Since the concept of using macromolecular contrast agents for MRI appeared more than a decade ago (1), various macromolecular gadolinium (Gd) (III) complexes have been reported in the literature. These complexes are prepared either by the conjugation of Gd(III) chelates to biomedical polymers including poly(amino acids) (2, 3), polysaccharides (4, 5), dendrimers (6, 7), and proteins (8), or by the copolymerization of DTPA dianhydride with diamines and the complexation with Gd(III) (9–11). Macromolecular Gd(III) complexes have demonstrated superior contrast enhancement for MR imaging of neoplastic and vascular systems where currently available agents are less effective (12–15). Unfortunately, the clinical application of macromolecular agents, including those prepared from typical biodegradable polymers (e.g., dextrans, polylysine, etc.), is limited by their slow excretion after the MRI examination and consequent long-term tissue accumulation of toxic Gd(III) ions (2, 16, 17). Until now, no macromolecular Gd(III) complex was available for clinical applications due to this safety concern. The limitation of these agents is that the degradation of the biomedical polymers is an enzymatic process, occurring in the cellular enzyme compartments. The degradation rate often dramatically decreases with chemical modification (18). Innovative approaches are necessary to develop safe, effective macromolecular agents for clinical use.
The disulfide-thiol exchange reaction is present in many biological systems. Thiols are important endogenous biomolecules, with a concentration of ∼15 μM in human blood plasma (19, 20). We hypothesize that the incorporation of disulfide bonds into the backbone of Gd(III)-DTPA containing macromolecules can result in macromolecular agents that can be broken down into smaller excretable complexes by endogenous thiols. The endogenous thiols, including cysteine and glutathione, can readily reach the disulfide bonds in the polymer backbones, resulting in extracellular degradation of the macromolecules via the disulfide-thiol exchange reaction. Because the plasma thiol concentration is quite low, the degradation of the macromolecules is a slow process, and thus can provide an acceptable time window for effective contrast enhancement in MRI. Thus, the macromolecular agent will be gradually degraded into low molecular weight Gd(III) chelates that can be readily cleared via renal glomerular filtration.
We designed novel extracellular degradable macromolecular Gd(III) complexes based on this hypothesis using polydisulfide to facilitate the excretion of Gd(III) chelates of macromolecular agents. Biodegradable polydisulfide Gd(III)-DTPA complexes were first synthesized and characterized. The degradability of the polydisulfide agent was evaluated by incubation with 15 μM cysteine under physiological conditions. The possible cross-reaction of the agent with the cysteine-containing proteins was investigated by incubation with human serum albumin (HSA), which has a cysteine-34 residual. The blood pool contrast enhancement and renal clearance of the agents with two different molecular weights (35,000 and 17,700 Da) were studied by MRI in rats with Gd-(DTPA-BMA) as a control. The metabolic products in rat urine were characterized with mass spectrometry.
MATERIALS AND METHODS
DTPA was purchased from J.K. Baker (Philipsburg, NJ). Cystamine dichloride was purchased from Lancaster Synthesis (Pelham, NH). GdCl3 was purchased from Aldrich Chemical Co. (Milwaukee, WI). HSA was purchased from Sigma (St. Louis, MO). All chemicals were used without further purification. DTPA-dianhydride was prepared according to the literature (21).
