- Email: [email protected]
Protein Binding of Xanthine Derivatives to Guinea Pig Serum Albumin
Protein Binding of Xanthine Derivatives to Guinea Pig Serum Albumin TAKAAKI HASEGAWA*', KENZOTAKAGI*, MASAYUKI NADAI*, AND KENIGHIMIYAMOTOs Received February 20, 1990, from the *Department of Hospital Pharmacy and the *Second Department of lnternal Medicine, Nag0 a University School of Medicine, Nagoya 466, Japan, and the 9 Research Laboratory for Development of Medicine, School of Pharmacy, Hokurilu University, Kanazawa 920- 7 7, Japan. Accepted for publication May 18, 1990. Abstract 0 Binding of the bronchodilators N '-alkylxanthine and N 'alkyl-N'-methylxanthine derivatives to guinea pig serum albumin was investigated in vitro using the ultrafiltration method. A marked difference in the binding parameters of xanthine derivatives was observed, and binding was shown to be concentration dependent. Significant relations were observed among their binding parameter, dissociation constant (&), and hydrophobicity (log PC). The extent of binding of xanthine derivatives was increased both when a N 'methyl group was replaced by a longer alkyl chain and when a N 3-alkylxanthine molecule was additionally replaced by a methyl group. Reversed-phase HPLC retention, as an index of hydrophobicity of xanthine derivatives, was also determined. Significant relationships were found between the adjusted retention time data for each xanthine derivative and their hydrophobicity or biological activities, such as their abilities to cause muscle relaxation and cyclic AMP phosphodiesterase (PDE) inhibition. These findings indicate that the difference in the extent of binding among xanthine derivatives is related to hydrophobicity,which is an important determinant of their biological activities.
In the course of studies on the structure-activity relationship of N-alkylxanthine derivatives, 1-methyl-3-propylxanthine (MPX) was selected as one of the most interesting compounds synthesized in our laboratory.l-4 In our recent studies, the relaxant effects and inhibiting activities on cyclic AMP phosphodiesterase (PDE) of various xanthine derivatives, using tracheal smooth muscle isolated from guinea pigs, were estimated. The ability of xanthine derivatives to inhibit cyclic AMP PDE activity contributes, at least in part, to the mechanism of bronchodilatory action of these derivatives. The hydrophobicity of the derivatives is an important determinant in biological activities such as relaxation induction and cyclic AMP PDE inhibition.5.6 Finally, we have proposed that hydrophobicity is useful for chemical modifications of more potent xanthine derivatives; for example, as a smooth muscle relaxant. In general, drug permeability through biological membranes is considered to depend on factors such as protein binding in serum, hydrophobicity, and molecular size of the drug. Our previous study suggested that the potency of the relaxant effect of the xanthine molecule depends on cell membrane permeability and affinity for cyclic AMP PDE based on its hydrophobic property.6 In particular, the extent of binding of the drug to serum protein, which is related to its hydrophobicity, is an important factor in permeation since it is well known that only the unbound drug affects pharmacological activity and the capacity of diffusion across biological membranes. However, though the relation between hydrophobicity and biological activities of xanthine derivatives has been clarified, to date little is known about the protein binding characteristics of the derivatives. In vitro studies using guinea pig isolated tracheal smooth muscle preparations are commonly performed in order to evaluate the relaxant effects of drugs. There has been little 0022-3549/91/0400-0349$0 1 .OO/O 0 7991, American Pharmaceutical Association
information, however, on the protein binding characteristics of xanthine derivatives in guinea pig serum proteins. The aim of the present study was to evaluate these protein binding characteristics of xanthine derivatives in isolated guinea pig serum albumin and determine their role in biological activities such as relaxation induction and cyclic AMP PDE inhibition.
