This protocol describes the preparation and characterization of a dendrimeric magnetic resonance imaging (MRI) contrast agent that carries cyclen-based macrocyclic chelates coordinating paramagnetic gadolinium ions. In a series of MRI experiments in vitro, this agent produced an amplified MRI signal when compared to the commercially available monomeric analogue.
Paramagnetic complexes of gadolinium(III) with acyclic or macrocyclic chelates are the most commonly used contrast agents (CAs) for magnetic resonance imaging (MRI). Their purpose is to enhance the relaxation rate of water protons in tissue, thus increasing the MR image contrast and the specificity of the MRI measurements. Current clinically approved contrast agents are low molecular weight molecules that are rapidly cleared from the body. The use of dendrimers as carriers of paramagnetic chelators can play an important role in the future development of more efficient MRI contrast agents. Specifically, the increase in local concentration of the paramagnetic species results in a higher signal contrast. Furthermore, this CA provides a longer tissue retention time due to its high molecular weight and size. Here, we demonstrate a convenient procedure for the preparation of macromolecular MRI contrast agents based on poly(amidoamine) (PAMAM) dendrimers with monomacrocyclic DOTA-type chelators (DOTA – 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetate). The chelating unit was appended by exploiting the reactivity of the isothiocyanate (NCS) group towards the amine surface groups of the PAMAM dendrimer to form thiourea bridges. Dendrimeric products were purified and analyzed by means of nuclear magnetic resonance spectroscopy, mass spectrometry, and elemental analysis. Finally, high resolution MR images were recorded and the signal contrasts obtained from the prepared dendrimeric and the commercially available monomeric agents were compared.
Magnetic resonance imaging (MRI) is a powerful and non-ionizing imaging technique widely used in biomedical research and clinical diagnostics due to its noninvasive nature and excellent intrinsic soft-tissue contrast. The most commonly used MRI methods utilize the signal obtained from water protons, providing high-resolution images and detailed information within the tissues based on differences in the density of the water signals. The signal intensity and the specificity of the MRI experiments can be further improved using contrast agents (CAs). These are paramagnetic or superparamagnetic species that affect the longitudinal (T1) and transverse (T2) relaxation times, respectively1,2.
Complexes of the lanthanide ion gadolinium with polyamino polycarboxylic acid ligands are the most commonly used T1 CAs. Gadolinium(III) shortens the T1 relaxation time of water protons, thus increasing the signal contrast in MRI experiments3. However, ionic gadolinium is toxic; its size approximates that of calcium(II), and it seriously affects calcium-assisted signaling in cells. Therefore, acyclic and macrocyclic chelates are employed to neutralize this toxicity. Various multidentate ligands have been developed so far, resulting in gadolinium(III) complexes with high thermodynamic stability and kinetic inertness1. Those based on the 12-membered azamacrocycle cyclen, in particular its tetracarboxylic derivative DOTA (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetate) are the most investigated and applied complexes of this CA class.
Nevertheless, GdDOTA-type CAs are low molecular weight systems, displaying certain disadvantages such as low contrast efficiency and fast renal excretion. Macromolecular and multivalent CAs may be a good solution to these problems4. Since CA biodistribution is mainly determined by their size, macromolecular CAs display much longer retention times within tissues. Equally important, the multivalency of these agents results in an increased local concentration of the monomeric MR probe (e.g., GdDOTA complex), substantially improving the acquired MR signal and the measurement quality.
Dendrimers are amongst the most preferred scaffolds for the preparation of multivalent CAs for MRI4,5. These highly branched macromolecules with well-defined sizes are prone to various coupling reactions on their surface. In this work, we report the preparation, purification, and characterization of a dendrimeric CA for MRI consisting of a generation 4 (G4) poly(amidoamine) (PAMAM) dendrimer coupled to GdDOTA-like chelates (DCA). We describe the synthesis of the reactive DOTA derivative and its coupling to the PAMAM dendrimer. Upon complexation with Gd(III), the standard physicochemical characterization procedure of DCA was performed. Finally, MRI experiments were performed to demonstrate the ability of DCA to produce MR images with a stronger contrast than those obtained from low molecular weight CAs.