DTPA-Cystamine Copolymers (I)
DTPA dianhydride (8.0 g, 22 mmol), cystamine dihydrochloride (5.0 g, 22 mmol) and triethylamine were stirred in 100 ml DMSO at room temperature for 48 hr (9). The polymers were precipitated by adding the reaction mixture dropwise in acetone with stirring. The precipitate was collected by filtration, washed with acetone, and dried under vacuum. The polymers were dissolved in deionized water, dialyzed with a membrane with a molecular weight cutoff of 6000–8000 Da against deionized water, and then lyophilized. Yield: 3.60 g, 32%. The molecular weights of the copolymers were determined by size exclusion chromatography (SEC) on a Superose 6 column (HR 30/10, Pharmacia, Gaithersburg, MD) equipped with a refractive index detector. Standard poly[N-(2-hydroxypropyl)methacrylamide]s were used for the molecular weight calibration. Proton NMR (ppm, D2O): 2.79 (t, 4H, NHCH2CH2-S-), 3.04 (d, 4H, N-CH2-CO-NH), 3.25 (m, 8H, N-CH2CH2-N), 3.39 (s, 4H, N-CH2COOH), 3.48 (t, 4H, NH-CH2CH2-S-), 3.65 (s, 2H, NCH2COOH).
(Gd-DTPA)-Cystamine Copolymers (II)
DTPA-cystamine copolymers (1.1 g) were dissolved in 5 ml deionized water and GdCl3 (0.6 g, excess based on the monomer chelate) was added to the solution with stirring. An excess of GdCl3 was used to assure complete complexation of DTPA moieties in the copolymers. Xylenol orange solution was added as an indicator to monitor the complexation of Gd(III) ions with the polymeric ligands. The pH of the solution was adjusted to 5–5.5 with dilute NaOH solution. Excess Gd(III) ions were removed by SEC with the Sephadex G-25 media (Pharmacia), and then eluted with deionized water. The polymer fraction was collected and lyophilized. Yield: 0.5 g, 35%. The Gd content was determined by ICP-OES (Perkin Elmer, Norwalk, CT, Optima 3100XL).
DTPA-Cystamine Copolymers (III)
DTPA dianhydride (1.71 g, 4.8 mmol), cystamine (0.72 g, 4.8 mmol), and triethylamine were stirred in 5 ml DMSO at room temperature for 48 hr (9). The polymers were isolated and characterized similarly as described for copolymers I. The copolymers were fractionated with SEC. The fraction of high molecular weight was collected. Yield: 0.3 g, 12%.
(Gd-DTPA)-Cystamine Copolymers (IV)
Copolymers III (255 mg) were dissolved in 2 ml deionized water and GdCl3 (140 mg, excess) was added to the solution with stirring. The copolymer complexes were isolated and characterized similar to that described for copolymers (II) Yield: 0.22 g, 66%.
In Vitro Degradation of (Gd-DTPA)-Cystamine Copolymers
(Gd-DTPA)-cystamine copolymers were incubated with 15 μM L-cysteine in PBS buffer (pH 7.4) at 37°C. The samples were taken at 5, 15, 60, 120, and 360 min and the molecular weight distribution of the copolymers was analyzed by SEC. In a separate experiment, (Gd-DTPA)-cystamine copolymers were incubated with an excess of cysteine in PBS at 37°C for 30 min. The reaction mixture was analyzed by MALDI-TOF mass spectrometry.
Incubation of (Gd-DTPA)-Cystamine Copolymers With HSA
(Gd-DTPA)-cystamine copolymers (II, 1.8 mg) and HSA (3.0 mg) were incubated with and without 15 μM cysteine, respectively, in 2.0 ml PBS buffer (pH 7.4) at 37°C. The reaction mixtures were characterized by SEC at 0, 1, 2, and 6 hr after incubation. The possible cross-reaction between the copolymers and HSA was analyzed based on the molecular weight distribution of the HSA and the copolymers in the incubation.
In Vivo Metabolism
Polydisulfide agents II and IV were intravenously injected into male Sprague-Dawley rats (500–550 g) at a dose of 0.1 mmol/kg via a tail vein after anesthetizing with sodium pentobarbital (35 mg/kg). The rats were placed in metabolic boxes and urine samples were collected at 24 hr postinjection. The samples were analyzed with MALDI-TOF mass spectrometry.