Experimental Section Chemical-The N-alkylxanthinederivatives 3-ethyl-(EX), 3-propyl-(PX),3-butyl-(BX),1-methyl-3-ethyl-(MEX),1-methyl-3-propyl(MPX), and 1-methyl-3-butylxanthines (MBX) were synthesized in our laboratory and were identical to those used previously.14 3-Methylxanthine (MX)and 1,3-dimethylxanthine (theophylline; TPH) were obtained from Sigma Chemical Company, St. Louis, MO. All other agents and reagents used in the experiment were obtained . commercially and were used without further purification. Isolated guinea pig serum albumin (GSA, Fraction V) was obtained from Sigma Chemical. Guinea pig serum was obtained from Hartley-strain male guinea pigs (300-500 g, Shizuoka Laboratory Animal Center, Shizuoka, Japan). Biological Activities and Physicochemical Parameter-The values of the PDE inhibition constant ( K J ,concentration producing 50%tracheal smooth muscle relaxation (EC,,), and apparent partition coefficient (PC)between n-octanol and pH 7.4 phosphate-buffered saline for various xanthine derivatives were taken from our previous literature.6 Biochemical Analysis-Total protein concentration was determined by the method of Lowry et al.7 Serum albumin was determined with the bromcresol green method using a commercial kit (Iatron Albumin Kit, Iatron Laboratories,Tokyo, Japan). Estimation of Protein Binding-Protein binding of each compound was determined by the ultrafiltration method using a commercially available device UFC3LGC (Japan Millipore Limited, Tokyo, Japan). The membrane filter of the UFC3LGC can retain compounds with molecular weights of >10 000. Guinea pig serum albumin (GSA)was dissolved in an isotonic phosphate buffersolution (pH 7.4). In a preliminary experiment, protein leakage and adsorption of each compound to the device or to the membrane were negligible. An aliquot (0.2 mL) of compound-spiked serum obtained from guinea pigs or -3% GSA solution with appropriate concentration ranges for each compound was poured into the upper reservoir cup and centrifuged at room temperaturefor 15 min at 2000 x g.Total and unbound (ultrafiltrated) compound concentrations were measured by HPLC. In all experiments, each xanthine derivative was dried in a methanol solution prior to adding guinea pig serum, GSA solution, and the mobile phase for HPLC. Samples thus obtained were shaken gently for 2 h at room temperature. The theoretical concentration of albumin was determined by the method described above. Values were calculated on the basis of human serum albumin with a molecular weight of 69 000. Assuming that albumin is the sole binding protein for these compounds and that only one binding site in serum exists for each compound, protein binding data were fitted to the following equation using the nonlinear least-squaresmethod program MULTI, written by Yamaoka et al.? Journal of Pharmaceutical Sciences I 349 Vol. 80, NO. 4, April 7997
Cb = Kd -k cf
(1)
where Cb and Cfare the concentrations of the bound drug and the unbound drug, respectively, nP is the binding capacity of the first class of binding sites, and Kd is the dissociation constant. High-Performance Liquid Chromatography Assay-Xanthine derivativeconcentrations in the GSA solution and in the ultrafiltrate were determined by HPLC assay. The HPLC apparatus used was a Shimadzu LC-6A system (Shimadzu Company, Kyoto, Japan) which consists of an LC-6A liquid pump, an SPD-6AV UV-VIS spectrophotometric detector, and an SIL-6A autoinjector. A cosmosil 5C,, column (NacalaiTesque, Kyoto, Japan) was used. Mobile phases were 30 mM of KH,PO, (pH 5.O):methanol(60:40v/v% for MPX and MBX; 80:20 v/v% for MX, EX, PX, BX, TPH, and MEX). Flow rate (0.8 to 1.5 mL) and column temperature (35-55 "C) were selected for each compound. The effluent column was monitored at 274 nm. In a 1.5-mL centrifuge plastic tube, 50 pL of samples and 0.2 mL of the internal standard solution (PX at a concentration of 1 pg/mL for MX, EX, TPH, MEX, and BX; BX at a concentration of 2 Fg/mL for PX; MPX and MBX at a concentration of 1 pg/mL for MBX and MPX, respectively) were vortexed and centrifuged at 6000 x g for 2 min. To a glass culture tube was added 0.2 mL of the supernatant.The supernatantwas then evaporated to dryness under a gentle stream of nitrogen at 40 "C, and the residue was reconstituted with 0.2 mL of the respective mobile phase. The reconstituted solution was then injected into the column. Each xanthine derivative was measured over a range 0.04-100 pg/mL. The detection limit for each drug was 0.03 jg/mL, with a linear detection range of up to 100 pg/mL. The coefficients of variation for the HPLC assay were <6%.
Results and Discussion The present study addresses the protein binding characteristics of various xanthine derivatives in guinea pig serum albumin. We previously reported on the existence of the specific binding sites of MPX in albumin in human serum proteins and on the difference in the binding profile of MPX among experimental animal sera, indicating a species difference.3 In order to clarify whether the specific binding sites of xanthine derivatives in guinea pig serum proteins exist in albumin, the binding of TPH, PX, and MPX to guinea pig serum and to guinea pig serum albumin were also investigated. In the present study, a concentration of guinea pig serum albumin was determined to be at -3% since preliminary experiments indicated that albumin content in guinea pig serum obtained from Hartley-strain male guinea pigs was 2.94 0.12%(n = 3). The degree of binding of these drugs to serum was nearly equal to that of serum albumin, indicating that albumin plays an important role in the protein binding of xanthine derivatives. This observation is reasonable since it is well known that a number of acidic drugs bind mainly to serum albumin and this supports our previous findings.3 Protein binding profiles for each of the studied xanthine derivatives in guinea pig serum albumin are shown in Figure 1. As can be seen in Figure 1, the simulated curves are shifted to the left with the prolongation of the alkyl chain length of the xanthine molecule. Protein binding behavior varied among the xanthine derivatives, with the order of the degree of binding in guinea pig serum albumin being MBX > MPX> PX > BX > MEX > EX > TPH > MX. As expected, binding of each of the N'-methyl-W-alkylxanthine derivatives (TPH, MEX, MPX, and MBX) to guinea pig serum albumin was stronger than that of the corresponding W-alkylxanthine derivatives (MX, EX, PX, and BX). These results reveal that the degree of protein binding of xanthine derivatives is increased both when a W-methyl group is replaced by a longer W-alkyl chain and when the "-position of a W alkylxanthine molecule is additionally replaced by a methyl group. It may be suggested that the methyl group in the N'-position for each W-alkylxanthine molecule has a steric
*
350 / Journal of Pharmaceutical Sciences Vol. 80, No. 4, April 1997
r
300
.=a c 0 .-
4 (0
L
200
4
C QJ U
C
0
u m 3
L 73
100
-0
C
3 0
m
0
100
200
300
400
500
Free drug c o n c e n t r a t i o n ( p M ) Flgure 1-Protein
binding profiles for xanthine derivatives in guinea pig serum albumin. Solid lines represent computer-fitted curves taken from eq 1. Symbols represent experimental values: (0) MX; (0)EX; (0)PX; (A) BX; (W) TPH; (0)MEX; (e)MPX; (A)MBX.
