Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Although cytochrome P450 2D6 (CYP2D6) is a minor component (~2%) of total cytochrome P450 content in human liver, a significant proportion (~30-40%) of drugs currently in clinical use are metabolized by CYP2D6. Furthermore, the genetic polymorphism associated with CYP2D6-mediated metabolism in humans has heightened awareness of the potential adverse drug reactions following impaired clearance of CYP2D6 substrates in individuals designated as poor metabolizers. Therefore, several efforts have been directed to better understand structure-function relationships of CYP2D6 substrates/inhibitors.

Despite overall structural diversity, most tightly bound CYP2D6 ligands usually contain a protonated basic amine nitrogen thought to be essential for electrostatic interactions with an active site Asp301 residue (Ellis et al., 1995; Mackman et al., 1996; Lewis et al., 1997; de Groot et al., 1999a,b; Hanna et al., 2001). Binding of substrate is generally followed by oxidation 5 to 7 Å from this interaction (Modi et al., 1996; Lin et al., 1997; Wu et al., 1997; Bach et al., 1999; Miller et al., 2001). Based on the findings from numerous quantitative structure-activity relationship (QSAR¹) studies on structurally diverse CYP2D6 ligands (Ferrari et al., 1991; Koymans et al., 1992; Strobl et al., 1993; Venhorst et al., 2000; Upthagrove and Nelson, 2001), lipophilicity and amine basicity have emerged as crucial determinants of binding. The observation that more basic amines possess a higher CYP2D6 affinity is also consistent with the proposal involving the formation of a "tighter" ion-pair with Asp301 (Upthagrove and Nelson, 2001).

With relatively few exceptions (Wu et al., 1997), most of the QSAR studies performed to date have been limited to secondary amine CYP2D6 ligands, and therefore important physiochemical properties governing CYP2D6 affinity of cyclic tertiary and quaternary amines remain poorly understood. In the present study, the correlation between lipophilicity and CYP2D6 affinity of cyclic tertiary and quaternary amines was examined for a series of N-substituted-4-phenyl-1,2,3,6-tetrahydropyridines and positively charged N-substituted-4-phenylpyridiniums, respectively. The N-substituted-4-phenyl-1,2,3,6-tetrahydropyridine scaffold was chosen based on the reasonably good and selective CYP2D6 substrate properties of the cyclic tertiary amine-containing Parkinsonian neurotoxin N-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP, see Table 2) (Kalgutkar et al., 2001). Besides the monoamine oxidase-B-mediated bioactivation of MPTP to the positively charged mitochondrial neurotoxin N-methyl-4-phenylpyridinium (MPP⁺, see Table 1), CYP2D6 metabolizes MPTP to the corresponding non-neurotoxic N-4-(4'-hydroxyphenyl)-N-methyl-1,2,3,6-tetrahydropyridine and 4-phenyl-1,2,3,6-tetrahydropyridine (PTP) metabolites (Coleman et al., 1996; Modi et al., 1997).

An additional reasoning behind the proposed QSAR analysis on 1,2,3,6-tetrahydropyridines and pyridiniums was due to their presence in the structures of the haloperidol (HP) metabolites N-N-[4-(4-fluorophenyl)-4-oxobutyl]-4-(4-chlorophenyl)-1,2,3,6-tetrahydropyridine (HPTP, see Table 3) and 4-(4-chlorophenyl)-1-[4-(4-fluorophenyl)-4-oxobutyl]pyridinium (HPP⁺, see Table 1), respectively. Interestingly, HPTP and HPP⁺ are potent CYP2D6 inhibitors (Tyndale et al., 1991; Pan et al., 1997; Shin et al., 2000) but not substrates (Usuki et al., 1996; Fang et al., 2001). In this context, the lack of CYP2D6 inhibition by the less lipophilic MPTP and MPP⁺ analogs (A. Kalgutkar, unpublished observations) suggests that lipophilicity is an important contributor toward inhibition by 1,2,3,6-tetrahydropyridines and pyridiniums. Based on the finding that the less lipophilic MPTP, but not HPTP, is a CYP2D6 substrate, the role of lipophilicity as a determinant of CYP2D6 substrate properties of 1,2,3,6-tetrahydropyridines is less clear.

