Abstract
Recently, a novel nonfluorescent probe 3-[2-(N,N-diethyl-N-methylammonium)-ethyl]-7-methoxy-4-methylcoumarin (AMMC), which produces a fluorescent metabolite AMHC (3-[2-(N,N-diethyl-N-methylammonium)ethyl]-7-hydroxy-4-methylcoumarin) was used with microsomes containing recombinant enzymes (rCYP) to monitor CYP2D6 inhibition in a microtiter plate assay. This article describes the studies that were performed in human liver microsomes (HLM) to establish the selectivity of AMMC toward CYP2D6. Metabolism studies in HLM showed that AMMC was converted to one metabolite identified by mass spectrometry as AMHC. Kinetic studies indicated an apparent Km of 3 μM with aVmax of 20 pmol/min · mg of protein for the O-demethylation reaction. TheO-demethylation of AMMC in HLM was inhibited significantly in the presence of a CYP2D6 inhibitory antibody. Using a panel of various HLM preparations (n = 12), a good correlation (r2 = 0.95) was obtained between AMMC O-demethylation and bufuralol metabolism, a known CYP2D6 substrate, but not with probes for the other major xenobiotic metabolizing CYPs. Finally, only rCYP2D6 showed detectable metabolism in experiments conducted with rCYPs using AMMC at a concentration of 1.5 μM (near Km). However, at a concentration of 25 μM AMMC, rCYP1A also contributed significantly to the formation of AMHC. Knowing the experimental conditions under which AMMC was selective for CYP2D6, a microtiter assay was developed to study the inhibition of various compounds in HLM using the fluorescence of AMHC as an indication of CYP2D6 activity. The inhibition potential of various chemicals was found to be comparable to those determined using the standard CYP2D6 probe, bufuralol, which requires high-performance liquid chromatography separation for the analysis of its CYP2D6-mediated 1′-hydoxylated metabolite.
Cytochromes P450 (CYPs1) comprise a superfamily of hemoproteins that play an important role in the metabolism of a wide variety of xenobiotics and endogenous compounds. The human CYP genes have been characterized and classified into various families and subfamilies based on their structures. The gene families CYP1, CYP2, and CYP3 are currently thought to be the major CYP enzymes responsible for the oxidative metabolism of drugs or xenobiotics (Spatzenegger and Jaeger, 1995).
Inhibition of CYP-mediated metabolism, often the mechanism for drug-drug interactions, can limit the use of a drug because of adverse clinical effects. In past years, substantial progress has been made in the development of in vitro tools that can be used by the pharmaceutical industry to predict drug-drug interactions. The potential for CYP enzyme inhibition is addressed routinely by measuring the rates of metabolism of a probe biotransformation (a specific substrate to a specific metabolite) in human liver microsomes (HLM) or heterologously expressed enzymes (rCYP) in the presence and absence of new chemical entities (NCEs) (as summarized in Parkinson, 1996). Usually, multiple inhibitor and/or probe substrate concentrations are tested to generate quantitative inhibition parameters [apparent inhibition constant (Ki) or inhibitor concentration that produces 50% inhibition (IC50value)].
A limitation with most of the existing probes is that they require HPLC analysis for their metabolite quantification, which severely limits sample throughput. Also, it is not uncommon to observe interference of the inhibitor with the assay endpoint (e.g., coelution with the metabolite) when UV detection is employed. The use of LC with mass spectrometry (MS) detection obviously circumvents the selectivity issue,but the sample preparation, including the selection of appropriate internal standard, can still be problematic. In general, analytical methodology is the rate-limiting step in the acquisition of in vitro data.
Recently, efforts have been made to develop microtiter plate assays with nonfluorescent CYP probes that produce fluorescent metabolites (Crespi et al., 1997; Chauret et al., 1999). Most published assays refer to the use of these fluorogenic probes in microsomes expressing individual CYPs (rCYP) because the selectivity of the probes is poor or unknown (Crespi et al., 1997). In this regard, a novel fluorogenic probe 3-[2-(N,N-diethyl-N-methyl-ammonium)ethyl]-7-methoxy-4-methylcoumarin (AMMC; see Fig. 1) was reported to be a useful substrate to address CYP2D6 inhibition in a microtiter plate assay using rCYP2D6 microsomes (Crespi et al., 1999).
This article reports the studies that were conducted in HLM to address the selectivity of AMMC toward CYP2D6. Results generated with AMMC are compared with those obtained with bufuralol, the classical probe that is known to be hydroxylated at the 1′-position by CYP2D6 in HLM (Kronbach et al., 1987).
Experimental Procedures
Materials.
