- Email: [email protected]
Charge-Transfer Chromatographic Study of the Complex Formation of Some Steroidal Drugs with Carboxymethyl-γ-cyclodextrin
ANALYTICAL BIOCHEMISTRY ARTICLE NO.
246, 205–210 (1997)
AB969975
Charge-Transfer Chromatographic Study of the Complex Formation of Some Steroidal Drugs with Carboxymethyl-g-cyclodextrin Tibor Cserha´ti and Esther Forga´cs Central Research Institute for Chemistry, Hungarian Academy of Sciences, P.O. Box 17, H-1525 Budapest, Hungary
Received September 17, 1996
The interaction between 15 steroidal drugs and carboxymethyl-g-cyclodextrin (CM-g-CD) was studied by reversed-phase charge-transfer thin-layer chromatography and the relative strength of interaction was calculated. CM-g-CD formed inclusion complexes with each compound, the complex always being less hydrophobic than the uncomplexed drug. The inclusion-forming capacity of drugs differed considerably depending on their chemical structures. The linear correlation between the hydrophobicity and specific hydrophobic surface area of anticancer drugs indicated that they can be considered as a homologous series of compounds, although their chemical structures are different. Hydrophobicity of drugs significantly influenced the strength of interaction, indicating the involvement of hydrophobic forces in the binding of drugs to CM-g-CD. The marked influence of CM-g-CD on the hydrophobicity of drugs suggests that this interaction may modify the biological properties (adsorption, uptake, half-life, etc.) of drug–CM-g-CD complexes drug, resulting in modified efficacy. q 1997 Academic Press
Cyclodextrins (CDs)1 are cyclic oligosaccharides which can form inclusion complexes with a wide variety of organic and inorganic compounds (1, 2). The complex formation of CDs with drugs such as antineoplastic agents (3), luteinizing hormone-releasing hormone agonist buserelin (4), antisense oligonucleotides (5), dihydroergotamine (6), psoralen (7), and cefpodoxime proxetil (8) has been recently reported. Both the physicochemical and biological characteristics of the complexed drug may deviate from those of the uncomplexed one resulting in better application
parameters and improved efficacy (9). The formation of inclusion complexes with CDs enhanced the dissolution of indomethacin (10), protected against the nephrotoxicity of gentamycin in rats (11), decreased the ocular irritation of pilocarpine (12), modified the penetration of hydrocortisone into excised human skin (13), improved the delivery of acitrecin (14), and modified the degradation of steroidal drugs in aqueous solution (15) and the stability of bioactive peptides (16), etc. Many physicochemical methods have been used for the study of the formation of inclusion complexes. Thus, calorimetry (17), spectrophotometry (18), and various liquid chromatographic methods such as reversed-phase thin-layer (19) and high-performance liquid chromatography (20, 21) have been successfully used for the assessment of the various aspects of host – guest interactions. The objectives of this work were to study the interaction of some steroidal drugs with carboxymethylg-cyclodextrin (CM-g-CD), to determine the relative strength of interaction with the aid of reversedphase chromatography, to compare their inclusion forming capacity, and to elucidate the role of molecular parameters in the inclusion complex formation. The elucidation of the formation of inclusion complexes between the steroidal drugs and CM-g-CD may promote the development of new, more effective drug formulations with higher biological efficiencies and lower toxic side effects. Steroidal drugs were chosen as model compounds because they are produced in considerable quantity by the pharmaceutical industry and are extensively used in the medical praxis. MATERIALS AND METHODS
1 Abbreviations used: CDs, cyclodextrins; CM-g-CD, carboxymethyl-g-cyclodextrin.
Polygram UV254 (Macherey-Nagel, Du¨rren, Germany) plates were impregnated by overnight predevel205
0003-2697/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
AB 9975
/
6m27$$$121
02-12-97 01:00:20
aba
´ TI AND FORGA ´ CS CSERHA
206
FIG. 1. Chemical structures of steroidal drugs.
