Displacement and Nonlinear Chromatographic Techniques in the

The ka values determined in this study ranged from 19.7 to 10.5. µM-1sec-1, with a relative order of mecamylamine > dextromethorphan g ketamine > bup...
0 downloads 0 Views 110KB Size
Download PDF
Anal. Chem. 2002, 74, 4618-4624

Displacement and Nonlinear Chromatographic Techniques in the Investigation of Interaction of Noncompetitive Inhibitors with an Immobilized r3β4 Nicotinic Acetylcholine Receptor Liquid Chromatographic Stationary Phase Krzysztof Jozwiak,†,‡ Jun Haginaka,†,§ Ruin Moaddel,† and Irving W. Wainer*,†

Gerontology Research Center, National Institute on Aging, National Institutes of Health, Baltimore, Maryland, Department of Inorganic and Analytical Chemistry, Medical University of Lublin, Lublin, Poland, and Mukogawa Women’s University, Hyogo, Japan

A liquid chromatographic column containing immobilized r3β4 nicotinic acetylcholine receptors (r3β4-nAChRs) has been used to determine the equilibrium association constants (Ka), desorption rate constants (kd), and adsorption rate constants (ka) for the noncompetitive inhibitors: mecamylamine, ketamine, bupropion, and dextromethorphan. Displacement chromatography, with mecamylamine as the displacer, was used to verify that the four compounds bound to the same site on the immobilized r3β4-nAChRs. Nonlinear chromatographic techniques were then utilized to calculate the Ka, ka, and kd values associated with the formation of the noncompetitive inhibitor-r3β4-nAChR complexes. The ka values determined in this study ranged from 19.7 to 10.5 µM-1sec-1, with a relative order of mecamylamine > dextromethorphan g ketamine > bupropion. The kd values determined in this study indicated that dextromethorphan-induced inhibition should produce a longer recovery time than the other three NCIs. This was consistent with results from a previous in vitro study. The data from this study indicate that the immobilized r3β4nAChR column and nonlinear chromatography can be used in the study of NCIs at the r3β4-nAChR. The ideal chromatographic peak shape is Gaussian, but actual peak profiles in liquid chromatography are often asymmetrical. The observed asymmetry can arise from a variety of sources, including extracolumn effects, heterogeneity of the stationary phase, heterogeneous mass transfer, or a nonlinear isotherm.1 Although peak tailing (or fronting) is a problem in analytical separations, concentration-dependent asymmetry can be used with nonlinear chromatography (NLC) techniques to characterize the separation processes occurring on the column. * Corresponding author address: Gerontology Research Center, National Institute on Aging, NIH, 5600 Nathan Shock Drive, Baltimore, MD 21224-6825. Phone: (410) 558-8498. Fax: (410) 558-8409. E-mail: [email protected]. † National Institutes of Health. ‡ Medical University of Lublin. § Mukogawa Women’s University. (1) Fornstedt, T.; Zhong, G.; Guiochon, G. J. Chromatogr. A 1996, 742, 5568.

