Abstract
The current study demonstrates thatN-n-alkylnicotinium analogs with increasing n-alkyl chain lengths from 1 to 12 carbons have varying affinity (Ki = 90 nM–20 μM) for S-(−)-[3H]nicotine binding sites in rat striatal membranes. A linear relationship was observed such that increasing n-alkyl chain length provided increased affinity for the α4β2* nicotinic acetylcholine receptor (nAChR) subtype, with the exception ofN-n-octylnicotinium iodide (NONI). The most potent analog wasN-n-decylnicotinium iodide (NDNI;Ki = 90 nM). In contrast, none of the analogs in this series exhibited high affinity for the [3H]methyllycaconitine binding site, thus indicating low affinity for the α7* nAChR. The C8 analog, NONI, had low affinity forS-(−)-[3H]nicotine binding sites but was a potent inhibitor of S-(−)-nicotine-evoked [3H]dopamine (DA) overflow from superfused striatal slices (IC50 = 0.62 μM), thereby demonstrating selectivity for the nAChR subtype mediatingS-(−)-nicotine-evoked [3H]DA overflow (α3α6β2* nAChRs). Importantly, theN-n-alkylnicotinium analog with highest affinity for the α4β2* subtype, NDNI, lacked the ability to inhibitS-(−)-nicotine-evoked [3H]DA overflow and, thus, appears to be selective for α4β2* nAChRs. Furthermore, the present study demonstrates that the interaction of these analogs with the α4β2* subtype is via a competitive mechanism. Thus, selectivity for the α4β2* subtype combined with competitive interaction with the S-(−)-nicotine binding site indicates that NDNI is an excellent candidate for studying the structural topography of α4β2* agonist recognition binding sites, for identifying the antagonist pharmacophore on the α4β2* nAChR, and for defining the role of this subtype in physiological function and pathological disease states.
Neuronal nicotinic acetylcholine receptors (nAChRs) are members of a ligand-gated ion channel family of receptors, consisting of transmembrane proteins of pentameric structure with potentially diverse composition (Anand et al., 1991). S-(−)-Nicotine activates all the known subtypes of nAChRs, although with varying affinities (Parker et al., 1998). Heteromeric nAChRs exist as combinations of α and β subunits; however, the exact subunit composition, stoichiometry, and arrangement of the subunits of native nAChRs have not been elucidated conclusively (Lukas et al., 1999).
Great functional diversity is suggested by the identification of 12 genes encoding α2–α10 and β2–β4 subunits, and from results of in situ hybridization studies revealing discrete, but overlapping, CNS distribution of mRNAs encoding these subunits (Wada et al., 1989;Dineley-Miller and Patrick, 1992; Séguéla et al., 1993; Le Novère et al., 1996). In addition to heteromeric nAChRs, homomeric nAChRs also are present in the CNS and are believed to consist of α7, α8, or α9 subunits; the α7* nAChR is one of the most abundant nAChR subtypes in brain (Wada et al., 1989; Flores et al., 1992). The α7* nAChR subtype is sensitive to inhibition by α-bungarotoxin and methyllycaconitine (MLA) (Schoepfer et al., 1990;Orr-Urtreger et al., 1997; Davies et al., 1999); [3H]MLA has been reported to be a useful radioligand for probing this nAChR subtype (Davies et al., 1999).
The α4β2* subtype is also a predominant nAChR in the CNS and is probed by high- affinityS-(−)-[3H]nicotine binding. Binding of S-(−)-[3H]nicotine to nAChRs in homogenates of rodent brain is reversible and stereospecific, and represents a single class of high-affinity sites located at the interface of the α/β subunits (Lippiello et al., 1987; Reavill et al., 1988; Zhang and Nordberg, 1993). Greater than 90% of high-affinity S-(−)-[3H]nicotine binding sites are immunoprecipitated with anti-β2 antibody (Whiting and Lindstrom, 1987; Flores et al., 1992). Furthermore, mice lacking the β2 subunit do not exhibit high-affinityS-(−)-[3H]nicotine binding (Zoli et al., 1998). Taken together, these results strongly suggest that α4β2* is the nAChR subtype probed byS-(−)-[3H]nicotine binding in brain.
nAChR subunit composition is an important factor that determines the relative affinity of nAChR antagonists at theS-(−)-[3H]nicotine binding site (Harvey and Luetje, 1996; Harvey et al., 1996; Chavez-Noriega et al., 1997). The relative inhibitory potency of the classic nAChR antagonist, dihydro-β-erythroidine (DHβE), at multiple recombinant nAChRs has provided insight into the contribution of α and β subunits to antagonist sensitivity. DHβE inhibition of agonist-induced currents in Xenopus oocytes afforded the following rank order of sensitivity of expressed rat nAChRs: α4β4 > α4β2 = α3β2 > α2β2 > α2β4 ≫ α3β4 (Harvey and Luetje, 1996; Harvey et al., 1996). Similarly, the rank order for DHβE inhibition of expressed human nAChR was α4β4 > α4β2 > α2β2 = α3β2 = α2β4 ≫ α3β4 (Chavez-Noriega et al., 1997). As such, both α and β subunit N-terminal binding domains are important for sensitivity to DHβE (Harvey and Luetje, 1996; Harvey et al., 1996). With respect to native receptors in rat brain, [3H]DHβE competitively binds to nAChRs with high affinity (Kd ∼10 nM; Williams and Robinson, 1984). Taken together, these results indicate that DHβE binds to agonist recognition sites on α4β2* native nAChR receptors.