1. Preparation of DCA
2. In Vitro Characterization of Dendrimeric Products
3. In Vitro MRI; Comparison Between DCA and GdDOTA
The preparation of DCA consisted of two stages: 1) synthesis of the monomeric DOTA-type chelator (Figure 1) and 2) coupling of the chelator with the G4 PAMAM dendrimer and subsequent preparation of the dendrimeric Gd(III) complex (Figure 2). In the first stage, a cyclen-based DOTA-type chelator containing four carboxylic acids and an orthogonal group suitable for further synthetic modifications was prepared. The preparation commenced from 1 (DO3A-tert-butyl ester)7, which was alkylated with tert-butyl 2-bromo-4-(4-nitrophenyl) butanoate8 to provide DOTA-derivative 2. The palladium-catalyzed hydrogenation reduced the aromatic nitro group in 2 to yield the aniline 3. The conversion of 3 with thiophosgene resulted in the isothiocyanate 4, which was previously used as an amine-reactive agent for the preparation of dendrimeric CAs17.
In the following stage, the macrocycle 4 was used as the basic monomeric unit in a coupling reaction to the commercially available G4 PAMAM dendrimer. The amine surface groups of the dendrimer react with the isothiocyanate groups of the monomer 4 in the presence of a base. The excess of 4 was removed by size-exclusion chromatography using a lipophilic gel filtration medium with methanol as the eluent. The tert-butyl esters on the obtained dendrimer-macrocyclic conjugate 5 were hydrolyzed with formic acid to yield 6, which was then lyophilized and used in the next step without purification. The formation of Gd(III) complexes of DOTA-type macrocycles was performed by adding GdCl3·6H2O to an aqueous solution of 6 while maintaining the pH at about 7. The excess of Gd(III) was complexed with a common chelator ethylenediaminetetraacetic acid (EDTA). The GdEDTA complex and excess EDTA were removed from the system by size-exclusion chromatography using a hydrophilic gel filtration medium with water as the eluent. The remaining small-size impurities were removed from the solution by centrifugation using 3 kDa centrifugal filtration units.
Following the synthesis of the dendrimer-macrocycle conjugates, a combined analytical approach has been employed to characterize the products. To determine the surface-amine occupancy of 5 and 6, 1H NMR spectra have been analyzed. The results were compared and confirmed with the final product (DCA), where the loading of the dendrimer with macrocycles has been estimated using elemental analysis and MALDI-TOF mass spectrometry (Figure 3). A combination of these three methods resulted in an average of 49 macrocyclic units being conjugated to the G4 dendrimer, which corresponds to ~75% amine surface group occupancy.
Further characterization of the dendrimeric complex included determination of the relaxivity values, resulting in 6.2 ± 0.1 mM-1sec-1 per Gd(III) (or roughly around 300 mM-1sec-1 per dendrimer) for the longitudinal relaxivity and 30.5 ± 0.6 mM-1sec-1 per Gd(III) (almost 1,500 mM-1sec-1 per dendrimer) for the transverse relaxivity. DLS measurements indicated a hydrodynamic diameter of 7.2 ± 0.2 nm for DCA (Figure 4).
Finally, to demonstrate the effect of the dendrimeric MRI contrast agent, MR imaging was performed on two sets of phantoms with DCA and the clinically available GdDOTA for comparison (Figure 5). The first set of phantoms were prepared for the purpose of comparing these two contrast agents at identical Gd(III) concentrations, while the second set was designed to demonstrate the effect at comparable molecule concentrations of the dendrimeric and monomeric contrast agents, respectively.
Figure 1: Synthesis of the macrocyclic DOTA-type chelator 4. Reagents, conditions, and isolated yields: (i) tert-butyl 2-bromo-4-(4-nitrophenyl)butanoate, K2CO3, DMF, 45 °C, 16 hr, 72%; (ii) H2, Pd/C, EtOH, RT, 16 hr, 95%; (iii) CSCl2, Et3N, RT, 2 hr, 53%. Please click here to view a larger version of this figure.