4618 Analytical Chemistry, Vol. 74, No. 18, September 15, 2002

NLC theory applies primarily to zonal affinity chromatography. In this approach, the injections of finite concentrations of a ligand produce asymmetric peak profiles with tailing proportional to the ligand concentrations. The model assumes a limited number of active sites on the column and a nonlinear isotherm. Moreover, interphase solute transfer cannot be considered as infinitely fast, becaue the kinetic rates of adsorption and desorption are assumed to be the primary source of band broadening and peak skew. The first formulation of nonlinear conditions in chromatography was derived by Thomas.2 Since then, several different approaches have been introduced for both frontal and zonal chromatography, as well as methods for the numerical integration of the chromatographic differential mass balance equation.3,4,5 In 1987 Wade et al.6 developed the impulse input solution for the mass balance equation. This approach is based upon the observation that when adsorption/desorption rates are slow, band broadening is insensitive to a moderate degree of column overload. In contrast to numerical integration methods, this approach uses the analytical solution, which can be applied directly to fit experimental peak profiles. The impulse input equation has been included in commercially available deconvolution software and can be easily applied to NLC studies. The nicotinic acetylcholine receptor (nAChR) is a family of ligand-gated ion channels that contain two cholinergic agonist binding sites.7 These sites are key targets for drug discovery in a variety of diseases, including Alzheimer’s, Parkinson’s, epilepsy, and Tourette’s Syndrome.8,9 The nAChR also contains two other binding sites at which noncompetitive inhibitors (NCIs) bind.9 One site is a high-affinity site located in the channel lumen within the pore. At this site, NCIs either inhibit nAChR function by reversible (2) Thomas, H. C. J. Am. Chem. Soc. 1944, 66, 1664. (3) Jaulmes, A.; Vidal-Madjar, C.; Colin, H.; Guiochon, G. J. Phys. Chem. 1986, 90, 207-215. (4) Vidal-Madjar, C.; Jaulmes, A.; Racine, M.; Sebille, B. J. Chromatogr. 1988, 458, 13-25. (5) Felinger, A.; Guiochon, G. Trends Anal. Chem. 1995, 14, 6-10. (6) Wade, J. L.; Bergold, A. F.; Carr, P. W. Anal. Chem. 1987, 59, 1286-1295. (7) Hucho, F.; Tsetlin, V. I.; Machold, J. Eur. J. Biochem. 1996, 239, 539-55. (8) Holladay, M. W.; Dart, M. J.; Lynch, J. K. J. Med. Chem. 1997, 40, 41694194. (9) Lloyd, K. G.; Williams, M. J. Pharmacol. Exp. Ther. 1999, 292, 461-467. 10.1021/ac0202029 Not subject to U.S. Copyright. Publ. 2002 Am. Chem. Soc.

Published on Web 08/20/2002

channel blockade or shorten the channel opening time in a voltagesensitive manner.10 The NCIs that bind to this site include mecamylamine,9,11 ketamine,12 bupropion,13 and dextromethorphan.14 A second set of low-affinity sites is located at the interface of the lipid membrane and the nAChR protein.9 NCIs acting at this site include quinacrine, estrogens, and xenoestrogens.15 NCI activity is experimentally determined by measuring concentration-dependent effects on whole-cell currents or 86Rb+ efflux, yielding IC50 values.12,14,16,17 These studies are time-consuming and exacting. In addition, the results can differ from laboratory to laboratory and from cell line to cell line. Previous studies with a liquid chromatographic column containing an immobilized R3β4-nAChR stationary phase have shown that this phase can be used to determine the binding affinities of competitive agonists and antagonists to the R3β4-nAChR. In the current study, this stationary phase was used to study the interaction of the NCIs with the immobilized R3β4-nAChR at the high-affinity site. In this study, the NCIs utilized were mecamylamine, ketamine, bupropion, and dextromethorphan, and displacement chromatographic techniques were initially used to verify that these compounds bound at the same site on the receptor. The NLC technique was then used to determine the relationship between peak deterioration and applied concentration6 and to calculate the equilibrium constant for adsorption (Ka), desorption rate constant (kd), and adsorption rate constant (ka) for each NCI-nAChR interaction. The data from this study indicate that the immobilized R3β4-nAChR column and nonlinear chromatography can be used in the study of NCIs at the R3β4-nAChR. EXPERIMENTAL SECTION Materials. (()-Epibatidine dihydrochloride, (()-mecamylamine hydrochloride, (()-ketamine hydrochloride, (()-bupropion hydrochloride, (S)-nicotine, and dextromethorphan hydrobromide monohydrate were purchased from Sigma Chemical (St. Louis, MO). HPLC grade methanol, ammonium acetate and 0.1 M ammonium hydroxide solution were purchased from Fisher Scientific (Pittsburgh, PA). The R3β4-nAChR column was prepared as previously described.18,19 Chromatographic System. Displacement Studies. The HPLC system used in the displacement studies was composed of a P2000 pump, a UV6000LP spectrophotometer, a FL3000 fluorometer, and an AS3000 autoinjector (all from ThermoQuest, San Jose, CA). All data were recorded and integrated using ChromQuest (ThermoQuest). The mobile phase was composed of 10 mM ammonium (10) Lena, C.; Changeux J.-P. Trends. Neurosci. 1993, 16, 181-186. (11) Xiao, Y. X.; Smith, R. D.; Caruso, F. S.; Kellar, K. J. J. Pharmacol. Exp. Ther. 2001, 299, 366-371. (12) Yamakura, T.; Chavez-Noriega, L. E.; Harris, R. A. Anesthesiology 2000, 92, 1144-1153. (13) Fryer, J. D.; Lukas, R. J. J. Pharmacol. Exp. Ther. 1999, 288, 88-92. (14) Hernandez, S. C.; Bertolino, M.; Xiao, Y. X.; Pringle, K. E.; Caruso, F. S.; Kellar, K. J. J. Pharmacol. Exp. Ther. 2000, 293, 962-967. (15) Nakazawa, K.; Ohno, Y. Eur. J. Pharmacol. 2001, 430, 175-183. (16) Chiodini, F.; Charpantier, E.; Muller, D.; Tassonyi, E.; Fuchs-Buder, T.; Bertrand, D. Anesthesiology 2001, 94, 643-651. (17) Furuya, R.; Oka, K.; Watanabe, I.; Kamiya, Y.; Itoh, H.; Andoh, T. Anesth. Analg. 1999, 88, 174-180. (18) Zhang, Y. X.; Xiao, Y. X.; Kellar, K. J.; Wainer, I. W. Anal. Biochem. 1998, 264, 22-25. (19) Wainer, I. W.; Zhang, Y. X.; Xiao, Y. X.; Kellar, K. J. J. Chromatogr. B 1999, 724, 65-72.

