Polymeric Conjugates of Gd3+−Diethylenetriaminepentaacetic Acid

The LALLS detector was a Chromatix KMX6 (Milton Roy, Riviera Beach, FL). The refractometric detector was a Waters R 410 (Milford, MA). The SEC columns...
0 downloads 0 Views 188KB Size
Bioconjugate Chem. 1997, 8, 605−610

605

TECHNICAL NOTES Polymeric Conjugates of Gd3+-Diethylenetriaminepentaacetic Acid and Dextran. 1. Synthesis, Characterization, and Paramagnetic Properties Richard Rebizak,† Michel Schaefer,‡ and E Ä dith Dellacherie*,† Laboratoire de Chimie-Physique Macromole´culaire, URA CNRS 494, ENSIC-INPL, B.P. 451, 54001 Nancy Cedex, France, and Guerbet-GCA, B.P. 15, 93601 Aulnay-sous-Bois, France. Received August 5, 1996X

Macromolecular conjugates of dextran and diethylenetriaminepentaacetic acid (DTPA), aimed to complex gadolinium, were synthesized to obtain contrast agents for nuclear magnetic resonance imaging with good paramagnetic properties and long intravascular persistence. These conjugates h w ) 43 kg/mol), which was first carboxywere prepared from dextran 40 (M h n ) 38 kg/mol and M methylated. Then amines were introduced by reacting ethylenediamine with dextran carboxylic acid groups in the presence of 2-ethoxy-1-(ethoxycarbonyl)-1,2-dihydroquinoline. DTPA was then covalently linked to aminated dextran by using three different coupling procedures (DTPA bisanhydride, dicyclohexylcarbodiimide/N-hydroxysuccinimide, and isobutyl chloroformate). The different final products were compared in terms of DTPA contents, molecular masses, and sizes, and it was proved that the last synthesis step led to a small fraction of cross-linked chains as M h n was between 128 and 166 kg/mol and M h w between 332 and 371 kg/mol. In spite of this partial cross-linking which theoretically decreases the complexation capacity of the dextran-linked DTPA molecules, the Gd3+complexed conjugates exhibited relaxivities at 20 MHz/mol of gadoliniums2.5 times as great as that of free GdDTPA2-.

INTRODUCTION

Today, the complexes most commonly used in humans for nuclear magnetic resonance imaging are small molecules such as gadolinium-diethylenetriaminepentaacetic acid (GdDTPA2-) and gadolinium-tetraazacyclododecanetetraacetic acid (GdDOTA-) (1). These complexes are very strong, and their thermodynamic stability constants are, respectively, 1022.5 (2) and 1028 L mol-1 (3). This high stability hampers the release in body, of gadolinium , Gd3+, which is a very toxic ion. However, because of their molecular size, these complexes are totally and rapidly excreted via urine. In man, GdDTPA2has a 90 min biological half-life, and 5 min after its injection, its blood concentration is only about 30% of the initial dose (4). Therefore, for prolonged clinical tests, injection of several doses of GdDTPA2- is necessary. Recently, a number of investigations have aimed to bind GdDTPA2- and GdDOTA- to polymers or macromolecules such as proteins (albumin, immunoglobulins) (5-8), polylysine (8), or polysaccharides such as dextran (9-13). The use of polymers as carriers of gadolinium complexes has two advantages. The first is that, by influencing the complex rotation speed, the r1 longitudinal and r2 transversal relaxivities are increased. For example, GdDTPA2- r1 relaxivity is 3.7 mM-1 s-1 at 20 MHz, while for polymers, r1 relaxivity can be >10 mM-1 s-1 (14). The second advantage is the increase in the plasma persistence. A polylysine-GdDTPA-, studied by * Author to whom correspondence should be addressed [telephone (33) 383 17 52 21; fax (33) 383 37 99 77; e-mail [email protected]]. † ENSIC-INPL. ‡ Guerbet-GCA. X Abstract published in Advance ACS Abstracts, June 1, 1997.