The T1 and T2 of the macromolecular complexes of different concentrations were measured on a 1.5 T Sigma GE NV/CVi scanner with a LX 8.4 operating system using inversion recovery and spin echo techniques, respectively. T1 measurements were obtained using a B1 homogeneity-corrected Look-Locker technique (22). Nonlinear curve fitting was performed using CURVEFIT in IDL (Interactive Data Language v. 5.3, 1999, Research Systems, Boulder, CO). The CURVEFIT function uses a gradient-expansion algorithm to compute a nonlinear least squares fit. T2 measurements were obtained using a 4-echo CPMG technique with very wide refocusing pulses to minimize signal loss. The data was found to be linear on a semilog plot out to the 60 ms echo. The IDL LINFIT function, which minimizes the chi-square error statistics, was used to fit the T2 data. The T1 and T2 relaxivities were calculated based on the slope of the plot of 1/Ti vs. Gd(III) concentrations.
Contrast-Enhanced MRI in Rats
The in vivo MRI vascular contrast enhancement of the degradable macromolecular agents was investigated in male Sprague-Dawley rats (180–200 g). The rats were anesthetized by the intraperitoneal administration of sodium pentobarbital at a dose of 35 mg/kg. A solution of (Gd-DTPA)-cystamine copolymers (II or IV) was injected i.v. via a tail vein. A control experiment was also performed with the injection of Gd-(DTPA-DMA) at a dose of 0.1 mmol/kg. Four rats were used for each experimental group. MRIs were acquired at 2, 7, 12, 17, 22, and 27 min after the injection on a 1.5 T Sigma GE NV/CVi MR scanner with an LX 8.4(m4) operating system, using a 3D spoiled gradient echo (SPGR) pulse sequence. The system body coil was used for RF excitation and a custom-built two-element phased array coil was placed beneath the rat for RF reception. Each coil loop is 4.5 × 7 cm (23). Imaging parameters used were 1.8 msec echo time (TE), 8 msec repetition time (TR), 60° RF tip angle, 31.25 kHz receiver bandwidth, 180 mm field of view, one average, 1.6 mm sagittal slice thickness and a 256 × 256 × 32 acquisition matrix interpolated to 512 × 512 × 64. MR images were evaluated on a GE (Madison, WI) Advantage Workstation with v. 3.1 software. Regions of interest (ROIs) were set on the aorta, urinary bladder, and kidney. In each animal the same ROIs were investigated at various time points after contrast.
Properties of (Gd-DTPA)-Cystamine Copolymers
The number average (Mn) and weight average (Mw) molecular weights of the polydisulfide DTPA ligands were 20,600 and 32,000 Da for copolymers I, and 68,300 and 94,200 Da for copolymers III. The Mn and Mw of the polydisulfide Gd(III)-DTPA complexes were 15,000 Da and 17,700 Da for copolymers II, and 27,000 and 35,000 Da for copolymers IV. The molecular weights of the (Gd-DTPA)-cystamine copolymers II and IV were smaller because they are nonionic after Gd(III) complexation and have a smaller hydrodynamic volume than the ionic DTPA-cystamine copolymers I and III. All DTPA-cystamine copolymers and [Gd(III)DTPA]-cystamine copolymers are water soluble. The Gd content was 22.2% and 22.0% for the copolymers II and IV, respectively, which were very close to the calculated content (22.92%) based on the formula of [(Gd-DTPA · H2O)-cystamine]n. The proton T1 and T2 relaxivities were 4.42 and 5.62 mM−1s−1 per complexed Gd(III) ion for copolymers II, and 6.28 and 5.74 mM−1s−1 for copolymers IV, respectively.