Quinidine, dextromethorphan, perhexeline, fluoxetine, imipramine, sparteine, norfluoxetine, and quinine were purchased from Sigma/RBI (St. Louis, MO). The compounds AMMC, bufuralol, 3-[2-(N,N-diethylamino)ethyl]-7-hydroxy-4-methylcoumarin (AHMC), and 1′-OH-bufuralol were from GENTEST Corp. (Woburn, MA) and BD Biosciences (San Jose, CA). Inhibitory monoclonal antibodies against CYP2D6 prepared in mouse ascites were obtained from Merck (Whitehouse Station, NJ) (Shou et al., 2000). All other reagents were of highest purity commercially available or HPLC grade.
Tissue and Microsomes.
Human tissues were obtained from various sources (F.P. Guengerich, Vanderbilt University School of Medicine, Nashville, TN; IIAM, Exton, PA; Quebec Transplant, Montreal, QC, Canada). Human liver microsomes were either from GENTEST Corp., BD Biosciences, or prepared from frozen (−80°C) tissue as described in the literature (Lu and Levin, 1972). Protein concentrations of the microsomal fractions were determined by the method of Lowry et al. (1951) using bovine serum albumin as a standard. Microsomes prepared from lymphoblast or baculovirus/insect cells cDNA expressing individual cytochrome (Supersomes) were obtained from GENTEST Corp and BD Biosciences.
Conditions for Individual Incubations.
Microsomal incubations were prepared on ice in Eppendorf polypropylene tubes containing an appropriate amount of human liver microsomal protein, the substrate, and a typical NADPH-regenerating system (100 mM phosphate buffer at pH 7.4 containing 20 mM glucose 6-phosphate, 2.0 mM NADP+, 2.0 mM magnesium chloride, and 4 units/ml glucose-6-phosphate dehydrogenase) in a total volume of 500 μl. The substrate, AMMC, was dissolved in acetonitrile in such concentrations that the total organic solvent content did not exceed 1% when added to microsomal incubations, because it is known that organic solvents can affect the activity of the enzyme (Chauret et al., 1998; Hickman et al., 1998; Busby et al., 1999). Incubations in phosphate buffer were used as controls, and incubations with no AMMC were used as blanks. The tubes were transferred then to a water bath set at 37°C. After a specific period of time, the reactions were quenched by adding an equivalent volume of acetonitrile to precipitate the proteins. The incubation mixtures were then centrifuged for 10 min at 13,000 rpm in an Eppendorf centrifuge 5415C. An aliquot of the supernatant was analyzed by HPLC/UV, HPLC/MS, or HPLC/fluorescence. To evaluate the linearity of the reaction rate for AMMC, the amount of microsomal protein and incubation time varied from 0 to 2 mg/ml and 0 to 90 min, respectively. To determine the kinetic parameters (apparentKm and Vmaxusing Lineweaver-Burk calculations), incubations containing 0.5 mg/ml microsomal protein (obtained from pooled donors) were conducted for 45 min with various concentrations AMMC (0–30 μM).
Incubations with bufuralol were conducted similar to those described above except that 0.8 mg/ml microsomal protein was used with 25 μM bufuralol, and the incubation time was 20 min. The reactions were stopped by adding 50 μl of aqueous perchloric acid (70%, v/v). The incubation mixtures were then centrifuged, and the supernatants were analyzed by HPLC/fluorescence.
Inhibition assays with CYP2D6 antibody were conducted using 0.5 mg/ml microsomal protein (individual or pooled donors) or rCYP2D6 (10 pmol) and 10 μl/ml CYP2D6 monoclonal antibody or vehicle. The microsomes and the antibody were kept at room temperature for 15 min before the incubations were conducted with 6 or 25 μM substrate AMMC as described above.
Incubations with the recombinant CYPs contained 10 pmol of the rCYP (except CYP2D6 where the enzyme content was 1.5 pmol CYP) in 0.2-ml total volume, 1.5 or 25 μM AMMC with a low NADPH-regenerating system consisting of 0.0081 mM NADP+, 0.41 mM glucose 6-phosphate, and 0.4 units/ml glucose-6-phoshate dehydrogenase. The incubation time was 1 h.
HPLC Analysis of Incubations.