Thus, we evaluated the importance of lipophilicity on the CYP2D6 inhibitory/substrate properties for a series of MPTP and MPP⁺ analogs containing N-alkyl substituents of increasing chain length against the bufuralol-1'-hydroxylase activity of the enzyme. For comparative purposes, these studies were also extended to include HP and its metabolites including HPTP and HPP⁺ as well as the reduced HP and reduced HPTP metabolites (±)-4-(4-chlorophenyl)-1-[4-(4-fluorophenyl)-4-hydroxybutyl]-4-piperdinol [(±)-RHP] and (±)-4-(4-chlorophenyl)-1-[4-(4-fluorophenyl)-4-hydroxybutyl]-1,2,3,6-tetrahydropyridine [(±)-RHPTP], respectively, that are known inhibitors of the CYP2D6-catalyzed dextromethorphan O-demethylation reaction (Shin et al., 2000).

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Caution. MPTP is a known nigrostriatal neurotoxin, and therefore compounds of this class should be handled using disposable gloves in a properly ventilated hood. Detailed procedures for the safe handling of MPTP have been reported (Markey and Schmiff, 1986).

Chemistry. Chemical yields are unoptimized-specific examples of one preparation. All chemicals and solvents used in synthesis were purchased from Aldrich (Milwaukee, WI). ¹H NMR spectra in DMSO-d₆ were recorded on a Varian Unity M-400 MHz spectrometer (Varian Medical Systems, Palo Alto, CA); chemical shifts are expressed in parts per million (ppm, ) calibrated to the deuterium lock signal for DMSO-d₆. Spin multiplicities are given as s (singlet), d (doublet), bd (broad doublet), dd (doublet of doublets), t (triplet), q (quartet), and m (multiplet). Coupling constants (J) are given in hertz (Hz). Positive ion electrospray ionization and collision-induced dissociation (CID) mass spectra were obtained on a Sciex API model 2000 liquid chromatography/tandem mass spectrometry (LC/MS/MS) triple quadrupole mass spectrometer (Sciex, Thornhill, ON, Canada). MPP⁺ · I , MPTP · (COOH)₂, N-propyl-4-phenylpyridinium iodide (2 · I), and N-propyl-4-phenyl-1,2,3,6-tetrahydropyridine oxalate salt [12 · (COOH)₂] were synthesized as described previously (Kalgutkar and Castagnoli, 1992). HP was obtained from Sigma-Aldrich (St. Louis, MO) whereas reduced HP, HPTP, reduced HPTP, and HPP⁺ were synthesized via procedures outlined in previously published reports (Usuki et al., 1996). cLog P for the test compounds was calculated from the Hansch and Leo's Pomona College Medicinal Chemistry Project MedChem software (clog P 4.0) distributed by BioByte Corp. (Claremont, CA) whereas pKa's for tetrahydropyridines were estimated from the ACDLogD Suite program (Advanced Chemistry Development Inc., Toronto, ON, Canada). Experimental and calculated log P values were essentially identical for MPTP (2.71 versus 2.74) and N-benzyl-4-phenyl-1,2,3,6-tetrahydropyridine (4.0 versus 4.5) (Altomare et al., 1992).

General Procedure for the Preparation of N-Alkyl- or N-Arylalkyl-4-phenylpyridinium Salts. A reaction mixture containing 4-phenylpyridine (1, 6.45 mmol), alkyl- or arylalkyl halide (25 mmol) in anhydrous acetone (10 ml) was stirred overnight at room temperature, during which time the crude pyridinium salt separated out of the reaction mixture. In some instances, ethyl ether was added to the solution to precipitate the pyridinium salt. Recrystallization from acetone/ethyl ether gave the desired product as a crystalline solid.

General Procedure for the Preparation of the Oxalate Salts of N-Alkyl- or N-Arylalkyl-4-phenyl-1,2,3,6-tetrahydropyridines. NaBH₄ (1.5 Eq) was added portionwise to the solution of the pyridinium salt (1 Eq) in anhydrous methanol (10 ml) at 0°C. After stirring for 30 min, the solvent was removed under vacuo, and the residue in 20 ml of H₂O was extracted with ethyl ether (2 × 25 ml). The combined organic layer was dried over MgSO₄, filtered, and evaporated, and the residue in 25 ml of dry ethyl ether was treated with oxalic acid (2 Eq). The precipitated oxalate salt was recrystallized from methanol/ethyl ether.