acetate adjusted to pH 7.4 with 0.1 M ammonium hydroxide. The buffer was modified with methanol at a 95:5 ratio of buffer/ methanol (v/v). The mobile phase was delivered at a flow rate of 0.2 mL/min. Epibatidine, ketamine, and bupropion were measured using UV detection at λ ) 220 nm, and dextromethorphan was measured using fluorescence detection with excitation and emission wavelengths of λ ) 276 and 310 nm, respectively. Nonlinear Studies. The nonlinear studies were conducted using a LC-MS system composed of a LC10AD pump (Shimadzu, Columbia, MD), ESA 540 autoinjector (ESA, Inc, Chelmsford, MA) and Micromass Q TOF mass spectrometer (Micromass, Beverly, MA). The data were recorded and processed using MassLynx v. 3.5 (Micromass). The mobile phase was composed of 10 mM ammonium acetate adjusted to pH 7.4 with 0.1 M ammonium hydroxide modified with methanol at a ratio of 9:1 (v/v) buffer/ methanol for bupropion and dextromethorphan and at a ratio of 95:5 (v/v) for ketamine and mecamylamine. The flow rate of mobile phase was 0.2 mL/min. Mecamylamine (Mw ) 167.29), ketamine (Mw ) 237.72), bupropion (Mw ) 239.74), and dextromethorphan (Mw ) 271.39) were analyzed in the positive ion mode (ESI+). The compounds were detected using single-ion monitoring of the corresponding [M + H]+ ions. The nebulization gas flow rate and desolvation gas flow rate were 45 L/h and 250 L/h, respectively. The capillary voltage was 2.5 kV for dextromethorphan and 2.7 kV for the other compounds; the cone voltage was 42 V for dextromethorphan and 27 V for the other compounds. The temperature of the source block was 120 °C, and the desolvation temperature was 350 °C. Chromatographic Procedures. Displacement Studies. All of the chromatographic experiments were carried out in duplicate. The initial retention times of the marker ligands ketamine, bupropion, and dextromethorphan were determined using the mobile phase composed of ammonium acetate (10 mM, pH 7.4)/ methanol (95:5, v/v). After the initial studies, the displacer ligand, mecamylamine, was added to the mobile phases in increasing concentrations of 10, 20, 40, 100, 150, and 200 µM. The column was equilibrated before use by passing the mobile phase through the column until a steady baseline was achieved at λ ) 220 nm. After equilibration of the column (approximately 1 h), 50-µL aliquots of 10 µM aqueous solutions of ketamine, bupropion, or dextromethrophan were independently injected, and their retention times determined. At the end of the series of mecamylamine concentrations, the initial mobile phase (without added mecamylamine) was passed through the column for 24 h. The marker ligands were then independently injected, and their retention times were determined. A 100 µM concentration of mecamylamine was then added to the mobile phase. After equilibration of the column (approximately 1 h), 50-µL aliquots of 10 µM aqueous solutions of ketamine, bupropion, and epibatidine were independently injected, and their retention times were determined. At the end of this series of displacement experiments, the initial mobile phase was passed through the column for 24 h. The marker ligands were then independently injected, and their retention times were determined. A 40 µM concentration of (S)-nicotine was then added to the mobile phase. After equilibration of the column (approximately 1 h), 50-µL aliquots of 10 µM aqueous solutions of ketamine, bupropion, and epibatidine were independently injected, and their Analytical Chemistry, Vol. 74, No. 18, September 15, 2002