effect on protein binding behaviors in guinea pig serum albumin. Computer estimates of binding parameters as calculated by the nonlinear least-squares method are summarized in Table I, in addition to data of biological activities (-log Ki and -log EC50)and physicochemical parameter (log PC). Considerable interstructural differences were observed in binding parameters. The binding affinity (Kd)was the highest for MBX (4.86 pM) and the lowest for MX (515.60 pM). Capacity of the first class (nP)ranged from 244.76pM for MX to 547.85 p M for PX, indicating a high affinity and low capacity for MBX in guinea pig serum albumin. These results suggest that interstructural differences in protein binding of the xanthine derivatives evaluated in this study may be due to differences in the binding affinity andlor capacity. The protein binding parameters of xanthine derivatives correlated positively with their hydrophobicity. As can be seen in Figure 2, a significant correlation (r = 0.919, p < 0.01) was obtained from the logarithmic plot of the partition coefficient (log PC)against the negative logarithmic values of the dissociation constant (-log K J . These results indicate that the avidity of protein binding of xanthine derivatives is related to the hydrophobic property, which increases with the alkyl chain length at the W-position of the Nl-methylxanthine molecule. These results were also supported by findings of a relation between the extent of protein binding and hydrophobicity of structurally similar drugs.9 Our previous studies also demonstrated that this increase in hydrophobicity of the xanthine molecule enhances the potency of biological activities such as the bronchodilating effect and PDE inhibition.6 We confirm that hydrophobicity expressed as a logarithmic partition coefficient (log PC) is an important key in both the protein binding behavior and biological activities of xanthine derivatives. A number of articles concerning the relations between HPLC retention time data and hydrophobicity or biological activity have been published.10-15 In the present study, in order to evaluate the usefulness of HPLC retention time data in predicting the structure-activity relation of xanthine derivatives, simultaneous separation of each xanthine derivative was performed using the reversed-phase column ((2-18) which is useful as an indicator for hydrophobicity. The optimal condition to separate each xanthine derivative se-
Table I-Chemical Constitution, Biological Activities (-log ECS0and -log K, Values), Physiological Parameter (-log K,,), and Physlcochemlcal Parameters (log PC and log RT) of Each Xanthine Derivative n
-log &, Mb
H H H H
Methyl
Me Me Me Me
Methyl
3.91 4.20 4.38 4.49 4.25 4.59 5.73 5.92
3.85 4.16 5.14 5.08 4.55 5.33 6.01 5.87
Ethyl
Propyl Butyl Ethyl
Propyl BUM
3.29 3.48 4.29 4.13 3.41 4.12 5.07 5.31
log PC"
log RT, minC
-0.72 -0.11 0.33 0.84
0.18 0.28 0.82 1.26 0.60 0.90 1.27 1.73
-0.04 0.52 1.02 1.29
a Each value was cited from a previous report (ref 6),where EC,, = the concentration producing 50% relaxation for the resting tone in guinea pig isolated tracheal muscle, K, = the inhibition constant for cyclic AMP phosphodiesterase on guinea pig isolated tracheal muscle, and PC represents apparent partition coefficients in rkoctanol. Each value represents the mean of the computer-estimatedparameter, where K,, represents the apparent dissociation constant for guinea pig serum albumin. Each value represents the mean of three measurements, where RT represents the adjusted retention time under the HPLC conditions used.
h
3 -0
Y
m 0
3-1 -1
' "0
1
2
log PC Figure 2-Correlation between apparent partition coefficients (log PC) of xanthine derivatives and their apparent dissociation constants (-log Kd) for guinea pig serum albumin. The log PC values are cited from a previous report (ref 6).The regression line (solid line) was calculated by the least-squaresmethod (dotted lines indicate 95% confidence limits for the regression line: symbols as defined in Figure 1).