m-Chlorobenzoate salt of N-Heptyl-4-phenyl-1,2,3,6-tetrahydropyridine-N-oxide (22). The oxalate salt of N-heptyl-4-phenyl-1,2,3,6-tetrahydropyridine (15, 0.3 g, 0.87 mmol) was neutralized with 1 N NaOH, and this aqueous solution was extracted with ethyl ether (2 × 10 ml). The combined organic solution was washed with water (2 × 25 ml), dried (MgSO₄), filtered, and then concentrated under reduced pressure. The oily residue was treated with 55% m-chloroperoxybenzoic acid (172 mg, 1 mmol) in dry methylene chloride (10 ml) at 0°C. The reaction mixture was allowed to stir at that temperature for 3 h and then concentrated under reduced pressure. The residue was triturated with ethyl ether (3 × 10 ml) and then stored at 0°C overnight to afford the m-chlorobenzoate salt of 22 as a crystalline white solid (300 mg, 88%). ¹H NMR (CDCl₃)  8.01-8.02 (m, 1H, ArH), 7.9-7.3 (m, 1H, ArH), 7.28-7.45 (m, 7H, ArH), 5.94 (s, 1H, C5 H), 4.61-4.65 (bd, 1H, CH), 4.22-4.25 (bd, 2H, CH₂), 3.66-3.76 (m, 3H, CH₂ and CH), 3.11-3.15 (bd, 1H, CH), 2.70-2.74 (bd, 1H, CH), 1.94-1.95 (m, 2H, CH₂), 1.16-1.32 (m, 8H, CH₂), 0.83-0.84 (t, 3H, CH₃). LC/MS/MS analysis revealed a single peak (R_t = 16.7 min) with a protonated molecular ion at 275 and a base fragment ion at 60.

Enzymology. Recombinant human CYP2D6 was expressed in-house using a baculovirus/Sf9 cell expression system. NADPH was purchased from Sigma-Aldrich, whereas (±)-bufuralol and 1'-hydroxy-(±)-bufuralol were obtained from BD Gentest Corp. (Woburn, MA).

Inhibition of Bufuralol  1'-Hydroxybufuralol Activity in Human Recombinant CYP2D6. In all experiments, the pyridinium salts and their corresponding tetrahydropyridine derivatives were dissolved and diluted serially in methanol. The final concentration of these compounds ranged from 0.05 to 100 µM and that of methanol was less than 0.1% in 200 µl of reaction volume. Each inhibition study was performed in triplicate. Incubation mixtures (200 µl) contained (±)-bufuralol (10 µM) and human recombinant CYP2D6 (10 nM) in 0.1 M phosphate buffer (pH 7.4). The reaction mixtures were prewarmed at 37°C for 2 min before adding NADPH (1.3 mM), then incubated for 15 min. Reactions were stopped by the addition of 0.4 ml of acetonitrile containing metoprolol (1 µM) as internal standard, and samples were centrifuged at 3000g for 15 min. 1'-Hydroxybufuralol formation was linear with respect to protein concentration and time. 1'-Hydroxybufuralol formation was monitored on a Sciex API model 3000 LC/MS/MS triple quadrupole mass spectrometer. Analytes were chromatographically separated using a Hewlett Packard Series 1100 HPLC system (Phenomenex Primesphere 5 µ C₁₈-HC 30 × 2.0 mm column; Phenomenex, Torrance, CA) using a 2- and 3-min binary gradient consisting of a mixture of 95% H₂O/5% acetonitrile with 0.1% acetic acid (solvent A) and 95% acetonitrile/5% H₂O with 0.1% acetic acid (solvent B) and a flow rate of 1.0-1.5 ml/min). Ionization was conducted in the positive ion mode at the ionspray interface temperature of 400°C, and N₂ was used as a nebulizing and heating gas. 1'-Hydroxybufuralol and metoprolol were analyzed using multiple reaction monitoring at mass ranges m/z 278  186 and m/z 268  116, respectively. For 1'-hydroxybufuralol, this reaction follows the protonated parent mass MH⁺ = 278 to its corresponding collision-induced dissociated fragment at m/z 186, which corresponds to dehydration and loss of the N-isopropylamine moiety. IC₅₀ values were determined by fitting the data in Deltagraph (version 4.5; SPSS Science Inc., Chicago, IL). The data listed represent the average values from three determinations.

Determination of Apparent K_m for the CYP2D6 Catalyzed Oxidation of N-Alkyl-4-phenyl-1,2,3,6-tetrahydropyridines. Apparent K_m values were determined using the substrate depletion method (Obach and Reed-Hagen, 2002). Reaction mixtures (0.65 ml) contained recombinant CYP2D6 (0.37 to 4.6 pmol), tetrahydropyridine derivative (0-100 µM), NADPH (0.5 mM), and MgCl₂ (3.3 mM) in 100 mM potassium phosphate buffer (pH 7.4). Reactions were initiated by the addition of NADPH and incubated for 20 min with 75-µl aliquots taken at t = 0, 5, 10, 15, 20, and 30 min. Aliquots were added to 200 µl of acetonitrile containing internal standard (0.5 µg/ml) and centrifuged at 3000g for 15 min. For control experiments, NADPH was omitted. Tetrahydropyridine depletion was measured by LC/MS/MS (Sciex API-3000) analysis of the supernatant (20 µl) using general analytical conditions described for the conversion of bufuralol to 1'-hydroxybufuralol. Analyte/internal standard peak height ratios were determined and normalized to the value obtained at t = 0. The percentage remaining versus time at each substrate concentration were fitted to first order decay functions to determine initial substrate depletion rate constants (k_dep). If substrate decline demonstrated non-linearity on log percentage remaining versus time curves, only those initial timepoints wherein log-linearity was observed were used to determine depletion rate constants.