In Vitro Degradation of (Gd-DTPA)-Cystamine Copolymers
The degradability of the (Gd-DTPA)-cystamine copolymers was investigated by the incubation of copolymers II with L-cysteine under physiological conditions. The changes of molecular weight distribution of the copolymers are shown in Figs. 1 and 2. The L-cysteine concentration was 15 μM, mimicking the thiol concentration in the plasma. A low concentration (0.42 mM-Gd) (Fig. 1) and a high concentration (1.4 mM-Gd) (Fig. 2) of copolymers II were selected in the degradation study, respectively, representing the initial Gd(III) plasma concentrations at an effective low dose (0.03 mmol-Gd/kg) and a normal administration dose (0.1 mmol-Gd/kg) of a contrast agent. The initial plasma concentrations of Gd(III) complexes were estimated based on an average American male weighing 75 kg. The peak intensity of small molecules around 55 min elution time increased with the concomitant decrease of the polymer, indicating the formation of smaller molecules. The decrease of the intensity of this peak in the 6-hr incubation mixtures was caused by signal interference of a following negative peak. The copolymers degraded relatively quickly at the low concentration. The Mw was 12,800, 11,800, and 9,300 Da for copolymers II of the low concentration at 5, 15, and 60-min incubation times, respectively (Fig. 1). The Mw was 14,200, 13,400, and 10,100 Da for the copolymers of the high concentration at 5, 15, and 60 min, respectively (Fig. 2). With a longer incubation time, up to 6 hr, most of the copolymers were broken into oligomers and small complexes.
The incubation of (Gd-DTPA)-cystamine copolymers II with an excess of cysteine for 30 min at 37°C resulted in complete degradation of the copolymers into the smallest repeating units (Scheme 1), as shown by MALDI-TOF mass spectrometry. Complexes V (m/z = 667, M++1) and VI (m/z = 786, M++1) were identified in the mass spectrum and no oligomers were recognized. The peak intensity of complex V was much higher than that of complex VI, suggesting more complex V was formed with excess cysteine. The characterization of the degradation products confirmed the structure of the polydisulfide-based Gd-DTPA complexes and the degradation mechanism via the disulfide-thiol exchange reaction.
Cross-Reaction of (Gd-DTPA)-Cystamine Copolymers With HSA
No change of the molecular weight distribution was observed for the incubation of human serum albumin (HSA) with an excess of the polydisulfide agent II (Fig. 3). The incubation in the presence of 15 μM cysteine resulted in the degradation of the copolymers and the molecular weight distribution of HSA remained the same (Fig. 4). These results suggest that no cross-reaction occurred between HSA and the copolymers.
In Vivo Metabolism of (Gd-DTPA)-Cystamine Copolymers
The in vivo degradation of the polydisulfide agents was more complicated than their in vitro degradation. Several metabolites of (Gd-DTPA)-cystamine copolymers were recognized in the mass spectra of the rat urine samples after administration of agents II and IV (Fig. 5). The major metabolites have mass (m/z) of 705 [V + 39 (K+)], 988 [V + 322 (glutathione – 2H + O + H+)], 1250 (unknown) and 1369 [V2 + 39 (K+)]. The spectra of all the metabolites exhibited the mass distribution pattern of the stable Gd isotopes. The metabolites were clearly the in vivo degradation products of the polydisulfide agents, confirming that the agents were metabolized into low molecular complexes that were excreted via renal filtration.
Contrast-Enhanced MRI of Rats
Figure 6 shows sagittal MRI contrast-enhanced images of rat hearts and aortas after the i.v. administration of (Gd-DTPA)-cystamine copolymers II at doses of 0.1 and 0.03 mmol-Gd/kg, and IV at doses of 0.06 and 0.03 mmol-Gd/kg and Gd-(DTPA-BMA) at a dose of 0.1 mmol-Gd/kg at various time points postinjection. The macromolecular agents produced significant contrast enhancement, especially in the heart and aorta, at high doses (0.1 mmol-Gd/kg for agent II and 0.06 mmol-Gd/kg for agent IV) 2 min after injection. Following this, gradual reduction of the signal intensity was observed in the heart and aorta for both agents. Agent IV did not show prolonged blood pool contrast enhancement 7 min postinjection, although it has a larger size than agent II. No significant difference in blood pool contrast enhancement was observed for both agents II and IV at high doses. Slight contrast enhancement was also observed for agent IV 2 min after injection at a low dose of 0.03 mmol-Gd/kg. However, no significant contrast enhancement was observed for the macromolecular agent II at a dose of 0.03 mmol-Gd/kg and Gd-(DTPA-BMA) (0.1 mmol-Gd/kg). The macromolecular agents of both molecular weights and Gd-(DTPA-BMA) produced considerable contrast enhancement in the kidney at high doses (Fig. 7). The contrast enhancement of the macromolecular agents at a dose of 0.03 mmol-Gd/kg was less significant. In all cases, the contrast enhancement in the kidney was strong at 2 min postinjection and gradually decreased.