S1043-1802(97)00062-1 CCC: $14.00

Schuhmann-Giampieri et al. (15), exhibited a rabbit plasma half-life of 1.9 h against 0.6 h for GdDTPA2-. These new macromolecular contrast agents can thus be used as blood pool tracers. This paper describes the synthesis and properties of dextran-linked DTPA. Dextran 40 (Figure 1, M h w about 40 kg/mol) was chosen because of its biocompatibility and biodegradability. It is commonly used as a plasma expander to improve blood circulation and to prevent blood platelet aggregation and as the basis of oxygencarrier blood substitutes (16). Different syntheses were investigated to bind DTPA to previously aminated dextran 40 (Figure 1), and the resulting DTPA-containing polymers were compared in terms of molecular mass and DTPA content. The aim was to obtain highly substituted polymers with molecular mass as low as possible, because high molecular mass polymers remain a long time in the intravascular space before being excreted, which can produce allergies. To limit the molecular mass, it was important to hamper the cross-linking reactions between amines of different polysaccharide chains and several carboxylic groups of one DTPA molecule. Furthermore, this reaction is also undesirable because the stability constant of the gadolinium-DTPA complex decreases when the number of DTPA carboxylic groups substituted by amine functions increases (17-19). EXPERIMENTAL PROCEDURES

Syntheses. Synthesis of Carboxymethyldextran (CMD). Ten grams of dextran 40 (Pharmacia LKB, Uppsala, Sweden) was dissolved in 82.5 mL of 6 N NaOH previously cooled in an ice bath. Then, 20.4 g (0.216 mol) of chloroacetic acid (Aldrich-France, St Quentin-Fallavier, France) was introduced. The mixture was maintained at 60 °C for 50 min and then cooled at room temperature. © 1997 American Chemical Society

606 Bioconjugate Chem., Vol. 8, No. 4, 1997

Then the polymer was precipitated with methanol. CMD was recovered by filtration and dried under vacuum at 50 °C. Synthesis of Aminated Carboxymethyldextran (CMDA) by Means of 2-Ethoxy-1-(ethoxycarbonyl)-1,2-dihydroquinoline (EEDQ). Ten grams of CMD (31 mmol of COONa) was dissolved in 100 mL of water and the pH was adjusted to 3 with 1 N HCl. Then an EEDQ solution prepared with 15.4 g (62 mmol) of EEDQ (Fluka, Buchs, Switzerland) and 233 mL of methanol was added dropwise. The resulting solution was added to 20.7 mL (310 mmol) of ethylenediamine (Aldrich) under stirring. After 4 h, CMDA was precipitated with methanol, then recovered by filtration, and finally dried under vacuum at 50 °C. Coupling of DTPA to Aminated Carboxymethyldextran (CMDA-DTPA). (a) Method Using DTPA Bisanhydride (DTPAba). DTPAba was previously prepared according to a method deriving from that of Eckelman et al. (20). One gram of CMDA (1.7 mmol of NH2) was dissolved in 10 mL of water. Then, 1/20 of this CMDA solution and 1 /20 of the total DTPAba amount to be used (11 g, 30.8 mmol) were successively added to 20 mL of water. This step was carried out by keeping the pH constant at 10 with 6 N NaOH. This operation was repeated until consumption of all the reagents. The final mixture was stirred for 4 h and then dialyzed against water for 3 days. Finally, the polymer was recovered by freeze-drying. The cellulosic dialysis tubes employed (Polylabo, Paris, France) had a molecular mass cutoff of 6000-8000 g/mol. (b) Method Using DTPA Succinimidic Ester (DTPAONSu). One gram (2.54 mmol) of DTPA (Aldrich) was dissolved in 10 mL of acetonitrile in the presence of triethylamine (Aldrich) at 55 °C. After DTPA was completely dissolved, the solution was cooled at room temperature and then 0.371 g (1.80 mmol) of dicyclohexylcarbodiimide (DCC; Aldrich) and 0.207 g (1.80 mmol) of N-hydroxysuccinimide (HONSu; Aldrich) were simultaneously added. The DCC/HONSu and DTPA/ DCC molar ratios were, respectively, 1 and 1.4. The solution was stirred for 90 min. The precipitated urea was removed by filtration, and the resulting solution containing DTPA-ONSu was dropped into a solution prepared with 1 g of CMDA (1.7 mmol of NH2/g) and 10 mL of water. This operation was carried out by maintaining a constant pH of 10. The solution was stirred for 24 h, and then acetonitrile was removed by rotative evaporation. Finally, the solution was dialyzed against water for 3 days, and the final polymer was recovered by freeze-drying. (c) Method Using DTPA Mixed Anhydride. First of all, 6.7 g (17 mmol) of DTPA was dissolved in 10 mL of acetonitrile at 55 °C in the presence of triethylamine. The solution was cooled at -10 °C and then dropped into 1.12 mL (8.6 mmol) of isobutyl chloroformate (IBCF; Aldrich). This mixture was stirred for only 1 min and then was added to a CMDA solution composed of 1 g of CMDA (1.7 mmol of NH2/g) and 10 mL of carbonate buffer (0.1 M, pH 8.4) at room temperature. After 4 h at room temperature, acetonitrile was removed by rotative evaporation. The resulting solution was dialyzed against water for 3 days, and the final polymer was recovered by freezedrying. Preparation of CMDA-GdDTPA- from CMDA-DTPA. CMDA-GdDTPA- was prepared as follows: CMDADTPA was mixed at pH 6.5 with GdCl3‚6H2O (Aldrich) using a molar ratio Gd3+/DTPA ) 1. Residual uncomplexed CMDA-DTPA (or Gd3+) concentration was determined, and when it was >1% of the CMDA-GdDTPAconcentration, further GdCl3‚6H2O (or CMDA-DTPA)