Metabolite Identification. N-Alkylpyridinium (150 µM) or N-alkyltetrahydropyridine derivatives (30 µM) were incubated with human recombinant CYP2D6 (50 nM) in 0.1 M potassium phosphate buffer (pH 7.4) at 37°C. The reaction mixtures were prewarmed at 37°C for 2 min before adding NADPH (1.3 mM) and then incubated for an additional 30 min. In control experiments, NADPH was omitted from the incubation mixtures. Reactions were terminated by addition of ice-cold acetonitrile (2 volumes) and centrifuged at 3000g for 15 min. The supernatant was evaporated to dryness under a steady nitrogen stream and then reconstituted in 200 µl mobile phase (10 mM ammonium formate, 0.1% formic acid/acetonitrile; 75:25) and analyzed for metabolite formation by LC/MS/MS. All incubations were conducted in triplicate. Assessments of pyridinium and tetrahydropyridine metabolism were conducted on a Sciex API model 2000 LC/MS/MS triple quadrupole mass spectrometer in conjunction with an LDC Analytical SpectroMonitor 3200 variable wavelength UV-detector. Analytes were chromatographically separated using a Hewlett Packard Series 1100 HPLC system (Palo Alto, Ca). An autosampler was programmed to inject 50 µl of the sample on a Zorbax Rx-C₈ 4.6 × 150 mm column using a binary gradient consisting of a mixture of 10 mM ammonium formate, 0.1% formic acid (solvent A) and acetonitrile (solvent B) at a flow rate of 1 ml/min. The LC gradient was programmed as follows: solvent A to solvent B ratio was held at 100:0 (v/v) for 3 min and then adjusted from 100:0 (v/v) to 10:90 (v/v) for 20 min and from 10:90 (v/v) to 100:0 (v/v) from 20 to 25 min. The column was re-equilibrated for 5 min prior to the next analytical run. Postcolumn flow was split such that mobile phase was introduced into the mass spectrometer via an ion spray interface at a rate of 50 µl/min. The remaining flow was diverted to the UV detector positioned in line to provide simultaneous UV detection ( = 254 nM) and total ion chromatogram. Ionization was conducted in the positive ion mode at the ionspray interface temperature of 150°C and using nitrogen for nebulizing and heating gas. Ion spray voltage was 4.5 kV, and the orifice voltage was optimized at 40 eV. Initial Q1 scans were performed between m/z 50 to 500. Metabolites were identified by comparing t = 0 samples to t = 30 min samples (with or without NADPH), and structural information was generated from collision-induced dissociation of the corresponding protonated molecular ions.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Chemistry. All of the N-substituted-4-aryl-1,2,3,6-tetrahydropyridine derivatives were prepared by sodium borohydride reduction of the corresponding N-substituted-4-arylpyridinium intermediates, which in turn were obtained by quaternization of commercially available 4-phenylpyridine with alkyl halides (Fig. 1). N-Substituted-4-aryl-1,2,3,6-tetrahydropyridines were isolated and characterized as their corresponding oxalate salts. N-oxidation of N-heptyl-4-phenyl-1,2,3,6-tetrahydropyridine (15) free base with m-chloroperoxybenzoic acid afforded the corresponding N-oxide 22 that was isolated and characterized as the corresponding m-chlorobenzoate salt (see Fig. 1).

View larger version (9K): [in this window] [in a new window]   Fig. 1.   Synthesis of N-alkyl-4-phenyl-pyridinium and -1,2,3,6-tetrahydropyridine derivatives and the putative N-oxide metabolite of N-heptyl-4-phenyl-1,2,3,6-tetrahydropyridine.