Recently, exciting developments in drug discovery have indicated that nAChR agonists may be useful for the treatment of cognitive dysfunction, neurodegeneration, and other CNS diseases (Lloyd and Williams, 2000; Glennon and Dukat, 2000). However, relatively little attention has focused on the development of nAChR antagonists as drug candidates (Dwoskin et al., 2000; Dwoskin and Crooks, 2001). Our previous research has discovered a new class of nAChR antagonists resulting from N-n-alkylation of theS-(−)-nicotine molecule (Dwoskin et al., 1999). TheseS-(−)-nicotine analogs exhibit potent and competitive inhibition of the nAChR subtype (α3α6β2* subtype) mediatingS-(−)-nicotine-evoked dopamine (DA) release from dopaminergic nerve terminals in striatum (Wilkins et al., 2002). Structure-activity relationships (SARs) reveal that analogs withN-n-alkyl chains ranging from C7 to C12 were the most potent antagonists of native α3α6β2* nAChRs. The C10 analog,N-n-decylnicotinium iodide (NDNI), was unique in this respect, since it did not inhibitS-(−)-nicotine-evoked DA release. To determine the nAChR-subtype selectivity of this new class ofN-n-alkylnicotinium antagonists, the current study evaluated the ability of these antagonists to inhibit high-affinity S-(−)-[3H]nicotine binding to rat striatal membranes (α4β2* subtype) and to inhibit [3H]MLA binding to rat whole brain membranes (α7* subtype). Analog affinity was compared with that of the classic antagonist, DHβE, and the mechanism of interaction of twoN-n-alkylnicotinium analogs, NDNI andN-n-octylnicotinium iodide (NONI), with α4β2* nAChRs was also determined.
Materials and Methods
Materials.
DHβE, S-(−)-nicotine di-d-tartrate, HEPES, Tris[hydroxymethyl]aminomethane hydrochloride (Trizma HCl), Tris[hydroxymethyl]aminomethane base (Trizma), and polyethylenimine (PEI) were purchased from Sigma/RBI (Natick, MA). R-(+)-Nicotine was purchased from Toronto Research Chemicals (Toronto, ON, Canada).S-(−)-[3H]Nicotine (S-(−)[N-methyl-3H]; specific activity, 80 Ci/mmol and (±)-[3H]methyllycaconitine ([1α,4(S),6β,14α,16β]-20-ethyl-1,6,14,16-tetramethoxy-4[[[2-([3-3H]methyl-2,5-dioxo-1-pyrrolidinyl)benzoyl]oxy]methyl]aconitane-7,8-diol); specific activity, 25.4 Ci/mmol; [3H]MLA) were purchased from PerkinElmer Life Sciences (Boston, MA) and Tocris Cookson Ltd. (Bristol, U.K.), respectively. Scintillation cocktail 3a70B was purchased from Research Products International Corp. (Mt. Prospect, IL). Remaining chemicals used in the buffers were obtained from Fisher Scientific (Pittsburgh, PA). N-Methylnicotinium iodide (NMNI), N-n-propylnicotinium iodide (NPNI), N-n-butylnicotinium iodide (NnBNI), and NONI were prepared as described by Crooks et al. (1995).N-Ethylnicotinium iodide (NENI),N-n-pentylnicotinium iodide (NPeNI),N-n-hexylnicotinium iodide (NHxNI),N-n-heptylnicotinium iodide (NHpNI),N-n-nonylnicotinium iodide (NNNI), NDNI,N-n-undecylnicotinium iodide (NUNI), andN-n-dodecylnicotinium iodide (NDDNI) were prepared from S-(−)-nicotine and the appropriaten-alkyl iodide using the general procedure described by Rui et al. (2002). All compounds were fully characterized by elemental analysis and determined to be free from S-(−)-nicotine utilizing spectroscopic (1H and13C nuclear magnetic resonance, and fast atom bombardment mass spectroscopy), thin layer chromatographic (silica gel), and combustion analysis procedures. Structures of theN-n-alkylnicotinium analogs are shown in Fig.1.
Subjects.
Male Sprague-Dawley rats (200–250 g) were obtained from Harlan Laboratories (Indianapolis, IN) and were housed two per cage with free access to food and water in the Division of Laboratory Animal Resources at the College of Pharmacy, University of Kentucky. Experimental protocols involving animals were in strict accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee at the University of Kentucky.
S-(−)-[3H]Nicotine Saturation Binding.
For each experiment, striata from two to four rats were homogenized using a Tekmar Polytron in 10 volumes of ice-cold modified Krebs-HEPES buffer (20 mM HEPES, 118 mM NaCl, 4.8 mM KCl, 2.5 mM CaCl2, 1.2 mM MgSO4, pH 7.5). Homogenates were incubated (5 min at 37°C) and centrifuged (29,000g for 20 min at 4°C). Tissue pellets were resuspended in 10 volumes of ice-cold Milli-Q water (Millipore Corp., Bedford, MA), incubated (5 min at 37°C), and centrifuged (29,000g for 20 min at 4°C). The tissue pellets were again resuspended in 10 volumes of ice-cold 10% Krebs-HEPES buffer, then incubated and centrifuged as described. Final tissue pellets were stored at −70°C in fresh 10% Krebs-HEPES buffer until use. Upon assay, pellets were resuspended in 10% Krebs-HEPES buffer, incubated, and centrifuged as previously described. Final pellets were resuspended in 2.0 ml of ice-cold Milli-Q water, and the amount of protein (∼200 μg of protein/100 μl of membrane suspension) was determined (Bradford, 1976).