Rebizak et al.

was added. Gd3+ concentration was determined following a colorimetric titration method with arsenazo, a compound that leads to a complex with Gd3+ absorbing at 654 nm (21). The concentration of uncomplexed CMDADTPA was determined using the potentiometric titration method described in the next part for the measurement of the DTPA content of CMDA-DTPA samples. Physicochemical Methods. The carboxylate content of CMD and the carboxylate and amine contents of CMDA were determined according to a pH-metric method. The equipment was composed of a titroprocessor (Metrohm E636, Herisau, Switzerland) coupled to a silver-glass combined electrode. DTPA contents were obtained by a complexometric reverse dosage. CMDA-DTPA was dissolved in distilled water with an excess of Gd3+ with respect to the expected DTPA concentration. Then, the amount of uncomplexed Gd3+ was dosed with EDTA. During this operation, the potentiel fluctuations between a copper electrode and a glass one were registered. The substitution of CMDA amines by DTPA carboxylic groups was checked by 13C NMR chemical shifts of amino R-carbon upon being converted to an amido R-carbon. The 13C NMR spectra were recorded in a water-D O mixture 2 (90/10 mL/mL) at room temperature on a Bruker AM 200 MHz spectrometer (Karlsruhe, Germany) with sodium 3-(trimethylsilyl)-1-propanesulfonate (DSS) as internal reference. h w were determined by size exclusion chroM h n and M matography coupled to low-angle laser light scattering (SEC-LALLS). The LALLS detector was a Chromatix KMX6 (Milton Roy, Riviera Beach, FL). The refractometric detector was a Waters R 410 (Milford, MA). The SEC columns were TSK PW G 4000 and G 6000 (Touzart et Matignon, Courtaboeuf, France; 7.8 mm × 30 cm), connected in series, and the eluent was a 0.15 M NaNO3 water solution. The polymer hydrodynamic size distribution profiles were determined by high-performance size exclusion chromatography (HP-SEC). The experiments were carried out on a system equipped with a refractometric detector (Waters R 410) and with a 500 Ultrahydrogel column (Waters; 7.8 mm × 30 cm). Eluent was a 0.05 M phosphate buffer, pH 7.2. Longitudinal (r1) relaxivity values were determined in water at 10 and 20 MHz, at 37 °C with an IBM Research field cycling relaxometer (Mons University, Belgium) using an inversion recovery pulse sequence. RESULTS AND DISCUSSION

Synthesis. The CMDA-DTPA synthesis was carried out in three steps (Figure 1). The first step consisted of introducing carboxylic functions on dextran 40 to obtain carboxymethyldextran (CMD) without degrading the polysaccharidic backbone. For this purpose, chloroacetic acid, which can react with one (or more) of the three hydroxylic groups of the glucopyranose unit, was used. In the second step, the CMD carboxylic groups were allowed to react with ethylenediamine by means of a coupling agent, 2-ethoxy-1-(ethoxycarbonyl)-1,2-dihydroquinoline (EEDQ), to give aminated carboxymethyldextran (CMDA). Finally, three different experimental procedures were investigated to activate one of the DTPA carboxylic functions so that it could react with CMDA amines. The most commonly used method consists of using DTPA bisanhydride (DTPAba). However, in most cases, the reaction of DTPAba with aminated polymers leads to polymers with a high molecular mass (22, 23). DTPA can also be activated in the form of succinimidic ester (DTPA-ONSu) by employing both dicyclohexylcar-

Bioconjugate Chem., Vol. 8, No. 4, 1997 607

Technical Notes

Figure 1. Synthesis of CMDA-DTPA. EEDQ, 2-ethoxy-1-(ethoxycarbonyl)-1,2-dihydroquinoline. Table 1. Influence of the EEDQ/COONa Molar Ratio on the Amine Content of CMDA

Table 2. DTPA Contents of CMDA-DTPA Samples Synthesized According to Three Different Methodsa