0
20
40
60
Retentlon tlme (mln)
Figure &Typical HPLC chromatogram for each xanthine derivative studied: (A) MX; (6)EX; (C) TPH; (D) PX; (E) BX; (F) MEX; (G) MPX; (H) MBX. The HPLC conditions are described in the text.
order is in agreement with that of log PC. A very good linear relationship was found between log P C and log RT values (r = 0.971, p < 0.01; Figure 4). Similarly, a significant lected was determined at the mobile phase which consisted of correlation (r = 0.916, p < 0.01) was also obtained between methano' added to a 30 mM KH2P0, solution (pH 5.0) in a -log Kd and log RT values (data not shown). As shown in ratio of 80:ZO. It has been reported, however, that the use of Figure 5, a significant correlation was also seen between log HPLC retention time data measured by a single mobile phase RT values and the negative logarithmic values for concencomposition in order to characterize hydrophobicity of a drug trations producing 50% of tracheal smooth muscle relaxation is unsuitable.12J3 The mobile phase composition and flow rate in vitro (-log EC,,; r = 0.915, p < 0.01). In the same way, a (1mL/min) were adjusted for a total chromatographic retensignificant correlation between log RT values and the negation time of 60 min. The temperature was at 40 "C, and 50 p L tive logarithmic values of the PDE inhibition constant (-log of sample containing each xanthine derivative was injected in Ki) was also obtained (r = 0.867, p < 0.01; data not shown). triplicate. A typical HPLC chromatogram obtained under These results indicate that the retention time data under the these conditions is shown in Figure 3. The mobile phase conditions in this study may be a suitable model for predicting hold-up time was 2.00 min, and the mean adjusted retention biological activities such as relaxation and cyclic AMP PDE times (RT) for each xanthine derivative were shown to be 1.51,1.90,6.57,18.11,3.95,7.94,18.83,and53.34minforMX, inhibition, since significant correlations were found to exist between hydrophobicity characterized by the partition coefEX, PX, BX, TPH, MEX, MPX, and MBX, respectively. This Journal of Pharmaceutical Sciences I 351 Vol. 80, No. 4, April 7997
-E
n
.-C U
3
c
u
m
W
cn 0
ul 0
c
c
I
o! -1
’ 0
1
2
Figure 4-Correlation between apparent partition coefficients (log PC) and adjusted HPLC retention time data (log RT) for each xanthine derivative.The log PCvalues are cited from a previous report (ref 6).The regression line (solid line) was calculated by the least-squares method (dotted lines indicate 95% confidence limits for the regression line; symbols as defined in Figure 1).
ficient (log PC)and the biological activities of the xanthine derivatives (r = 0.944 and r = 0.872, respectively).s Our previous pharmacokinetic studies using animals have demonstrated marked differences in the pharmacokinetic characteristics among various xanthine derivatives.1.2 In addition, we have demonstrated that protein binding of some xanthine derivatives is affected by various physiological factors such as pH, albumin concentration, and free fatty acids.4 Thus, the information obtained here concerning protein binding characteristics of xanthine derivatives, in addition to their hydrophobic data, may be useful in interpreting the pharmacokinetics and pharmacologic effects of a new xanthine derivative with a strong relaxant effect.
0
1. Takagi, K.; Hasegawa, T.; Kuzuya, T.; 0 awa, K.; Watanabe, T.;
f.;
Satake, T.; Miyamoto, K.; Wakusawa, Koshiura, R. Jpn. J . Pharmacol. 1988,46,373-378. 2. Apichartpichean, R.; Takagi, K.; Nadai, M.; Kuzuya, T.; Ogawa, K.; Miyamoto, K.; Hasegawa, T. Jpn. J . Pharmacol. 1988, 48, 341347. 3. Apichartpichean, R.; Takagi, K.; Kuzuya, T.; Nadai, M.; Oh-
352 I Journal of Pharmaceutical Sciences Vol. 80, No. 4, April 1991
2 (mid
Figure SCorrelation between adjusted HPLC retention time data (log RT) for each xanthine derivative and their relaxant effects (-log EC,,)
on isolated guinea pig tracheal smooth muscle. The -log EC,, values are cited from a previous report (ref 6).The regression line (solid line)was calculated by the least-squares method (dotted lines indicate 95% confidence limits for the regression line; symbols as defined in Figure 1).
4. 5. 6. 7. 8.
9. 10.
References and Notes
1 log RT
11. 12. 13. 14. 15.
shima, T.; Suzuki, K.; Horiuchi, T.; Miyamoto, K.; Hasegawa, T. Int. J . Clin. Pharmacol. Ther. Toxicol. 1989,27, 320-323. Hasegawa, T.; Apichartpichean, R.; Kuzuya, T.; Nadai, M.; Tsunekawa, Y.; Horiuchi, T.; Miyamoto, K. Int. J . Pharm. 1989, 56,235-241. Ogawa, K.; Takagi, K.; Satake, T. Br. J . Pharmacol. 1989, 97, 542-546. Miyamoto, K.; Takagi, K.; Sakai, R.; Wakusawa, S.; Koshiura, R.; Nadai, M.; Apichartpichean, R.; Hasegawa, T. J. Pharm. Pharmacol. 1989,41, 844-2341. Lowry, 0. H.; Rosebrough, N. J.; Farr, A. L.; Randall, R. J. J . Biol. Chem. 1951,193, 265-275. Yamaoka, K.; Tanigawara, Y.; Nakagawa, T.; Uno, T. J . Pharmacobic+Dyn. 1981,4, 879-885. Laznicek. M.: Kvetina. J.: Mazak., J.:. Krch. V. J . Pharm. Pharmacol. 1987,39, 79-83. ’ Henry, D.; Block, J. H.; Anderson, J. L.; Carlson, G. R. J. Med. Chem. 1976.19.619-626. Barker, J. K.;Rauls, D. 0.;Bourne, R. F. J . Med. Chen. 1979,22, 1301-1306. Tayer, N. E.; van de Waterbeemd, H.; Testa, B. J . Chromatogr. 1985,320, 293304. Thus, J. L. G.; Kraak, J. C. J . Chromatogr. 1985,320,271-279. Braumann, T. J . Chromatogr. 1986,373, 191-225. Miyake, K.; Mizuno, N.; Terada, H. J . Chromatogr. 1988, 439, 227-235.