opment in n-hexane:paraffin oil, 95:5 (v/v). The chemical structures of steroidal drugs are shown in Fig. 1 (1, hydrocortisone; 2, hydrocortisone acetate; 3, prednisolone; 4, prednisolone 21-acetate; 5, 16a-hydroxyprednisolone 21-acetate; 6, 16a-hydroxyprednisolone; 7, d-14–16-a-hydroxyprednisolone 21-acetate; 8, budesonide; 9, d-14-budenoside; 10, 16a-prednisolone16a,17-acetonide; 11, 9a-fluoro-16a-prednisolone-16a, 17-acetonide; 12, 11-desoxy-16a-prednisolone-16a,17acetonide; 13, 9a-fluoroprednisolone; 14, 9,11-epoxi16a-prednisolone-16a,17-acetonide; 15, 11-desoxy-17hydroxycortisone; 16, cholesterin; 17, cholesterin 3-acetate). Drugs were the gift of Professor Sa´ndor Go¨ro¨g (Gedeon Richter, Ltd., Budapest, Hungary). The drugs were separately dissolved in methanol at a concentration of 3 mg/ml and 2 ml of the solutions was
AID
AB 9975
/
6m27$$$122
02-12-97 01:00:20
plotted on the plates. Water–methanol mixtures were used as eluents, the methanol concentration ranging from 0 to 30 vol% in steps of 5 vol%. Because the objective was to study the complex formation between the solutes and CM-g-CD and not the study of the effect of CM-g-CD on the separation of solutes, they were separately spotted on the plates. In this way the competition between the steroidal drugs for the binding sites of CM-g-CD was excluded. Methanol was chosen as the organic solvent miscible with water because it forms only weak inclusion complexes with cyclodextrins (22, 23). CM-g-CD was the product of Cyclolab Research and Development Laboratory, Budapest, Hungary. CM-g-CD was added to the eluents in the concentration range of 0–15 mg/ml in steps of 2.5 mg/ml. Developments were carried out in sandwich chambers (22 1
aba
CHROMATOGRAPHIC STUDY OF STEROID COMPLEX FORMATION
207
FIG. 2. Videogram of reversed-phase thin-layer chromatographic plates under UV light. (A) Water:methanol (3:2, v/v); (B) water:methanol (3:2, v/v) / 15 mg/ml carboxymethyl-g-cyclodextrin. Steroidal drugs 1–15 appear consecutively from left to right. Steroids 16 and 17 do not move in these eluent systems.
22 1 3 cm) at room temperature, with the distance of development at about 16 cm. After development the plates were dried at 1057C and the spots of steroidal drugs were revealed by their UV spectra and by iodine vapor. Each experiment was run in quadruplicate. Some plates were evaluated by Densitometer CD-60 (Desaga, Heidelberg, Germany) using reflectance mode and 255 nm detection wavelength. Videodocumentation (VD 40, Desaga) was also used in some instances. The RM value characterizing the molecular hydrophobicity in reversed-phase thin-layer chromatography was calculated for each drug in each eluent: RM Å log(1/Rf 0 1).
[1]
When the coefficient of variation of the parallel determinations was higher than 8%, the RM value was omitted from the following calculations. To separate the effects of methanol and CM-g-CD on the hydrophobicity of steroidal drugs the equation was fitted to the experimental data RM Å RM0 / b1rC1 / b2rC2 ,
where RM Å RM value for a drug determined at given methanol and CM-g-CD concentrations; RM0 Å RM value extrapolated to zero methanol and CM-g-CD concentrations; b1 is the decrease in the RM value caused by 1% increase in methanol concentration in the eluent (related to the specific hydrophobic surface area of drugs (24); b2 is the decrease in the RM value caused by 1 mg/ml concentration change of CM-g-CD in the eluent (related to the relative strength of interaction); C1 and C2 are the concentrations of methanol and CMg-CD, respectively. Equation [2] was applied separately for each steroidal drug. To test the validity of the hypothesis that in the case of homologous series of solutes the slope and intercept values (b 1 and RM0 in Eq. [2]) are strongly intercorrelated (25, 26), the linear correlation was calculated between the two physicochemical parameters: RM0 Å A / Brb1 .
FIG. 3. Densitograms of compounds 3 and 15. (A) Water:methanol (3:2, v/v); (B) water:methanol (3:2, v/v) / 15 mg/ml carboxymethylg-cyclodextrin.
AID
AB 9975
/
6m27$$$122
02-12-97 01:00:20
[2]
[3]
To find the physicochemical parameters of steroidal drugs significantly influencing their complexforming capacity, stepwise regression analysis was applied (27). The relative strength of interaction (b2)
aba
´ TI AND FORGA ´ CS CSERHA
208
TABLE 1
Parameters of Linear Correlations between the Hydrophobicity (RM) of Steroidal Drugs and the Methanol (C1 vol%) and Carboxymethyl-g-Cyclodextrin (C2 mg/mL) in the Eluent RM Å RM0 / b1rC1 / b2rC2 No. of steroidal drugs Parameter na RM0 0102xb1 103xsb1b 0102xb2 103xsb2b b=1 (%)c b=2 (%)c r2d Fcalce
1
2
3
4
5
6
7
8
18 2.06 4.22 1.76 2.98 2.93 70.10 29.90 0.9782 337.29
15 3.17 5.65 3.92 5.03 4.88 58.31 41.69 0.9502 114.48
18 2.17 4.41 2.03 3.24 3.36 69.27 30.73 0.9741 282.20
22 2.75 4.76 1.29 4.81 4.43 77.25 22.75 0.9876 754.73
16 2.66 5.54 3.61 3.62 4.77 66.90 33.10 0.9501 123.86
18 2.00 4.91 2.33 2.48 3.86 76.63 23.37 0.9701 243.60
17 2.50 5.03 1.83 3.43 2.68 68.23 31.77 0.9837 422.00
15 2.84 4.11 2.69 2.86 3.35 64.13 35.87 0.9519 118.66
No. of steroidal drugs Parameter na RM0 0102xb1 103xsb1b 0102xb2 103xsb2b b=1 (%)c b=2 (%)c r2 d Fcalce
9
10
11
12
13
14
15
15 2.98 4.70 3.33 2.41 4.15 70.82 29.18 0.9432 99.71
20 3.53 5.48 1.72 5.71 5.82 76.41 23.59 0.9868 636.47
16 2.34 4.01 2.58 4.97 3.41 51.57 48.43 0.9647 177.52
20 3.69 5.53 2.07 7.64 6.99 70.96 29.04 0.9795 406.22
17 2.13 4.17 2.22 5.43 3.25 52.90 47.10 0.9758 281.60
15 2.80 5.10 2.67 3.95 3.32 61.34 38.36 0.9697 191.71
15 2.59 4.34 3.83 5.79 4.77 48.27 51.73 0.9416 96.74
Note. Numbers refer to steroidal drugs in Fig. 1. b1 , decrease in the RM value caused by 1% increase in methanol concentration in the eluent (related to the specific hydrophobic surface area of drugs; b2 , decrease in the RM value caused by 1 mg/ml concentration change of CM-t-CD in the eluent (related to the relative strength of interaction). a Number of data points. b Standard deviations of b1 and b2 . c Standard partial regression coefficients of b1 and b2 , which are normalized to unity. d Coefficient of determination. e Calculated F value indicating the fitness of Eq. [2] to the experimental data.