Figure 8 shows images of the rat urinary bladder at various time points after the administration of the contrast agents. In the first 2 min postinjection images, no contrast enhancement was observed for Gd-(DTPA-BMA) and the macromolecular agents. Contrast enhancement in the bladder was first observed 7 min postinjection, with a progressive increase in the signal intensity. The increase of the signal intensity indicated the accumulation of the Gd(III) chelates in the bladder in all cases. The signal intensity in the bladder increased in a similar pharmacokinetic pattern for both the macromolecular agents and Gd-(DTPA-BMA). This indicated that the Gd(III) chelates of the macromolecular agent excreted into the urinary bladder via renal filtration, similar to Gd-(DTPA-BMA).
Extracellular degradable macromolecular Gd(III) complexes have been designed and prepared by incorporating cleavable disulfide bonds in the backbones of macromolecules to facilitate the excretion of the Gd(III) chelates. The in vitro degradation studies demonstrated the degradability of (Gd-DTPA)-cystamine copolymers via the disulfide-thiol exchange with cysteine at the plasma thiol concentration. The size reduction rate of the polydisulfides was concentration-dependent. At the fixed thiol concentration, the size of the copolymers reduced more rapidly at the low dose compared to the higher dose (Figs. 1, 2). The copolymers at both doses degraded into small, excretable oligomers and complexes with a longer incubation time, up to 6 hr, indicating their biodegradability by the endogenous thiols.
There might be a concern over the possible cross-reaction of cysteine-containing proteins with the polydisulfide agent. However, no such reaction was observed between the cysteine-34 residue on HSA and the polydisulfide. In the absence of cysteine, the molecular distribution of the copolymers and HSA did not change after 6 hr of incubation (Fig. 3). In the presence of 15 μM cysteine, the copolymers degraded over time, and no change was observed for HSA (Fig. 4). The location of the cysteine residue on the protein may affect the cross-reaction with the polydisulfide. The steric effect may prevent the reaction of the copolymers with HSA. If there were an interaction between the copolymers with cysteine residues on proteins, the endogenous thiols would eventually remove the attached Gd(III) complexes from the proteins by the disulfide-thiol exchange.
(Gd-DTPA)-cystamine copolymers showed dose-dependent contrast enhancement for the blood pool (Fig. 6). No contrast was observed in the heart and vascular systems for the low molecular weight copolymers II at the low dose (0.03 mmol-Gd/kg) and for Gd-(DTPA-BMA) at 0.1 mmol-Gd/kg. Only slight contrast enhancement was observed for the high molecular weight copolymers IV at the low dose. The rapid reduction of the size of the macromolecular agents at the low dose, as shown in the in vitro degradation studies, may result in rapid extravasation of Gd(III) chelates such that by 2 min postinjection the blood concentration of the contrast agents was not high enough for effective contrast enhancement, especially in the case of the high molecular agent IV. Although the molecular weight of agent IV is ∼35,000 Da, the linear macromolecular agent has a similar hydrodynamic volume to HSA, as shown in size exclusion chromatography. However, it did not result in prolonged contrast enhancement compared to a prototype blood pool contrast agent based on HSA at relatively low doses (24, 25). Significant cardiac and aortic contrast enhancement was observed at the relatively high doses 2 min after contrast injection and the signal intensity then rapidly decreased for both agents II and IV. Rapid decrease of blood pool contrast enhancement might be expected for agent II because it has a small hydrodynamic volume and can be categorized as a rapid clearance blood pool agent (26). Although agent IV was expected to show more prolonged blood pool contrast enhancement due to its larger hydrodynamic volume (26), no significant difference was observed between the agents of different size. This may be attributed to degradability of the agent and rapid extravasation after degradation.