In the course of studies on the structure-activity relationship of N-alkylxanthine derivatives, 1-methyl-3-propylxanthine (MPX) was selected as one of the most interesting compounds synthesized in our laboratory.l-4 In our recent studies, the relaxant effects and inhibiting activities on cyclic AMP phosphodiesterase (PDE) of various xanthine derivatives, using tracheal smooth muscle isolated from guinea pigs, were estimated. The ability of xanthine derivatives to inhibit cyclic AMP PDE activity contributes, at least in part, to the mechanism of bronchodilatory action of these derivatives. The hydrophobicity of the derivatives is an important determinant in biological activities such as relaxation induction and cyclic AMP PDE inhibition.5.6 Finally, we have proposed that hydrophobicity is useful for chemical modifications of more potent xanthine derivatives; for example, as a smooth muscle relaxant. In general, drug permeability through biological membranes is considered to depend on factors such as protein binding in serum, hydrophobicity, and molecular size of the drug. Our previous study suggested that the potency of the relaxant effect of the xanthine molecule depends on cell membrane permeability and affinity for cyclic AMP PDE based on its hydrophobic property.6 In particular, the extent of binding of the drug to serum protein, which is related to its hydrophobicity, is an important factor in permeation since it is well known that only the unbound drug affects pharmacological activity and the capacity of diffusion across biological membranes. However, though the relation between hydrophobicity and biological activities of xanthine derivatives has been clarified, to date little is known about the protein binding characteristics of the derivatives. In vitro studies using guinea pig isolated tracheal smooth muscle preparations are commonly performed in order to evaluate the relaxant effects of drugs. There has been little 0022-3549/91/0400-0349$0 1 .OO/O 0 7991, American Pharmaceutical Association
information, however, on the protein binding characteristics of xanthine derivatives in guinea pig serum proteins. The aim of the present study was to evaluate these protein binding characteristics of xanthine derivatives in isolated guinea pig serum albumin and determine their role in biological activities such as relaxation induction and cyclic AMP PDE inhibition.
Experimental Section Chemical-The N-alkylxanthinederivatives 3-ethyl-(EX), 3-propyl-(PX),3-butyl-(BX),1-methyl-3-ethyl-(MEX),1-methyl-3-propyl(MPX), and 1-methyl-3-butylxanthines (MBX) were synthesized in our laboratory and were identical to those used previously.14 3-Methylxanthine (MX)and 1,3-dimethylxanthine (theophylline; TPH) were obtained from Sigma Chemical Company, St. Louis, MO. All other agents and reagents used in the experiment were obtained . commercially and were used without further purification. Isolated guinea pig serum albumin (GSA, Fraction V) was obtained from Sigma Chemical. Guinea pig serum was obtained from Hartley-strain male guinea pigs (300-500 g, Shizuoka Laboratory Animal Center, Shizuoka, Japan). Biological Activities and Physicochemical Parameter-The values of the PDE inhibition constant ( K J ,concentration producing 50%tracheal smooth muscle relaxation (EC,,), and apparent partition coefficient (PC)between n-octanol and pH 7.4 phosphate-buffered saline for various xanthine derivatives were taken from our previous literature.6 Biochemical Analysis-Total protein concentration was determined by the method of Lowry et al.7 Serum albumin was determined with the bromcresol green method using a commercial kit (Iatron Albumin Kit, Iatron Laboratories,Tokyo, Japan). Estimation of Protein Binding-Protein binding of each compound was determined by the ultrafiltration method using a commercially available device UFC3LGC (Japan Millipore Limited, Tokyo, Japan). The membrane filter of the UFC3LGC can retain compounds with molecular weights of >10 000. Guinea pig serum albumin (GSA)was dissolved in an isotonic phosphate buffersolution (pH 7.4). In a preliminary experiment, protein leakage and adsorption of each compound to the device or to the membrane were negligible. An aliquot (0.2 mL) of compound-spiked serum obtained from guinea pigs or -3% GSA solution with appropriate concentration ranges for each compound was poured into the upper reservoir cup and centrifuged at room temperaturefor 15 min at 2000 x g.Total and unbound (ultrafiltrated) compound concentrations were measured by HPLC. In all experiments, each xanthine derivative was dried in a methanol solution prior to adding guinea pig serum, GSA solution, and the mobile phase for HPLC. Samples thus obtained were shaken gently for 2 h at room temperature. The theoretical concentration of albumin was determined by the method described above. Values were calculated on the basis of human serum albumin with a molecular weight of 69 000. Assuming that albumin is the sole binding protein for these compounds and that only one binding site in serum exists for each compound, protein binding data were fitted to the following equation using the nonlinear least-squaresmethod program MULTI, written by Yamaoka et al.? Journal of Pharmaceutical Sciences I 349 Vol. 80, NO. 4, April 7997