Inhibition of (±)-Bufuralol  (±)-1'-Hydroxybufuralol Activity in Human Recombinant CYP2D6. IC₅₀ values for inhibition of CYP2D6-mediated (±)-bufuralol hydroxylation by N- and C-4-substituted pyridiniums and 1,2,3,6-tetrahydropyridines were measured in triplicate using recombinant enzyme and are displayed in Table 1 (pyridiniums), Table 2 (1,2,3,6-tetrahydropyridines), and Table 3 (HP and its metabolites), along with the corresponding clog P and pKa values where applicable. Plots of clog P versus log[IC₅₀] for pyridiniums (Fig. 2, panel A) and 1,2,3,6-tetrahydropyridines (Fig. 2; panel B) are also shown. Under the present experimental conditions, the specific CYP2D6 inhibitor quinidine inhibited bufuralol-1'-hydroxylase activity with an IC₅₀ value of 0.05 µM, consistent with that reported in previous publications (Mankowski, 1999).

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Plot of log [IC50] for the inhibition of CYP2D6 catalyzed bufuralol-1'-hydroxylation by straight chain N-alkyl- and N-aryl-4-phenylpyridinium derivatives () and -1,2,3,6-tetrahydrpyridine () derivatives versus clog P.  Log [IC50] = 1.76 clog P + 1.13; r2 = 0.78 (pyridiniums)

CYP2D6 Inhibition by N-Alkyl- and N-Arylalkyl-4-phenylpyridiniums. The CYP2D6 inhibitory potency of HPP⁺, the pyridinium metabolite of HP, against (±)-bufuralol-1'-hydroxylase activity was ~6-fold greater than that reported for the inhibition of the corresponding CYP2D6-mediated dextromethorphan O-demethylation [IC₅₀(bufuralol-1'-hydroxylase) = 2 µM; IC₅₀(dextromethorphan O-demethylation = 0.34 µM] (Shin et al., 2000). In contrast, the comparatively less lipophilic MPP⁺ did not inhibit the bufuralol-1'-hydroxylase activity of CYP2D6 at the concentration range employed (IC₅₀ > 100 µM).

CYP2D6 Inhibition by N-Alkyl- and N-Arylalkyl-4-phenyl-1,2,3,6-tetrahydropyridines. The good correlation between lipophilicity and inhibitory potency observed with the straight chain N-alkyl-4-phenylpyridiniums was maintained only with the short  medium straight chain N-alkyl-4-phenyl-1,2,3,6-tetrahydropyridines (Table 2). Thus, less lipophilic tetrahydropyridines such as MPTP and PTP as well as the neutral 4-phenylpyridine did not inhibit CYP2D6, but replacement of the N-methyl group in MPTP with a N-propyl substituent (compound 12) resulted in a significant increase in potency (MPTP, IC₅₀ > 100 µM; N-propyltetrahydropyridine (12), IC₅₀ = 10 µM). Unlike N-alkyl-4-phenylpyridinium SAR, where introduction of a terminal polar hydroxyl group was detrimental for inhibition, the corresponding N-(3-hydroxypropyl)-4-phenyl-1,2,3,6-tetrahydropyridine derivative 13 retained the inhibitory potency of 12. That tetrahydropyridine 13 inhibited CYP2D6 also was somewhat surprising considering that its lipophilicity was comparable with MPTP and PTP, which were inactive as inhibitors. Although increasing the alkyl chain length in 12 also resulted in increased inhibition, the optimal alkyl chain length in the tetrahydropyridine series comprised of 5 carbons as opposed to 11 carbon atoms in the pyridinium series and further increases in alkyl chain length did not appear to significantly influence inhibition.

CYP2D6 Inhibition by Haloperidol and its Metabolites. Consistent with previous reports on the inhibition of the CYP2D6-catalyzed dextromethorphan O-demethylation reaction (Shin et al., 2000), HP and several of its metabolites including (±)-RHP, HPTP, and (±)-RHPTP also inhibited the CYP2D6 catalyzed bufuralol-1'-hydroxylation pathway (Table 3). The IC₅₀ values for the inhibition of the bufuralol-1'-hydroxylase activity by HP (clog P = 3.85), (±)-RHP (clog P = 3.28), HPTP (clog P = 5.54), and (±)-RHPTP (clog P = 4.81) were 6.5, 0.4, 5.7, and 0.08 µM, respectively (see Table 3), whereas those for the inhibition of dextromethorphan O-demethylation by HP, (±)-RHP, and (±)-RHPTP were 5.7, 0.89, and 1.34 µM, respectively. Interestingly, RHPTP was ~17-fold more potent as a inhibitor of bufuralol-1'-hydroxylation when compared with its inhibitory effects against the dextromethorphan O-demethylase activity of the enzyme. Furthermore, increased lipophilicity of HP metabolites did not necessarily correlate with inhibitory potency since the more lipophilic HP metabolite, HPTP, was less potent as a CYP2D6 inhibitor than the relatively less lipophilic (±)-RHP and (±)-RHPTP metabolites. Also of some interest were the findings that the reduced HP and HPTP metabolites were ~6.0- and 71-fold more potent than the corresponding oxidized compounds (HP and HPTP).