CMDA composition NH2 content

COONa content

molar ratio EEDQ/COONa

mmol/g

mol/mol of glucose

mmol/g

mol/mol of glucose

0.5 1 2

1.0 1.3 1.7

0.15 0.27 0.37

2.0 1.5 1.0

0.44 0.31 0.21

a Initial CMD contained 0.59 mol of COONa/mol of glucose unit. The COONa and NH2 contents were determined by a pH-metric measurement.

bodiimide (DCC) and N-hydroxysuccinimide (HONSu) (8, 24) or in the form of a mixed anhydride by means of IBCF (25, 26). CMD was prepared according to the literature (27) from dextran 40. At 60 °C, in one step and for a 50 mn reaction time, a carboxymethylated polymer was obtained that contained 2.8 mmol of COONa/g, i.e. 0.59 mol of COONa/mol of glucose unit. In the synthesis of CMDA, an increase in the amine content was observed when the EEDQ/COONa molar ratio increased (Table 1). However, the anhydride resulting from the reaction between EEDQ and CMD carboxylic groups is sensitive to hydrolysis. This side reaction consumes reagents and hampers the synthesis of highly aminated polymers. Accordingly, it can be seen in Table 1 that an excess of EEDQ with regard to COONa did not lead to the amidification of all the COONa functions. Indeed, the CMDA sample prepared with EEDQ/COONa ) 2 mol/mol still contained residual unsubstituted COONa groups (0.21 mol of COONa/mol of glucose unit). The 13C NMR spectrum of CMDA (not shown) exhibited two peaks at 39 and 41.7 ppm, corresponding to, respectively, the amino R-carbon and the amido R-carbon, while,

no.

nature of the activated DTPA

1 2 3

DTPAba DTPA succinimidic ester DTPA mixed anhydride

activated DTPA/NH2 DTPA content molar ratio used in of polymer the reaction (mmol/g) 17 8 5

0.74 0.76 0.74

a The DTPA contents were determined by complexometric titration. The initial CMDA contained 0.35 mol of NH2/mol of glucose unit.

in the case of CMDA-DTPA and whatever the synthesis procedure (DTPA bisanhydride, activated ester, or mixed anhydride), the peak corresponding to the amino R-carbon disappeared. It can thus be concluded that all of the CMDA amino groups were substituted by DTPA. To obtain a high substitution rate of CMDA amines by DTPA in spite of the reagent hydrolysis (coupling agent and activated DTPA), an excess of activated DTPA with regard to the amine concentration was used. This excess was different depending on the synthesis procedure (Table 2). Under these conditions, the three CMDADTPA samples, synthesized from the same CMDA (0.35 mol of NH2/mol of glucose unit), exhibited about the same DTPA content, i.e. 0.75 mmol of DTPA/g (Table 2). This value corresponds only to the dextran-monoamidified DTPA since it was determined by a back-titration procedure using EDTA. In fact, the constant of complexation of EDTA with Gd3+ is about 1017 L mol-1 (28), i.e. 100 times smaller than that of monoamidified DTPA [about 1019 L mol-1 (5, 17)], and the complex of Gd3+ with dextran-monoamidified DTPA cannot be dissociated by a low concentration of EDTA. On the other hand, it is not the same situation with Gd3+ complexed to dextrandiamidified DTPA, since the complex stability constant is only about 1016 L mol-1 (17), i.e. smaller than that of EDTA-complexed Gd3+, which leads to the removal of

608 Bioconjugate Chem., Vol. 8, No. 4, 1997

Rebizak et al.

Table 3. Average Molecular Masses of Successive Polymeric Derivativesa M h n (kg/mol) polymer dextran 40 CMD CMDA CMDA-DTPA 1b CMDA-DTPA 2b CMDA-DTPA 3b

M h w (kg/mol)

measd calcd measd calcd 38 45 45 128 166 137

49 50 95

43 56 67 345 332 371

56 57 142

polydispersity index, I measd 1.1 1.2 1.5 2.7 2.0 2.7

calcd 1.1 1.1 1.5

a Molecular masses were determined by SEC-LALLS as described under Experimental Procedures. The CMDA-DTPA numbers refer to those of Table 2. The theoretical CMD and CMDA molecular masses (M h n and M h w) were calculated using those of initial dextran and the experimental chemical compositions of CMD (0.59 mol of COONa/mol of glucose unit) and CMDA (0.24 mol of COONa and 0.35 mol of NH2/mol of glucose unit). b The theoretical molecular masses of CMDA-DTPA samples were calculated using those of parent CMDA and the COONa, amine, and DTPA contents as determined by the CMDA-DTPA analysis.