4619

retention times were determined. At the end of this series of displacement experiments, the initial mobile phase was passed through the column for 24 h. The marker ligands were then independently injected, and their retention times were determined. Nonlinear Studies. The column was equilibrated with the mobile phase, ammonium acetate (10 mM, pH 7.4)/methanol (either 95:5 or 90:10, v/v), for 2 h. After equilibration, 50-µL aliquots of aqueous solutions of one of the experimental ligands, mecamylamine, ketamine, bupropion or dextromethorphan, were independently injected into the column, and the elution profiles of the ligands were monitored by MS detection. A series of aqueous concentrations of each ligand was sequentially injected onto the chromatographic system. The concentrations used in this study were: 1, 2, 5, 10, 20, 50, 100, 200, 500, and 1000 µM. The only exception was bupropion, for which the 2 µM concentration was not studied. At the end of each series, the column was flushed with the mobile phase for 2 h. Data Analysis. Displacement Studies. The retention factor (k′) of each marker ligand was calculated using eq 1, where tR and t0 are retention times of retained and unretained compounds, respectively. The retention time of the unretained compound, t0, was determined from the negative peak corresponding to the retention time of the water, which served as the diluent for the standards.

k′ ) (tR - t0)/t0

(1)

The relationship between the k′ of a marker ligand and the mobile phase concentration of the displacer was expressed using eq 2.20,21

1/(k′ - X) ) VMK2[D]/K3mL + VM/K3mL

(2)

where VM is the void volume of the column; K2 and K3 are equilibrium constants for the binding of the displacer and marker ligand, respectively; mL represents moles of marker ligand bound to the stationary phase; and X is the residual k′ resulting from binding at sites unaffected by the displacer. Nonlinear Studies. The data from the nonlinear studies was analyzed using PeakFit v4.11 for Windows software (SPSS Inc., Chicago, IL). The mathematical approach was the nonlinear chromatography (NLC) model derived from impulse input solution and described by eq 3 (PeakFit User’s Manual, pp 8-25)

y)

[

( )]

a0 a3 1 - exp a3 a2

[

x

( ) (

( )[

]

) ( )]

a1 2xa1x -x - a1 I1 exp x a2 a2 a3 a1 x 1-T , 1 - exp a2 a2 a2

(3)

where y ) intensity of signal, x ) reduced retention time, and (20) Kaliszan, R.; Wainer, I. W. Combination of Biochromatography and Chemometrics: A Potential New Research Strategy in Molecular Pharmacology and Drug Design in Chromatographic Separations based on Molecular Recognition; Jinno, K., Ed.; Wiley-VCH: New York, 1997. (21) Hage, D. S.; Austin, J. J. Chromotogr. B 2000, 739, 39-54.

4620

Analytical Chemistry, Vol. 74, No. 18, September 15, 2002

T(u, v) ) exp(-v)