Figure 2: Synthesis of the dendrimeric MRI contrast agent DCA. Reagents and conditions: (i) 4, Et3N, DMF, 45 °C, 48 hr, 91%; (ii) formic acid, 60 °C, 24 hr, quant; (iii) GdCl3∙6H2O, pH 7.0, RT, 24 hr, 71%. Please click here to view a larger version of this figure.
Figure 3: Characterization of the dendrimeric product by means of MALDI-TOF mass spectrometry. A typical MALDI-TOF mass spectrum obtained for DCA. Please click here to view a larger version of this figure.
Figure 4: Characterization of the dendrimeric product by means of dynamic light scattering (DLS). DLS measurement of DCA (HEPES, pH 7.4). Please click here to view a larger version of this figure.
Figure 5: In vitro MRI experiments on tube phantoms at 7 T magnetic field. (a,b) T1-weighted and (c,d) T2-weighted MRI of DCA and GdDOTA. Each MRI experiment was performed with two different concentrations of the contrast agent: (a,c) with comparable Gd(III) concentrations (HEPES, pH 7.4); (b,d) with a DCA:GdDOTA concentration ratio of 1:5 (HEPES, pH 7.4). The concentrations are expressed per molecule and the SNR values are displayed in the parentheses. The parameters used in these experiments were: field-of-view (FOV) = 40 x 40 mm2, slice thickness = 0.5 mm, number of excitations (NEX) = 30; (a) matrix size (MTX) = 256 x 256, repetition time (TR) = 100 msec, echo time (TE) = 2.95 msec, flip angle (FA) = 90°, acquisition time (TA) = 12 min 48 sec; (b) MTX = 256 x 256, TR/TE = 20/2.95 msec, FA = 90°, TA = 2 min 34 sec; (c) MTX = 512 x 512, TR/TE = 10,000/130 msec, Rare factor (RF) = 16, TA = 26 min 40 sec; (d) MTX = 512 x 512, TR/TE = 10,000/100 msec, RF = 16, TA = 26 min 40 sec. Please click here to view a larger version of this figure.
Preparation of the dendrimeric MRI contrast agent requires appropriate selection of the monomeric unit (i.e., the chelator for Gd(III)). They reduce the toxicity of this paramagnetic ion and, to date, a wide variety of acyclic and macrocyclic chelators serve this purpose1-3. Among these, macrocyclic DOTA-type chelators possess the highest thermodynamic stability and kinetic inertness and, hence, are the most preferred choice for the preparation of inert MRI contrast agents1,18. Furthermore, they are prone to various synthetic transformations, which result in bifunctional chelators, capable of linking to various functional molecules (e.g., targeting vectors or nano-carriers) while still forming stable Gd(III) complexes19. To this end, the DOTA-type monomeric unit described in this procedure was prepared from DO3A-tert-butyl ester, the common and readily available precursor, and the bromide derivative of the 4-(4-nitrophenyl) butanoic acid. This molecule is derived from DOTA and possesses a similar structure to coordinate Gd(III). The synthetic modification aims to make this chelator prone to coupling reactions to various functional molecules and carriers. Namely, the preparation of the DOTA-modified molecule results in a chelator still with four carboxylic groups available for coordination to Gd(III) to form an inert complex and an orthogonal nitrophenyl group, which upon conversion attaches this chelator to the dendrimer surface. This procedure also allows for flexibility in the choice of the orthogonal reactive group (e.g., NH2 or COOH), which can serve to couple the Gd(III) chelator to a desired carrier in a preferred manner.
The obtained bifunctional chelator can be coupled to other molecules in two different ways (i.e., synthetic procedures). When the nitro group is reduced to an amino group, the resulting aniline can undergo a condensation reaction with the carboxylic acid group of the other molecule8. Moreover, an aromatic primary amine functional group in the presence of thiophosgene can be easily converted into an isothiocyanate, a group which readily reacts with amines in polar organic solvents as well as water, offering more reaction possibilities for the coupling of monomeric units to dendrimers17,20,21.