was the dependent variable, whereas the hydrophobicity (RM0), specific hydrophobic surface area (b1) of Eq. [2], and the complex hydrophobicity parameter R M0/b1 (28) were the independent variables, respectively. The number of accepted independent variables was not limited and the acceptance limit was set to the 95% significance level. In the common multivariate regression analysis the presence of independent variables exerting no significant influence on the change of dependent variable considerably decreases the significance level of the equation. Stepwise regression analysis eliminates from the selected equation the dependent variables having no significant impact on the dependent variable, increasing in this manner the reliability of the calculation. Software of stepwise regression analysis was prepared by CompuDrug Ltd., Budapest, Hungary.
AID
AB 9975
/
6m27$$$122
02-12-97 01:00:20
RESULTS AND DISCUSSION
The simultaneous effect of methanol and CM-g-CD concentrations on the RM values of drugs are shown in Figs. 2 and 3. An increase in the CM-g-CD concentration causes a decrease in RM values, indicating complex (probably inclusion complex) formation. Interaction of the more hydrophilic CM-g-CD with the steroidal drugs decreases the hydrophobicity of the latter. This finding suggests that the biological properties (adsorption, uptake, half-life, etc.) of drug–CM-g-CD complexes may be different from those of an uncomplexed drug resulting in modified effectivity. The data further prove that the addition of CM-g-CD to the eluent does not cause peak distortion or peak widening, lessening the accuracy and reliability of the determination of the relative strength of interaction.
aba
CHROMATOGRAPHIC STUDY OF STEROID COMPLEX FORMATION
The parameters of Eq. [2] are compiled in Table 1. The equation fits the experimental data well, the significance levels in each instance being over 99.9% (see calculated F values). The ratios of variance explained were about 94–99% (see r2 values). Each steroidal drug interacted with CM-g-CD (b2 values differ significantly from zero), meaning that in pharmaceutical formulations containing both steroidal drugs and CM-g-CD their possible interaction must be taken into consideration. The parameters of Eq. [2] show high variations between the drugs, proving that the hydrophobicity (RM0), specific hydrophobic surface area (b1), and their capacity to form inclusion complexes with CM-g-CD (b2) differ considerably. This finding also suggests that the inclusion complex formation may influence differently the biological effect of individual steroidal drugs. The path coefficients (b* values) indicate that the impact of the change of methanol and CM-g-CD concentrations on the reversed-phase mobility of steroidal drugs is commensurable; that is, the retention of steroidal drugs can be equally modified by changing either the methanol or the CM-g-CD concentration in the eluent. Significant linear correlation was found between the intercept (hydrophobicity) and slope (specific hydrophobic surface area) values of steroidal drugs (Fig. 4). This finding indicates that from a chromatographic point of view these drugs behave as a homologous series of compounds, although their chemical structures are different. Significant relationships (significance level over 95%) were found between the hydrophobicity parame-
209
FIG. 5. Relationship between the hydrophobicity (RM0) of steroidal drugs and their capacity to interact with carboxymethyl-g-cyclodextrin (b2). Results of stepwise regression analysis (rcalc is the coefficient of correlation indicating the fitness of the selected equation to the experimental data in the case of one independent variable, and r95% is the tabulated value of the coefficient of correlation related to 95% significance level).