∫ exp(-t)I (2 xvt)dt 0

u

0

The T function acts as a “switching” function, which produces the skew in the peak profile when the column is overloaded. I0() and I1() are modified Bessel functions, a0 ) area parameter, a1 ) center parameter, reveal to true thermodynamic capacity factor, a2 ) width parameter, and a3 ) distortion parameter. Obtained chromatograms (for each compound at each concentration) were smoothed with MassLynx software (Micromass) and then extracted to an Excel worksheet as a set of two columns, retention time and signal intensity. The input data was further processed using the NLC function (Peakfit software) for which Pearson VII Limit Minimization was chosen as the minimization method. The minimization procedure was repeated at least three times in order to find the global minimum. The set of NLC parameters (a0, a1, a2, and a3) was collected for each profile. The NLC parameters were further processed for the calculation of the k′, Ka, ka, and kd values. These parameters were calculated using the following relationships (PeakFit User’s Manual, pp 8-26): k′ ) a1, the real thermodynamic capacity factor; kd ) 1/a2t0, the solute desorption constant rate (t0 is the dead time of a column); Ka ) a3/C0, the equilibrium constant for adsorption (C0 is the concentration of the solute injected multiplied by the width of the injection pulse (as a fraction of column dead volume)). ka ) kdKa, the solute adsorption constant rate. RESULTS AND DISCUSSION Displacement Studies. Chromatographic retention (k′) is the sum of specific and nonspecific interactions between the solute and the stationary phase. When the stationary phase contains an immobilized receptor, such as the nAChR, the specific interactions occur at the binding site or sites on the receptor. If the binding site(s) have been characterized, then specific interactions can be identified and characterized by competitive displacement experiments. Nonspecific interactions also occur with the immobilized protein and the chromatographic backbone, and these can be differentiated from specific interactions by the displacement techniques. The nAChR contains a well-characterized competitive binding site for agonists and antagonists.7,22 Previous studies with liquid chromatographic columns containing immobilized R3β4-nAChR or R4β2-nAChR have probed this site using the competitive agonist [3H]epibatidine as the marker ligand and nicotine, carbachol, A85380, and atropine as the competitive displacers.18,19 The results demonstrated that the marker and displacer ligands bound at the same site on the receptor and that the approach could be used to determine the binding affinities of competitive agonists and antagonists. In this study, binding to the high affinity NCI site within the channel lumen was examined using mecamylamine as the displacer ligand and ketamine, bupropion and dextromethorphan as (22) Cheng Y.-C.; Prusoff W. H. Biochem. Pharm. 1973, 22, 3099-3108.

Table 1. Chromatographic Retentions, Expressed as Retention Factor k′, of Bupropion, Ketamine, and Dextromethorphan as a Function of the Mobile Phase Concentrations of Mecamylamine concn mecamylamine (µM) 0 10 20 40 100 150 200 R2 (1/k′) vs C

bupropion 24.8 22.1 21.3 19.4 16.1 15.9 14.0 0.949

retention factor (k′) ketamine dextromethorphan 9.4 8.5 8.1 7.5 6.4 6.7 6.3 0.818

171.1 140.0 135.7 114.5 96.7 87.1 70.5 0.970

the marker ligands. Mecamylamine was used as the displacer because it is an extensively studied NCI known to bind primarily to the high affinity NCI site.9,11 Mecamylamine is also transparent under the UV and fluorescence conditions employed in the study, and concentrations up to 200 µM of the substance could be added to the mobile phase without affecting detection of the marker ligands. The average (n ) 2) retention factors (k’s) produced by the serial concentrations of mecamylamine in the mobile phase are presented in Table 1. The k′ values were calculated using eq 1. In these calculations, the retention time of water was determined for each injection, and that value was used as the marker for the unretained solute (t0). For each marker ligand, the addition of mecamylamine to the mobile phase reduced the retention time (tR). The relationship between chromatographic retention of the marker ligands and the concentration of mecamylamine in the mobile phase was determined by plotting 1/k′ versus mecamylamine concentration, eq 2. All correlations produced linear relationships, with R2 values ranging from 0.833 to 0.970 (Table 1). The X value (from eq 2) for each of these calculations was 0, indicating that the marker and displacer compete for the same binding site, and there was no residual specific binding, that is, single site binding. Previous whole cell inhibition studies with mecamylamine,14 ketamine,12 bupropion,13 and dextromethorphan14 have demonstrated that these compounds primarily bind at the high-affinity NCI site on the nAChR. In this study, linear relationships existed between the 1/k’s of the marker ligands and the mobile phase concentrations of mecamylamine (Table 1). The results are consistent with the binding of all four ligands at the same site on the R3β4-nAChR, the channel lumen within the pore. Thus, for the NCIs used in this study, the interaction of these compounds with the immobilized R3β4-nAChR reflects the same process observed in the whole-cell studies. The specificity of the competitive displacements was demonstrated by the effects of mecamylamine, a NCI, and (S)-nicotine, a competitive agonist, on the k’s of epibatidine (competitive agonist), ketamine (NCI), and bupropion (NCI). With the standard mobile phase, the k’s for epibatadine, ketamine, and bupropion were 10.8, 9.4, and 24.8, respectively. When 100 µM mecamylamine was added to the mobile phase, there was no significant effect on the k′ of epibatidine (10.6), but the k’s of ketamine and