Saturation binding ofS-(−)-[3H]nicotine was performed in duplicate in a final volume of 200 μl of Krebs-HEPES buffer containing 250 mM Tris (pH 7.5, at 4°C). Reactions were initiated by addition of 100 μl of membrane suspension to tubes containing 50 μl of Krebs-HEPES buffer and 50 μl ofS-(−)-[3H]nicotine (0.625–20 nM, final concentration). Nonspecific binding at eachS-(−)-[3H]nicotine concentration was determined in duplicate in the presence of 10 μMS-(−)-nicotine. Following incubation (90 min at 4°C), reactions were terminated by dilution with ice-cold Krebs-HEPES buffer followed by immediate filtration through Whatman GF/B glass fiber filters (presoaked in 0.5% PEI) using a Brandel cell harvester (Biomedical Research and Development Laboratories, Inc., Gaithersburg, MD). Filters were rinsed three times with 3 ml of ice-cold Krebs-HEPES buffer and transferred to scintillation vials, 3 ml of scintillation cocktail were added, and radioactivity was determined by liquid scintillation spectroscopy (Tri-Carb 2100 TR Liquid Scintillation Analyzer, PerkinElmer Life Sciences).
Inhibition of S-(−)-[3H]Nicotine Binding.
Striatal membranes were prepared as previously described. Inhibition of specificS-(−)-[3H]nicotine binding by synthetic N-n-alkylnicotinium analogs was assessed using a previously described method (Crooks et al., 1995). Briefly, assays were performed in triplicate in a final volume of 200 μl of Krebs-HEPES buffer containing 250 mM Tris buffer (pH 7.5, 4°C). Reactions were initiated by the addition of 100 μl of membrane suspension to tubes containing 50 μl of Krebs-HEPES buffer or one of at least seven concentrations (0.1 nM–1 mM, final concentration) of S-(−)-nicotine,R-(+)-nicotine, DHβE orN-n-alkylnicotinium analog and 50 μl ofS-(−)-[3H]nicotine (3 nM, final concentration). Nonspecific binding was determined in triplicate in the presence of 10 μM S-(−)-nicotine. Following incubation (90 min at 4°C), reactions were terminated by dilution of samples with ice-cold Krebs-HEPES buffer followed by immediate filtration through Whatman GF/B glass fiber filters (presoaked in 0.5% PEI) using the cell harvester. Filters were processed and radioactivity was determined as previously described.
Inhibition of [3H]MLA Binding.
Whole rat brain (minus cortex, striatum, and cerebellum) was homogenized in 20 volumes of ice-cold hypotonic buffer (2 mM HEPES, 14.4 mM NaCl, 0.15 mM KCl, 0.2 mM CaCl2, and 0.1 mM MgSO4, pH 7.5). Homogenates were incubated at 37°C for 10 min and centrifuged (25,000g for 15 min at 4°C). Pellets were washed three times by resuspension in 20 volumes of the same buffer, followed by centrifugation using the above parameters. Final pellets were resuspended in incubation buffer to provide ∼150 μg of protein/100 μl membrane suspension. Binding assays were performed in duplicate, in a final vol of 250 μl of incubation buffer, containing 20 mM HEPES, 144 mM NaCl, 1.5 mM KCl, 2 mM CaCl2, 1 mM MgSO4, and 0.05% bovine serum albumin, pH 7.5. Assays were initiated by the addition of 100 μl of membrane suspension to 150 μl of sample containing 2.5 nM [3H]MLA and one of at least six concentrations (30 nM–100 μM, final concentration) of analog, and incubated for 2 h at room temperature. Nonspecific binding was determined in the presence of 10 μM MLA. Assays were terminated by dilution with 3 ml of ice-cold incubation buffer followed by immediate filtration through Schleicher and Schuell (Keene, NH) #32 glass fiber filters (presoaked with 0.5% PEI) using the cell harvester. Filters were processed as described above in theS-(−)-[3H]nicotine binding assay.
Mechanism of Analog Inhibition ofS-(−)-[3H]Nicotine Binding.
Striatal membrane homogenates were prepared as previously described. Saturation of specific S-(−)-[3H]nicotine binding was determined in the absence and presence of three concentrations of NDNI or NONI. Reactions were initiated by addition of 100 μl of membrane suspension to tubes containing 50 μlS-(−)-[3H]nicotine (0.625–20 nM, final concentration) and 50 μl of Krebs-HEPES buffer (control, i.e., absence of analog) or one of three concentrations of NDNI or NONI. Concentrations of NDNI (0.6, 2, and 6 μM) and NONI (0.66, 2.2, and 22 μM) were chosen based on results obtained from the inhibition isotherms. Nonspecific binding at eachS-(−)-[3H]nicotine concentration was determined in duplicate in the presence of 10 μMS-(−)-nicotine. Following incubation (90 min at 4°C), reactions were terminated and filters processed as previously described.
Analysis of S-(−)-[3H]Nicotine Binding Data.
ForS-(−)-[3H]nicotine saturation binding isotherms, specificS-(−)-[3H]nicotine binding was expressed as femtomoles per milligram of protein and plotted as a function of S-(−)-[3H]nicotine concentration. Data were fitted by one- and two-site hyperbolic functions using weighted (1/Y2minimized), nonlinear least-squares regression, since the nonweighted fit minimizing Y2 was not unique. One-site binding was modeled using the equation, Y = (Bmax ·X)/(Kd + X), where Y = specificS-(−)-[3H]nicotine binding,X =S-(−)-[3H]nicotine concentration,Bmax = maximum binding density, andKd = the dissociation binding constant. Two-site binding was modeled using the equation,Y = (Bmax1 ·X)/(Kd1 + X) + (Bmax2 ·X)/(Kd2 + X), where Bmax1 andBmax2 = maximum binding density for sites 1 and 2, respectively, and Kd1and Kd2 = dissociation binding constants for sites 1 and 2, respectively. Fits were compared using theF statistic, and the one-site model was chosen unless the two-site model provided a significantly (p < 0.05) better fit.
