI) of 0.16, PAA/2Gd(III) and PAA/3Gd(III) were on average 54 nm and 85 nm in diameter with PDI values of 0.25 and 0.39, respectively.

Figure 3.

Size distribution of PAA/1Gd(III), PAA/2Gd(III) and PAA/3Gd(III) nanogels as measured by dynamic light scattering.

Nanogels were then visualized (Figure 4) by transmission electron microscopy (TEM). Spherical nanogel particles with smaller sizes were observed, which is in accordance with the hydrodynamic sizes measured by DLS. Spherical voids seen on the TEM microphotographs are most likely generated by water escaping from the nanogel upon placement under vacuum and irradiation by the electron beam..

Figure 4.

TEM microphotographs of polyacrylamide nanogels with Gd³⁺-chelating crosslinkers PAA/2Gd(III)..

To evaluate the effect of nanogels on cell viability, murine macrophages were incubated with PAA/1Gd(III), PAA/2Gd(III) and PAA/3Gd(III) at various concentrations, and cytotoxicity was measured by MTT assay (Figure S1). From the MTT assay results, it appears that PAA/1Gd(III), PAA/2Gd(III) and PAA/3Gd(III) are well tolerated by the cells at concentrations up to 100 μg/ml.

Relaxivity studies

In order to test our central hypothesis, the relaxivity values of all Gd³⁺ complexes (crosslinkers alone and nanogels incorporating crosslinkers 1, 2, and 3) were calculated. T₁ was measured at 60 MHz (1.41 T), 37 °C in PBS buffer and exact concentrations of Gd³⁺ were measured by inductively coupled plasma atomic emission spectroscopy (ICP-AES). Under these conditions, the clinically used contrast agents Magnevist® (DTPA-Gd³⁺ complex) and Dotarem® (DOTA-Gd³⁺ complex) had relaxivities of 3.6 ± 0.3 mM^−1s−1 and 2.9 ± 0.3 mM^−1s−1, respectively. Proton NMRD profiles were also recorded at 37 °C and analyzed using the classical inner sphere and outer sphere theories (Figure 5). As expected from its larger molecular weight, the complexed crosslinker 1Gd(III) has a larger relaxivity than Magnevist®. The larger value of the fitted water residence time agrees with previous results showing that replacement of carboxylic groups with amides reduces the water exchange rate. ^{1, 41} The incorporation of the complexed crosslinker 1Gd(III) into nanogels resulted in an increase of relaxivity from 4.8 ± 0.2 mM⁻¹s⁻¹ for the crosslinker alone to 9.7 ± 0.5 mM⁻¹s⁻¹ for the corresponding PAA nanogel at 60 MHz and 37°C. This corresponds to a 2-fold improvement in relaxivity over Magnevist® after incorporation of the DTPA as a crosslinker. The decrease in mobility of the Gd complex in the nanogels was further confirmed by the NMRD profiles showing a “hump” between 10 and 100 MHz. The r₁ increase is lower than expected for PAA/1Gd(III) due to the slow water exchange rate resulting from the presence of the amide chelating groups.

Figure 5.

NMRD profiles of the chelated crosslinkers a) 1Gd(III), c) 2Gd(III) e) 3Gd(III) and their corresponding PAA nanogels b) PAA/1Gd(III), d) PAA/2Gd(III) and f) PAA/3Gd(III). The relaxivity measurements were performed at 37 °C. Data points shown in rhombus and fitted curve in solid line.