ters of steroidal drugs and their capacity to interact with CM-g-CD (Fig. 5). The fact that hydrophobicity (RM0) of steroidal drugs exerts a significant influence on the strength of interaction indicates that hydrophobic forces are involved in the interaction. This result supports the assumption that steroidal drugs enter into the lipophilic cyclodextrin cavity, and they are retained by hydrophobic forces. However, these hydrophobicity explains only a fairly low ratio of the total variance, indicating that other molecular parameters may account for the strength of interaction. Unfortunately, other molecular parameters of steroidal drugs have never been determined; therefore, it was impossible to include them into the calculation. It can be concluded from the data that steroidal drugs form complexes (probably inclusion complexes) with CM-g-CD. The strength of complex formation markedly depends on the structure of the drug molecule and is significantly influenced by the hydrophobicity of drugs. It is probable that the complex formation of drugs with CM-g-CD modifies the various biological parameters (uptake, transfer, decomposition rate, etc.) and—consequently—the biological efficacy of steroidal drugs in living organisms. REFERENCES
FIG. 4. Relationship between the hydrophobicity (RM0) and specific hydrophobic surface area (b1) of steroidal drugs. (rcalc is the coefficient of correlation indicating the fitness of Eq. [3] to the experimental data in the case of one independent variable, and r99% is the tabulated value of the coefficient of correlation related to 99% significance level).
3. Cserha´ti, T. (1995) Anal. Biochem. 225, 328–332.
AID
aba
AB 9975
/
6m27$$$122
02-12-97 01:00:20
1. Szejtli, J. (1982) Cyclodextrins and Their Inclusion Complexes, Akade`miai Kiado´, Budapest, Hungary. 2. Szejtli, J. (1988) Cyclodextrin Technology, Kluwer Academic, Dordrecht, The Netherlands.
´ TI AND FORGA ´ CS CSERHA
210
4. Matsubara, K., Abe, K., Irie, T., and Uekama, K. (1995) J. Pharm. Sci. 84, 1295–1300. 5. Zhao, Q. Y., Temsamani, J., and Agrawal, S. (1995) Antisense Res. Dev. 5, 185–192. 6. Helm, H., Mu¨ller, B. W., and Waaler, T. (1995) Eur. J. Pharmacol. Sci. 3, 195–202. 7. Vincieri, F. F., Mazzi, G., Mulinacci, N., Bambagiotti-Alberti, M., Dallacqua, F., and Vedaldi, D. (1995) Farmaco 50, 543–548. 8. Hamaura, T., Ohtani, T., Yoda, K., Haruyama, H., Kusai, A., and Nishimura, K. (1995) STP Pharmacol. Sci. 5, 339–346. 9. Ammar, H. O., and El-Nahhas, A. (1995) Pharmazie 50, 408– 409. 10. Habib, M. J., Ghosh, T. K., Akogyeram, C. O., and Ahmadi, B. (1995) Drug Dev. Ind. Pharmacol. 21, 1815–1820. 11. Shiotani, K., Irie, T., Uekama, K., and Ishimaru, Y. (1995) Eur. J. Pharmacol. Sci. 3, 139–152. 12. Suhonen, P., Jarvinen, T., Lehmussaari, K., Reunamari, T., and Urtti, A. (1995) Pharmacol. Res. 12, 529–533. 13. Preiss, A., Mehnert, W., and Fromming, K. H. (1995) Pharmazie 50, 121–125. 14. Loftsson, T., Igurdardottir, A. M., and Olafsson, J. H. (1995) Int. J. Pharmacol. 119, 245–254.
AID
AB 9975
/
6m27$$$122
02-12-97 01:00:20
15. Loftsson, L., Jonsdottir, B., Baldvinsdottir, J., and Fridriksdottir, H. (1994) STP Pharmacol. Sci. 4, 354–358. 16. Irwin, W. J., Dwivedi, A. K., Holbrook, P. A., and Dey, M. J. (1994) Pharmacol. Res. 11, 1698–1703. 17. Castronuovo, G., Elia, V., Fessas, D., Giordano, A., and Velleca, F. (1995) Carbohydr. Res. 272, 31–40. 18. Loukas, Y. L., Vyza, E. A., and Valiraki, A. P. (1995) Analyst 120, 333–338. 19. Forga´cs, E., and Cserha´ti, T. (1994) Quant. Struct.-Act. Relat. 13, 38–42. 20. Sadlej-Sosnowska, N. (1995) J. Pharmacol. Biomed. Anal. 13, 701–704. 21. Sadlej-Sosnowska, N. (1996) J. Chromatogr. A 728, 89–95. 22. Buva´ri, A., Szejtli, J., and Barcza, L. (1983/1984) J. Incl. Phenom. 1, 151–157. 23. Harada, A., and Takahashi, S. (1984) Chem. Lett. 12, 2089– 2090. 24. Horva´th, C., Melander, W., and Molna´r, I. (1976) J. Chromatogr. 125, 129–156. 25. Cserha´ti, T. (1984) Chromatographia 18, 318–322. 26. Valko´, K. (1984) J. Liq. Chromatogr. 7, 1405–1424. 27. Mager, H. (1982) Moderne Regressionsanalyse, pp. 135–167, Salle, Sauerlander, Frankfurt am Main. 28. Valko´, K., and Sle`gel, P. (1993) J. Chromatogr. A 631, 49–61.