Figure 2. HP-SEC elution profiles of dextran (- ‚ -), CMD (- -), CMDA (s), and CMDA-DTPA 1 (- - -) prepared by means of DTPAba (see Table 2), on a 500 Ultrahydrogel column: eluent, 0.05 M phosphate buffer, pH 7.2; refractometric detection; flow rate, 0.7 mL/mn; void volume, V0 ) 6.4 mL; total permeation volume, Vt ) 12 mL.

Gd3+ from its complex with dextran-diamidified DTPA upon addition of EDTA. Hydrodynamic Volumes and Average Molecular Masses (M h n and M h w). The average molecular masses of different polymeric derivatives of the same family are shown in Table 3, and their hydrodynamic volume distributions are presented in Figure 2. As expected, CMD has average molecular masses higher than those of initial dextran (Table 3) as a result of the coupling of carboxymethyl groups, and the experimental values are close to those that can be theoretically calculated. Moreover, the experimental polydispersity index is similar to that of dextran 40. These results prove that the CMD chains are statistically substituted. The CMD hydrodynamic volume estimated by HP-SEC (Figure 2) is higher than that of dextran 40. By using the calibration plot of a 500 Ultrahydrogel column, obtained with dextrans of known M h w, the CMD hydrodynamic volume was found to be close to that of a dextran with a M h w of 100 kg/mol. The great difference between the molecular size of dextran 40 and that of CMD can be explained by the fact that, under the applied ionic strength conditions (0.05 M phosphate buffer, pH 7.2),

many repulsive electrostatic interactions exist inside the polysaccharidic chains between the anionic groups. During the CMDA synthesis, some cross-linking mechanisms can take place. Indeed, at the pH of CMDA synthesis (pH 11), a part of the ethylenediamine molecules is totally nonprotonated (pKa1 ) 7.5 and pKa2 ) 10.7) and can react with two activated carboxylic groups of two different CMD chains. Therefore, to limit these undesirable side reactions, an excess of ethylenediamine with respect to COONa concentration (diamine/COONa ) 10 mol/mol) was used. Table 3 shows that the CMDA molecular masses as measured by SEC-LALLS are a little higher than those which can be theoretically calculated from the CMDA composition. Moreover, the experimental polydispersity index of CMDA (I ) 1.5) is slightly greater than that of the parent CMD (I ) 1.2) and its molecular mass distribution profile is a little less symmetrical than that of CMD (not shown). All of these results prove that the CMDA sample is cross-linked. The CMDA average hydrodynamic molecular size as illustrated by the HP-SEC profile (Figure 2) is much smaller than that of CMD and close to that of initial dextran, although the CMDA M h w value is higher. In fact, the replacement of some CMD carboxylate groups by amines leads, under the experimental conditions at which the polymers are studied (0.05 M phosphate buffer, pH 7.2), to attractive electrostatic interactions between NH3+ and COO- inside the CMDA chains, which results in a shrinkage of the polymer chains. The last step of this synthesis concerned the coupling of DTPA with CMDA. As DTPA possesses five carboxylic acid functions, there was a risk that more than one of them would have been activated in the presence of a coupling agent. In this case one DTPA molecule could react with two (or more) amine functions carried by two (or more) different CMDA chains. Therefore, the coupling procedures (DTPAba, DCC/HONSu, and IBCF) were optimized to limit these cross-linking reactions. In the DTPAba procedure, to limit the probability that one DTPAba molecule would react with two CMDA amines, a very low molar amine concentration of about 3.4 × 10-3 mol/L was used. Concerning the procedures using both HONSu and DCC, Spanoghe et al. (8) and Paxton et al. (24) showed that it was possible to prepare a monoactivated DTPA by using DCC/DTPA and HONSu/DTPA molar ratios, respectively, of 1 and 0.7. These conditions were applied for the synthesis of CMDA-DTPA by means of DCC/HONSu. In the IBCF procedure, a IBCF/DTPA molar ratio of 0.5 mol/mol was used. The other reaction parameters (molar ratios of activated DTPA/NH2) are those given in Table 2. h w, and polydispersity index values Theoretical M h n, M of, respectively, 95 and 142 kg/mol and 1.5 were calculated for a CMDA-DTPA containing 0.35 mol of DTPA and 0.24 mol of COONa/mol of glucose unit. The theoretical DTPA content was assumed to be 0.35 mol/ mol of glucose unit since the parent CMDA contained 0.35 mol of NH2/mol of glucose unit and since according to the 13C NMR spectra all of the NH2 functions were substituted. The content in COONa was that of the hw parent CMDA. As shown in Table 3, the M h n and M experimental values of the three CMDA-DTPA samples are higher than those theoretically calculated. The same remark can be made for the polydispersity indices, which are >2 while that of the parent CMDA is 1.5. These results lead to the conclusion that, whatever the experimental procedure carried out to prepare the CMDADTPA samples, a part of the DTPA molecules is polyactivated, which gives rise to a few cross-linking reactions and therefore to a fraction of high molecular mass

