Bioconjugate Chem. 2000, 11, 397−400
397
Enhancing the “Stickiness” of Bile Acids to Cross-Linked Polymers: A Bioconjugate Approach to the Design of Bile Acid Sequestrants Lan-hui Zhang, Vaclav Janout, Jodi L. Renner, Maki Uragami, and Steven L. Regen* Department of Chemistry and Zettlemoyer Center for Surface Studies, Lehigh University, Bethlehem, Pennsylvania 18015. Received November 2, 1999; Revised Manuscript Received January 6, 2000
A series of cross-linked polymers have been synthesized by reaction of 2% cross-linked, chloromethylated polystyrene with a tertiary amine derivative (7) of cholic acid (prepared by reaction of cholic acid methyl ester with N,N-dimethyl-ethylenediamine), followed by exhaustive quaternization with trimethylamine. Increased loadings of 7 result in enhanced binding of taurocholate ion and a decrease in its rate of release upon exposure to 50 mM aqueous NaCl. Examination of particle-size effects indicates that film diffusion and particle diffusion both contribute to the observed kinetics. Interruption tests that were carried out confirmed that particle diffusion is kinetically important. The relevance of this bioconjugate approach to the design of bile acid sequestrants for the treatment of hypercholesterolemia is briefly discussed.
INTRODUCTION
Scheme 1
Conventional anion-exchange resins such as Dowex 1-X2 (Cholestyramine) serve as important therapeutic agents for the treatment of hypercholesterolemia (1). Although their mechanism of action is reasonably wellunderstood (i.e., they bind and remove bile acids from the intestines, thereby upregulating the conversion of cholesterol to bile acids in the liver), extensive efforts that have been made to improve resin efficiency have led to only modest advances. In an in-depth review of this area, Stedronsky has raised the intriguing possibility that the rate of release of bile acids from cross-linked polymers may be as important a factor in defining resin efficiency as its strength of binding, i.e., slow off-rates should help to avoid competitive removal of bile acids from the lumen through the wall of the ileum via active transport (2). It is noteworthy, in this regard, that nearly all studies involving bile acid sequestrants have focused on the thermodynamics of binding; only one investigation to date has examined the kinetics of release of bile acids (3, 4).
In this paper, we describe a bioconjugate approach to the synthesis of novel polymers that improves, for the first time, the kinetics as well as the thermodynamics of bile acid sequestration. Specifically, we show that the conjugation of a cholic acid derivative to an anionexchange resin increases the resin’s affinity toward taurocholate (T), and slows down its rate of release. This work was based on the hypothesis that “cloistering” of an ionically bound bile acid next to ones that are covalently attached should result in enhanced affinity due to van der Waals attraction and/or hydrogen bonding between nearest-neighbors (Scheme 1). * To whom correspondence should be addressed. Phone: (610) 758-4842. Fax: (610) 758-6560. E-mail:
[email protected].
EXPERIMENTAL PROCEDURES
Materials and Methods. Unless stated otherwise, all chemicals were purchased from commercial sources and used without further purification. Cross-linked polystyrene was chloromethylated using experimental procedures similar to those previously described (5). Specific protocols, which were used to load quaternized ionexchange resins with taurocholate have previously been reported (4). Typically, a taurocholate-loaded resin was prepared by suspending 1.0 g of a given resin in 400 mL of 1 mM sodium taurocholate solution. After equilibrating the resin for 72 h at room temperature, the concentration of sodium taurocholate that remained in the aqueous phase was determined by UV (205 nm), using an appropriate calibration curve. The resin was then separated from the free taurocholate in solution by filtration, and the resin rinsed with deionized water until no significant absorbance could be detected at 205 nm. The amount of resin-bound taurocholate was then calculated by subtracting the amount of residual sodium taurocholate in solution from the initial quantity that was present in the aqueous phase. Combining this data, with the total exchange capacity of the resin (see below), afforded the initial fraction of pendant ammonium groups (θo) having taurocholate as a counterion. The water used in these experiments was purified using a Millipore Milli-Qfiltering system containing one carbon and two ionexchange stages. All kinetics experiments were carried
10.1021/bc990151v CCC: $19.00 © 2000 American Chemical Society Published on Web 04/18/2000
398 Bioconjugate Chem., Vol. 11, No. 3, 2000
Zhang et al.