aba
246, 205–210 (1997)
AB969975
Charge-Transfer Chromatographic Study of the Complex Formation of Some Steroidal Drugs with Carboxymethyl-g-cyclodextrin Tibor Cserha´ti and Esther Forga´cs Central Research Institute for Chemistry, Hungarian Academy of Sciences, P.O. Box 17, H-1525 Budapest, Hungary
Received September 17, 1996
The interaction between 15 steroidal drugs and carboxymethyl-g-cyclodextrin (CM-g-CD) was studied by reversed-phase charge-transfer thin-layer chromatography and the relative strength of interaction was calculated. CM-g-CD formed inclusion complexes with each compound, the complex always being less hydrophobic than the uncomplexed drug. The inclusion-forming capacity of drugs differed considerably depending on their chemical structures. The linear correlation between the hydrophobicity and specific hydrophobic surface area of anticancer drugs indicated that they can be considered as a homologous series of compounds, although their chemical structures are different. Hydrophobicity of drugs significantly influenced the strength of interaction, indicating the involvement of hydrophobic forces in the binding of drugs to CM-g-CD. The marked influence of CM-g-CD on the hydrophobicity of drugs suggests that this interaction may modify the biological properties (adsorption, uptake, half-life, etc.) of drug–CM-g-CD complexes drug, resulting in modified efficacy. q 1997 Academic Press
Cyclodextrins (CDs)1 are cyclic oligosaccharides which can form inclusion complexes with a wide variety of organic and inorganic compounds (1, 2). The complex formation of CDs with drugs such as antineoplastic agents (3), luteinizing hormone-releasing hormone agonist buserelin (4), antisense oligonucleotides (5), dihydroergotamine (6), psoralen (7), and cefpodoxime proxetil (8) has been recently reported. Both the physicochemical and biological characteristics of the complexed drug may deviate from those of the uncomplexed one resulting in better application
parameters and improved efficacy (9). The formation of inclusion complexes with CDs enhanced the dissolution of indomethacin (10), protected against the nephrotoxicity of gentamycin in rats (11), decreased the ocular irritation of pilocarpine (12), modified the penetration of hydrocortisone into excised human skin (13), improved the delivery of acitrecin (14), and modified the degradation of steroidal drugs in aqueous solution (15) and the stability of bioactive peptides (16), etc. Many physicochemical methods have been used for the study of the formation of inclusion complexes. Thus, calorimetry (17), spectrophotometry (18), and various liquid chromatographic methods such as reversed-phase thin-layer (19) and high-performance liquid chromatography (20, 21) have been successfully used for the assessment of the various aspects of host – guest interactions. The objectives of this work were to study the interaction of some steroidal drugs with carboxymethylg-cyclodextrin (CM-g-CD), to determine the relative strength of interaction with the aid of reversedphase chromatography, to compare their inclusion forming capacity, and to elucidate the role of molecular parameters in the inclusion complex formation. The elucidation of the formation of inclusion complexes between the steroidal drugs and CM-g-CD may promote the development of new, more effective drug formulations with higher biological efficiencies and lower toxic side effects. Steroidal drugs were chosen as model compounds because they are produced in considerable quantity by the pharmaceutical industry and are extensively used in the medical praxis. MATERIALS AND METHODS
1 Abbreviations used: CDs, cyclodextrins; CM-g-CD, carboxymethyl-g-cyclodextrin.
Polygram UV254 (Macherey-Nagel, Du¨rren, Germany) plates were impregnated by overnight predevel205
0003-2697/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
AB 9975
/
6m27$$$121
02-12-97 01:00:20
aba
´ TI AND FORGA ´ CS CSERHA
206
FIG. 1. Chemical structures of steroidal drugs.