The analysis of the rat urine samples after injection with the polydisulfide agents of both low and high molecular weights demonstrated the in vivo degradation of the macromolecular agents. The agents with both molecular weights produced the same metabolites because they have the same basic structure. The detection of metabolites of smallest repeat unit (V), dimer of V (V2), cross-reaction product of monomer V with glutathione and other unknown metabolites confirmed that the polydisulfide agents were degraded into small, excretable complexes by the endogenous thiols. The in vivo degradation of the copolymers facilitated the excretion of the Gd(III) chelates via renal filtration. Consequently, the oligomers and small Gd(III) complexes gradually accumulated in the urinary bladder, resulting in increasing contrast enhancement over time (Fig. 8). The comparison of the bladder images enhanced by the copolymers and Gd-(DTPA-BMA) suggests that Gd(III) complexes from the copolymers of both molecular weights excrete in a similar pharmacokinetic pattern as Gd(III)-(DTPA-BMA).
Macromolecular contrast agents based on degradable biomedical polymers, including polylysine and dextrans, have been reported previously (2, 4, 8, 13, 16, 17). These agents were prepared by the conjugation of Gd(III) chelates into the polymers, which are primarily degraded by lysosomal enzymes. The cellular uptake of these macromolecules is a slow, passive process, and chemical modification of the polymers often inhibits enzymatic degradation (18). In addition, lysosomal metabolism may release toxic Gd(III) ions. In contrast, the polydisulfide Gd(III) complexes are degraded by the plasma thiols, which is an extracellular process. Since the plasma thiols can easily reach the disulfide bonds in the polymer chains, the degradation of the polydisulfide and subsequent excretion of Gd(III) complexes is a rapid process compared to the agents based on typical biodegradable polymers. The risk of the release of toxic Gd(III) ions due to enzymatic metabolism is also significantly reduced. The polydisulfide agent may have a potential of being repeatedly administered for initial diagnosis and subsequent assessment of the therapeutic response of the disease.
The degradability of our macromolecular contrast agents by extracellular endogenous thiols is critical to address the toxicity issues impeding the clinical development of macromolecular contrast agents. Further studies are required to fully understand the safety and effectiveness of the novel polydisulfide Gd(III) complexes as blood pool contrast agents. Detailed studies including pharmacokinetics and safety of the novel agents are currently ongoing.
The polydisulfide-based macromolecular Gd(III) complexes have demonstrated degradability by endogenous thiols, including cysteine. The degradation rate was relatively slow at the plasma thiol concentration, suggesting a slow breakdown of the macromolecular agents in the plasma. The identification of degradation products in rat urine samples after intravenous injection of the agents further confirmed the in vivo biodegradability of the polydisulfide agents. Preliminary in vivo studies in experimental animals have shown that the agent provides superior contrast enhancement for MR blood pool imaging to Gd(DTPA-BMA) and clears rapidly via renal filtration. This novel design of extracellular biodegradable macromolecular MRI contrast agents may have great potential to solve the safety concerns impeding the clinical application of macromolecular Gd(III) complexes.
We thank Mr. Chien-Wen Chang for technical assistance in the synthesis of the macromolecular Gd(III) complexes. Mass spectral data were acquired at the University of Utah Mass Spectrometry Facility.