To investigate the overall metabolism profile of AMMC, analysis of the microsomal incubations was carried out using LC/MS on a Waters Alliance 2690 HPLC interfaced to a Micromass Quattro LC triple quadrupole mass spectrometer (Waters Corp., Milford, MA). The sample (5 μl) was injected onto a YMC-AQ C18 column (4.6 × 50 mm, Waters Corp.). A gradient mobile phase consisting initially of 90:10, solvent A (0.1% formic acid in water) and solvent B (0.1% formic acid in acetonitrile) was brought to a composition of 60:40 A/B in 10 min at a flow rate of 1 ml/min. Positive ion electrospray over the range m/z 200 to 400 was used for detection. Selected incubations were analyzed also by UV (Waters Photodiode Array, model 994) or fluorescence (Shimadzu RF-551, Kyoto, Japan). UV detection was monitored at 325 nm and fluorescence was read at an excitation wavelength of 395 nm and an emission wavelength of 460 nm.
For quantitative analysis of the parent AMMC and metabolite AMHC (3-[2-(N,N-diethyl-N-methylammonium)ethyl]-7-hydroxy-4-methylcoumarin) in the kinetic experiments (Km andVmax determinations), a Waters Alliance 2790 HPLC coupled to a Micromass Quattro LC mass spectrometer was used. A gradient system similar to the one described above was chosen except that a Phenomenex Luna C18 column (2.0 × 50 mm; Torrance, CA), a flow rate of 1.5 ml/min, and an injection volume of 25 μl were used for analysis. Detection of substrate and metabolite used selected reaction monitoring of the compound specific transitions of m/z 304.2 to 217 andm/z 290.2 to 203, respectively. This corresponds to the loss of the diethyl methylamine moiety (Fig.2). An external calibration curve using the compound AHMC (refer to Fig. 1) was used for the indirect quantification of AMHC because this metabolite is not available commercially. Detection of AHMC used selected reaction monitoring of the compound specific transition of m/z 276.1 to 203.
Analysis of microsomal incubates (50 μl) of bufuralol was performed on a Waters system using a Zorbax C18 column (Chromatographic Specialties, Brockville, ON, Canada). A gradient mobile phase consisting initially of 85:15 (v/v), A (1 mM aqueous HClO4)/B (acetonitrile) was brought to a composition of 60:40, v/v, A/B in 10 min at a flow rate of 2 ml/min. Fluorescence detection (Shimadzu RF-551, Kyoto, Japan) was achieved at an excitation wavelength of 252 nm and an emission wavelength of 302 nm.
Conditions for Inhibition Studies.
A) Incubations using the fluoregenic probe AMMC
Incubations using HLM were conducted in 96-well microtiter plates using experimental conditions reported in Table1 and according to published methodology (Chauret et al., 1999). A Biomek 1000 automated laboratory workstation controlled with Nemesis version 1.0.5 software (Beckman Coulter, Inc., Fullerton, CA), was used for serial (1:3, v/v) dilutions of the test compounds and for the addition of all of the reagents in the various steps of the incubations (microsomes in buffer, substrate, inhibitor, and NADPH, 0.1-ml total volume). An enzymatic reaction, as described in Chauret et al. (1999), was performed at the end of the incubation to remove excess NADPH because the fluorescence of this reagent interferes with the detection of AMHC. Reactions were terminated by the addition 0.1 ml of 4:1 (v/v), acetonitrile/Tris base solution (0.5 M). Fluorescence of samples was monitored using a cytofluorimeter (Cytofluor II, Applied Biosystems, version 4.1 or 4.2; Foster City, CA) with excitation and emission filters set at 395/40 and 460/40 nm, respectively. Typically, ten concentrations of inhibitors were tested (0.03–100 μM) for all inhibitors except quinidine (0.0015–5 μM). The inhibitors were selected on the basis that they covered a wide range of IC50 values for CYP2D6 inhibition (Crespi et al., 1999). The analysis and calculations of the microtiter plate data to generate IC50values were achieved as described previously (Chauret et al., 1999).
Incubations using rCYP2D6 were conducted in 96-well microtiter plates using experimental conditions given in Table 1 and were based on reported methods (Crespi and Stresser, 2001). A Multiprobe II liquid handling station (Packard Instruments, Downers Grove, IL) was used for the serial 1:3 dilutions (v/v) of the test compounds in the low NADPH-regenerating system (as described above). The low NADP+ concentration was used to minimize interference due to the fluorescence of NADPH without the need to enzymatically eliminate the NADPH postincubation. Reactions were initiated by the addition 0.1 ml of prewarmed enzyme and substrate to 0.1 ml of the inhibitor solution containing the low NADPH-regenerating system and were terminated by the addition of 0.075 ml 4:1 (v/v) of acetonitrile/Tris base (0.5 M). Fluorescence readings were obtained using a FLUOstar model 403 fluorescence plate reader (BMG Lab Technologies Inc., Durham, NC) using an excitation and emission wavelength of 390 and 460 nm, respectively. Data were exported and analyzed using an Excel spreadsheet. The IC50values were calculated by linear interpolation.