Effect of Lipophilicity on the CYP2D6 Substrate Properties of N-Alkyl-4-phenyl-1,2,3,6-tetrahydropyridines. The effect of lipophilicity on the CYP2D6 substrate properties of several N-alkyl-4-phenyl-1,2,3,6-tetrahydropyridines was also assessed. The exercise was only limited to N-alkyl substituents of increasing chain length (increased clog P) and the N-phenethyl-4-phenyl-1,2,3,6-tetrahydropyridine analog 21 that had demonstrated potent CYP2D6 inhibition. Since authentic standards of all metabolites were unavailable, the apparent K_m values for N-alkyl/N-arylalkyl-4-phenyl-1,2,3,6-tetrahydropyridine derivatives as CYP2D6 substrates were assessed using the substrate depletion methodology (extension of the in vitro t_1/2 method traditionally used in intrinsic clearance predictions) (Obach and Reed-Hagen, 2002). Recombinant CYP2D6 was incubated with the candidate 1,2,3,6-tetrahydropyridine (0-100 µM) in the presence of NADPH, and aliquots were analyzed for substrate loss versus time. The percentage of the 1,2,3,6-tetrahydropyridine remaining versus time at each substrate concentration was fitted to first order decay functions to determine initial substrate depletion rate constants (k_dep). In theory, as the substrate concentration exceeds K_m, the depletion rate constant declines, and the inflection point of this relationship represents the apparent K_m value.

Structural Elucidation of Metabolites Derived from the CYP2D6-Catalyzed Oxidation of N-Alkyl-4-phenyl-1,2,3,6-Tetrahydropyridines and Pyridiniums. N-Alkyl-4-phenyl-1,2,3,6-tetrahydropyridines. Attempts were also made to identify the metabolites obtained following CYP2D6-mediated oxidation of N-alkyl-4-pheny-1,2,3,6-tetrahydropyridines (Table 4). Consistent with previous reports (Modi et al., 1997), CYP2D6-catalyzed oxidation of MPTP resulted in the formation of PTP (MH⁺ = 160, R_t = 9.40 min) and N-4-(4'-hydroxylphenyl)-N-methyl-1,2,3,6-tetrahydropyridine (MH⁺ = 190, R_t = 6.06 min), the structures of which were confirmed via CID of the respective protonated molecular ions and by comparison of the LC R_t with those of the synthetic standards. Likewise, LC/MS/MS analysis of the CYP2D6/N-propyl derivative (12, MH⁺ = 202, R_t = 10.4 min) mixture indicated that the protonated molecular mass of the more polar metabolite (MH⁺ = 218, R_t = 8.47 min) was 16 mass units higher than the parent tetrahydropyridine. The CID spectrum of parent and metabolite produced a common base fragment at m/z + H⁺ = 72 consistent with the molecular mass of the protonated N-propyliminium ion suggesting that parahydroxylation had most likely occurred on the C-4 aromatic ring in 12 as was observed with MPTP and in rat liver microsomal incubations with the related N-cyclopropyl-4-phenyl-1,2,3,6-tetrahydropyridine analog (Kuttab et al., 2001). The structure of the less polar metabolite corresponded to PTP. Besides aromatic hydroxylation and N-demethylation, trace quantities of N-propyl-4-phenylpyridinium (M⁺ = 190, m/z + H⁺ = 156; R_t = 9.84 min) were also discernible in the incubation mixture.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   CID spectra of the synthetic standard of N-heptyl-4-phenyl-1,2,3,6-tetrahydropyridine-N-oxide [MH+ = 274; m/z = 60 (100%)], the N-hydroxyheptyl-4-phenyl-1,2,3,6-tetrahydropyridine metabolite [MH+ = 274; m/z = 144 (100%)] derived from the CYP2D6-catalyzed oxidation of N-heptyl-4-phenyl-1,2,3,6-tetrahydropyridine, and parent N-heptyl-4-phenyl-1,2,3,6-tetrahydropyridine [MH+ = 258; m/z = 128 (100%)].