Specific S-(−)-[3H]nicotine binding was also plotted as a function of logS-(−)-[3H]nicotine concentration. To obtain the Bmax andKd values, nonlinear regression was performed using a variable-slope sigmoid function, holding minimum specific binding (Bt) at a constant value of zero, such thatY = Bt + ((Tp − Bt)/(1 + 10(log EC50−X) · n)) whereY = specificS-(−)-[3H]nicotine binding,X = logS-(−)-[3H]nicotine concentration, Bt and Tp = minimum and maximum specificS-(−)-[3H]nicotine binding densities, respectively, log EC50 = logS-(−)-[3H]nicotine concentration at 50% receptor occupancy, and nH = the Hill slope factor, an index of binding cooperativity. Simple linear regression on the Scatchard-transformed data were performed, such thatY = mX + b, whereY = B/F, X =B, m = slope, and b = extrapolated Y-intercept. TheBmax value was obtained from the extrapolated X-intercept value, and theKd value was obtained from −1/slope.
Analyses of Analog Binding Inhibition Data.
ForN-n-alkylnicotinium analog inhibition ofS-(−)-[3H]nicotine binding, data were expressed as specificS-(−)-[3H]nicotine bound as a percentage of control and plotted as a function of log analog concentration. Data were fit by a variable slope model, using nonlinear least-squares regression, such that Y = Bt + ((Tp − Bt)/(1 + 10)(log IC50−X) ·n)), where Y = specificS-(−)-[3H]nicotine binding,X = log [S-(−)-[3H]nicotine], Bt and Tp = minimum and maximum specificS-(−)-[3H]nicotine binding densities, respectively, log IC50 = log[compound], which decreasedS-(−)-[3H]nicotine receptor occupancy by 50%, and n = the pseudo-Hill coefficient. Analog affinity constants (Ki values) were calculated using the equation, Ki= IC50/(1 + concentration ofS-(−)-[3H]nicotine/Kd), where IC50 = the concentration of analog inhibiting S-(−)-[3H]nicotine binding by 50% and Kd = theS-(−)-[3H]nicotine dissociation constant determined from initial saturation binding experiments (Cheng and Prusoff, 1973).
For N-n-alkylnicotinium analog inhibition of [3H]MLA binding, data were expressed as specific [3H]MLA bound as a percentage of control, and plotted as a function of log analog concentration. Data were fit by a one-site binding competition model, using nonlinear least-squares regression, such that Y = Bt + ((Tp − Bt)/(1 + 10X − log IC50)), where Y = specific [3H]MLA binding (percentage of control),X = log [[3H]MLA], Bt = minimum [3H]MLA binding (held constant at a value of 0), Tp = the [3H]MLA binding density, and log IC50 = log [analog], which decreased [3H]MLA receptor occupancy by 50%. Analog affinity constants (Ki values) were calculated using the equation, Ki= IC50/(1 + concentration of [3H]MLA/Kd), where IC50 = the concentration of analog inhibiting [3H]MLA binding by 50%, andKd = the [3H]MLA dissociation constant (1.93 nM) determined from initial saturation binding experiments (Cheng and Prusoff, 1973).
The mechanism of analog interaction with high-affinityS-(−)-[3H]nicotine binding sites was assessed. Saturation binding ofS-(−)-[3H]nicotine was assessed in the absence and presence of three concentrations of NDNI and NONI. Minimum binding was held constant at a value of zero, and simple slope sigmoid fits were used since the F statistic revealed that the variable slope did not provide a significantly better fit to the data. For each analog, the saturation isotherms were assessed for parallelism by comparing simple and variable slope sigmoid curve fit. The saturation isotherms did not reach a clear asymptotic maximum in the presence of the highest concentrations of analog. No significant differences between maximumS-(−)-[3H]nicotine binding values extrapolated from curve fits to the data at each NDNI or NONI concentration were found by one-way ANOVA (NDNI,F3,14 = 0.929, p = 0.46; NONI, F3,16 = 1.61,p = 0.24). Therefore, the mean extrapolated value for maximum S-(−)-[3H]nicotine binding was used as the maximum binding constant. To further assess the mechanism of analog interaction with high-affinity receptors, saturation data were also analyzed using Scatchard analysis. Mean specific S-(−)-[3H]nicotine binding data were transformed and fit by linear regression, andBmax andKd values were derived. ApparentKd values were transformed to the respective negative log (−log Kd) value for parametric statistical analysis. One-way ANOVAs ofBmax and −logKd values were used to determine effects of analog concentration onS-(−)-[3H]nicotine binding parameters.
To further verify the mechanism of analog interaction with high affinity S-(−)-[3H]nicotine binding sites, the binding affinities derived from saturation analyses in the absence and presence of three concentrations of NDNI or NONI were negative log transformed (−log Kd) and plotted as a function of analog concentration (Lew and Angus, 1995). Three nonlinear regression models were fit to the data. The basic equation was a mathematical equivalent of a Schild regression indicative of competitive binding inhibition, Y = −1 · log ((X · 1e-6) + (10log Kb)) −P, where Y = −logKd for specificS-(−)-[3H]nicotine binding in the absence or presence of analog, X = concentration of analog (micromolar), log Kb = the logarithm of analog binding inhibition constant derived fromS-(−)-[3H]nicotine binding inhibition isotherms, and P = the negative logarithm of the constant C (−log C), where C= 0.50 fractional receptor occupancy in the absence and presence of analog). A “power departure” model was also used to fit the data and was defined by the equation Y = −1 · log (((X · 1e-6)slope) + (10log Kb)) −P. The power departure model differed from the basic equation by inclusion of a variable slope factor. A “quadratic departure” model was also used and was defined by the equationY = −1 · log((X · 1e-6) · (1 + ((slope · (X · 1e-6))/(10logKb))) + (10logKb)) − P. In addition to inclusion of the variable slope factor, the quadratic departure equation allows for the detection of more complex interactions with the binding site, such as a nonequilibrium steady state and/or heterogeneity of the receptor population. The basic equation defining competitive interaction was chosen as the best fit model, unless one of the more complex models provided a significantly (p < 0.05) better fit to the data. Regression and statistical analyses were performed using the commercially available software packages Prism v3.0 (GraphPad Software, Inc., San Diego, CA), whereas more complex statistical analyses were preformed using SPSS standard v9.0 (SPSS Science, Chicago, IL).