opment in n-hexane:paraffin oil, 95:5 (v/v). The chemical structures of steroidal drugs are shown in Fig. 1 (1, hydrocortisone; 2, hydrocortisone acetate; 3, prednisolone; 4, prednisolone 21-acetate; 5, 16a-hydroxyprednisolone 21-acetate; 6, 16a-hydroxyprednisolone; 7, d-14–16-a-hydroxyprednisolone 21-acetate; 8, budesonide; 9, d-14-budenoside; 10, 16a-prednisolone16a,17-acetonide; 11, 9a-fluoro-16a-prednisolone-16a, 17-acetonide; 12, 11-desoxy-16a-prednisolone-16a,17acetonide; 13, 9a-fluoroprednisolone; 14, 9,11-epoxi16a-prednisolone-16a,17-acetonide; 15, 11-desoxy-17hydroxycortisone; 16, cholesterin; 17, cholesterin 3-acetate). Drugs were the gift of Professor Sa´ndor Go¨ro¨g (Gedeon Richter, Ltd., Budapest, Hungary). The drugs were separately dissolved in methanol at a concentration of 3 mg/ml and 2 ml of the solutions was
AID
AB 9975
/
6m27$$$122
02-12-97 01:00:20
plotted on the plates. Water–methanol mixtures were used as eluents, the methanol concentration ranging from 0 to 30 vol% in steps of 5 vol%. Because the objective was to study the complex formation between the solutes and CM-g-CD and not the study of the effect of CM-g-CD on the separation of solutes, they were separately spotted on the plates. In this way the competition between the steroidal drugs for the binding sites of CM-g-CD was excluded. Methanol was chosen as the organic solvent miscible with water because it forms only weak inclusion complexes with cyclodextrins (22, 23). CM-g-CD was the product of Cyclolab Research and Development Laboratory, Budapest, Hungary. CM-g-CD was added to the eluents in the concentration range of 0–15 mg/ml in steps of 2.5 mg/ml. Developments were carried out in sandwich chambers (22 1
aba
CHROMATOGRAPHIC STUDY OF STEROID COMPLEX FORMATION
207
FIG. 2. Videogram of reversed-phase thin-layer chromatographic plates under UV light. (A) Water:methanol (3:2, v/v); (B) water:methanol (3:2, v/v) / 15 mg/ml carboxymethyl-g-cyclodextrin. Steroidal drugs 1–15 appear consecutively from left to right. Steroids 16 and 17 do not move in these eluent systems.
22 1 3 cm) at room temperature, with the distance of development at about 16 cm. After development the plates were dried at 1057C and the spots of steroidal drugs were revealed by their UV spectra and by iodine vapor. Each experiment was run in quadruplicate. Some plates were evaluated by Densitometer CD-60 (Desaga, Heidelberg, Germany) using reflectance mode and 255 nm detection wavelength. Videodocumentation (VD 40, Desaga) was also used in some instances. The RM value characterizing the molecular hydrophobicity in reversed-phase thin-layer chromatography was calculated for each drug in each eluent: RM Å log(1/Rf 0 1).
[1]
When the coefficient of variation of the parallel determinations was higher than 8%, the RM value was omitted from the following calculations. To separate the effects of methanol and CM-g-CD on the hydrophobicity of steroidal drugs the equation was fitted to the experimental data RM Å RM0 / b1rC1 / b2rC2 ,
where RM Å RM value for a drug determined at given methanol and CM-g-CD concentrations; RM0 Å RM value extrapolated to zero methanol and CM-g-CD concentrations; b1 is the decrease in the RM value caused by 1% increase in methanol concentration in the eluent (related to the specific hydrophobic surface area of drugs (24); b2 is the decrease in the RM value caused by 1 mg/ml concentration change of CM-g-CD in the eluent (related to the relative strength of interaction); C1 and C2 are the concentrations of methanol and CMg-CD, respectively. Equation [2] was applied separately for each steroidal drug. To test the validity of the hypothesis that in the case of homologous series of solutes the slope and intercept values (b 1 and RM0 in Eq. [2]) are strongly intercorrelated (25, 26), the linear correlation was calculated between the two physicochemical parameters: RM0 Å A / Brb1 .
FIG. 3. Densitograms of compounds 3 and 15. (A) Water:methanol (3:2, v/v); (B) water:methanol (3:2, v/v) / 15 mg/ml carboxymethylg-cyclodextrin.
AID
AB 9975
/
6m27$$$122
02-12-97 01:00:20
[2]
[3]
To find the physicochemical parameters of steroidal drugs significantly influencing their complexforming capacity, stepwise regression analysis was applied (27). The relative strength of interaction (b2)
aba
´ TI AND FORGA ´ CS CSERHA
208
TABLE 1
Parameters of Linear Correlations between the Hydrophobicity (RM) of Steroidal Drugs and the Methanol (C1 vol%) and Carboxymethyl-g-Cyclodextrin (C2 mg/mL) in the Eluent RM Å RM0 / b1rC1 / b2rC2 No. of steroidal drugs Parameter na RM0 0102xb1 103xsb1b 0102xb2 103xsb2b b=1 (%)c b=2 (%)c r2d Fcalce
1
2
3
4
5
6
7
8
18 2.06 4.22 1.76 2.98 2.93 70.10 29.90 0.9782 337.29
15 3.17 5.65 3.92 5.03 4.88 58.31 41.69 0.9502 114.48
18 2.17 4.41 2.03 3.24 3.36 69.27 30.73 0.9741 282.20
22 2.75 4.76 1.29 4.81 4.43 77.25 22.75 0.9876 754.73
16 2.66 5.54 3.61 3.62 4.77 66.90 33.10 0.9501 123.86
18 2.00 4.91 2.33 2.48 3.86 76.63 23.37 0.9701 243.60
17 2.50 5.03 1.83 3.43 2.68 68.23 31.77 0.9837 422.00
15 2.84 4.11 2.69 2.86 3.35 64.13 35.87 0.9519 118.66
No. of steroidal drugs Parameter na RM0 0102xb1 103xsb1b 0102xb2 103xsb2b b=1 (%)c b=2 (%)c r2 d Fcalce
9
10
11
12
13
14
15
15 2.98 4.70 3.33 2.41 4.15 70.82 29.18 0.9432 99.71
20 3.53 5.48 1.72 5.71 5.82 76.41 23.59 0.9868 636.47
16 2.34 4.01 2.58 4.97 3.41 51.57 48.43 0.9647 177.52
20 3.69 5.53 2.07 7.64 6.99 70.96 29.04 0.9795 406.22
17 2.13 4.17 2.22 5.43 3.25 52.90 47.10 0.9758 281.60
15 2.80 5.10 2.67 3.95 3.32 61.34 38.36 0.9697 191.71
15 2.59 4.34 3.83 5.79 4.77 48.27 51.73 0.9416 96.74
Note. Numbers refer to steroidal drugs in Fig. 1. b1 , decrease in the RM value caused by 1% increase in methanol concentration in the eluent (related to the specific hydrophobic surface area of drugs; b2 , decrease in the RM value caused by 1 mg/ml concentration change of CM-t-CD in the eluent (related to the relative strength of interaction). a Number of data points. b Standard deviations of b1 and b2 . c Standard partial regression coefficients of b1 and b2 , which are normalized to unity. d Coefficient of determination. e Calculated F value indicating the fitness of Eq. [2] to the experimental data.