As expected, since the substitution of the DOTA macrocycle would slow molecular tumbling, our complexed crosslinkers 2Gd(III) and 3Gd(III) exhibited higher relaxivities than Dotarem®, of 4.3 ± 0.2 mM⁻¹s⁻¹ and 5.0 ± 0.3 mM⁻¹s⁻¹, respectively, at 60 MHz. The slightly higher relaxivity of 3Gd(III) compared to 2Gd(III) could be explained by its higher rotational correlation time τ_R (Table S1) corresponding to a slightly slower tumbling frequency. As anticipated, the NMRD profile shows that the τ_M, τ_SO and τ_V values are quite similar for both complexes and comparable to those of Gd-DOTA (Table S1). The relaxivity of nanogels incorporating 2Gd(III) and 3Gd(III) was greater than that of the complexed crosslinkers alone, at 17.6 ± 0.9 mM⁻¹s⁻¹ and 14.8 ± 0.7 mM⁻¹s⁻¹, respectively. This 3 to 4-fold increase in relaxivity can be explained by the relative rigidity of the system after crosslinking. These r₁ values are 5.1 to 6.1-fold improvements over Dotarem®. While these enhancements are limited by the slower water exchange rate in the nanogels, the faster water exchange rate as compared to PAA/1Gd(III) results in a faster relaxation rate of these two DOTA-based nanogels.

Based on the density of our nanogels (1.016 g/mL; see ESI), the Gd concentration, and the average nanoparticle volume (calculated from the nanoparticle size measured by DLS), we calculated the number of Gd³⁺ per particle to be 134. This value allowed us to obtain the relaxivity per particle, which was 2352 mM⁻¹.s⁻¹ in the case of PAA/2Gd(III).

Although the tumbling frequency of the crosslinker in the nanogel is reduced because the attachment of the bismethacrylamide or bisacrylamide linkers restricts each chelate's rotation, the motions of the linkers are still rapid. This internal flexibility becomes the limiting factor for further reduction in tumbling frequency and increase in relaxivity.⁴² This also explains the similar r₁ among nanogels with different crosslinking densities (see Figure S2). These nanogels have an overall denser structure at a higher crosslinking density; however, locally, the flexibility of the linkers remains high. As a result, increasing crosslinking density has no significant effect on r₁.

To be clinically translatable, an MRI contrast agent should minimize release of toxic Gd, which occurs mainly through transmetallation of Gd with endogenous metal ions. ^{7, 43-44} This released Gd³⁺ can accumulate in the body, sometimes causing nephrogenic systemic fibrosis, especially in patients with reduced kidney function. ^43-44 Among the endogenous ions, Zn²⁺ is most likely to displace Gd³⁺ because of its relatively high concentration in vivo, high affinity to metal chelates, and similar radius to that of Gd³⁺. The stability of all the chelated crosslinkers and their corresponding nanogels toward transmetallation was evaluated by monitoring their paramagnetic longitudinal relaxation rate 1/T₁ over time in the presence of other metal ions (Figure 6). Specifically, we followed a method described in the literature,^44-45 incubating our complexes at a Gd³⁺ concentration of 2.5 mM with 2.5 mM zinc chloride in phosphate buffer ([PO₄^3-] = 25 mM) at 37 °C. The concentration of Zn²⁺ used is at least twenty times greater than in plasma; this high concentration makes any difference in transmetallation apparent more rapidly. Upon transmetallation of Gd³⁺ with Zn²⁺, free Gd³⁺ ions are released and an insoluble Gd(PO₄) complex is formed, preventing rechelation of Gd³⁺ to DOTA. As a control, complexes were incubated without zinc chloride.

Figure 6.

Stability of the Gd³⁺ complexes of a) DTPA-based crosslinker 1 and its corresponding PAA nanogel, b) DOTA-based crosslinker 2 and its corresponding PAA nanogel, and c) DOTA-based crosslinker 3 and its corresponding PAA nanogel. Experiments were carried out in phosphate buffer with 2.5 mM ZnCl₂ at 37 °C and pH 7.4. Control experiments were carried out in the absence of ZnCl₂ (see Figure S3).