bupropion were reduced by 32 and 35% to 6.4 and 16.1, respectively. When the mobile phase contained 40 µM (S)-nicotine, there were no significant changes in the k’s of ketamine and bupropion, but the k′ of epibatidine was reduced by 11% to 9.6. Thus, the NCI (mecamylamine) displaced the other NCIs (bupropion, ketamine) but not the agonist (epibatidine), whereas the agonist ((S)-nicotine) displaced another antagonist but not NCIs. Nonlinear Studies. The shape of a chromatographic peak is also a function of the specific and nonspecific interactions between the solute and the stationary phase. In particular, the kinetics involved in the formation and dissolution of the solute-stationary phase complex, that is, the association and dissociation rate constants, ka and kd, respectively. When the stationary phase contains an immobilized protein, the dissociation of a ligandprotein complex is usually slower than the rate of complex formation producing non-Gaussian peaks with tailing. The degree of deviation from a Gaussian distribution is a function of applied ligand concentration and the kinetics of ligandreceptor interactions occurring during the chromatographic process. Carr et al. developed the relationship between on-column ligand concentration and peak shape and defined it in the context of nonlinear chromatography (NLC).6 The NLC model assumes that (1) the stationary phase contains a limited number of immobilized active sites, (2) kinetic rates of adsorption and desorption are the primary source of band broadening and peak skew, and (3) dispersive and extracolumn effects are negligible. The effect of increasing ligand concentrations on the chromatographic profiles of the compounds used in this study is illustrated using mecamylamine, Figure 1. Each profile was smoothed and analyzed using eq 3, and the resulting NLC parameters, a0, a1, a2, and a3, were determined for each concentration of ketamine, bupropion, mecamylamine, and dextromethorphan. Table 2 presents the parameters (followed by standard errors of their determination) calculated for each concentration of ketamine (the tables with the full set of parameters for bupropion, mecamylamine and dextromethorphan can be found in Supporting Information). The relationships between the concentrations of injected NCIs and the calculated NLC parameters (a1, a2, and a3) were determined. According to the NLC approach, the parameters a1 and a2 should be independent of concentration, and this was observed between NCI concentrations of 1-20 µM. For NCI concentrations between 50 and 1000 µM, there is a significant discontinuity between the calculated a1 and a2 values and those determined for the 1-20 µM concentrations (Table 2). In addition, above 50 µM, the standard errors of all of the calculated NLC parameters (a0, a1, a2, a3) were relatively increased. In contrast to a1 and a2, a3 should have a linear relationship with the NCI concentration. This was observed at NCI concentrations between 1 and 20 µM (Table 2). However, when the concentrations were >50 µM, the effect of NCI concentration on a3 appeared to reach saturation and plateaued. The effect of concentration on the NLC parameters is consistent with the increasing bimodality of the peak shapes observed with the 50-1000 µM concentrations (Figure 1). The relationships between NCI concentration and the a1, a2, and a3 parameters and the bimodality of the peak profiles suggest that at NCI concentrations above 50 µM, all high-affinity (luminal) NCI active sites on Analytical Chemistry, Vol. 74, No. 18, September 15, 2002

4621

Figure 1. Effect of increasing concentrations of mecamylamine, from 1 to 1000 µM, on the chromatographic profiles of mecamylamine. Table 2. NLC Parameters Calculated for Each Peak Profile of Ketamine conc [µM]