For coupling the bifunctional chelator to the dendrimeric carrier, an appropriate dendrimeric scaffold should be selected. Several factors related to the final dendrimer conjugate structure and the desired application should be accounted for in this step. Due to wide commercial availability of dendrimeric carriers, products with different core structures, surface-reactive groups, or generations can be chosen. Consequently, the conjugation reaction will depend on the surface group of the dendrimer and the orthogonal group of the chelator, while the final conjugate may be neutral, charged, or have different sizes (up to 15-20 nm, depending on dendrimer generation)22. All these aspects should be taken into account prior to preparing the dendrimeric CA, since they may affect the solubility, relaxivity (MRI signal enhancement), diffusion, and other pharmacokinetic properties of the contrast agent, which can potentially endanger its application in MRI. For instance, cationic dendrimers may exhibit toxicity in biological systems. However, this effect can be reduced by conjugation of negatively charged groups on the dendrimer surface, thereby reducing their overall positive charge23.
In this protocol, we have prepared the dendrimeric contrast agent DCA using the procedure in which the isothiocyanate group of the monomeric macrocycle 4 was coupled to a commercial cystamine-core G4-PAMAM equipped with 64 primary amine surface groups. The initial purification of the hydrophobic dendrimeric product 5 was performed by gel chromatography using a column with a lipophilic gel filtration medium and methanol as the eluent in order to remove most of the unreacted monomeric units. The hydrolysis of t-butyl esters with formic acid is straightforward, resulting in a water-soluble dendrimeric product that can be purified with size-exclusion chromatography using a hydrophilic gel filtration medium. The complexation of the multimeric and dendrimeric chelators with Gd(III) was performed whilst maintaining the solution at a neutral pH in order to facilitate the complex formation. Otherwise, the complexation of Gd(III) (added as the chloride salt) reduces the pH, slowing down the reaction. Finally, it is worth noting that amine groups in the dendrimer core also tend to coordinate with Gd(III), but only with the excess that could not be chelated with the DOTA units. Avoiding the presence of Gd(III) outside the DOTA chelator is essential, since leakage of Gd(III) from the CA may have undesired effects; namely, it can induce toxicity in vivo18. The excess Gd(III) can be effectively removed by complexation with EDTA followed by ultrafiltration of GdEDTA and free EDTA using 3 kDa molecular weight cut-off (MWCO) filters. Lower MWCO filters might be used when the dendrimeric conjugates have lower molecular weights.
There are two major troubleshooting issues related to the preparation of DCA. Due to the large broadening effect of Gd(III) on NMR signals, the analysis of DCA by means of NMR spectroscopy is not informative. Instead, this analysis should be performed in earlier steps (compounds 5 and 6). Next, the conjugation of monomacrocyclic units to the dendrimer surface is never accomplished with 100% conversion, but it is likely to be between 50-90% (see below). Typically, the reaction yields can be increased by adding a second portion of the monomeric reactive unit after the first conjugation of dendrimer and monomeric unit is completed24. However, every preparation batch results in somewhat different average numbers of chelators conjugated on the dendrimer surface, even when identical dendrimer and DOTA units are used as materials for coupling. Although the final amount of Gd(III) present in DCA can be determined independently via the BMS method (see section 2.2), for better characterization of dendrimeric conjugates, it is necessary to perform the estimation of bound monomeric units each time a new batch of DCA is prepared (see 2.1 and discussion below).
The analytical characterization of the isolated dendrimeric products can be performed by means of 1H NMR spectroscopy (only on products 5 and 6), elemental analysis, and MALDI-TOF MS. Typical yields for the conversion of amino surface groups lie between 50-90%, depending on the dendrimer generation, the type of chelator, and the reaction conditions used (solvent and temperature)6,20,24,25. In this particular case, the calculated masses obtained from the combined analyses correspond to an average of 49 monomeric chelates being coupled to the dendrimer (i.e., ~75% occupancy of the dendrimer surface amines). Although a slight mismatch in the final number of reacted amino groups could be expected between these methodologies25, their direct comparison provides reasonable evidence for the formation of the desired DCA with a particular average number of attached chelating units.