B) Incubations with the standard probe, bufuralol.
Inhibition studies using bufuralol as a probe were performed similar to those described previously except that the specific experimental conditions summarized in Table 1 were used. Typically, five concentrations of inhibitors [0.1–100 μM for all inhibitors, except quinidine (0.01–10 μM)] were used to generate an IC50 value. Because both bufuralol and its 1′-hydroxy metabolite are fluorescent, HPLC analysis, as described previously, was used for the quantification of the metabolite.
Results
Metabolism Profile of AMMC.
Human liver microsome incubations containing AMMC resulted in the formation of one major metabolite as detected by UV, fluorescence, or MS analysis (Fig. 2). Based on MS analysis, the metabolite M1 (m/z = 290.2) corresponded to the loss of 14 atomic mass units from the parent compound (m/z = 304.2) consistent with O- or N-demethylation (–CH3 + H). Based on the fragmentation pattern, the amino alkyl side chain of the metabolite was not modified by metabolism, indicating that the demethylation had occured on the coumarin moiety. From this MS evidence, the metabolite M1 was identified as AMHC. The rate of metabolism of AMMC was linear with time and protein concentration up to 45 min and at least 1 mg of protein. Using a microsomal protein concentration of 0.5 mg/ml (obtained from pooled human liver donors), the apparent Km andVmax were determined to be ∼3 μM and 20 pmol/(min · mg of protein), respectively.
Involvement of CYP in the Metabolism of AMMC.
The involvement of CYP in the O-demethylation of AMMC was determined using several approaches including metabolism studies (using microsomes containing recombinant enzymes and human liver microsomes from various donors) and inhibition studies.
Metabolism studies with microsomes prepared from cell lines expressing a single CYP (Supersomes) are reported in Fig.3. At a substrate concentration of 25 μM, the major enzyme involved in the O-demethylation of AMMC was found to be CYP2D6 with minor contributions from CYP1 (15 and 8% relative to CYP2D6 for CYP1A1 and CYP1A2, respectively), CYP1B1 (10% relative to CYP2D6), and CYP2B6 (5% relative to CYP2D6). At a concentration of 1.5 μM substrate, CYP2D6 was the only observable enzyme involved in the metabolism of AMMC. All the other enzymes tested gave <1% AMHC relative to CYP2D6.
Inhibition studies with a specific inhibitory antibody against CYP2D6 are illustrated in Fig. 4. Significant inhibition of the AMMC O-demethylation activity was observed in the presence of the antibody (80 and 96% inhibition at 25 and 6 μM AMMC, respectively).
Phenotyped human liver microsomes (n = 12) were used to compare the AMMC-O-demethylase with the activity of selective CYP substrates. As shown in Fig.5, a good correlation was obtained between the rate of AMMC-O-demethylase and bufuralol-1′-hydroxylase activities (r2 = 0.95 and 0.91, at 1.5 and 25 μM AMMC, respectively). AMMC-O-demethylase activity at either substrate concentration did not correlate with the levels of phenacetin O-deethylation (CYP1A), coumarin 7-hydroxylation (CYP2A6), (S)-mephenytoin-hydroxylation (CYP2C19), chlorzoxazone 6-hydroxylation (CYP2E1), testosterone 6β-hydroxylation (CYP3A), or lauric acid 12-hydroxylation (CYP4A) (r2 values all at or below 0.3).
Inhibition of CYP2D6.
The 96-well plate assay, as described above, was used to evaluate the inhibitory potential of various chemical inhibitors using AMMC as a probe. With an HLM preparation high in CYP2D6 (as determined in previous studies; (Chauret et al., 1997), the signal corresponding to 100% activity (incubations containing HLM, AMMC, and NADPH) was approximately 3-fold above the signal corresponding to 0% activity (incubations containing HLM, AMMC without NADPH). In Table2, the IC50 values obtained for various inhibitors in HLM using the fluorescent probe AMMC are compared with those determined with the classical probe bufuralol. There was a good correlation between the values obtained with both probes (r2 = 0.98, n = 8). The IC50 values obtained with AMMC and bufuralol using rCYP2D6 as the enzyme source were also determined, and the results are presented in Table 2. Once again, there was a good correlation between the values obtained with both probes (r2 = 0.96, n = 7).
Discussion
This paper describes the metabolism studies that were performed in HLM to determine the selectivity of AMMC, a novel fluorescent probe (Crespi et al., 1999), toward CYP2D6. When AMMC was incubated in human liver microsomes, one major metabolite was detected using UV, fluorescence, or MS detection. It was identified as theO-desmethyl metabolite, AMHC, by LC-MS. The AMMC-O-demethylation in HLM occurred at an apparentKm of 3 μM and aVmax of 20 pmol/(mg of protein · min).