a0

st. err. a0

a1

st. err. a1

a2

st. err. a2

a3

st. err. a3

r2

F

1 2 5 10 20 50 100 200 500 1000

2420 5274 12400 23400 38978 73479 107700 137120 200780 261380

46.5 54.5 214 838 1810 4586 11716 14154 21069 202720

10.80 11.59 11.40 12.56 13.9 18.5 19.3 17.7 18.2 20.8

0.18 0.13 0.23 0.43 0.5 1.0 3.7 2.8 2.9 29

0.1291 0.1298 0.1566 0.1883 0.242 0.32 0.35 0.34 0.31 0.35

0.0062 0.0038 0.0050 0.0099 0.017 0.02 0.04 0.05 0.04 0.19

0.723 1.031 1.061 1.57 2.24 4.53 5.2 4.7 5.3 6.8

0.046 0.040 0.078 0.16 0.22 0.53 1.9 1.5 1.6 17

0.970 0.992 0.981 0.964 0.963 0.981 0.967 0.939 0.943 0.973

5401 21374 10999 11620 11064 22256 12609 6545 7076 15551

the column are saturated. At this point, binding to low-affinity (nonluminal) NCI active sites on the nAChR begins to affect the NLC profiles. Since the impulse input solution model assumes only one type of active site and one type of ligand-receptor interaction, the NLC model cannot be correctly applied at concentrations greater than 50 µM. Therefore, to remain below column saturation, the values for a1, a2, and a3 corresponding to the NCI concentrations of 1-20 µM were used to calculate the Ka, kd, and ka values (as presented in the Data Analysis Section), and the results were averaged (Table 3). The experimentally observed peak profiles produced by 10 µM injections of mecamylamine, ketamine, bupropion, and dextromethorphan are presented in Figure 2A-D, respectively. The corresponding fitted function shapes are also presented in Figure 2. The NLC parameters determined in this study describe the chromatographic interaction between the immobilized protein and the ligand.6 It is of interest to determine the pharmacological relevance of the chromatographic data. In previous studies utilizing the immobilized R3β4-nAChR stationary phase and competitive agonists and antagonists, the results demonstrated that the chromatographically obtained Kd values correlate with those obtained by standard membrane binding studies.18,19 Thus, the NLC-derived Ka values may be a relative probe of the interaction of NCIs with the R3β4-nAChR. 4622 Analytical Chemistry, Vol. 74, No. 18, September 15, 2002

Table 3. Nonlinear Chromatographic Parameters for the Interaction of Mecamylamine, Ketamine, Bupropion, and Dextromethorphan with the Immobilized r3β4-nAChR Stationary Phase, Expressed as Retention Factor (k ′(NLC)), Equilibrium Constants for Adsorption (Ka), and Association and Dissociation Rate Constants (ka and kd, Respectively) compd mecamylamine ketamine bupropion dextromethorphan

Ka ka kd k′(NLC) (µM-1) (µM-1 sec-1) (sec-1) log Pa 17.0 12.0 14.0 82.0

4.2 2.3 1.8 20.1

19.7 15.0 10.5 16.7

4.8 6.9 5.8 0.8

2.20 3.17 2.75 3.35

a Calculated using HyperChem Molecular Modeling System (HyperCube Inc, Gainseville, FL).

However, to our knowledge, for the NCIs used in this study, there are no membrane binding-affinity studies to which the chromatographically derived affinities, Ka (Table 3) can be compared. This is a result of the experimental procedures used to identify and characterize NCIs at the R3β4-nAChR, which produce IC50 values, not Kd or Ka (Kd ) 1/Ka) values. In some cases, Kd values can be calculated from IC50 values using the Cheng-Prusoff relationship.22 This approach requires the use of

Figure 2. Experimental peak profiles overlaid with fitted NLC functions (1) and fitted NLC functions alone following 50-µL injections of 10 µM solutions of mecamylamine (A), ketamine (B), bupropion (C), and dextromethorphan (D).

competitive agonists whose concentration and EC50 values are known and has been used to calculate Kd values for competitive antagonists.22 These data were not available for the NCIs utilized in this study, and the significance of the chromatographically calculated Ka values could not be evaluated. The reported IC50 values at the R3β4-nAChR for mecamylamine,14 bupropion,12 ketamine,13 and dextromethorphan14 are 1.0, 1.4, 9.5, and 8.9 µM, respectively. The relative order, from highest activity to lowest, is mecamylamine > bupropion > dextromethorphan > ketamine. Previous liquid chromatographic studies with the immobilized R3β4-nAChR stationary phase demonstrated that for competitive agonists and antagonists, relative chromatographic retention can be correlated to affinity; i.e., the lower the affinity, the lower the retention.18,19,23 In this study, the relative order of the chromatographic retentions, expressed as k′(NLC) (Table 3) was ketamine > bupropion > mecamylamine > dextromethorphan. The relative order of ketamine, bupropion, and mecamylamine is consistent with the pharmacological data, but dextromethorphan is clearly an outlier. Since hydrophobicity is often associated with chromatographic retention in a reversed-phase system, there is the possibility that this property plays a role in the failure to correlate the chromatographic and pharmacological data. In this study, when the NCI concentrations were 10 µM, the chromatographic retention of dextromethorphan (k′(NLC) ) 82.0) was ∼5 times greater than that of mecamylamine, (k′(NLC) ) 17.0), whereas the hydropho(23) Baynham, M. T.; Patel, S.; Moaddel, R.; Wainer, I. W. J. Chromatogr. B 2002, 772, 155-161.