The in vitro characterization aiming to assess the potential of DCA to enhance the contrast in MRI experiments consisted of DLS, relaxometric, and MRI experiments. The hydrodynamic diameter of DCA was determined to be 7.2 ± 0.2 nm by DLS measurements, which is in agreement with previously reported conjugates of this kind with G4 generation 4 PAMAM dendrimers26. Determination of the longitudinal relaxivity of DCA followed the previously described procedure15 and revealed the value of 6.2 ± 0.1 mM-1sec-1 per Gd(III). About 50% of the enhancement in the r1 of paramagnetic Gd(III) in DCA relative to small-size molecules of a similar type (e.g., GdDOTA) can be explained with the intermediate size of the dendrimeric contrast agent. Namely, the reduced motion of the Gd-chelates attached to the dendrimer surface increases the rotational correlation time and, hence, r1; this effect can still be observed at high magnetic fields for smaller nano-sized agents. Otherwise, the increase in rotational correlation time dominantly contributes to r1 enhancement at low magnetic fields27. On the other hand, the size of the dendrimeric contrast agent had a pronounced effect on the transverse relaxivity28, resulting in the value of 30.5 ± 0.6 mM-1sec-1 per Gd(III). In summary, the methods for in vitro assessment of DCA are straightforward and require only careful sample preparation, so no difficulties are expected when acquiring data and analyzing the results.
To demonstrate the performance of the dendrimeric contrast agent and its power to affect the image contrast, we performed MRI experiments on tube phantoms with the newly prepared contrast agent DCA. We also used a solution of a commercially available and clinically approved MRI contrast agent, GdDOTA, as a comparison and tubes with water as a control. In the first T1-weighted MRI experiment, when equal Gd(III) concentrations were used (0.5 or 1 mM of Gd(III) in DCA or GdDOTA), the SNR in the tubes with DCA was already up to 12% higher due to an increase of about 50% in longitudinal relaxivity of DCA compared to GdDOTA (Figure 5a). The second T1-weighted MRI experiment was designed to demonstrate the effect of DCA when the concentrations were calculated per molecule. Although 5 times less DCA was applied compared to GdDOTA (50 vs. 250 µM or 100 vs. 500 µM DCA vs. GdDOTA, respectively), a high loading of DCA with Gd(III) resulted in a significant increment in the image contrast, which in turn resulted in the observed SNR values being at least three times higher in the phantom tubes filled with DCA. Expectedly, both T2-weighted MRI experiments exhibited large (3-20 times) differences in the SNR between the phantom tubes filled with DCA and GdDOTA.
In conclusion, this protocol describes a convenient preparation of a dendrimeric CA for MRI using common synthetic procedures to provide DCA with improved properties compared to small-size CAs. DCA exhibits preferred thermodynamic stability and kinetic inertness when compared to its monomeric CA analogues. Nevertheless, the multivalency of DCA and, hence, the high local concentration of the paramagnetic species in the target region induces high contrast in the MR images. Considering the often preferable pharmacokinetic properties (e.g., longer tissue retention time) compared to their monomeric CA analogues, or the ability to carry further functionalities (e.g., targeted vectors), these dendrimeric-macrocycle conjugates represent a promising and valuable class of contrast agents for various future MRI and molecular imaging applications.
The authors have nothing to disclose.
The financial support of the Max-Planck Society, the Turkish Ministry of National Education (PhD fellowship to S. G.), and the German Exchange Academic Service (DAAD, PhD fellowship to T. S.) are gratefully acknowledged.