Correlation of N-n-Alkylnicotinium Inhibition of S-(−)-[3H]Nicotine Binding to Striatal Membranes and S-(−)-Nicotine-Evoked [3H]DA Overflow from Superfused Striatal Slices.
To evaluate the nAChR subtype selectivity of this novel series of analogs, log IC50 values for each analog to inhibitS-(−)-nicotine-evoked [3H]DA overflow, which were obtained from a previous report (Wilkins et al., 2002), were plotted as a function of the log IC50values derived fromS-(−)-[3H]nicotine binding inhibition curves (Fig. 2). Pearson's analysis of these data was used to detect the presence of a significant correlation between these nAChR sites.
Results
S-(−)-[3H]Nicotine Saturation Binding.
Binding ofS-(−)-[3H]nicotine to rat striatal membranes was saturable and specific (Fig. 2). Nonspecific binding was ∼12% of total binding at a S-(−)-nicotine concentration approximating the Kd. A one-site hyperbolic model provided the best fit to the data, suggesting that equilibrium binding ofS-(−)-[3H]nicotine represents interaction with a single population of binding sites. Values for affinity (Kd, 95% confidence interval) and maximum density (Bmax ± S.E.M.) were 1.94 (1.36, 2.51) nM and 97.6 (±4.7) fmol/mg protein, respectively. Variable slope sigmoid fit (R2 = 0.995) of specificS-(−)-[3H]nicotine binding as a function of log S-(−)-[3H]nicotine concentration provided similar estimates forKd andBmax of 1.56 (0.92, 2.64) nM and 89.8 (±5.9) fmol/mg protein, respectively, with a Hill coefficient of 0.963 (r2 = 1). Linear regression (r2 = 0.948) of the Scatchard-transformed data provided similar estimates ofKd andBmax values of 1.94 (1.31, 2.56) nM and 98.1 (±5.2) fmol/mg protein (Fig. 2, inset).
N-n-Alkylnicotinium Analog Inhibition of Specific S-(−)-[3H]Nicotine Binding.
Inhibition curves forN-n-alkylnicotinium analogs as well as the reference compounds S-(−)-nicotine,R-(+)-nicotine, and DHβE are shown in Fig.3. Comparisons were made between simple- and variable-slope sigmoid curve fits to the data for each compound and were found not different, with exceptions of DHβE (F1,5 = 18.5, p < 0.05) and S-(−)-nicotine (F1,6 = 28.9, p < 0.05), suggesting a more complex interaction of these two compounds with S-(−)-[3H]nicotine binding sites. Inhibition parameters and slope factors are provided in Table1. Similar to the reference compounds, the N-n-alkylnicotinium analogs completely inhibited S-(−)-[3H]nicotine binding. S-(−)-Nicotine had the highest affinity (Ki = 0.8 nM) of the compounds tested.R-(+)-Nicotine had an affinity (Ki = 49 nM) ∼60-fold lower than that of S-(−)-nicotine, indicating a high degree of enantioselectivity for nicotine at high-affinityS-(−)-[3H]nicotine binding sites. Analog affinities varied across a 230-fold range (Ki values ranged from 93 nM to 20 μM; Table 1). From the series, the C10 analog NDNI exhibited the greatest affinity, with aKi value of 93 nM, which was not different from either the C12 analog NDDNI (Ki = 140 nM) or DHβE (Ki = 143 nM). The remaining synthetic analogs exhibited lower affinities, with an overall rank order of the compounds tested: S-(−)-nicotine >R-(+)-nicotine > NDNI = NDDNI = DHβE > NHxNI > NNNI = NENI = NHpNI > NMNI > NnBNI > NONI = NPNI (Table 1).
N-n-Alkylnicotinium Analog Inhibition of [3H]MLA Binding.
Results from full concentration-response curves for the analogs to inhibit [3H]MLA binding revealed that only four analogs (NENI, NHxNI, NHpNI, and NONI) at the highest concentration of 100 μM inhibited binding by greater than 50% of total specific binding (Table2). Based on these results,Ki values were obtained for NENI, NHxNI, NHpNI, and NONI from one-site competition models of the data (Table 2). None of the analogs in this series exhibited high affinity for [3H]MLA binding sites.
Analog Affinity forS-(−)-[3H]Nicotine Binding Sites as a Function of n-Alkyl Chain Length.
A plot of the log transform of the S-(−)-[3H]nicotine binding inhibition constant (log Ki) as a function of n-alkyl chain length for each analog is presented in Fig. 4. Linear regression of analog affinity (Ki) for S-(−)-[3H]nicotine binding sites byn-alkyl chain length was performed with NONI (C8 analog) included and excluded from the analysis, because the Ki value for NONI appeared to deviate from the linear trend. With the logKi value for NONI included in the analysis, the linear model did not provide a good fit to the data (r2 = 0.338), and the slope of the regression line was not significantly different from a value of zero (F1,8 = 7.08, p < 0.078). When the log Ki value for NONI was excluded, the linear model provided a significantly better fit (r2 = 0.560), with a significantly non-zero slope (–0.156; F1,7 = 8.91,p < 0.05). Therefore, with the exception of NONI, analog affinity varied in a linear fashion with n-alkyl chain length, i.e., affinity increased with increasing chain length.