N-Alkyl-4-phenylpyridiniums. Interestingly, at high concentrations (200 µM), qualitative LC/MS/MS analysis of some of the lipophilic long-chain N-alkyl-4-phenylpyridinium derivatives (e.g., heptyl, octyl, nonyl, undecyl, and tetradecyl) also demonstrated the formation of trace levels of monohydroxylated metabolites following incubation with NADPH-supplemented recombinant CYP2D6. Thus, LC/MS/MS analysis of incubation mixtures revealed the formation of a single polar metabolite with molecular weight 16 mass units higher than the respective masses of the parent compounds. The CID spectrum indicated a base fragment at MH⁺ = 156 consistent with the molecular mass of protonated 4-phenylpyridine [4-(C₆H₅)-C₅H₅N⁺] suggesting that CYP2D6-mediated hydroxylation of the long-chain N-alkylpyridiniums had exclusively occurred on the long-chain alkyl substituent as observed with the corresponding long-chain alkyl tetrahydropyridines.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

The present study clearly establishes the importance of lipophilicity in influencing CYP2D6 inhibition by quaternary pyridinium derivatives. Results from the QSAR analysis indicate that the hydrophilic nature of MPP⁺ is mainly responsible for the lack of CYP2D6 inhibition, since chain length extension of its N-methyl group to higher alkyl homologues substantially increases inhibitory potency. As the size of the N-alkyl group increases so does lipid solubility that results in increased cell penetration and allows access to the active site of the enzyme. QSAR studies revealed that a critical chain length, equivalent to a minimal degree of lipophilicity (clog P  0), is essential for CYP2D6 inhibition and is further supported by a reasonably good correlation between the log IC₅₀ values and the calculated lipophilicities of N-alkyl-4-phenylpyridiniums as well as the positively charged pyridinium metabolite of haloperidol (r² = 0.78; Fig. 2). An outlier in this correlation was the N-phenethyl-4-phenylpyridinium analog 11 of identical lipophilicity as the N-pentyl-4-phenylpyridinium analog 4 but ~12-fold more potent as a CYP2D6 inhibitor, suggesting that subtle differences exist in the active site binding of straight chain N-alkyl- and N-arylalkylpyridiniums. Detailed QSAR studies comparing CYP2D6 inhibition by a series of N-alkyl and N-arylalkylpyridiniums of similar lipophilicity should help in resolving this discrepancy. Overall, the identification of lipophilicity as an important physiochemical parameter for CYP2D6 inhibition by cyclic quaternary amines is consistent with previous reports on secondary amine-containing -adrenergic blockers (Ferrari et al., 1991) and N-alkylamine-substituted warfarin O-methyl ethers (Venhorst et al., 2000), which also indicated strong positive correlations between increased lipophilicity and low IC₅₀ values for CYP2D6 inhibition.

The excellent correlation between lipophilicity and CYP2D6 inhibitory potency of the straight chain N-alkyl-4-phenylpyridiniums was also maintained with the straight chain N-alkyl-4-phenyl-1,2,3,6-tetrahydropyridines, where alkyl chain length extensions of the N-methyl group in MPTP revealed dramatic improvements in inhibitory potency. Noteworthy differences in CYP2D6 inhibition by pyridiniums and 1,2,3,6-tetrahydropyridines included the observation that the optimal alkyl chain length in the 1,2,3,6-tetrahydropyridine series comprised of 5 carbons as opposed to 11 carbon atoms in the pyridinium series. Furthermore, in contrast with the lack of CYP2D6 inhibition by N-(3-hydroxypropyl)-4-phenylpyridinium, the corresponding N-(3-hydroxypropyl)tetrahydropyridine analog was a potent CYP2D6 inhibitor. Collectively, these findings suggest that N-alkyl side chains in pyridiniums and 1,2,3,6-tetrahydropyridines must bind in different regions of the protein, a feature that could be explored via specific SAR studies on heteroatom-containing N-alkyl substituents in pyridiniums and 1,2,3,6-tetrahydropyridines. Furthermore, the observation that N-(3-hydroxypropyl)tetrahydropyridine inhibits CYP2D6 is surprising, considering that its lipophilicity is comparable with MPTP and PTP that are inactive as CYP2D6 inhibitors. Thus, it appears that, in addition to lipophilicity, CYP2D6 inhibition is also influenced by the nature of the N-alkyl substituent within the 1,2,3,6-tetrahydropyridines. Collectively, this is reflected in a poor correlation between lipophilicity and inhibitory potency of this class of cyclic tertiary amines. An additional example of this behavior is evident upon comparison of the CYP2D6 inhibitory properties of HP and HPTP versus those of reduced HP and reduced HPTP. The less lipophilic-reduced compounds were ~6.0- and 71-fold more potent than the corresponding oxidized ones suggesting that presence of a hydrogen bond donor (OH group) instead of hydrogen bond acceptor (C = O) adjacent to the 4-fluorophenyl ring increases inhibitory potency. Active site docking of candidate pyridiniums and 1,2,3,6-tetrahydropyridines in the CYP2D6 homology model, especially within the lipophilic pocket, that has been predicted to bind alkyl substituents in 4-N-alkylaminomethyl-7-methoxycoumarinyl derivatives (Venhorst et al., 2000), amphetamines (de Groot et al., 1999a), and the H₁ receptor antagonist terfenadine (Jones et al., 1998), may help resolve some of these issues.