Bioconjugate Chem., Vol. 8, No. 4, 1997 609

Technical Notes

Figure 3. Molecular mass distribution profile (logarithmic normal distribution) of CMDA-DTPA 1 (cf. Figure 2) obtained by SEC-LALLS: columns used, TSK PW G 6000 and 4000; eluent, 0.1 M NaNO3 aqueous solution.

the molecular size of CMDA-DTPA can be compared to that of a 140 kg/mol dextran, CMDA-GdDTPA- is now comparable to a 77 kg/mol dextran. This result confirms that uncomplexed CMDA-DTPA is really subjected to repulsive electrostatic interactions which can no longer exist in the final complex with Gd3+. Paramagnetic Properties. As CMDA-GdDTPAhas to be used as a contrast agent in NMR imaging, its capacity in enhancing contrast was estimated by determining its r1 relaxivity at 10 and 20 MHz, at 37 °C. For CMDA-GdDTPA- 1 (obtained from CMDA-DTPA 1, Table 2), the r1 values at 10 and 20 MHz were, respectively, 8.9 and 9.8 mM-1 s-1. These values are thus, respectively, 1.9 and 2.5 times as great as those of molecular complexes GdDTPA2- that are commonly used in NMR imaging. This enhancement of the r1 value of a paramagnetic element when bound to a macromolecule has been already described and explained (14, 29). This increase is caused by a “changeover” in the values of correlation times, τR (rotation correlation time) and τS (electron spin relaxation time), when the paramagnetic species becomes bound to a macromolecule, which brings about an increase in τC (effective correlation time) (14). CONCLUSION

Figure 4. HP-SEC elution profiles of CMDA-DTPA 1 (s) (cf. Figure 2) and of corresponding CMDA-GdDTPA- (- ‚ -), on a 500 Ultrahydrogel column. Conditions were as in Figure 2.

molecules. This cross-linking phenomenon is clearly proved in the CMDA-DTPA molecular mass distribution profiles. One example of these CMDA-DTPA profiles is given in Figure 3 (sample 1 synthesized by means of DTPAba), and it can be seen that the profile is widened toward the high molecular masses. The fact that all of the CMDA-DTPA samples present some cross-linked chains proves that the monoamidification is not the only reaction whatever the coupling procedure and conditions employed. Among the four types of polymers studiedsdextran, CMD, CMDA, and CMDA-DTPAsthe last one has, by far, the largest average hydrodynamic volume (Figure 2). For example, the CMDA-DTPA sample 1 (prepared by means of DTPAba) has a molecular size close to that of a dextran with a M h w of 140 kg/mol. This large increase in the molecular size results of course from the cross-linking reactions but above all from the presence on the polysaccharidic chains of a great amount of COO- groups (residual carboxymethyl and DTPA) which, in the used HP-SEC eluent, strongly extends the polymer structure. Figure 4 presents the HP-SEC profile of CMDA-DTPA 1 compared to that of CMDA-GdDTPA- obtained after complexation of the DTPA carboxylate functions with Gd3+. It shows that the Gd3+ complexation produces a decrease in the polymer molecular size. Indeed, while

Polymeric conjugates of Gd3+-DTPA and dextran described in this paper present some differences with those already reported elsewhere: first, they are more substituted than almost all of the dextran conjugates described in the literature (10-13). The possibility of using highly substituted polymers is important as it allows one to inject small amounts of polymer while achieving a good contrast and thus to limit the risks of allergies. Only the conjugates presented by Armitage et al. (9) possessed DTPA contents similar to those of this paper, but in Armitage’s work, DTPA was linked directly to the polysaccharide via an ester function, while in the conjugates here described, a short spacer arm was used and DTPA was linked to aminated dextran by an amide function. According to the analysis of the synthesized conjugates, a small fraction of polymeric chains is cross-linked, which means that a part of the dextran-linked DTPA molecules is diamidified. This could lead to in vivo toxicity because of the decrease of the complexation constant between Gd3+ and DTPA due to the decrease of the number of free carboxylate groups. In fact, while the thermodynamic stability constant of GdDTPA2- is 1022.5 L mol-1 (2), that of Gd3+-complexed monoamidified DTPA as determined by Lauffer et al. (5) and Sherry et al. (17) is, respectively, 1019.1 and 1019.7 L mol-1 and that of Gd3+-complexed diamidified DTPA is 1016.2 L mol-1 (17). To overcome the difficulty caused by the potential in vivo Gd3+ release from diamidified DTPA, the high molecular mass fraction of conjugates should be removed, for example, by ultrafiltration. Such experiments are now under investigation. Finally, the CMDA-GdDTPA2- r1 relaxivity values at 10 and 20 MHz are higher, per mole of Gd, than those of free ligand. Thus, during a clinical examination, the gadolinium concentration needed to obtain contrast will be lower with CMDA-GdDTPA- than with GdDTPA2-. Moreover, as CMDA-GdDTPA- has an average molecular size bigger than that of dextran 40, it is likely that its plasma half-life will be greater than that of dextran, which will be another advantage over GdDTPA2-. ACKNOWLEDGMENT