Mechanism of N-n-Alkylnicotinium Analog-Induced Inhibition ofS-(−)-[3H]Nicotine Binding.
Within the series of analogs examined, the C10 and C8 analogs, NDNI and NONI, respectively, represent extremes in affinity forS-(−)-[3H]nicotine binding sites. The mechanism by which NDNI and NONI interact with high-affinityS-(−)-[3H]nicotine binding sites was determined by assessing shifts in theS-(−)-[3H]nicotine concentration-binding curves in the presence of three concentrations of NDNI or NONI relative to control (in the absence of analog). NDNI and NONI concentrations were chosen based onKi values determined in the previous inhibition studies (Table 2).S-(−)-[3H]Nicotine saturation binding isotherms were observed to be parallel in the absence and presence of analog, as evidenced by the better fit of the data to the simple slope sigmoid model compared with the variable slope model (Table 3).
Scatchard transformations of high-affinityS-(−)-[3H]nicotine binding in the absence and presence of NDNI and NONI are shown in Fig.5. Values for affinity (Kd) and maximum density (Bmax) were derived from linear regressions of these data (Table 3). One-way ANOVA of the logKd values obtained in the absence and presence of three concentrations of each analog indicated a significant decrease in S-(−)-[3H]nicotine affinity for its binding site (NONI,F3,16 = 30.5, p < 0.001; NDNI, F3,14 = 21.1,p < 0.001). One-way ANOVA also revealed no significant differences in Bmax values (NONI,F3,16 = 0.07, p = 0.97; NDNI, F3,14 = 0.34,p = 0.80), indicating that both NDNI and NONI interact with high-affinityS-(−)-[3H]nicotine binding sites in rat striatum via a competitive mechanism.
To further evaluate the nature of the interaction of NDNI and NONI with high-affinity S-(−)-[3H]nicotine binding sites, affinity values forS-(−)-[3H]nicotine in the absence and presence of NDNI or NONI were plotted as a function of analog concentration and nonlinear regression curves generated (Fig.6). In agreement with the Scatchard analysis, a competitive mode of interaction of each analog with high-affinity S-(−)-[3H]nicotine binding sites was indicated by best fit of the data to the simple model (NDNI, R2 = 0.952; NONI,R2 = 0.985). Neither the power departure model (NDNI, p = 0.40; NONI,p = 0.20) nor the quadratic departure model (NDNI,p = 0.33; NONI, p = 0.36) provided a significantly better fit to the data for either NDNI or NONI.
Correlation of N-n-Alkylnicotinium Inhibition of S-(−)-[3H]Nicotine Binding to Striatal Membranes and S-(−)-Nicotine-Evoked [3H]DA Overflow from Superfused Striatal Slices.
To evaluate the nAChR subtype selectivity of this series of analogs, data from both the presentS-(−)-[3H]nicotine binding assays and previously reported S-(−)-nicotine-evoked [3H]DA overflow assays (Wilkins et al., 2002) were analyzed by correlation analysis, and the results are shown in Fig. 7. NDNI did not inhibitS-(−)-nicotine-evoked [3H]DA overflow and, as such, the highest concentration (100 μM) examined in the current study was included in the correlation analysis. Correlation of IC50 values for inhibition ofS-(−)-nicotine-evoked [3H]DA overflow and of IC50 values for inhibition ofS-(−)-[3H]nicotine binding was not significant (Pearson r = 0.255, p > 0.05). Another correlation analysis was performed subsequently, in which the data for NDNI and NONI were excluded as outliers. Upon exclusion of NDNI and NONI, a significant correlation was obtained (Pearson r = 0.855, p < 0.05). Therefore, the lack of correlation observed when data for NDNI and NONI were included in the analysis suggests that inhibition ofS-(−)-nicotine-evoked [3H]DA overflow is not well correlated with inhibition ofS-(−)-[3H]nicotine binding. This interpretation is supported by the observations that NDNI did not inhibit S-(−)-nicotine-evoked [3H]DA overflow but potently inhibitedS-(−)-[3H]nicotine binding. Also, NONI did not inhibitS-(−)-[3H]nicotine binding but potently inhibited nicotine-evoked [3H]DA overflow. Interestingly, when the correlation analysis was performed excluding NDNI and NONI, a relationship was revealed suggesting that the remaining analogs interact with common nAChR sites.
Discussion
The current study demonstrates thatN-n-alkylnicotinium analogs withn-alkyl chain lengths from 1 to 12 carbons, which have been previously shown to act as antagonists at α3α6β2* nAChRs mediating S-(−)-nicotine-evoked [3H]DA overflow (Wilkins et al., 2002), have varying affinity forS-(−)-[3H]nicotine binding sites in striatum. A linear relationship was observed such that with increasingn-alkyl chain length, affinity for the α4β2* nAChR subtype increased. The most potent analog was NDNI (C10 analog; Kivalue = 90 nM). In contrast, the analogs did not exhibit high affinity for [3H]MLA binding sites, indicating low affinity for α7* nAChRs. Furthermore, the present study demonstrates that the interaction of these analogs with high affinityS-(−)-[3H]nicotine binding sites on the α4β2* subtype is via a competitive mechanism. Importantly, NDNI did not inhibit S-(−)-nicotine-evoked [3H]DA overflow from superfused striatal slices and, thus, appears to be selective for the α4β2* subtype. In contrast, NONI (C8 analog) had low affinity forS-(−)-[3H]nicotine binding sites but potently inhibited S-(−)-nicotine-evoked [3H]DA overflow (IC50 = 0.62 μM), demonstrating selectivity for α3α6β2* nAChRs mediating S-(−)-nicotine-evoked [3H]DA overflow.