was the dependent variable, whereas the hydrophobicity (RM0), specific hydrophobic surface area (b1) of Eq. [2], and the complex hydrophobicity parameter R M0/b1 (28) were the independent variables, respectively. The number of accepted independent variables was not limited and the acceptance limit was set to the 95% significance level. In the common multivariate regression analysis the presence of independent variables exerting no significant influence on the change of dependent variable considerably decreases the significance level of the equation. Stepwise regression analysis eliminates from the selected equation the dependent variables having no significant impact on the dependent variable, increasing in this manner the reliability of the calculation. Software of stepwise regression analysis was prepared by CompuDrug Ltd., Budapest, Hungary.
AID
AB 9975
/
6m27$$$122
02-12-97 01:00:20
RESULTS AND DISCUSSION
The simultaneous effect of methanol and CM-g-CD concentrations on the RM values of drugs are shown in Figs. 2 and 3. An increase in the CM-g-CD concentration causes a decrease in RM values, indicating complex (probably inclusion complex) formation. Interaction of the more hydrophilic CM-g-CD with the steroidal drugs decreases the hydrophobicity of the latter. This finding suggests that the biological properties (adsorption, uptake, half-life, etc.) of drug–CM-g-CD complexes may be different from those of an uncomplexed drug resulting in modified effectivity. The data further prove that the addition of CM-g-CD to the eluent does not cause peak distortion or peak widening, lessening the accuracy and reliability of the determination of the relative strength of interaction.
aba
CHROMATOGRAPHIC STUDY OF STEROID COMPLEX FORMATION
The parameters of Eq. [2] are compiled in Table 1. The equation fits the experimental data well, the significance levels in each instance being over 99.9% (see calculated F values). The ratios of variance explained were about 94–99% (see r2 values). Each steroidal drug interacted with CM-g-CD (b2 values differ significantly from zero), meaning that in pharmaceutical formulations containing both steroidal drugs and CM-g-CD their possible interaction must be taken into consideration. The parameters of Eq. [2] show high variations between the drugs, proving that the hydrophobicity (RM0), specific hydrophobic surface area (b1), and their capacity to form inclusion complexes with CM-g-CD (b2) differ considerably. This finding also suggests that the inclusion complex formation may influence differently the biological effect of individual steroidal drugs. The path coefficients (b* values) indicate that the impact of the change of methanol and CM-g-CD concentrations on the reversed-phase mobility of steroidal drugs is commensurable; that is, the retention of steroidal drugs can be equally modified by changing either the methanol or the CM-g-CD concentration in the eluent. Significant linear correlation was found between the intercept (hydrophobicity) and slope (specific hydrophobic surface area) values of steroidal drugs (Fig. 4). This finding indicates that from a chromatographic point of view these drugs behave as a homologous series of compounds, although their chemical structures are different. Significant relationships (significance level over 95%) were found between the hydrophobicity parame-
209
FIG. 5. Relationship between the hydrophobicity (RM0) of steroidal drugs and their capacity to interact with carboxymethyl-g-cyclodextrin (b2). Results of stepwise regression analysis (rcalc is the coefficient of correlation indicating the fitness of the selected equation to the experimental data in the case of one independent variable, and r95% is the tabulated value of the coefficient of correlation related to 95% significance level).