Scheme 2
Table 1. Influence of Covalently-Attached 7 on Equilibrium Binding Constants and Initial Release Rates resin 1 2 3 4 5 6 Aa Ba a
percent ring substitution
diameter (mm)
0 8 15 0 9 18 8 15
0.13 ( 0.03 0.87 ( 0.13 0.87 ( 0.09 0.84 ( 0.11 0.87 0.87
θo
K
[-d(θ/θo)/dt]o
0.275 ( 0.005 0.314 ( 0.005 0.300 ( 0.005 0.265 ( 0.005 0.270 ( 0.005 0.215 ( 0.005
18 ( 1 19 ( 2 28 ( 2
0.145 ( 0.010 0.081 ( 0.005 0.066 ( 0.005 0.0109 ( 0.0005 0.0064 ( 0.0008 0.0013 ( 0.0002 0.0069 ( 0.0005 0.0031 ( 0.0003
Initial rates estimated by interpolation of the data from 4, 5, and 6.
out at 22 ( 1 °C, and UV measurements were made using a Milton Roy Spectronic 1201 spectrometer. N2,N2-Dimethylamino-N1-2-ethyl-choleamide. A mixture of methylcholeate (2.70 g, 6.3 mmol) and N,Ndimethylamino ethylamine (2.4 g, 27 mmol) was stirred at 110 °C under a nitrogen atmosphere for 16 h, and the excess of amine then removed under reduced pressure to give crude product. The desired N2,N2-(dimethylamino)-N1-2-ethyl-choleamide (7) was recrystallized (two times) from H2O/CH3OH (10:1, v/v) to give 2.91 g (90.6%) of product having a mp 192-3 °C. 1H NMR (CD3OD, 360 MHz): δ 3.93 (s, 1H), 3.79 (s, 1 H), 3.36 (m, 1H), 3.20 (m, 2 H), 2.45 (t, 2 H), 2.27 (s, 6 H), 0.9-2.25 (m, 30 H), 0.69 (s, 3 H). Anal. calcd (C28H50N2O4): C, 70.24; H, 10.53; N, 5.85. Found: C, 70.14; H, 10.47; N, 5.79. Quaternization of Chloromethyl-poly(styren-98co-divinylbenzene-2) with N2,N2-Dimethylamino-N12-ethyl-choleamide and trimethylamine (Polymer 5). To a suspension of fully swollen chloromethylated polystyrene [3.116 g, 19.6 mmol Cl (95% ring substitution), 2% DVB] in 20 mL of anhydrous THF and 6.0 mL of anhydrous N,N-dimethylacetamide (DMAA) was added a solution of N2,N2-(dimethylamino)-N1-2-ethyl-choleamide (0.924 g, 1.96 mmol) in 8 mL of anhydrous THF plus 3 mL of DMAA. After the suspension was heated (70 °C) for 64 h in a sealed flask, the polymer was separated, washed with 5 × 20 mL of THF. A small sample (ca. 20 mg) was then dried (23 °C, 0.5 Torr, 48 h) and subjected to elemental analysis. The resulting analysis (N, 1.31; Cl, 15.75) indicated that 9% of the chloromethylene groups were quaternized by the sterol. The main portion of the polymer was then introduced into 44 mL of a 25% aqueous solution of trimethylamine
(196 mmol) containing 12 mL of THF. After the mixture was heated (50 °C) in sealed flask for 5 days, the polymer was separated, washed with 600 mL of water, and dried (23 °C, 0.5 Torr, 48 h) to constant weight. Analysis of the resulting resin by use of a modified Volhard titration indicated that the remaining 85% of the chloromethylene groups was quaternized with trimethylamine to give 5, i.e., the chloride ion content was 2.47 mmol/g of dry resin (6). Similar procedures were used to synthesize the other target polymers. RESULTS AND DISCUSSION
Polymer-Bioconjugates. Three resins that were designed to test our “cloistering” hypothesis were 1, 2, and 3 (Scheme 2). Analogous resins having a larger bead diameter (4, 5, and 6) were also chosen as synthetic targets, to examine the influence of particle diffusion on the kinetics of exchange of a resin-bound bile acid. Thus, chloromethylation of 2% cross-linked polystyrene, followed by quaternization with excess trimethylamine, afforded 1 and 4. Alternatively, quaternization with limited quantities of 7, followed by exhaustive quaternization with trimethylamine yielded 2, 3, 5, and 6; actual loadings of 7 for each resin are shown in Scheme 2. Each of these resins was then subjected to ion exchange with sodium taurocholate. The fraction of pendant ammonium groups (θo) having taurocholate (T) as a counterion for each polymer is listed in Table 1. Kinetics and Thermodynamics of Bile Acid Release. To measure the rate of release of T from each target resin, 0.1 g samples were gently agitated in 100