Affinity and maximum binding density estimates (1.94 nM and 97.6 fmol/mg protein, respectively) obtained fromS-(−)-[3H]nicotine saturation binding analysis were modeled best by a one-site hyperbolic function. Linear Scatchard transformation and Hill coefficient of 0.963 indicated an interaction with a single class of binding sites. Current parameter estimates were within the range of reported values, using either striatal or whole brain membranes (Kd= 0.4–14 nM; Bmax = 55–200 fmol/mg protein; Lippiello et al., 1987; Martino-Barrows and Kellar, 1987;Reavill et al., 1988). Variability inKd andBmax estimates may be attributed to methodological differences in assays employed (e.g., inclusion of protease or cholinesterase inhibitors in tissue preparation buffers, variation in Mg2+ and Ca2+concentrations in binding buffers). Validation of current assay conditions was provided by comparableKi values forS-(−)-nicotine, R-(+)-nicotine, and DHβE (Ki = 0.8, 50, and 140 nM, respectively), as reported by others (Martino-Barrows and Kellar, 1987;Reavill et al., 1988). Thus, the present assay reliably measured interaction with high-affinityS-(−)-[3H]nicotine binding sites in striatum.
Inhibition of specificS-(−)-[3H]nicotine binding byS-(−)-nicotine, R-(+)-nicotine, DHβE, and eachN-n-alkylnicotinium analog was modeled using a variable-slope sigmoid equation. With the exception ofS-(−)-nicotine and DHβE, slope factors derived were near unity for R-(+)-nicotine and each of the analogs examined, indicating a simple competitive interaction with high-affinityS-(−)-[3H]nicotine binding sites. In contrast, DHβE and S-(−)-nicotine exhibited shallow slopes (n < 1), suggesting that these compounds recognize either more than one agonist binding site on a single α4β2* receptor, or more than one conformational state of the α4β2* nAChR, or more than one nAChR subtype. Interestingly, multiple binding sites for [3H]DHβE in rat cortical membranes have been detected (Williams and Robinson, 1984), suggesting that DHβE distinguishes more than one binding site or more than one state of the α4β2* nAChR. In the latter study, a pseudo-Hill coefficient of 0.9 for S-(−)-nicotine inhibition of [3H]DHβE binding was observed, suggesting that S-(−)-nicotine interacts with one of multiple [3H]DHβE sites.
N-Alkylation of the S-(−)-nicotine molecule converts this potent agonist into analogs that selectively and competitively interact with α4β2* nAChRs.N-n-Alkylnicotinium analogs exhibited affinities for S-(−)-[3H]nicotine binding sites across an ∼200-fold concentration range, from ∼90 nM (NDNI) to ∼20 μM (NONI). The relationship between analog affinity forS-(−)-[3H]nicotine binding sites was a linear function of n-alkyl chain length, with the exception of NONI. These N-n-alkylnicotinium analogs are larger molecules than S-(−)-nicotine, and those with longer chain lengths (C9, C10, and C12) were more potent inhibitors ofS-(−)-[3H]nicotine binding than those with shorter chain lengths (C1–C7). The higher affinity of the longer n-alkyl chain analogs may reflect a stronger association with agonist binding sites on α4β2* nAChRs, due to increased lipophilic interaction of the carbon chain with a hydrophobic amino acid-rich region of the protein near the binding pocket. As such, this lipophilic interaction may stabilize the analog-receptor complex, thereby increasing the affinity for the agonist recognition site by long-chain analogs.
The mechanism of N-n-alkylnicotinium analog interaction with S-(−)-[3H]nicotine binding sites was determined using two analogs exhibiting extremeKi values (Ki = 90 nM and 20 μM for NDNI and NONI, respectively). NDNI and NONI differ structurally by only two methylene units in the alkyl chain length. Scatchard analyses ofS-(−)-[3H]nicotine saturation binding in the absence and presence of NDNI or NONI demonstrated that affinity of S-(−)-[3H]nicotine for its binding sites decreased in the presence of increasing concentrations of analog, with no change inBmax value. The competitive interaction of NDNI and NONI with high affinityS-(−)[3H]nicotine binding sites was assessed further using the Lew and Angus analysis, which alleviates concern regarding variability in the estimate ofBmax. The simple, one-site model best fit the data, corroborating the interpretation of a competitive interaction of NDNI and NONI with high-affinityS-(−)-[3H]nicotine binding sites. These results suggest that NDNI and NONI competitively interact with either specific amino acid residues directly involved inS-(−)-[3H]nicotine binding or with nearby residues allowing for steric hindrance of the interaction of S-(−)-[3H]nicotine with its high-affinity binding site.
To evaluate nAChR subtype selectivity of theN-n-alkylnicotinium analogs, correlation analysis of data from both the presentS-(−)-[3H]nicotine binding assays and previously reported S-(−)-nicotine-evoked [3H]DA overflow assays (Wilkins et al., 2002) revealed that NDNI and NONI stand out as selective analogs for α4β2* and α3α6β2* nAChR subtypes, respectively. NDNI exhibited high affinity forS-(−)-[3H]nicotine binding sites but did not inhibit S-(−)-nicotine-evoked [3H]DA overflow. On the other hand, NONI was a potent inhibitor of S-(−)-nicotine-evoked [3H]DA overflow but exhibited low affinity forS-(−)-[3H]nicotine binding sites. Thus, NDNI and NONI appear to be excellent lead compounds for probing agonist recognition sites on α4β2* and α3α6β2* nAChRs, respectively.