ters of steroidal drugs and their capacity to interact with CM-g-CD (Fig. 5). The fact that hydrophobicity (RM0) of steroidal drugs exerts a significant influence on the strength of interaction indicates that hydrophobic forces are involved in the interaction. This result supports the assumption that steroidal drugs enter into the lipophilic cyclodextrin cavity, and they are retained by hydrophobic forces. However, these hydrophobicity explains only a fairly low ratio of the total variance, indicating that other molecular parameters may account for the strength of interaction. Unfortunately, other molecular parameters of steroidal drugs have never been determined; therefore, it was impossible to include them into the calculation. It can be concluded from the data that steroidal drugs form complexes (probably inclusion complexes) with CM-g-CD. The strength of complex formation markedly depends on the structure of the drug molecule and is significantly influenced by the hydrophobicity of drugs. It is probable that the complex formation of drugs with CM-g-CD modifies the various biological parameters (uptake, transfer, decomposition rate, etc.) and—consequently—the biological efficacy of steroidal drugs in living organisms. REFERENCES
FIG. 4. Relationship between the hydrophobicity (RM0) and specific hydrophobic surface area (b1) of steroidal drugs. (rcalc is the coefficient of correlation indicating the fitness of Eq. [3] to the experimental data in the case of one independent variable, and r99% is the tabulated value of the coefficient of correlation related to 99% significance level).
3. Cserha´ti, T. (1995) Anal. Biochem. 225, 328–332.
AID
aba
AB 9975
/
6m27$$$122
02-12-97 01:00:20
1. Szejtli, J. (1982) Cyclodextrins and Their Inclusion Complexes, Akade`miai Kiado´, Budapest, Hungary. 2. Szejtli, J. (1988) Cyclodextrin Technology, Kluwer Academic, Dordrecht, The Netherlands.
´ TI AND FORGA ´ CS CSERHA
210
4. Matsubara, K., Abe, K., Irie, T., and Uekama, K. (1995) J. Pharm. Sci. 84, 1295–1300. 5. Zhao, Q. Y., Temsamani, J., and Agrawal, S. (1995) Antisense Res. Dev. 5, 185–192. 6. Helm, H., Mu¨ller, B. W., and Waaler, T. (1995) Eur. J. Pharmacol. Sci. 3, 195–202. 7. Vincieri, F. F., Mazzi, G., Mulinacci, N., Bambagiotti-Alberti, M., Dallacqua, F., and Vedaldi, D. (1995) Farmaco 50, 543–548. 8. Hamaura, T., Ohtani, T., Yoda, K., Haruyama, H., Kusai, A., and Nishimura, K. (1995) STP Pharmacol. Sci. 5, 339–346. 9. Ammar, H. O., and El-Nahhas, A. (1995) Pharmazie 50, 408– 409. 10. Habib, M. J., Ghosh, T. K., Akogyeram, C. O., and Ahmadi, B. (1995) Drug Dev. Ind. Pharmacol. 21, 1815–1820. 11. Shiotani, K., Irie, T., Uekama, K., and Ishimaru, Y. (1995) Eur. J. Pharmacol. Sci. 3, 139–152. 12. Suhonen, P., Jarvinen, T., Lehmussaari, K., Reunamari, T., and Urtti, A. (1995) Pharmacol. Res. 12, 529–533. 13. Preiss, A., Mehnert, W., and Fromming, K. H. (1995) Pharmazie 50, 121–125. 14. Loftsson, T., Igurdardottir, A. M., and Olafsson, J. H. (1995) Int. J. Pharmacol. 119, 245–254.
AID
AB 9975
/
6m27$$$122
02-12-97 01:00:20
15. Loftsson, L., Jonsdottir, B., Baldvinsdottir, J., and Fridriksdottir, H. (1994) STP Pharmacol. Sci. 4, 354–358. 16. Irwin, W. J., Dwivedi, A. K., Holbrook, P. A., and Dey, M. J. (1994) Pharmacol. Res. 11, 1698–1703. 17. Castronuovo, G., Elia, V., Fessas, D., Giordano, A., and Velleca, F. (1995) Carbohydr. Res. 272, 31–40. 18. Loukas, Y. L., Vyza, E. A., and Valiraki, A. P. (1995) Analyst 120, 333–338. 19. Forga´cs, E., and Cserha´ti, T. (1994) Quant. Struct.-Act. Relat. 13, 38–42. 20. Sadlej-Sosnowska, N. (1995) J. Pharmacol. Biomed. Anal. 13, 701–704. 21. Sadlej-Sosnowska, N. (1996) J. Chromatogr. A 728, 89–95. 22. Buva´ri, A., Szejtli, J., and Barcza, L. (1983/1984) J. Incl. Phenom. 1, 151–157. 23. Harada, A., and Takahashi, S. (1984) Chem. Lett. 12, 2089– 2090. 24. Horva´th, C., Melander, W., and Molna´r, I. (1976) J. Chromatogr. 125, 129–156. 25. Cserha´ti, T. (1984) Chromatographia 18, 318–322. 26. Valko´, K. (1984) J. Liq. Chromatogr. 7, 1405–1424. 27. Mager, H. (1982) Moderne Regressionsanalyse, pp. 135–167, Salle, Sauerlander, Frankfurt am Main. 28. Valko´, K., and Sle`gel, P. (1993) J. Chromatogr. A 631, 49–61.
aba