Enhancing the “Stickiness” of Bile Acids
Bioconjugate Chem., Vol. 11, No. 3, 2000 399
Figure 1. Fraction of pendant ammonium groups (θ) that maintain taurocholate as a counterion as a function of time of exposure to 50 mM NaCl at 23 °C. (A) (9) 1, (2) 2, and (b) 3. (B) (inset) (0) 4, (4) 5, and (O) 6; values of θ that were measured after 10 days for 4, 5, and 6 were 0.120, 0.150, and 0.168.
mL of 50 mM NaCl via “hand rocking”. Aliquots (1 mL) of the aqueous phase were then withdrawn over a period of 30 min and analyzed directly by UV (205 nm) using appropriate calibration curves. Specific experimental protocols that were followed were similar to those reported elsewhere (4). For the smaller beads, rocking was carried out in chromatography column (25 × 530 mm), equipped with a fine filter and 150 mL of solution. Mild agitation was specifically used in order to simulate physiologically relevant flow conditions. The fraction of pendant ammonium groups that maintain taurocholate as a counterion, θ, was determined as function of time by use of the following equation:
θ ) θ0 - CV/FW Here, θ0 is the fraction of pendant ammonium groups having taurocholate as a counterion at the start of the exchange, C is the concentration of taurocholate ion that has been released into the solution, V is the volume of the solution, F is the initial moles of ammonium groups per gram of resin, and W is the initial weight of the resin. The overall relative error of a single measurement of θ is estimated to be less than 2%. Specific release profiles of T that were observed using resins 1, 2, and 3 are shown in Figure 1A. As is readily apparent, increased loadings of 7 resulted in stronger binding of T, i.e., final equilibrium values of θ were higher. Equilibrium binding constants (K) for T, which were calculated by use of eq 1, were listed in Table 1; here XCl and XT are the mole fractions of the resin that are in the chloride
K)
XT[Cl-] XCl[T-]
(1)
and taurocholate forms, respectively, and [Cl-] and [T-] are the concentrations of chloride and taurochlolate anions in the bulk aqueous phase. An analysis of the initial rate of ion exchange [-d(θ/θo)/dt]o showed a steady decrease on going from 1 to 2 to 3 (Table 1). Initial rates were estimated by fitting the rate data (first 30 min) according to the following equation: θ ) ae-bt + C, where ab/θo is taken as the initial rate of ion exchange. Such a decline strongly suggests that diffusion of ions within the polymer (particle diffusion) is at least partly rate limiting,
Figure 2. Plot of θ as a function of time of exposure to 50 mM NaCl at 23 °C for (b) the first 10 min as well as a 20 min period immediately following a 49-min interruption, and for (- ‚ -) an uninterrupted 30-min period for 5. The open circles represent data obtained after the interruption, which have been “moved to the left” by 49 min.