Cb = Kd -k cf
(1)
where Cb and Cfare the concentrations of the bound drug and the unbound drug, respectively, nP is the binding capacity of the first class of binding sites, and Kd is the dissociation constant. High-Performance Liquid Chromatography Assay-Xanthine derivativeconcentrations in the GSA solution and in the ultrafiltrate were determined by HPLC assay. The HPLC apparatus used was a Shimadzu LC-6A system (Shimadzu Company, Kyoto, Japan) which consists of an LC-6A liquid pump, an SPD-6AV UV-VIS spectrophotometric detector, and an SIL-6A autoinjector. A cosmosil 5C,, column (NacalaiTesque, Kyoto, Japan) was used. Mobile phases were 30 mM of KH,PO, (pH 5.O):methanol(60:40v/v% for MPX and MBX; 80:20 v/v% for MX, EX, PX, BX, TPH, and MEX). Flow rate (0.8 to 1.5 mL) and column temperature (35-55 "C) were selected for each compound. The effluent column was monitored at 274 nm. In a 1.5-mL centrifuge plastic tube, 50 pL of samples and 0.2 mL of the internal standard solution (PX at a concentration of 1 pg/mL for MX, EX, TPH, MEX, and BX; BX at a concentration of 2 Fg/mL for PX; MPX and MBX at a concentration of 1 pg/mL for MBX and MPX, respectively) were vortexed and centrifuged at 6000 x g for 2 min. To a glass culture tube was added 0.2 mL of the supernatant.The supernatantwas then evaporated to dryness under a gentle stream of nitrogen at 40 "C, and the residue was reconstituted with 0.2 mL of the respective mobile phase. The reconstituted solution was then injected into the column. Each xanthine derivative was measured over a range 0.04-100 pg/mL. The detection limit for each drug was 0.03 jg/mL, with a linear detection range of up to 100 pg/mL. The coefficients of variation for the HPLC assay were <6%.
Results and Discussion The present study addresses the protein binding characteristics of various xanthine derivatives in guinea pig serum albumin. We previously reported on the existence of the specific binding sites of MPX in albumin in human serum proteins and on the difference in the binding profile of MPX among experimental animal sera, indicating a species difference.3 In order to clarify whether the specific binding sites of xanthine derivatives in guinea pig serum proteins exist in albumin, the binding of TPH, PX, and MPX to guinea pig serum and to guinea pig serum albumin were also investigated. In the present study, a concentration of guinea pig serum albumin was determined to be at -3% since preliminary experiments indicated that albumin content in guinea pig serum obtained from Hartley-strain male guinea pigs was 2.94 0.12%(n = 3). The degree of binding of these drugs to serum was nearly equal to that of serum albumin, indicating that albumin plays an important role in the protein binding of xanthine derivatives. This observation is reasonable since it is well known that a number of acidic drugs bind mainly to serum albumin and this supports our previous findings.3 Protein binding profiles for each of the studied xanthine derivatives in guinea pig serum albumin are shown in Figure 1. As can be seen in Figure 1, the simulated curves are shifted to the left with the prolongation of the alkyl chain length of the xanthine molecule. Protein binding behavior varied among the xanthine derivatives, with the order of the degree of binding in guinea pig serum albumin being MBX > MPX> PX > BX > MEX > EX > TPH > MX. As expected, binding of each of the N'-methyl-W-alkylxanthine derivatives (TPH, MEX, MPX, and MBX) to guinea pig serum albumin was stronger than that of the corresponding W-alkylxanthine derivatives (MX, EX, PX, and BX). These results reveal that the degree of protein binding of xanthine derivatives is increased both when a W-methyl group is replaced by a longer W-alkyl chain and when the "-position of a W alkylxanthine molecule is additionally replaced by a methyl group. It may be suggested that the methyl group in the N'-position for each W-alkylxanthine molecule has a steric
*
350 / Journal of Pharmaceutical Sciences Vol. 80, No. 4, April 1997
r
300
.=a c 0 .-
4 (0
L
200
4
C QJ U
C
0
u m 3
L 73
100
-0
C
3 0
m
0
100
200
300
400
500
Free drug c o n c e n t r a t i o n ( p M ) Flgure 1-Protein
binding profiles for xanthine derivatives in guinea pig serum albumin. Solid lines represent computer-fitted curves taken from eq 1. Symbols represent experimental values: (0) MX; (0)EX; (0)PX; (A) BX; (W) TPH; (0)MEX; (e)MPX; (A)MBX.