For all three DTPA-based contrast agents, r₁ decreases over time, indicating transmetallation with Zn²⁺. However, the transmetallation in the nanogel is much slower than that of Magnevist® and 1Gd(III), taking three to five times as long to decrease to 80% of initial r₁, respectively. This indicates that our nanogel PAA/1Gd(III) has a slower rate of transmetallation than the other two chelates. After 66 h, the nanogel's r₁ decreased by 23%, compared to ~40% and ~80% decreases for Magnevist® and 1Gd(III), respectively, suggesting that PAA/1Gd(III) transmetallates to a lesser extent. This stability likely results from greater structural rigidity³⁵ and reduced Zn²⁺ accessibility to chelating sites. In a control experiment without Zn²⁺, r₁ did not change appreciably for any of the three contrast agents (see Fig. S3). These results suggest that our nanogel, compared to Magnevist® or 1Gd(III), is more inert against dechelation and is potentially a safer MRI contrast agent.

DOTA-based crosslinkers 2 and 3 and their corresponding nanogels all appear highly stable, with 2Gd(III) and 3Gd(III) losing less than 5% of Gd³⁺ over three days. The nanogels’ relaxation rate did not decrease at all and instead increased slightly, likely because of evaporation of buffer during the three days of incubation. These results confirm that the incorporation of the crosslinkers into nanogels leads to enhanced stability, which is maximal in the case of the DOTA-based nanogels.

We also examined the contrast provided by all our Gd(III) complexes at 7 T of 1Gd(III), 2Gd(III), 3Gd(III) and their corresponding nanogels PAA/1Gd(III), PAA/2Gd(III), PAA/3Gd(III) by phantom imaging, with Magnevist® and Dotarem® as controls (see Figure S4). At identical gadolinium concentrations, better contrast was obtained with the nanogels compared to either chelated crosslinkers alone or the clinical contrast agents. Therefore, by using such systems, smaller doses of gadolinium would provide similar contrast to that obtained with clinically used contrast agents.

Conclusions

In this article, we have shown that incorporation of gadolinium chelate crosslinkers into nanogels leads to a significant (3 to 4-fold) increase in relaxivity. The relaxivities of the complexed DTPA crosslinker 1Gd(III) and DOTA crosslinkers 2Gd(III) and 3Gd(III) are 4.8 mM⁻¹s⁻¹, 4.3 mM⁻¹s⁻¹, and 5.0 mM⁻¹s⁻¹, respectively, at 60 MHz and 37 °C. When formulated into nanogels, these values rise to 9.7 mM⁻¹s⁻¹, 17.6 mM⁻¹s⁻¹, and 14.8 mM⁻¹s⁻¹, respectively. Importantly, DTPA-based nanogels are thermodynamically and kinetically more inert against transmetallation than Magnevist®, suggesting that crosslinker chelates may represent an important approach towards stable metal-chelating biomedical agents. Moreover, we have reported here the first example of nanogels incorporating a DOTA-based contrast agent as a crosslinker itself. This strategy yields a high relaxivity agent with better contrast than commercial agents and maximal stability toward transmetallation. We are now working on the incorporation of our crosslinkers into nanogels with a biodegradable polymer backbone to allow breakdown into biocompatible small molecules to facilitate clearance. We also plan to use different metals in these chelating crosslinkers to yield nanogels for theranostics combining different imaging modalities such as MRI (Gd³⁺) and PET (⁶⁴Cu) with radiotherapy (¹⁷⁷Lu).

Abbreviations

DCM

dichloromethane

DCC

N,N′-Dicyclohexylcarbodiimide

DLS

Dynamic Light Scattering

DMAP

4-Dimethylaminopyridine

DMF

dimethylformamide

GBCA

gadolinium-based contrast agents

HR ES-MS

high resolution electrospray mass spectrometry

LC-MS

Liquid Chromatography-Mass Spectrometry

MeOH

Methanol

NMR

nuclear magnetic resonance

ES-MS

electrospray mass spectrometry

NMRD

Nuclear Magnetic Relaxation Dispersion

PAA

polyacrylamide

PTA

Phosphotungstic Acid

TEA

triethylamine

TEM

Transmission Electronic Microscopy

TLC

thin layer chromatography