since the rate of processes controlled by film diffusion (diffusion of ions through a thin quiet liquid layer surrounding the polymer beads) are independent of resin composition. To confirm kinetic contributions from particle diffusion, similar release profiles were obtained using larger-sized particles, i.e., 4, 5, and 6 (Figure 1B). For a particle diffusion-controlled exchange, the observed rate is known to be inversely proportional to the square of the radius of the bead. Thus, a ca. 7-fold increase in radius, on going from 1-3 to 4-6 should lead to a ca. 49-fold reduction in rate. For film diffusion, where the rate is inversely proportional to the radius, the observed rate should decrease by a factor of 7. Comparison of the initial rates of release of T from 1 and 4, which are devoid of 7, showed a 13-fold reduction with the larger beads. Similarly, a 12-fold reduction in rate can be estimated on going from 2 to a resin that is ca. 7 times larger, having the identical loading of 7 (see 2 and A in Table 1). When the loading of 7 reaches 15% ring substitution, the larger beads (see 3 and B in Table 1) now show a ca. 21-fold reduction in rate. These results imply that both film and particle diffusion are rate limiting and that particle diffusion increases in importance when the loading of 7 along the polymer backbone becomes sufficiently high. Further evidence that particle diffusion contributes to the observed kinetics was obtained by carrying out an “interruption test” with 4 and 5 (7, 8). Thus, after the ion exchange was allowed to proceed for 10 min, the resin was quickly removed from the aqueous phase. Reimmersion after ca. 50 min resulted in a small but distinct increase in rate, indicating that concentration gradients within the beads are leveling out during the interruption period (Figure 2). Such a finding indicates that particle diffusion is kinetically important in these systems. The increase in the equilibrium binding constant for T, which accompanies the introduction of 7, provides compelling evidence for the existence of cloistering. The extent to which cloistering contributes to the observed kinetics, however, is less certain. In principle, three factors could influence the contributions made from particle diffusion: (i) cloistering, (ii) deswelling of the resin, leading to a more compact state, and/or (iii) steric hindrance to diffusion, i.e., filling up the solvent-swollen channels with permanently attached bile acids. Since an examination of water-swollen 4, 5, and 6 by optical microscopy revealed bead diameters were the same, within experimental error, swelling changes cannot account for these observed differences. Although we believe
400 Bioconjugate Chem., Vol. 11, No. 3, 2000
that both cloistering and steric hindrance to diffusion are important factors, we are unable to separate their kinetic contributions at the present time. The modulation of the kinetics and the thermodynamics of bile acid sequestration, via the conjugation of 7 reported herein, represents a fundamentally new approach to a problem that has taunted a large number of polymer/medicinal chemists over the past 40 years. Whether or not such a strategy can lead to improved therapeutic agents, however, remains to be established. Efforts aimed at addressing this question are currently in progress. ACKNOWLEDGMENT
We thank the National Science Foundation (Grant CHE-9612702) for support of this research. We are also grateful to Dr. Ken Foster (Dow Chemical Co., Midland, MI) for a sample of cross-linked polystyrene, which was used as starting material. LITERATURE CITED (1) Ast, M., and Frishman, W. H. (1990) Bile Acid Sequestrants. J. Clin. Pharmacol. 30, 99-106.
Zhang et al. (2) Stedronsky, E. R. (1994) Interaction of bile acids and cholesterol with nonsystemic agents having hypocholesterolemic properties. Biochim. Biophys. Acta 210, 255-287. (3) Johns, W. H., and Bates, T. R. (1970) Quantification of the Binding Tendencies of Cholestyramine II: Mechanism of Interaction with Bile Salt and Fatty Acid Salt Anions. J. Pharm. Sci. 59, 329-333. (4) Regen, S. L., Stedronsky, E. R., Zhang, L. H., and Janout, V. (1998) Kinetics of exchange of a resin-bound bile acid by chloride ion under mild flow conditions. Macromolecules 31, 5542-5545. (5) Hodge, P., and Sherrington, D. C. (1980) Chloromethylation of cross-linked polystyrene. Polymer-Supported Reactions in Organic Synthesis, p 477, John Wiley & Sons, New York. (6) Stewart, J. M., and Young, J. D. (1969) Chloride analysis by the modified Volhard method. Solid-Phase Peptide Synthesis, pp 55-56, W. H. Freeman, San Francisco. (7) Hellferich, F. (1962) Kinetcs. Ion Exchange, pp 250-322, McGraw-Hill, New York. (8) Kressman, T. R. E., and Kitchener, J. A. (1949) Cation exchange with a synthetic phenolsulphonate resin. Discuss. Faraday Soc. 7, 90-104.
BC990151V