Cyclen | CheMatech | C002 | |
tert-Butyl bromoacetate | Alfa Aesar | A14917 | |
N,N-Dimethylformamide | Fluka | 40248 | |
Potassium carbonate | Sigma-Aldrich | 209619 | |
4-(4-Nitrophenyl)butryic acid | Aldrich | 335339 | |
Thionyl chloride | Acros Organics | 382662500 | Note: Corrosive substance; toxic if inhaled |
Bromine | Acros Organics | 402841000 | Note: causes severe skin burns, fatal if inhaled |
Diethyl ether | any source | ||
Sodium sulphate | Acros Organics | 196640010 | |
Chloroform | VWR Chemicals | 22711.29 | |
tert-Butyl 2,2,2-trichloroacetimidate | Aldrich | 364789 | Note: flammable substance; irritrant to skin and eyes |
Boron trifluoride etherate | Acros Organics | 174560250 | 48 % BF3. Note: Flammable substance; causes skin burns, fatal if inhaled |
Sodium bicarbonate | Acros Organics | 424270010 | |
Ethyl-acetate | any source | For column chromatography | |
n-Hexane | any source | For column chromatography | |
Bulb-to-bulb (Kugelrohr) distillation apparatus | Büchi | Model type: Glass oven B-585 | |
Silicagel | Carl Roth GmbH | P090.2 | |
Methanol | any source | For column chromatography | |
Dichloromethane | any source | For column chromatography | |
Ethanol | VWR Chemicals | 20821.296 | |
Ammonia | Acros Organics | 428381000 | 7N Solution in Methanol |
Palladium | Aldrich | 643181 | 15 % wet |
Hydrogenation apparatus PARR | PARR Instrument Company | ||
Celite 503 | Aldrich | 22151 | |
Sintered glass funnel | any source | ||
Thiophosgen | Aldrich | 115150 | Note: irritrant to skin; toxic if inhaled |
Triethylamine | Alfa Aesar | A12646 | |
Dichloromethane | Acros Organics | 348460010 | Extra dry |
Magnetic stirrer | any source | ||
PAMAM G4 Dendrimer | Andrews ChemService | AuCS – 297 | 10 % wt. solution in MeOH |
Lipophylic Sephadex LH-20 | Sigma | LH20100 | |
Thin-layer chromatography plates | Merck Millipore | 1.05554.0001 | |
Formic acid | VWR Chemicals | 20318.297 | |
Lophylizer | any source | ||
Gadollinium(III) chloride hexahydrate | Aldrich | G7532 | |
Sodium hydroxide | Acros Organics | 134070010 | |
pH meter | any source | ||
Ethylenediaminetetraacetic acid disodium salt dihydrate | Aldrich | E5134 | |
Mass spectrometer (ESI) | Agilent | Ion trap SL 1100 | |
Acetate buffer | any source | pH 5.8 | |
Xylenol orange | Aldrich | 52097 | 20 μM in acetate buffer |
Hydrophylic Sephadex G-15 | GE Healthcare | 17-0020-01 | |
Amicon Ultra-15 Centrifugal Filter Unit | Merck Millipore | UFC900324 | Ultracel-3 membrane (MWCO 3000) |
Centrifuge | any source | ||
NMR spectrometer | Bruker | Avance III 300 MHz | |
Topspin | Bruker | version 2.1 | |
Combustion analysis instrument | EuroVector SpA | EuroEA 3000 Elemental Analyser | |
MALDI-ToF MS instrument | Applied Biosystems | Voyager-STR | |
Deuteriumoxid | Carl Roth GmbH | 6672.3 | |
tert-Butyl alcohol | Carl Roth GmbH | AE16.1 | |
Vortex mixer | any source | ||
Norell NMR tubes | Deutero GmbH | 507-HP-7 | |
NMR coaxial tube | Deutero GmbH | coaxialb-5-7 | |
DLS instrument | Malvern | Zetasizer Nano ZS | |
0.20 μm PTFE filter | Carl Roth GmbH | KC94.1 | |
HEPES | Fisher BioReagents | BP310 | |
Plastic tube vials | any source | ||
Dotarem | Guerbet | NDC 67684-2000-1 | |
MRI scanner | Bruker | BioSpec 70/30 USR magnet (7 T). Note: potential hazards related to high magnetic fields | |
RF coil | Bruker | dual frequency volume coil (RF RES 300 1H/19F 075/040 LIN/LIN TR) | |
Paravision (software) | Bruker | Version 5.1 |