The α3α6β2* subtype has been suggested to mediateS-(−)-nicotine-evoked DA release primarily based on sensitivity to the α3α6β2-selective antagonists, neuronal bungarotoxin (Schulz and Zigmond, 1989; Grady et al., 1992) and α-conotoxin MII (Cartier et al., 1996). The α3α6β2 selectivity of these antagonists is indicated by activity in recombinant systems expressing specific nAChR subtypes (Luetje et al., 1990; Cartier et al., 1996) and by results from nAChR knockout mice studies (Champtiaux et al., 2002; Picciotto and Corrigall, 2002). Importantly, α-conotoxin MII only partially inhibited nicotine-evoked DA release (Kulak et al., 1997), suggesting the potential involvement of other nAChR subtypes in this response, such as α4-, β2-, and β4-containing nAChRs (Picciotto et al., 1998; Sharples et al., 2000). Rat substantia nigra neurons express mRNA for α3, α4, α5, α6, α7, β2, β3, and β4 subunits (Wada et al., 1989; Dineley-Miller and Patrick, 1992; Charpantier et al., 1998). Inasmuch as high levels of α6 and β3 mRNAs are expressed in substantia nigra DA neurons (Le Novère et al., 1996; Goldner et al., 1997; Charpantier et al., 1998), their potential combination with α3 and β2 subunits in the mediation of S-(−)-nicotine-evoked DA release in striatum is probable, but has not been established conclusively.
The current results show that N-n-alkylnicotinium analogs interact with both α4β2* and α3α6β2* nAChR subtypes, but not with the α7* nAChR subtype. As was observed in the interaction of these analogs with α3α6β2* (Crooks et al., 1995;Wilkins et al., 2002), the current SAR reveals that an increase in affinity at α4β2* nAChRs is dependent on increasingn-alkyl chain length. Whereas the current receptor binding assays do not provide information as to whether theseN-n-alkylnicotinium analogs function as agonists or antagonists at α4β2* receptors, recently, these analogs have been observed to inhibit S-(−)-nicotine-evoked86Rb+ efflux from preloaded rat thalamic synaptosomes, a functional assay for α4β2* receptors, suggesting an antagonist mechanism of action (L. H. Wilkins, D. K. Miller, J. T. Ayers, P. A. Crooks and L. P. Dwoskin, manuscript submitted for publication). The results suggest that the binding site on the α4β2* subtype that normally accommodates S-(−)-nicotine also accommodates these charged, more sterically bulky molecules, perhaps in a unique binding mode. As such, the unprotonated form of these analogs was proposed previously to interact with the α3α6β2* subtype, in a manner in which the roles of the pharmacophoric nitrogen-containing moieties are reversed (Crooks et al., 1995; Wilkins et al., 2002). Moreover, comparison of the SAR of the analogs at both the α4β2* and α3α6β2* nAChR subtypes reveals that the selectivity of the subtype interaction cannot be explained simply by lipophilicity alone. Thus, the relative lack of interaction of NONI with α4β2* and the lack of interaction of NDNI with α3α6β2* suggest that each of these molecules exists in a unique molecular conformation that is recognized by one subtype but is not compatible with the other. Based on our current knowledge, it is likely that the α subunit plays a critical role in this surprising selective recognition profile of NDNI and NONI.
In summary, a series of N-n-alkylnicotinium analogs exhibited a wide range of affinity forS-(−)-[3H]nicotine binding sites representing α4β2* nAChRs in striatum. When the n-alkyl substituent ranged from C1 to C12, a linear relationship betweenn-alkyl chain length and analog affinity was found, with the exception of the C8 analog, NONI. The ability of NONI to potently inhibit S-(−)-nicotine-evoked [3H]DA overflow from superfused striatal slices, combined with its low affinity forS-(−)-[3H]nicotine and [3H]MLA binding sites, suggests selectivity for the α3α6β2* nAChR subtype. The C10 analog, NDNI, exhibited the highest affinity for the α4β2* subtype; however, this analog did not interact with either α3α6β2* or α7* subtypes. Selectivity for the α4β2* subtype combined with competitive interaction with S-(−)-nicotine binding sites indicates that NDNI is an excellent candidate for studying the structural topography of agonist recognition sites on α4β2* nAChRs, for establishing the antagonist pharmacophore for this subtype, and for defining its role in physiological function and pathological disease states.
Footnotes
-
↵1 Current address: Targacept, Inc., 200 East First Street; Suite 300, Winston-Salem NC 27101-4165.
-
↵2 Current address: AstraZeneca, 1800 Concord Pike, P.O. Box 15437, Wilmington DE 19850-5437.
-
This study was supported by National Institutes of Health Grants DA00399 and DA10934.
-
DOI: 10.1124/jpet.102.043349
- Abbreviations:
- nAChR
- nicotinic acetylcholine receptor
- CNS
- central nervous system
- ANOVA
- analysis of variance
- DHβE
- dihydro-β-erythroidine
- [3H]MLA
- [3H]methyllycaconitine
- NDNI
- N-n-decylnicotinium iodide
- DA
- dopamine
- SAR
- structure-activity relationship
- NDDNI
- N-n-dodecylnicotinium iodide
- NENI
- N-ethylnicotinium iodide
- NHpNI
- N-n-heptylnicotinium iodide
- NHxNI
- N-n-hexylnicotinium iodide
- NMNI
- N-methylnicotinium iodide
- NnBNI
- N-n-butylnicotinium iodide
- NNNI
- N-n-nonylnicotinium iodide
- NONI
- N-n-octylnicotinium iodide
- NPNI
- N-n-propylnicotinium iodide
- NPeNI
- N-n-pentylnicotinium iodide
- NUNI
- N-n-undecylnicotinium iodide
- PEI
- polyethylenimine
- ∗
- putative nicotinic acetylcholine receptor subtype designation
- Received August 16, 2002.
- Accepted September 20, 2002.
- The American Society for Pharmacology and Experimental Therapeutics