effect on protein binding behaviors in guinea pig serum albumin. Computer estimates of binding parameters as calculated by the nonlinear least-squares method are summarized in Table I, in addition to data of biological activities (-log Ki and -log EC50)and physicochemical parameter (log PC). Considerable interstructural differences were observed in binding parameters. The binding affinity (Kd)was the highest for MBX (4.86 pM) and the lowest for MX (515.60 pM). Capacity of the first class (nP)ranged from 244.76pM for MX to 547.85 p M for PX, indicating a high affinity and low capacity for MBX in guinea pig serum albumin. These results suggest that interstructural differences in protein binding of the xanthine derivatives evaluated in this study may be due to differences in the binding affinity andlor capacity. The protein binding parameters of xanthine derivatives correlated positively with their hydrophobicity. As can be seen in Figure 2, a significant correlation (r = 0.919, p < 0.01) was obtained from the logarithmic plot of the partition coefficient (log PC)against the negative logarithmic values of the dissociation constant (-log K J . These results indicate that the avidity of protein binding of xanthine derivatives is related to the hydrophobic property, which increases with the alkyl chain length at the W-position of the Nl-methylxanthine molecule. These results were also supported by findings of a relation between the extent of protein binding and hydrophobicity of structurally similar drugs.9 Our previous studies also demonstrated that this increase in hydrophobicity of the xanthine molecule enhances the potency of biological activities such as the bronchodilating effect and PDE inhibition.6 We confirm that hydrophobicity expressed as a logarithmic partition coefficient (log PC) is an important key in both the protein binding behavior and biological activities of xanthine derivatives. A number of articles concerning the relations between HPLC retention time data and hydrophobicity or biological activity have been published.10-15 In the present study, in order to evaluate the usefulness of HPLC retention time data in predicting the structure-activity relation of xanthine derivatives, simultaneous separation of each xanthine derivative was performed using the reversed-phase column ((2-18) which is useful as an indicator for hydrophobicity. The optimal condition to separate each xanthine derivative se-
Table I-Chemical Constitution, Biological Activities (-log ECS0and -log K, Values), Physiological Parameter (-log K,,), and Physlcochemlcal Parameters (log PC and log RT) of Each Xanthine Derivative n
-log &, Mb
H H H H
Methyl
Me Me Me Me
Methyl
3.91 4.20 4.38 4.49 4.25 4.59 5.73 5.92
3.85 4.16 5.14 5.08 4.55 5.33 6.01 5.87
Ethyl
Propyl Butyl Ethyl
Propyl BUM
3.29 3.48 4.29 4.13 3.41 4.12 5.07 5.31
log PC"
log RT, minC
-0.72 -0.11 0.33 0.84
0.18 0.28 0.82 1.26 0.60 0.90 1.27 1.73
-0.04 0.52 1.02 1.29
a Each value was cited from a previous report (ref 6),where EC,, = the concentration producing 50% relaxation for the resting tone in guinea pig isolated tracheal muscle, K, = the inhibition constant for cyclic AMP phosphodiesterase on guinea pig isolated tracheal muscle, and PC represents apparent partition coefficients in rkoctanol. Each value represents the mean of the computer-estimatedparameter, where K,, represents the apparent dissociation constant for guinea pig serum albumin. Each value represents the mean of three measurements, where RT represents the adjusted retention time under the HPLC conditions used.
h
3 -0
Y
m 0
3-1 -1
' "0
1
2
log PC Figure 2-Correlation between apparent partition coefficients (log PC) of xanthine derivatives and their apparent dissociation constants (-log Kd) for guinea pig serum albumin. The log PC values are cited from a previous report (ref 6).The regression line (solid line) was calculated by the least-squaresmethod (dotted lines indicate 95% confidence limits for the regression line: symbols as defined in Figure 1).
0
20
40
60
Retentlon tlme (mln)
Figure &Typical HPLC chromatogram for each xanthine derivative studied: (A) MX; (6)EX; (C) TPH; (D) PX; (E) BX; (F) MEX; (G) MPX; (H) MBX. The HPLC conditions are described in the text.