It was clear from the study using rCYP (Fig. 3) that, at a concentration of AMMC close to its Km, CYP2D6 was the major enzyme involved in the formation of AMHC. This was further confirmed in experiments where the AMMC-O-demethylation was completely abolished (96%) in the presence of a CYP2D6 antibody. At a concentration of AMMC well above its Km (25 μM), other enzymes, especially the CYP1 family, also contributed to the metabolism of AMMC as shown by the study with rCYPs. An experiment with inhibitory antibody (Fig. 4), where only partial inhibition was obtained with donor 2, provided further evidence of the role of P4501A in the metabolism of AMMC because it had been shown, in a previous study, that this particular donor was high in CYP1A2 (data not shown).
The AMMC-O-demethylase activity in HLM obtained from various donors was highly correlated with the 1′-hydroxylation of bufuralol, a highly selective CYP2D6 probe (Fig. 5). It did not correlate with any other selective CYP substrates. At a concentration of 1.5 μM AMMC, the linear regression line passed through the origin and no activity was observed in two specimens of hepatic microsomes that were deficient in CYP2D6 [as determined by Western blot and bufuralol-1′-hydroxylase activity (results not shown)]. These observations imply that there was a minimal contribution of other enzymes to AMMC demethylation under these conditions. In contrast, at a concentration of 25 μM AMMC, the linear regression line did not pass through the origin and significant activities were observed in the two specimens of hepatic microsomes that were deficient in CYP2D6. This is in accordance with the fact that there is a contribution from other enzymes (putatively CYP1A2) to AMMC demethylation at this higher concentration.
Knowing the experimental conditions under which AMMC was selectively metabolized by CYP2D6 in HLM, a CYP2D6 inhibition assay was developed in a 96-well plate format using the fluorescence of AMHC as the endpoint. Two approaches were used to eliminate the interference originating from the fluorescence of NADPH. In HLM, an enzymatic NADPH removal step using glutathione reductase and oxidized glutathione was used (Chauret et al., 1999), whereas for the rCYP, an NADPH-regenerating system was refined to minimize the concentration of NADPH. In both cases, the interference of NADPH was minimal. The assay was performed in an HLM rich in CYP2D6 and at a concentration of AMMC equal to approximately double the Km to achieve a good signal-to-noise ratio (at this concentration, AMMC-O-demethylation is selective for CYP2D6). Similar experimental conditions were adopted for inhibition using bufuralol as a probe. The IC50 values obtained for eight inhibitors were all within 4-fold using the two probe substrates, which is good considering that the inhibitors chosen covered a range of greater than 5 log units of IC50 values. Similar results were obtained in experiments conducted using rCYP2D6 as the enzyme source. In general, the values obtained with AMMC were slightly lower than those obtained with bufuralol. The fact that the amount of protein and the incubation time in the inhibition protocols with AMMC in HLM were higher may account for the difference, especially if the test compounds are extensively metabolized by microsomal proteins.
In conclusion, it has been clearly demonstrated that, under proper experimental conditions, the novel fluorogenic substrate, AMMC, can be used as a selective CYP2D6 probe in HLM. It has been recommended that new chemical entities be evaluated in different in vitro systems for a better understanding of the CYP2D6 inhibition phenomenon (Palamanda et al., 1998). This present study demonstrates that AMMC can be added to the arsenal of currently accepted CYP2D6 probes such as bufuralol for studying CYP2D6 inhibition in human liver microsomes. The great advantage of AMMC over standard probes is that HPLC analysis is not required for metabolite quantification, increasing sample throughput.
Footnotes
- Abbreviations used are::
- CYPs
- cytochromes P450
- rCYP
- recombinant cytochrome P450
- HLM
- human liver microsome
- AMMC
- 3-[2-(N,N-diethyl-N-methylammonium)-ethyl]-7-methoxy-4-methylcoumarin
- AMHC
- 3-[2-(N,N-diethyl-N-methylammonium)ethyl]-7-hydroxy-4-methylcoumarin
- AHMC
- 3-[2-(diethylamino)ethyl]-7-hydroxy-4-methylcoumarin
- NCE
- new chemical entity
- MS
- mass spectrometry
- LC
- liquid chromatography
- HPLC
- high-performance liquid chromatography
- Received March 23, 2001.
- Accepted May 23, 2001.
- The American Society for Pharmacology and Experimental Therapeutics