This work was supported by a CIFRE grant (Socie´te´ Guerbet-GCA, Aulnay-sous-Bois, France, and ANRT,

610 Bioconjugate Chem., Vol. 8, No. 4, 1997

France). We thank Mrs. T. Geoffroy and Mr. A. Vicherat for helpful technical assistance, respectively, in LALLS and RMN analyses. We also thank Prof. R. N. Muller (Mons University, Belgium) for helpful discussions on the interpretation of relaxivities. LITERATURE CITED (1) Niendorf, H. P., Alhassan, A., Haustein, J., Clauss, W., and Cornelius, I. (1993) Safety and risk of gadolinium-DTPA: extended clinical experience after more than 5,000,000 applications. Adv. MRI Contrast 2, 12. (2) Martell, A. E., and Smith, R. M. (1986) Critical Stability Constants, Vol. 6, Plenum Press, New York. (3) Desreux, J. F. (1980) Nuclear magnetic resonance spectroscopy of lanthanide complexes with a tetraacetic tetraaza macrocycle. Unusual conformation properties. Inorg. Chem. 19, 1319. (4) Schmiedl, U., Moseley, M. E., Ogan, M. D., Chew, W. D., and Brasch, R. C. (1987) Comparison of initial biodistribution patterns of Gd-DTPA and albumin-(Gd-DTPA) using rapid spin echo MR imaging. J. Comput. Assist. Tomogr. 11, 306. (5) Lauffer, R. B., Brady, T. J., Brown III, R. D., Baglin, C., and Koenig, S. H. (1986) 1/T1 NMRD profiles of solutions of Mn2+ and Gd3+ protein-chelate conjugates. Magn. Reson. Med. 3, 541. (6) Paajanen, H., Reisto, T., Hemmila, I., Komu, M., Niemi, P., and Kormano, M. (1990) Proton relaxation enhancement of albumin, immunoglobulin G and fibrinogen labeled with GdDTPA. Magn. Reson. Med. 13, 38. (7) Niemi, P., Reisto, T., Hemmila, I., and Kormano, M. (1991) Magnetic field dependence of longitudinal relaxation rates of solutions of various protein-gadolinium3+ chelate conjugates. Invest. Radiol. 26, 820. (8) Spanoghe, M., Lanens, D., Dommisse, R., Van Der Linden, A., and Alderweireldt, F. (1992) Proton relaxation enhancement by means of serum albumin and poly-L-lysine labeled with DTPA-Gd3+: relaxivities as a function of molecular weight and conjugation efficiency. Magn. Reson. Imaging 10, 913. (9) Armitage, F. E., Richardson, D. E., and Li, K. C. P. (1990) Polymeric contrast agents for magnetic resonance imaging: synthesis and characterization of gadolinium diethylenetriaminepentaacetic acid conjugated to polysaccharides. Bioconjugate Chem. 1, 365. (10) Wang, S. C., Wikstroem, M. G., White, D. L, Klaveness, J., Holtz, E., Rongved, P., Moseley, M. E., Michael, E., and Brash, R. C. (1990) Evalution of the gadolinium-DTPA-labeled dextran as an intravascular MR contrast agent: imaging characteristics in normal rat tissues. Radiology 175, 483. (11) Bligh, S. W. A., Harding, C. T., Sadler, P. J., Bulman, R. A., Bydder, G. M., Pennoock, J. M., Kelly, J. D., Latham, I. A., and Marriott, J. A. (1991) Use of paramagnetic chelated derivatives of polysaccharides and spin-labeled polysaccharides as contrast agents in magnetic resonance imaging. Magn. Reson. Med. 17, 516. (12) Rongved, P., and Claveness, J. (1991) Water soluble polysaccharides as carriers of paramagnetic contrast agents for magnetic resonance imaging : Synthesis and relaxation properties. Carbohydr. Res. 214, 315. (13) Brash, R. C. (1991) Rationale and applications for macromolecular Gd-based contrast agent. Magn. Reson. Med. 22, 282. (14) Lauffer, R. B., and Brady, T. J. (1985) Preparation and water relaxation properties of proteins labeled with paramagnetic metal chelates. Magn. Reson. Imaging 3, 11. (15) Schuhmann-Giampieri, G., Schmitt-Willich, H., Frenzel, T., Press, W-R., and Weinmann, H-J. (1991) In vivo and in