order is in agreement with that of log PC. A very good linear relationship was found between log P C and log RT values (r = 0.971, p < 0.01; Figure 4). Similarly, a significant lected was determined at the mobile phase which consisted of correlation (r = 0.916, p < 0.01) was also obtained between methano' added to a 30 mM KH2P0, solution (pH 5.0) in a -log Kd and log RT values (data not shown). As shown in ratio of 80:ZO. It has been reported, however, that the use of Figure 5, a significant correlation was also seen between log HPLC retention time data measured by a single mobile phase RT values and the negative logarithmic values for concencomposition in order to characterize hydrophobicity of a drug trations producing 50% of tracheal smooth muscle relaxation is unsuitable.12J3 The mobile phase composition and flow rate in vitro (-log EC,,; r = 0.915, p < 0.01). In the same way, a (1mL/min) were adjusted for a total chromatographic retensignificant correlation between log RT values and the negation time of 60 min. The temperature was at 40 "C, and 50 p L tive logarithmic values of the PDE inhibition constant (-log of sample containing each xanthine derivative was injected in Ki) was also obtained (r = 0.867, p < 0.01; data not shown). triplicate. A typical HPLC chromatogram obtained under These results indicate that the retention time data under the these conditions is shown in Figure 3. The mobile phase conditions in this study may be a suitable model for predicting hold-up time was 2.00 min, and the mean adjusted retention biological activities such as relaxation and cyclic AMP PDE times (RT) for each xanthine derivative were shown to be 1.51,1.90,6.57,18.11,3.95,7.94,18.83,and53.34minforMX, inhibition, since significant correlations were found to exist between hydrophobicity characterized by the partition coefEX, PX, BX, TPH, MEX, MPX, and MBX, respectively. This Journal of Pharmaceutical Sciences I 351 Vol. 80, No. 4, April 7997
-E
n
.-C U
3
c
u
m
W
cn 0
ul 0
c
c
I
o! -1
’ 0
1
2
Figure 4-Correlation between apparent partition coefficients (log PC) and adjusted HPLC retention time data (log RT) for each xanthine derivative.The log PCvalues are cited from a previous report (ref 6).The regression line (solid line) was calculated by the least-squares method (dotted lines indicate 95% confidence limits for the regression line; symbols as defined in Figure 1).
ficient (log PC)and the biological activities of the xanthine derivatives (r = 0.944 and r = 0.872, respectively).s Our previous pharmacokinetic studies using animals have demonstrated marked differences in the pharmacokinetic characteristics among various xanthine derivatives.1.2 In addition, we have demonstrated that protein binding of some xanthine derivatives is affected by various physiological factors such as pH, albumin concentration, and free fatty acids.4 Thus, the information obtained here concerning protein binding characteristics of xanthine derivatives, in addition to their hydrophobic data, may be useful in interpreting the pharmacokinetics and pharmacologic effects of a new xanthine derivative with a strong relaxant effect.
0
1. Takagi, K.; Hasegawa, T.; Kuzuya, T.; 0 awa, K.; Watanabe, T.;
f.;
Satake, T.; Miyamoto, K.; Wakusawa, Koshiura, R. Jpn. J . Pharmacol. 1988,46,373-378. 2. Apichartpichean, R.; Takagi, K.; Nadai, M.; Kuzuya, T.; Ogawa, K.; Miyamoto, K.; Hasegawa, T. Jpn. J . Pharmacol. 1988, 48, 341347. 3. Apichartpichean, R.; Takagi, K.; Kuzuya, T.; Nadai, M.; Oh-
352 I Journal of Pharmaceutical Sciences Vol. 80, No. 4, April 1991
2 (mid
Figure SCorrelation between adjusted HPLC retention time data (log RT) for each xanthine derivative and their relaxant effects (-log EC,,)
on isolated guinea pig tracheal smooth muscle. The -log EC,, values are cited from a previous report (ref 6).The regression line (solid line)was calculated by the least-squares method (dotted lines indicate 95% confidence limits for the regression line; symbols as defined in Figure 1).
4. 5. 6. 7. 8.
9. 10.
References and Notes
1 log RT
11. 12. 13. 14. 15.
shima, T.; Suzuki, K.; Horiuchi, T.; Miyamoto, K.; Hasegawa, T. Int. J . Clin. Pharmacol. Ther. Toxicol. 1989,27, 320-323. Hasegawa, T.; Apichartpichean, R.; Kuzuya, T.; Nadai, M.; Tsunekawa, Y.; Horiuchi, T.; Miyamoto, K. Int. J . Pharm. 1989, 56,235-241. Ogawa, K.; Takagi, K.; Satake, T. Br. J . Pharmacol. 1989, 97, 542-546. Miyamoto, K.; Takagi, K.; Sakai, R.; Wakusawa, S.; Koshiura, R.; Nadai, M.; Apichartpichean, R.; Hasegawa, T. J. Pharm. Pharmacol. 1989,41, 844-2341. Lowry, 0. H.; Rosebrough, N. J.; Farr, A. L.; Randall, R. J. J . Biol. Chem. 1951,193, 265-275. Yamaoka, K.; Tanigawara, Y.; Nakagawa, T.; Uno, T. J . Pharmacobic+Dyn. 1981,4, 879-885. Laznicek. M.: Kvetina. J.: Mazak., J.:. Krch. V. J . Pharm. Pharmacol. 1987,39, 79-83. ’ Henry, D.; Block, J. H.; Anderson, J. L.; Carlson, G. R. J. Med. Chem. 1976.19.619-626. Barker, J. K.;Rauls, D. 0.;Bourne, R. F. J . Med. Chen. 1979,22, 1301-1306. Tayer, N. E.; van de Waterbeemd, H.; Testa, B. J . Chromatogr. 1985,320, 293304. Thus, J. L. G.; Kraak, J. C. J . Chromatogr. 1985,320,271-279. Braumann, T. J . Chromatogr. 1986,373, 191-225. Miyake, K.; Mizuno, N.; Terada, H. J . Chromatogr. 1988, 439, 227-235.