An interesting observation was also noted in the course of inhibition studies with HP metabolites in that the IC₅₀ values for inhibition of the bufuralol-1'-hydroxylase activity of CYP2D6 by (±)-RHPTP and HPP⁺ were ~16- and 6-fold higher than previously reported for inhibition of the dextromethorphan O-demethylase activity. In contrast, IC₅₀ values for CYP2D6 inhibition of HP and (±)-RHPTP were substrate-independent. This finding raises the possibility that CYP2D6 inhibition is substrate-dependent in a manner similar to that reported for the CYP3A4 isozyme (Wang et al., 2000; Schrag and Wienkers, 2001). Comparing the inhibitory potencies of CYP2D6 inhibitors against diverse CYP2D6 substrates can confirm this hypothesis.

As observed in the course of the inhibition studies, increasing the length of the N-alkyl substituent in MPTP vastly improved the CYP2D6 substrate properties of the resulting analogs such that N-heptyl-4-phenyl-1,2,3,6-tetrahydropyridine demonstrated a >100-fold decrease in K_m compared with the parent. It is interesting to point out that in many cases, 1,2,3,6-tetrahydropyridines were better CYP2D6 inhibitors than substrates as reflected by the significantly higher K_m compared with the IC₅₀ values. Similar observations have been noted with other CYP2D6 substrates and inhibitors including the antiemetic drug metoclopramide, which is a better inhibitor (IC₅₀ = 1 µM) than substrate (K_m = 68 µM) (Desta et al., 2002). A potential explanation for this anomaly could be that 1,2,3,6-tetrahydropyridines inhibit CYP2D6 in a partially competitive fashion as opposed to purely competitive inhibition that generally results in similar inhibition (K_i) and Michaelis-Menten (K_m) constants. To verify this proposal, detailed kinetic characterization of CYP2D6 inhibition by 1,2,3,6-tetrahydropyridines and pyridiniums is currently underway.

Although, the lipophilicities of HPTP and reduced HPTP were comparable with the N-heptyltetrahydropyridine derivative, the HP metabolites were inactive as CYP2D6 substrates. This observation suggests that steric constraints rather than lipophilicity are responsible for the lack of CYP2D6 substrate properties of cyclic tertiary amines tethered to bulky N-arylalkyl substituents such as those found in HP and its metabolites. This phenomenon appears to be a common theme among cyclic tertiary amine-containing anti-depressants such as trazodone (Rotzinger et al., 1998), nefazodone (von Moltke et al., 1999), buspirone (Lilja et al., 1998), and gepirone (von Moltke et al., 1998) (Fig. 4). All of these drugs contain bulky N-alkyl substituents and are inactive as CYP2D6 substrates. Like HP and its metabolites, these compounds are exclusively metabolized by CYP3A4. Whether this property could be exploited as a general strategy for designing central nervous system agents devoid of CYP2D6 substrate properties remains to be explored.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Examples of cyclic tertiary amines that are inactive as CYP2D6 substrates.

Finally, metabolite identification studies indicated that N-alkyl-4-phenyl-1,2,3,6-tetrahydropyridines underwent monohydroxylations on the aromatic ring and on the N-alkyl substituent. Given the clear preference of CYP2D6 for aromatic hydroxylation (5-7 Å from the basic amine nitrogen), hydroxylation on the N-alkyl substituent (medium to long-chain N-alkyltetrahydropyridines) or N-demethylation (MPTP) seems to be a paradox, because it suggests that 1,2,3,6-tetrahydropyridines can bind in the completely opposite or "upside-down" orientation in a single active site model. An alternate explanation for CYP2D6-mediated N-dealkylation/N-alkylhydroxylation reactions is that the protonated amine binds to Asp301 but is rapidly deprotonated to yield an uncharged amine for single electron oxidation (Grace et al., 1994). Our preliminary observation on the weak CYP2D6 substrate properties of the long-chain N-alkylpyridinium compounds suggest that additional SAR on the N- and C-4 substituent could potentially provide a good quaternary amine-based CYP2D6 substrate, which should prove valuable in investigating these proposals.