bicity (represented as log P value) for dextromethorphan is greater than that for mecamylamine (3.35 versus 2.2) (Table 3). Thus, it appears that the difference in chromatographic retention may, indeed, be a function of hydrophobicity rather than affinity. However, the log P value for ketamine is 3.17, but its chromatographic retention (k′(NLC) ) 12.0) was lower than mecamylamine’s. The data suggests that the 5-7-fold differences in the retention factors of dextromethorphan relative to mecamylamine, ketamine, and bupropion are a function of the specific interaction of these compounds with the R3β4-nAChR. This is supported by the kd values (Table 3), which indicate that the source of dextromethorphan’s significantly greater chromatographic retention is the slower dissociation of the dextromethorphan-nAChR complex relative to the nAChR complexes formed by the other 3 NCIs. A key problem in the direct comparison of the chromatographic data determined in this study and the functional data derived from the literature is the fact that binding affinities do not necessarily directly translate into pharmacological activities. To obtain a valid comparison between chromatographic and functional data, it is necessary to reduce as many variables as possible. One step is to use the same cell line in both the activity studies and the preparation of the affinity column. Another approach is to control the relative magnitudes of the IC50 values by using the same cell lines and the same experimental approaches for all of the test compounds. In this study, only the IC50 values for mecamylamine and dextromethorphan were measured in the same cell line used to prepare the affinity column, Analytical Chemistry, Vol. 74, No. 18, September 15, 2002

4623

the KXR3β4R2 cell line. In addition, both IC50 values were determined using the same experimental approach, the measurement of nicotine-stimulated 86Rb+ efflux.14 For mecamylamine and dextromethorphan, neither the k′ nor Ka values correlate to the relative IC50 values. However, k′ and Ka are functions of the rate constants, ka and kd, and these values may be better indicators of pharmacological activity. Since ka is a reflection of the bimolecular interaction between the NCI and the immobilized nAChR, it may be the way to compare chromatotographic and IC50 values. For mecamylamine and dextromethorphan, the relative order of the ka and IC50 values were consistent. The kd values reflect the first-order rate constants for the dissolution of the NCI-nAChR complexes. Therefore, to a first approximation, the kd value can be associated with the duration of the inhibitory effect. Thus, one would not assume that the kd values would correlate with the IC50 values, and such a correlation was not observed. However, the kd values determined in this study indicate that dextromethorphan induced inhibition should produce a longer recovery time than the other three NCIs. The only comparable study utilized mecamylamine and dextromethorphan in the KXR3β4R2 cell line.14 In the cell-based study, the recovery period for receptor blockade by dextromethorphan was significantly longer than that for mecamylamine. The authors suggested that dextromethorphan bound to a site deep within the receptor channel and, even though the IC50 of dextromethorphan may be higher than that of mecamylamine, it does not dissociate from the channel as easily.14 The chromatographic data from the present study is consistent with this hypothesis. The calculated kd value for dextromethorphan is sixfold slower than that of mecamylamine (Table 3) indicating a significant difference in the dissociation rates of the two NCI-nAChR complexes. CONCLUSIONS The results of this study are the first to demonstrate that an immobilized nAChR stationary phase can be used to examine the

4624

Analytical Chemistry, Vol. 74, No. 18, September 15, 2002

binding of compounds to the high-affinity NCI-binding site on the immobilized receptor. The data have been used to establish NLC parameters for the specific interactions between the NCIs used in this study and the immobilized R3β4-nAChR. These parameters can now be applied to the calculation of the Ka, ka, and kd values for other compounds using only a single chromatographic experiment. The results have also indicated that the chromatographic data may be useful in the determination of functional properties, such as IC50 values. However, the validation of this application requires the chromatographic analysis of a larger cassette of compounds and comparison of the derived data with IC50 values obtained using the KXR3β4R2 cell line. These experiments are currently being pursued, and the results will be presented elsewhere. ACKNOWLEDGMENT The authors would like to thank Peter Carr for his assistance in the development and analysis of the nonlinear chromatography experiments and Ken Kellar for his knowledge and guidance regarding noncompetitive inhibitors at the nAChR. SUPPORTING INFORMATION AVAILABLE A listing of NLC parameters (accompanied by standard errors of their estimation) for each peak profile of ketamine, bupropion, mecamylamine, and dextromethorphan is available as Supporting Information. This material is available free of charge via the Internet at http://pubs.acs.org.

Received for review March 28, 2002. Accepted July 10, 2002. AC0202029