Rebizak et al. vitro evaluation of Gd-DTPA-polylysine as a macromolecular contrast agent for resonance magnetic imaging. Invest. Radiol. 26, 969. (16) Dellacherie, EÄ . (1996) Polysaccharides in oxygen-carrier blood substitutes. In Polysaccharides in Medicinal Applications (S. Dumitriu, Ed.) p 525, Marcel Dekker, New York. (17) Sherry, A. D., Cacheris, W. P., and Kuan, K-T. (1988) Stability constants for Gd3+ binding to model DTPA-conjugates and DTPA-proteins: implications for their use as magnetic resonance contrast agents. Magn. Reson. Med. 8, 180. (18) Aime, S., Anelli, P. L., Botta, M., Fedeli, F., Grandi, M., Paoli, P., and Ugeri, F. (1992) Synthesis, characterization, and 1/T1 NMRD profiles of gadolinium(III) complexes of monoamide derivatives of DOTA-like ligands. X-ray structure of the 10-[2-[[2-hydroxy-1-(hydroxymethyl)ethyl]amino]-1[(phenylmethoxy)methyl]-2-oxoethyl]-1,4,7,10-tetra-azacyclododecane-1,4,7-triacetic acid-gadolinium(III) complex. Inorg. Chem. 31, 2422. (19) Konings, M. S., Dow, W. C., Love, D. B., Raymond, K. N., Quay, S. C., and Rocklage, S. M. (1990) Gadolinium complexation by a new DTPA-amide ligand. Amide oxygen coordination. Inorg. Chem. 29, 1488. (20) Eckelman, W. C., Karesh, S. M., and Reba, R. C. (1975) New compounds: fatty acid and long chain hydrocarbon derivatives containing a strong chelating agent. J. Pharm. Sci. 64, 704. (21) Onishi, H., and Sekine, K. (1972) Reagent for the determination of Th, U, Zr and rare earth metals. Talanta 19, 473. (22) Paik, C. H., Ebbert, M. A., Murphy, P. R., Lassman, C. R., Reba, R. C., Eckelman, W. C., Pak, K. Y., Powe, J., Steplewski, Z., and Koprowski, H. (1983) Factors influencing DTPA conjugation with antibodies by cyclic DTPA anhydride. J. Nucl. Med. 24, 1158. (23) Maisano, F., Gozzini, L., and de Haen, C. (1992) Coupling of DTPA to proteins : a critical analysis of the cyclic dianhydride method in the case of insulin modification. Bioconjugate Chem. 3, 212. (24) Paxton, R. J., Jakowatz, J. G., Beatty, J. D., Beatty, B. G., Vlahos, W. G., Williams, L. E., Clark, B. R., and Shively, J. E. (1985) High-specific-activity 111In-labeled anticarcinoembryonic antigen monoclonal antibody : improved method for the synthesis of diethylenetriaminepentaacetic acid conjugates. Cancer Res. 45, 5694. (25) Li, M., and Meares, C. F. (1993) Synthesis, metal chelate stability studies and enzyme digestion of a peptide-linked DOTA derivative and its corresponding radiolabeled immunoconjugates. Bioconjugate Chem. 4, 275. (26) Sieving, P. F., Watson, A. D., and Rocklage, S. M. (1990) Preparation and characterization of paramagnetic polychelates and their protein conjugates. Bioconjugate Chem. 1, 65. (27) Bouttemy, M. (1960) Contribution a` l'e´tude des carboxyme´thylcelluloses. Pre´paration des carboxyme´thylcelluloses de hauts degre´s de substitution (Contribution to the study of carboxymethylcelluloses. Preparation of carboxymethylcelluloses with a high degree of substitution). Bull. Soc. Chim. Fr. 1750. (28) Cacheris, W. P., Nickle, S. K., and Sherry, A. D. (1987) Thermodynamic study of lanthanide complexes of 1,4,7triazacyclononane-N,N′,N′′-triacetic acid and 1,4,7,10-tetraazacyclododecane-N,N′,N′′,N′′′-tetraacetic acid. Inorg. Chem. 26, 958. (29) Lauffer, R. B. (1987) Paramagnetic metal complexes as water proton relaxation agents for NMR imaging: theory and design. Chem. Rev. 87, 901.

BC970062N